The Lancet Oncology

Editorial

www.thelancet.com/oncology Vol 13 June 2012 559

SGO

/Scie

nce

Phot

o Li

brar

y

For more on The Lancet Oncology’s call for GPs to receive better education see Leading Edge Lancet Oncol 2009; 10: 97

For more on variation in GP referral patterns for patients with cancer see Articles Lancet Oncol 2012; 13: 353–65

For more on the partnership between Cancer Research UK and the Royal College of General Practitioners see News Lancet Oncol 2012: 12: e232

For more on QCancer risk calculators see http://qcancer.org

For more on the challenges of supporting patients with multiple morbidities see Articles Lancet 2012; published online May 10. DOI:10.1016/S0140-6736(12)60240-2

Cancer detection and primary care…revisited3 years ago, The Lancet Oncology called for general practitioners (GPs) in the UK to undertake more rigorous training and better continued education to identify the key symptoms of complex diseases such as cancer. As gatekeepers of the health-care system, it is crucial that GPs are able to triage patients to secondary care as soon as possible. Results of a survey released in May, 2012, by the UK Teenage Cancer Trust (TCT) show progress is still urgently needed. A third of all young cancer patients reported their GPs took no action despite presentation with common cancer symptoms and a quarter of patients had to visit the GP four or more times before their symptoms were taken seriously.

Patients need to be able to trust their family doctor and be confi dent that they will be treated accurately and with suffi cient priority. The TCT survey is disturbing—misdiagnoses were frequent and some patients were labelled as attention seekers. Rationally, Simon Davies, TCT chief executive, believes, “young people need GPs to take a ‘three strikes’ approach. If a young person presents with the same symptoms three times, GPs should automatically refer them for further investigation”.

Although the TCT survey was small (collating the opinions of only 300 patients), the fi ndings mirror those of Lyratzopoulos and colleagues published in The Lancet Oncology in April, 2012, that analysed more than 41 000 patients with 24 types of cancer. In that study, researchers found patients typically needed three or more consultations with their GP before a referral was made and the probability of an increased number of consultations was higher among young patients. Additionally, the number of consultations varied by cancer type, further indicating a lack of recognition of classic cancer symptoms.

So what can be done to restore trust? Usefully, Cancer Research UK and the Royal College of General Practitioners have launched an initiative to support GPs by putting together models of best practice, and by reviewing care pathways and thresholds for further investigation to ensure GPs have better access to diagnostics and secondary care. The initiative has also appointed a national GP clinical lead to coordinate eff orts. Additionally, the Department of Health has announced a pilot project within GP practices of cancer-risk prediction

tools (QCancer risk calculators) developed by researchers at the University of Nottingham.

These partnerships are good examples of engagement between policy makers and physicians with organisations that have a perceptive understanding of the patient viewpoint and of research realities and possibilities. Whether these initiatives will be successful in transforming the eff ectiveness of the GP and improving patient care will take time to assess, but it is unlikely that they will prove to be a broad panacea. It is more likely that even greater engagement between traditional and less traditional partners will be needed to develop innovative solutions. It is becoming increasingly clear, for example, that the UK health-care system is not designed to cope with multiple comorbidities—a common situation among patients with cancer—and in the future GPs will need to take a central and proactive role in coordinating patient care throughout their entire journey within the National Health Service. This will require rethinking of the current infrastructure, and might require adjustments to GPs’ case burdens to ensure suffi cient time is available for more thorough consultations, especially in socioeconomically deprived areas.

While the role of the GP in cancer diagnosis is undeniably important, it is essential not to forget interdependency on improved patient education, screening, secondary care and access to latest treatments, supportive and palliative care, and coordinated long-term follow-up. The GP, therefore, cannot be blamed entirely for cancer survival in the UK lagging behind other high-income countries. Improved understanding of the factors contributing to the diff erences between the UK’s cancer outcomes and those of other countries will provide important clues and solutions.

800 000 people visit a GP every day in the UK, but questions are increasingly being asked about the competency of those doctors that undermine patient trust. This is unfortunate given the UK Government is about to hand over considerably more responsibility to primary care physicians as part of the controversial health-care reform bill. However, implementation of this bill could be a fresh start in a process of restoring trust and ensuring GPs have access to the best tools necessary to provide a fi rst-class service and to guarantee all patients receive the best possible care. ■ The Lancet Oncology

Comment

560 www.thelancet.com/oncology Vol 13 June 2012

Capecitabine in the treatment of rectal cancerThe fl uorouracil prodrug capecitabine was developed as an oral substitute for intravenous fl uorouracil in the 1990s. Since then, many phase 2 and 3 trials have investigated capecitabine in diff erent tumour types and stages, at various doses, and as a single agent or multiagent therapy.1,2 Most phase 3 trials that compared the two drugs reported that capecitabine was at least as eff ective as fl uorouracil, and capecitabine was approved by the US Food and Drug Administration (FDA) for treatment of metastatic breast cancer in 1998, for metastatic colorectal cancer in 2001, and as adjuvant therapy for colon cancer in 2005.

Fluorouracil-based chemoradiation is standard treat ment for many solid tumours, and substituting fl uorouracil with capecitabine is attractive because of the ease of administration and mimicking of a continuous infusion.3 Capecitabine has been assessed in several phase 1 and 2 trials of adjuvant or neoadjuvant chemoradiotherapy for rectal cancer, as monotherapy or in combination with oxaliplatin, irinotecan, or targeted therapies; however, until

now, capecitabine was never formally compared with fl uorouracil in a randomised trial.1 In The Lancet Oncology, Hofh einz and colleagues4 report results of their trial testing non-inferiority for overall survival with capecitabine versus fl uorouracil, as part of neoadjuvant chemoradiotherapy and as single-agent adjuvant systemic therapy. Overall survival with capecitabine was non-inferior to fl uorouracil, and, in fact, slightly better at 5 years. These fi ndings mirror those of the large X-ACT trial5 of adjuvant capecitabine in colon cancer, which led to FDA approval in 2005. The results of these two trials4,5 seem to warrant replacement of fl uorouracil with capecitabine for adjuvant therapy of rectal cancer. Substitution of capecitabine for fl uorouracil in combination regimens is also logical, and is being assessed in ongoing trials of rectal cancer registered with ClinicalTrials.gov.

Although use of adjuvant systemic therapy in rectal cancer is widespread, the evidence base for this approach is not as strong as in colon cancer,6 which can raise the question of how solid the evidence for a specifi c treatment should be.7 The post-hoc exploratory fi nding of improved survival with capecitabine over fl uorouracil in the present study adds to the large body of circumstantial evidence supporting a benefi t for adjuvant therapy in rectal cancer.

Hofh einz and colleagues’ study began in 2002 as a trial to assess postoperative chemoradiation, but was changed in 2005 to include patients receiving preoperative chemoradiation, after publication of the German CAO/ARO/AIO-94 study8 showed improved local control with neoadjuvant chemoradiotherapy. This amendment presented some methodological diffi culties, since the two cohorts could not be directly compared. Whereas in the adjuvant cohort the inclusion of stage II–III disease was based on histological staging, inclusion in the neoadjuvant cohort was necessarily based on clinical staging. In the CAO/ARO/AIO-94 trial, such clinical staging meant that 18% of patients had stage I disease.8 Therefore, better survival might be expected in the neoadjuvant compared with adjuvant cohort of the present trial; however, the reverse was true. This is an intriguing result and might be related to lower compliance with adjuvant chemotherapy after preoperative chemoradiation and surgery. So

vere

ign,

ISM

/Scie

nce

Phot

o Li

brar

y

See Articles page 579

Comment


In the neoadjuvant cohort, capecitabine chemo-radiation provided a better response than fl uorouracil chemoradiation; there were more pathological complete responses and more downstaging. This does not necessarily translate into better local control, because with optimum total mesorectal excision after chemoradiotherapy the number of local recurrences should already be very low. Better local control could, however, be benefi cial with the current interest in organ-saving treatment of rectal cancer.

It is anticipated that the results of the NSABP R-04 trial (NCT00058474), expected at the end of 2013, will show, in accordance with the present study, that capecitabine is at least as eff ective as fl uorouracil for neoadjuvant chemoradiotherapy, confi rming capecitabine as the basis for systemic therapy in the treatment of colorectal cancer. Future trials should focus on the role of chemoradiotherapy in organ-saving treatment, and on improving the cure of micrometastatic disease, possibly by treating earlier in a neoadjuvant setting.

Geerard L Beets, *Regina G H Beets-Tan Department of Surgery (GLB) and Department of Radiology (RGHB-T), Maastricht University Medical Center, Maastricht, [email protected]

We declare that we have no confl icts of interest.

1 Hirsch BR, Zafar SY. Capecitabine in the management of colorectal cancer. Cancer Manag Res 2011; 3: 79–89.

2 Bang YJ. Capecitabine in gastric cancer. Expert Rev Anticancer Ther 2011; 11: 1791–806.

3 Rich TA, Shepard RC, Mosley ST. Four decades of continuing innovation with fl uorouracil: current and future approaches to fl uorouracil chemoradiation therapy. J Clin Oncol 2004; 22: 2214–32.

4 Hofh einz R-D, Wenz F, Post S, et al. Chemoradiotherapy with capecitabine versus fluorouracil for locally advanced rectal cancer: a randomised, multicentre, non-inferiority, phase 3 trial. Lancet Oncol 2012; 13: 579–88.

5 Twelves C, Wong A, Nowacki MP, et al. Capecitabine as adjuvant treatment for stage III colon cancer. N Engl J Med 2005; 352: 2696–704.

6 Bujko K, Glynne-Jones R, Bujko M. Does adjuvant fl uoropyrimidine-based chemotherapy provide a benefi t for patients with resected rectal cancer who have already received neoadjuvant radiochemotherapy? A systematic review of randomised trials. Ann Oncol 2010; 21: 1743–50.

7 Glimelius B. Adjuvant chemotherapy in rectal cancer—an issue or a nonissue? Ann Oncol 2010; 21: 1739–41.

8 Sauer R, Becker H, Hohenberger W, et al. Preoperative versus postoperative chemoradiotherapy for rectal cancer. N Engl J Med 2004; 351: 1731–40.

Finally, a substantial role for radiotherapy in melanoma

Jam

es S

teve

nson

/Scie

nce

Phot

o Li

brar

y

Published OnlineMay 9, 2012DOI:10.1016/S1470-2045(12)70165-1


Does adjuvant radiotherapy have a well-defi ned role in the defi nitive management of high-risk malignant melanoma? For decades, the answer to this question has been murky and contentious.1

Early reports gave confl icting results, but the data were clouded by variability in target fi eld sizes, radiation doses, and fractionation schemes. In The Lancet Oncology, Bryan Burmeister and colleagues2 present an important intergroup randomised trial showing that adjuvant nodal basin radiotherapy, when used carefully and systematically, signifi cantly improved regional lymphatic control for high-risk patients compared with no further treatment after lymphadenectomy (20 relapses among 109 patients in the adjuvant radiotherapy group vs 34 among 108 patients in the observation group, hazard ratio [HR] 0·56, 95% CI 0·32–0·98; p=0·041). They show that widely accepted risk stratifi cation measures, such as the number and size of involved nodes and the presence of extracapsular disease, might be used to identify patients at high risk of regional lymphatic failure, and that the treatment of these

patients with a radiation dose of 48 Gy in 20 fractions will signifi cantly improve local control. Although Burmeister and colleagues showed a signifi cant improvement in risk of local relapse within the aff ected nodal basins, unfortunately, overall survival did not diff er signifi cantly (59 vs 47 deaths, HR 1·37, 95% CI 0·94–2·01; p=0·12). Toxic eff ects were generally mild and manageable, much the same as in previous studies.

Where do we go from here, and how do we build on this work? Many new, promising targeted pharmaceuticals and immunomodulating compounds with clear activity against melanoma have been introduced.3 These compounds were developed on the basis of a wealth of preclinical data for melanoma cell-cycle regulatory circuits, signal transduction control, and immune system activation signals.4 Some of this work relates specifi cally to the identifi cation of mutations that activate oncogenes that are present in a large proportion of melanoma specimens and—perhaps more importantly—the synthesis and testing of small molecule inhibitors of these aberrant gene

Comment


pathways. Thus, the RAS and MAPK pathways might fall victim to sporadic activating mutations, such as V600E, which causes constitutive activation of the BRAF oncogene. This misregulated pathway accelerates cellular proliferation and also induces prolonged activation of key antiapoptotic cell survival pathways.5 To repair this dangerous loss of cellular homoeostasis, several new drugs targeting this and similar mutations—primarily in the RAF, MEK, and KIT pathways—have been developed.6 For the BRAF mutation, the inhibitor vemurafenib targets the locus aff ected by the mutation, and clinical trials with this drug showed unprecedented early responses, eff ectively doubling survival for mutation-sensitive patients. This triumph of rational drug design and molecularly targeted therapy in early trials resulted in rapid FDA approval for metastatic melanoma.7 Other compounds that address genetic mutations in one of the aff ected pathways are now in late phase testing or early post-approval assessment.

Could we reasonably add one or more of these new targeted agents to local-field radiotherapy? Preclinical studies exploring such potential combinations have been published. For example, Sambade and colleagues8 showed that BRAF-positive cells were radiosensitised after pretreatment with vemurafenib. Full clinical trials of this and related combination therapies are in progress. In immunomodulation, the humanised monoclonal antibody ipilimumab recognises and blocks the CTLA 4 site on lymphocytes, thus changing the activation of T regulatory and helper cell subsets.9 Studies combining radiotherapy with ipilimumab are underway.

Postow and colleagues10 reported details of immunological analyses of a patient with widespread metastatic melanoma who was receiving both external radiotherapy and ipilimumab. The patient had clear in-fi eld responses and delayed responses

out-of-fi eld (abscopal) to localised radiotherapy. The kinetics of the cellular immune changes and the time to clinical response are consistent with the hypothesis that the radiation-induced abscopal eff ect relates to opsonisation of melanoma target antigens with heightened presentation or processing of cellular immune system targets. Further confi rmatory studies are awaited.

The results of Burmeister and coworkers’ study fi nally give us a foundation on which to build. Radiotherapy is unlikely to have a dominant role in melanoma treatment, but we can at least fi nally assign it a reserved place in the treatment arsenal.

Roger MacklisRadiation Oncology, Cleveland Clinic, Cleveland, OH 44195, [email protected]

I declare that I have no confl icts of interest.

1 Khan N, Khan M, Almasan A, Singh A, Macklis R. The evolving role of radiation therapy in the management of malignant melanoma. Int J Rad Oncol Biol Phys 2011; 80: 645–54.

2 Burmeister BH, Henderson MA, Ainslie J, et al. Adjuvant radiotherapy versus observation for patients at risk of lymph node fi eld relapse after therapeutic lymphadenectomy for melanoma: a randomised trial. Lancet Oncol 2012; published online May 8. DOI:10.1016/S1470-2045(12)70138-9.

3 Sekulic A, Haluska P, Miller AJ, et al. Malignant melanoma in the 21st century: the emerging molecular landscape. Mayo Clin Proc 2008; 83: 825–46.

4 Fecher L A, Cummings SD, Keefe MJ, Alani RM. Toward a molecular classifi cation of melanoma. J Clin Oncol 2007; 25: 1606–20.

5 Miller AJ, Mihm MC. Mechanisms of disease: melanoma. N Engl J Med 2006; 355: 51–65.

6 Smyth EC, Carvajal RD. Treatment of metastatic malignant melanoma: a new world opens. The Skin Cancer Foundation Journal (New York) 2011: 46–48.

7 Mouawad R, Sebert M, Michels, et al. Treatment for metastatic melanoma: old drugs and new strategies. Crit Rev Oncol Hematol 2010; 74: 27–39.

8 Sambade MJ, Peters EC, Thomas NE, Kaufmann WK, Kimple RJ, Shields JM. Melanoma cells show a heterogeneous range of sensitivity to ionizing radiation and are radiosensitized by inhibition of B-RAF with PLX-4032. Radiother Oncol 2011; 98: 394–99.

9 Hodi FS, O’Day SJ, McDermott DF, et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med 2010; 363: 711–23.

10 Postow MA, Callahan MK, Barker CA, et al. Immunologic correlates of the abscopal eff ect in a patient with melanoma. N Engl J Med 2012; 366: 925–31.

Lynch syndrome: new tales from the cryptPublished Online

May 1, 2012DOI:10.1016/S1470-

2045(12)70134-1


Lynch syndrome is a hereditary disease that predisposes to cancer and is caused by germline mutations in DNA mismatch repair (MMR) genes.1 The disease is inherited in a classic Mendelian dominant way, but tumours act like a recessive disease—a somatic mutation is needed

to inactivate the uninvolved allele. Individuals with Lynch syndrome are phenotypically normal at birth, but carry a silent germline mutation. At some point in the individual’s development, the so-called second hit occurs, which initiates the progression towards

Comment


neoplasia. In the case of familial adenomatous polyposis, the germline mutation occurs in the APC gene. A somatic mutation in the wild-type APC allele leads to the emergence of a dysplastic aberrant crypt focus, which is a very small adenoma. However, Lynch syndrome is diff erent. The paired mutations do not create a neoplastic phenotype directly, but rather they create a defi ciency in DNA MMR, which allows the accumulation of mutations and microsatellite instability.2

One unanswered question in Lynch syndrome regards what is the timing and sequence of genetic events through which colorectal cancers evolve. One theory is that adenomatous polyps develop as they do in sporadic neoplasia, MMR defi ciency occurs within the adenoma, which accelerates the accumulation of mutations, and is followed by step-wise neoplastic progression. An alternative idea is that defi ciency of MMR occurs fi rst, and the adenoma evolves from the MMR-defi cient cell. Data exist to support either model but no consensus has been reached on which one is most representative. Matthias Kloor and colleagues3 in The Lancet Oncology provide new insights, and show that MMR-defi cient crypts can be found in Lynch syndrome before the aff ected cells seem to be neoplastic. Kloor and colleagues3 did immunohistochemistry for MMR proteins on non-neoplastic colonic tissues that were resected along with colorectal cancer tissue. The investigators identifi ed colonic crypts that were near-normal in appearance and did not express MMR proteins. Moreover, the DNA showed microsatellite instability, suggesting strongly that mismatch repair activity was functionally defi cient in these crypts. The absent MMR protein matched the germline mutation, and no MMR-defi cient crypts were found in the colons of individuals with sporadic colorectal cancer.

MMR-defi cient crypts are prevalent in colons of patients with Lynch syndrome. About one MMR-defi cient crypt was recorded per cm² of mucosa, or about one per 10 000 crypts in the colon, and slightly fewer were recorded in the small intestine of patients with Lynch syndrome. At least one MMR-defi cient crypt was noted in eight of ten colons of patients with Lynch syndrome. Subtle nuclear abnormalities were noted at the bottoms of these crypts, but the architecture was not identifi ably neoplastic. MMR-defi cient crypts are not the equivalent of dysplastic aberrant crypt

foci, which already have a neoplastic phenotype.4 In fact, what happens to all the MMR-defi cient crypts is not yet apparent. Unfortunately, these lesions are too small and subtle to be relied upon clinically to suggest a diagnosis of Lynch syndrome. Importantly, the investigators acknowledge that, although they noted these lesions occurred frequently, most patients with Lynch syndrome will develop zero to two cancers, and typically only a few adenomatous polyps, through their lifetimes.5 Small bowel cancers occur in no more than 1–4% of patients with Lynch syndrome,6 yet that organ has one MMR-defi cient crypt per 2 cm² of mucosa. Obviously, most of these lesions do not develop into cancer. So, what happens to them?

Kloor and colleagues3 speculate that immunoediting detects and eradicates these lesions because micro-satellite instability permits frameshift mutations and predictable types of cancer-associated neoantigens.7,8 The authors propose that the generation of these neoantigens sensitise the immune system against tumours, and that patients with Lynch syndrome auto-vaccinate themselves against cancer.

What are the limitations of this work? First, the authors have clearly found MMR-defi cient crypts, and the data strongly suggest they are Lynch syndrome-specifi c, but they did not show that these lesions can grow, are neoplastic, or are precursors to neoplasms. They could be irrelevant to tumour formation. Something restrains or eliminates these lesions, and it might be that MMR-defi cient cells simply migrate up the crypt and are eliminated like their MMR-profi cient siblings. MMR-defi cient crypts do not populate the entire colon, so some mechanism must therefore limit their expansion. In a study,9 MMR-defi ciency was uniformly present in adenomatous polyps >9 mm in diameter in Lynch syndrome patients, but was frequently absent in smaller lesions, suggesting that the MMR defect occurs after the evolution of the adenoma in some instances, rather than as a primary event.9

The authors did not report immune infi ltrates in the MMR-defi cient crypts, so one might thus anticipate future studies to determine if an immunological reaction exists against these lesions. The missing link in this work is the contrast between the large number of MMR-defi cient crypts and the relatively small number of clinically relevant neoplasms in this disease.

Step

hani

e Sc

hulle

r/Sc

ienc

e Ph

oto

Libr

ary

Comment


There were an estimated 12·7 million new cases of cancers which accounted for 7·6 million deaths globally in 2008, with about two-thirds of cancer deaths in less developed countries.1 Two previous comprehensive analyses of infections and cancers reported that in 1990 and 2002 about one in six cancer cases worldwide could be attributed to infectious agents.2,3

In The Lancet Oncology, de Martel and colleagues4 provide an updated systematic analysis of the proportion of cancer cases attributable to infection globally and by region in 2008. Compared with the two previous analyses, the absolute number of cancer cases due to infection increased by about half a million since 1990, whereas the proportion of cancer cases attributable to infection remained stable at about 16–18%. Most of the infection-attributable cases occurred in less developed countries and were due to preventable or treatable infections (hepatitis B and C virus [HBV and HCV], human papillomavirus [HPV], and Helicobacter pylori). The estimated attributable fraction for all infections combined was 16·1%—by comparison, in 2004 WHO estimated the attributable fraction for the combined eff ect of nine lifestyle and environmental risk factors as 35%.5

To estimate the population attributable fractions (PAFs), de Martel and colleagues used data on cancer incidence from the GLOBOCAN project,1 along with the most recent epidemiological evidence on the causal eff ect of each infection on cancer and the most reliable

data on prevalence of infection among cases. For HBV and HCV, where reliable data on prevalence were not available for many countries, logistic regression models were fi tted using data on cancer incidence to estimate prevalence of infection in cancer cases. This method might have led to overestimation of prevalence when cross-country variations are due to other causes of cancer (eg, alcohol use for liver cancer). Future investigations might benefi t from collating all available evidence on exposure to infection, including cancer registry data and population surveys. Such a comprehensive approach might require the use of statistical models to incorporate potential sources of bias and variability among diff erent measures of exposure.

In any global analysis of risk factors and diseases, the choice of geographical regions and pooling methods can have a substantial eff ect on estimates, particularly when the data are somewhat sparse, as in the present analysis. To aggregate country-level data into regional estimates and account for missing data in some countries, the authors weighted the country-level data by sample size and incidence of corresponding cancer cases. This method gives more weight to larger studies and studies done in high-risk countries, and can lead to overestimation of regional infection prevalence.

Another methodological consideration is that a case of cancer might be attributable to more than one infection. Ideally, the joint eff ect of multiple infections

Global burden of infection-related cancer revisited

Published OnlineMay 9, 2012

DOI:10.1016/S1470-2045(12)70176-6


Jam

es C

aval

lini/S

cienc

e Ph

oto

Libr

ary

That said, this work is highly novel, underscores the diff erences between Lynch syndrome and sporadic colorectal cancers, and raises a fresh group of important questions to be addressed.

C Richard Boland GI Cancer Research Laboratory, Division of Gastroenterology, Department of Internal Medicine, Sammons Cancer Center and Baylor Research Institute, Baylor University Medical Center, Dallas, TX, [email protected]


1 Boland CR. Evolution of the nomenclature for the hereditary colorectal cancer syndromes. Fam Cancer 2005; 4: 211–18.

2 Boland CR, Goel A. Microsatellite instability in colorectal cancer. Gastroenterology 2010; 138: 2073–87.

3 Kloor M, Huth C, Viogt AY, et al. Prevalence of mismatch repair-defi cient crypt foci in Lynch syndrome: a pathological study. Lancet Oncol 2012; published online May 1. DOI:10.1016/S1470-2045(12)70109-2.

4 Gupta AK, Pretlow TP, Schoen RE. Aberrant crypt foci: what we know and what we need to know. Clin Gastroenterol Hepatol 2007; 5: 526–33.

5 Edelstein DL, Axilbund J, Baxter M, et al. Rapid development of colorectal neoplasia in patients with Lynch syndrome. Clin Gastroenterol Hepatol 2011; 9: 340–43.

6 Schulmann K, Brasch FE, Kunstmann E, et al. HNPCC-associated small bowel cancer: clinical and molecular characteristics. Gastroenterology 2005; 128: 590–99.

7 Duval A, Hamelin R. Mutations at coding repeat sequences in mismatch repair-defi cient human cancers: toward a new concept of target genes for instability. Cancer Res 2002; 62: 2447–54.

8 Bauer K, Michel S, Reuschenbach M, Nelius N, von Knebel DM, Kloor M. Dendritic cell and macrophage infi ltration in microsatellite-unstable and microsatellite-stable colorectal cancer. Fam Cancer 2011; 10: 557–65.

9 Yurgelun MB, Goel A, Hornick JL, et al. Microsatellite instability and DNA mismatch repair protein defi ciency in Lynch syndrome colorectal polyps. Cancer Prev Res 2012; 5: 574–82.

Comment

www.thelancet.com/oncology Published online May 9, 2012 DOI:10.1016/S1470-2045(12)70176-6 1

Global burden of infection-related cancer revisitedThere were an estimated 12·7 million new cases of cancers which accounted for 7·6 million deaths globally in 2008, with about two-thirds of cancer deaths in less developed countries.1 Two previous comprehensive analyses of infections and cancers reported that in 1990 and 2002 about one in six cancer cases worldwide could be attributed to infectious agents.2,3

In The Lancet Oncology, de Martel and colleagues4 provide an updated systematic analysis of the proportion of cancer cases attributable to infection globally and by region in 2008. Compared with the two previous analyses, the absolute number of cancer cases due to infection increased by about half a million since 1990, whereas the proportion of cancer cases attributable to infection remained stable at about 16–18%. Most of the infection-attributable cases occurred in less developed countries and were due to preventable or treatable infections (hepatitis B and C virus [HBV and HCV], human papillomavirus [HPV], and Helicobacter pylori). The estimated attributable fraction for all infections combined was 16·1%—by comparison, in 2004 WHO estimated the attributable fraction for the combined eff ect of nine lifestyle and environmental risk factors as 35%.5

To estimate the population attributable fractions (PAFs), de Martel and colleagues used data on cancer incidence from the GLOBOCAN project,1 along with the most recent epidemiological evidence on the causal eff ect of each infection on cancer and the most reliable data on prevalence of infection among cases. For HBV and HCV, where reliable data on prevalence were not available for many countries, logistic regression models were fi tted using data on cancer incidence to estimate prevalence of infection in cancer cases. This method might have led to overestimation of prevalence when cross-country variations are due to other causes of cancer (eg, alcohol use for liver cancer). Future investigations might benefi t from collating all available evidence on exposure to infection, including cancer registry data and population surveys. Such a comprehensive approach might require the use of statistical models to incorporate potential sources of bias and variability among diff erent measures of exposure.

In any global analysis of risk factors and diseases, the choice of geographical regions and pooling methods

can have a substantial eff ect on estimates, particularly when the data are somewhat sparse, as in the present analysis. To aggregate country-level data into regional estimates and account for missing data in some countries, the authors weighted the country-level data by sample size and incidence of corresponding cancer cases. This method gives more weight to larger studies and studies done in high-risk countries, and can lead to overestimation of regional infection prevalence.

Another methodological consideration is that a case of cancer might be attributable to more than one infection. Ideally, the joint eff ect of multiple infections should be estimated using data on prevalence of co-infection and the combined relative risks. Since these data are unavailable, the authors summed the number of attributable cancers across diff erent infections, assuming that diff erent infections do not interact to cause cancer. This assumption is necessary but not suffi cient to make the eff ect additive. A cancer might be caused by more than one infectious agent even if the infections do not share biological pathways; therefore, the joint eff ect of multiple infections is usually less than the sum of individual eff ects.

Notwithstanding these limitations, which are some-what imposed by the available data, de Martel and colleagues have provided the most up-to-date worldwide estimates of the role of infectious agents in causing cancer. Their estimates show the potential for preventive and therapeutic programmes in less developed countries to signifi cantly reduce the global burden of cancer and the vast disparities across regions and countries.6,7 Since eff ective and relatively low-cost vaccines for HPV8 and HBV9 are available, increasing vaccine coverage should be a priority for health systems in high-burden countries. De Martel and colleagues’ analysis also highlights gaps in our knowledge of the prevalence and eff ect of carcinogenic infections, particularly in less developed countries. Improvi ng the quality and quantity of such information is essential for prioritising preventive programmes and monitoring their eff ectiveness.

Goodarz DanaeiHarvard School of Public Health, Boston, MA, [email protected]



See Online/ArticlesDOI:10.1016/S1470-2045(12)70137-7

Jam

es C

aval

lini/S

cienc

e Ph

oto

Libr

ary

Comment

2 www.thelancet.com/oncology Published online May 9, 2012 DOI:10.1016/S1470-2045(12)70176-6

1 Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010; 127: 2893–917.

2 Pisani P, Parkin DM, Munoz N, Ferlay J. Cancer and infection: estimates of the attributable fraction in 1990. Cancer Epidemiol Biomarkers Prev 1997; 6: 387–400.

3 Parkin DM. The global health burden of infection-associated cancers in the year 2002. Int J Cancer 2006; 118: 3030–44.

4 de Martel C, Ferlay J, Franceschi S, et al. Global burden of cancers attributable to infections in 2008: a review and synthetic analysis. Lancet Oncol 2012; published online May 9. DOI:10.1016/S1470-2045(12)70137-7.

5 WHO. Global health risks: mortality and burden of disease attributable to selected major risks. Geneva: World Health Organization, 2004.

6 Jemal A, Center MM, DeSantis C, Ward EM. Global patterns of cancer incidence and mortality rates and trends. Cancer Epidemiol Biomarkers Prev 2010; 19: 1893–907.

7 Kamangar F, Dores GM, Anderson WF. Patterns of cancer incidence, mortality, and prevalence across fi ve continents: defi ning priorities to reduce cancer disparities in diff erent geographic regions of the world. J Clin Oncol 2006; 24: 2137–50.

8 Goldie SJ, Kohli M, Grima D, et al. Projected clinical benefi ts and cost-eff ectiveness of a human papillomavirus 16/18 vaccine. J Natl Cancer Inst 2004; 96: 604–15.

9 Kim SY, Salomon JA, Goldie SJ. Economic evaluation of hepatitis B vaccination in low-income countries: using cost-eff ectiveness aff ordability curves. Bull World Health Organ 2007; 85: 833–42.

Comment


should be estimated using data on prevalence of co-infection and the combined relative risks. Since these data are unavailable, the authors summed the number of attributable cancers across diff erent infections, assuming that diff erent infections do not interact to cause cancer. This assumption is necessary but not suffi cient to make the eff ect additive. A cancer might be caused by more than one infectious agent even if the infections do not share biological pathways; therefore, the joint eff ect of multiple infections is usually less than the sum of individual eff ects.

Notwithstanding these limitations, which are some what imposed by the available data, de Martel and colleagues have provided the most up-to-date worldwide estimates of the role of infectious agents in causing cancer. Their estimates show the potential for preventive and therapeutic programmes in less developed countries to signifi cantly reduce the global burden of cancer and the vast disparities across regions and countries.6,7 Since eff ective and relatively low-cost vaccines for HPV8 and HBV9 are available, increasing vaccine coverage should be a priority for health systems in high-burden countries. De Martel and colleagues’ analysis also highlights gaps in our knowledge of the prevalence and eff ect of carcinogenic infections, particularly in less developed countries. Improvi ng the

quality and quantity of such information is essential for prioritising preventive programmes and monitoring their eff ectiveness.

Goodarz DanaeiHarvard School of Public Health, Boston, MA, [email protected]



2 Pisani P, Parkin DM, Munoz N, Ferlay J. Cancer and infection: estimates of the attributable fraction in 1990. Cancer Epidemiol Biomarkers Prev 1997; 6: 387–400.


4 de Martel C, Ferlay J, Franceschi S, et al. Global burden of cancers attributable to infections in 2008: a review and synthetic analysis. Lancet Oncol 2012; published online May 9. DOI:10.1016/S1470-2045(12)70137-7.

5 WHO. Global health risks: mortality and burden of disease attributable to selected major risks. Geneva: World Health Organization, 2004.

6 Jemal A, Center MM, DeSantis C, Ward EM. Global patterns of cancer incidence and mortality rates and trends. Cancer Epidemiol Biomarkers Prev 2010; 19: 1893–907.

7 Kamangar F, Dores GM, Anderson WF. Patterns of cancer incidence, mortality, and prevalence across fi ve continents: defi ning priorities to reduce cancer disparities in diff erent geographic regions of the world. J Clin Oncol 2006; 24: 2137–50.

8 Goldie SJ, Kohli M, Grima D, et al. Projected clinical benefi ts and cost-eff ectiveness of a human papillomavirus 16/18 vaccine. J Natl Cancer Inst 2004; 96: 604–15.

9 Kim SY, Salomon JA, Goldie SJ. Economic evaluation of hepatitis B vaccination in low-income countries: using cost-eff ectiveness aff ordability curves. Bull World Health Organ 2007; 85: 833–42.

Exploring the limits in low rectal cancer Investigations into the best possible surgery for cancer have generally lacked vigorous and scientifi c analysis and surgeons have been challenged to question the notion that surgical technique cannot be subjected to the rigours of scientifi c enquiry.1 In The Lancet Oncology, Shin Fujita and colleagues2 report the early results of a randomised controlled trial that aims to show whether or not mesorectal excision alone (standard treatment in Europe and North America) is non-inferior to mesorectal excision with lateral lymph node dissection (standard treatment in Japan) for patients with clinical stage II or stage III lower rectal cancer. The primary non-inferiority analysis is planned for 2015, but the researchers present results for operation time, blood loss, postoperative morbidity (grade 3 or 4), and hospital mortality. This investigation might seem unimportant to some readers, even to those

who provide modern multidisciplinary care in the form of staging, neoadjuvant and adjuvant therapy, surgery, and pathological assessment of the excised specimen. However, the detailed nature of this study means that it off ers something for everyone involved in cancer care.

To put the report into context, treatment of rectal cancer has been one of the great oncological triumphs of the past 30 years. Treatment has been built around the surgical concept of total mesorectal excision, involving precision surgery in an anatomically defi ned plane3 with a reduction in margin involvement and perforation of the excised specimen.4 These achievements have been enhanced by large randomised trials of total mesorectal excision, with preoperative therapy in selected cases,5 although the modest reductions in local recurrence have not



Stev

e Gsc

hmei

ssne

r/Sc

ienc

e Ph

oto

Libr

ary

Comment


translated into improved survival. The evolution of MRI staging has helped to bring together these initiatives and aids selection and individualised treatments.6

In the background is the widely, but not universally,7 held belief that rectal cancer might be diff erent in patients in Japan compared with those in Europe and North America, mainly because of the incidence of lateral pelvic sidewall nodal disease. In Japan, lateral nodal involvement is considered common and amenable to surgical resection. In Europe and North America, involved sidewall nodes are considered indicative of the presence, or imminent predictors, of systemic disease. Surgical excision is deemed futile, with the best possible treatment judged to be a combination of radiotherapy and chemotherapy.7 Both strategies have merit, particularly when nodes are visualised at preoperative staging, but both involve over-treatment, mainly because of staging inaccuracy.

Fujita and colleagues randomly assigned 350 patients to mesorectal excision alone and 351 to mesorectal excision with lateral lymph node dissection. As expected, the addition of lateral lymph node dissection resulted in a signifi cantly longer operation time (median 360 min [IQR 296–429] vs 254 min [210–307]; p<0·0001), and signifi cantly higher blood loss (576 mL [IQR 352–900] vs 337 mL [170–566]; p<0·0001). Postoperative complications of grade 3 and grade 4 were also higher in the group who underwent lateral lymph node dissection, but diff erences were not signifi cant.

Patients were defi ned as having clinical stage II or stage III disease, but in this context radiological might be a more appropriate term because CT and MRI are now standard. Unsurprisingly, several patients were over-staged or under-staged. Thus use of well established ultrasound terminology by prefi xing the TNM preoperative staging with the technique used, such as “mr” or “ct”, is recommended.8 This investigation will help correlation between staging and ultimate pathology because no patients had neoadjuvant therapy and therefore the degree of under-staging and over-staging can be estimated. The pathological fi nding that 7% of patients with early tumours had involved pelvic sidewall nodes in the group who underwent lateral lymph node dissection is important. Follow-up will show whether removing

these nodes has been diagnostic or therapeutic, and whether the extra morbidity, costs, and potential mortality is justifi able if lateral lymph node dissection was applied to populations in Europe and North America.

The Japanese defi nition of lower rectal cancer is below the peritoneal refl ection, a variable entity that is diffi cult to categorise accurately in the preoperative setting7 and Fujita and colleagues importantly report the height of the tumour from the anal verge in their table 1. Such information, in combination with the English Low Rectal Cancer Development Programme MRI-based defi nition, ought to become the standard by which treatments for low rectal cancer are measured and compared. However, a crucial aspect that is not reported is whether patients with a low anastomosis had a defunctioning stoma. The overall higher rate of complications in the mesorectal excision with lateral lymph node dissection group is unsurprising but rates of anastomotic leakage (13% in the mesorectal excision alone group vs 11% in the mesorectal excision with lateral lymph node dissection group) are diffi cult to comprehend and suggests increased use of defunctioning stoma with more complex surgery. Stoma rates should be reported in future updates.

Despite these issues, this will be a landmark study when the full results are available, and it will advance knowledge in many ways. The true strength of this investigation is the exciting opportunity to explore aspects of optimum staging and surgery, and to enlighten all clinicians in the matter of the pelvic sidewall, one of the most treacherous and technically challenging aspects of modern low rectal cancer management.

Brendan MoranColorectal Research Unit, North Hampshire Hospital, Basingstoke, Hampshire RG24 9NA, [email protected]


1 Horton R. Surgical research or comic opera: questions, but few answers. Lancet 1996; 347: 984–85.

2 Fujita S, Akasu T, Mizusawa J, et al, a, on behalf of the Colorectal Cancer Study Group of Japan Clinical Oncology Group. Postoperative morbidity and mortality after mesorectal excision with and without lateral lymph node dissection for clinical stage II or stage III lower rectal cancer (JCOG0212): results from a multicentre, randomised controlled, non-inferiority trial. Lancet Oncol 2012; published online May 15. DOI:10.1016/S140-2045(12)70158-4.

For the English Low Rectal Cancer Development

Programme see http://www.lorec.nhs.uk

Comment


Conventional treatment of prostate cancer is blindfolded; the disease is typically treated on the basis of histological evidence obtained through random sampling of the prostate gland. Focal therapy, however, is changing this.1 With the aid of advanced imaging techniques, such as multiparametric MRI, we are now often able to visualise prostate cancer,2 thus enabling organ-sparing approaches by precise targeting of the cancerous foci as opposed to radical treatment of the entire gland.

Advances in imaging as well as ablative techniques have allowed the development of minimally invasive alternatives. Development of new techniques and refi nement of old techniques has not ceased and we are likely to witness greater integration of these technologies in the management schemes for prostate cancer. The study by Hashim Ahmed and colleagues3 represents the fi rst rigorously designed study of the early outcomes of focal therapy (ie, the application of an organ-sparing approach for a disease that has been traditionally treated with radical, whole-gland therapies).

This study should be considered as an early report, in view of the constant development of both imaging and ablative technology. In fact, in this study ablation was not guided by MRI but by conventional ultrasound whereby the operator had the challenging task of reconstructing the MRI images (and biopsy results) and applying these images to the ultrasound image while accounting for tissue deformation and other factors. With newer technologies, such as MRI and ultrasound fusion platforms this task becomes automated, potentially increasing the accuracy of therapy. Despite these factors, the results support excellent functional outcomes that compare well with those of conventional treatment. This reinforces

the notion that tissue-sparing techniques translate to a reduction in adverse events and thus better preservation of physiological functions that could result in improved quality of life.

What will the future of focal therapy be and how would it refl ect on prostate cancer treatment at large? Although focal therapy remains a developmental approach, evidence is being gathered suggesting that the exploration of such techniques is worthwhile in view of acceptable oncological outcomes paired with excellent preservation of physiological functions, namely urinary and sexual function domains.1,4 Focal therapy needs standardisation, defi nition of consensual treatment schemes and nomenclature, as well as optimisation of patient selection criteria.5 With these issues addressed, focal therapy could potentially become a mainstream therapy for localised prostate cancer that would off er personalised treatment tailored to the patient and his disease characteristics as opposed to conventional, stereotyped, whole-gland treatment.

Furthermore, focal therapy has allowed for imaging of the disease in situ. Modern imaging studies suggest that prostate cancer can not only be seen but also diff erentiated from benign tissue and provide a degree of characterisation of the disease.2,6 This aspect could lead to an entirely new framework for diagnosis of prostate cancer. What if imaging could represent a substitute for random histological sampling of the prostate? What if, upon suspicion of prostate cancer we would undertake an imaging study that would guide the biopsy (similar to most other malignancies)? Would random biopsies be obsolete? Would the same imaging modality be able to diagnose, characterise, and guide treatment of prostate cancer? These questions will need straightforward answers in

Prostate cancer treatment unblinded

Published OnlineApril 17, 2012DOI:10.1016/S1470-2045(12)70136-5


Mau

ro F

erm

arie

llo/S

cienc

e Ph

oto

Libr

ary

3 Heald RJ, Moran BJ, Ryall RD, Sexton R, MacFarlane JK. Rectal cancer: the Basingstoke experience of total mesorectal excision 1978–1997. Arch Surg 1998; 133: 894–99.

4 Quirke P, Durdey P, Dixon MF, Williams NS. Local recurrence of rectal adenocarcinoma due to inadequate surgical resection. Histopathological study of lateral tumour spread and surgical excision. Lancet 1996; 328: 996–99.

5 Kapiteijn E, Marijnen CA, Nagtegaal ID, et al. Preoperative radiotherapy combined with total mesorectal excision for resectable rectal cancer. N Engl J Med 2001; 345: 638–46.

6 Taylor FGM, Quirke P, Heald RJ, et al. Preoperative high-resolution magnetic resonance imaging can identify good prognosis stage I, II, and III rectal cancer best managed by surgery alone. Ann Surg 2011; 253: 711–19.

7 Yano H, Moran BJ. The incidence of lateral pelvic side-wall nodal involvement in low rectal cancer may be similar in Japan and the West. Br J Surg 2008; 95: 33–49.

8 Moran BJ, Brown G, Cunningham D, et al. Clarifying the TNM staging of rectal cancer in the context of modern imaging and neoadjuvant therapy: “y” “u” and “p” need “mr” and “ct”. Colorectal Dis 2008; 10: 242–43.

Comment


The main principle of traditional Chinese medicine—which began more than 4000 years ago and is summarised in The Yellow Emperor’s Canon of Medicine—is the notion of syndrome diff erentiation and treatment. This notion aims to analyse, induce, synthesise, judge, and summarise the clinical data of symptoms and signs collected with four diagnostic methods (inspection, listening and smelling, inquiry, and pulse reading and palpation) into a diagnosis of one syndrome. Therapeutic strategies are then decided according to the result of syndrome diff erentiation. Thus, although people living hundreds or thousands of years ago did not have access to next-generation sequencers, they were nevertheless very astute observers of nature and human beings.

A study1 published in 2010 showed that diff erent radiological signs of nasopharyngeal carcinoma (eg, cervical lymph node metastases, skull-base erosions, and radiosensitivity) could be correlated with diff erent syndrome types from traditional Chinese medicine. In another study2 of 706 individuals, Chen and colleagues showed a possible genetic basis for the classifi cation of physical constitution in traditional Chinese medicine. Therefore, syndrome diff erentiation as per traditional Chinese medicine might provide a subtle phenotypic subclassifi cation within an allopathic disease diagnosis, suggesting an underlying genotypic diff erence.

Fast forward to modern allopathic medicine and the TNM system, which has been in existence since the 1940s, and use of classifi cation from anatomical

prognostic factors has withstood the test of time. Such anatomical assessment remains the basis of personalised medicine.

In nasopharyngeal carcinoma, a group from Hong Kong successfully tailored therapy by reducing the radiotherapy dose to the pituitary fossa and neck for patients without cancer involvement of those structures. This group’s next attempt at personalised medicine was confi rmation of the usefulness of addition of concurrent chemotherapy to radiotherapy in two trials. The fi rst trial (NCT00563927) tested the addition of chemotherapy for patients at highest risk of development of distant metastases and the second (NCT00563862) had a component of accelerated radiation in a subset of patients at increased risk for local failure. The same group is presently assessing the use of Epstein-Barr virus (EBV) DNA levels after radiotherapy to identify patients who would benefi t from additional adjuvant treatment (NCT00370890).

Use of molecular signatures has proven to be important for the prediction of outcomes of many cancers. A group from Guangzhou (Guangdong, China) led an international consortium3 to identify an eight signature classifi er for prediction of survival in nasopharyngeal carcinoma. MicroRNAs (miRNAs) make up a large class of small non-protein-coding RNAs that are very infl uential in gene regulation and cellular function. Complex interactions between miRNAs, and their dysregulation, can amplify diverse pathogenic processes that contribute to initiation of tumour growth, proliferation, and metastases.

Ricc

ardo

Cas

siani

-Ingo

ni/S

PL


DOI:10.1016/S1470-2045(12)70115-8

This online publication has been corrected. The corrected

version fi rst appeared at thelancet.com/oncology on

May 28, 2012


Personalised medicine in nasopharyngeal cancer

the near future and the urological and oncological communities have to be alert to this rapidly changing fi eld of prostate cancer management, and specifi cally, focal therapy.

Matvey Tsivian, Michael R Abern, *Thomas J PolascikDivision of Urology, Department of Surgery, and Duke Cancer Institute, Duke University Medical Center, Durham, NC, USA (MT, MRA, TJP)[email protected]


1 de la Rosette J, Ahmed H, Barentsz J, et al. Focal therapy in prostate cancer-report from a consensus panel. J Endourol 2010; 24: 775–80.

2 Yakar D, Debats OA, Bomers JG, et al. Predictive value of MRI in the localization, staging, volume estimation, assessment of aggressiveness, and guidance of radiotherapy and biopsies in prostate cancer. J Magn Reson Imaging 2012; 35: 20–31.

3 Ahmed HU, Hindley RG, Dickinson L, et al. Focal therapy for localised and multifocal prostate cancer: a prospective development study. Lancet Oncol 2012; published online April 17. DOI:10.1016/S1470-2045(12)70121-3.

4 Ahmed HU, Akin O, Coleman JA, et al. Transatlantic Consensus Group on active surveillance and focal therapy for prostate cancer. BJU Int 2011; published online Nov 11. DOI:10.1111/j.1464-410X.2011.10633.x.

5 Abern MR, Tsivian M, Polascik TJ. Focal therapy of prostate cancer: evidence-based analysis for modern selection criteria. Curr Urol Rep 2012; published online Feb 2. DOI:10.1007/s11934-012-0241-5.

6 Hoeks CM, Barentsz JO, Hambrock T, et al. Prostate cancer: multiparametric MR imaging for detection, localization, and staging. Radiology 2011; 261: 46–66.

Comment


miRNAs have prognostic signifi cance in various haematological and solid tumours. In The Lancet Oncology, Liu and colleagues4 report a distinct miRNA profi le derived from nasopharyngeal carcinoma tumour tissue, and a fi ve miRNA signature that independently predicts disease-free and overall survival and correlates with TNM staging. Moreover, in combination with TNM staging this miRNA signature improves the prognostic value versus TNM alone in terms of survival. This signature-based scoring could improve oncologists’ ability to identify patients who will likely have poorer outcomes and therefore plan improved therapeutic interventions. Liu and colleagues should be commended for their retrospective analysis of systematically annotated specimens, because biopsy material is usually small in nasopharyngeal carcinoma compared with other cancers in which surgery is defi nitively required. Their study is also the fi rst comprehensive assessment of miRNAs as prognostic biomarkers in this cancer. Such a study potentially allows for increased probing of the biological mechanisms underpinning the role of miRNA in the oncogenesis of this EBV-transformed cancer. Ideally, investigators will eventually be able to replicate and validate this tumour tissue 5-miRNA signature in patients’ serum, which is easier to access and contains miRNA that is as stable as that of tissue. Another report has suggested that serum EBV miRNA of patients with nasopharyngeal carcinoma correlates positively with cellular copy numbers of EBV miRNAs.5 The same group reported that EBV miRNAs were able to inhibit tumour suppressor genes such as PTEN and deregulate pathways including Wnt signalling.

As pointed out by Liu and colleagues, elucidation of the intricate network and roles of miRNA in

development of nasopharyngeal carcinoma could eventually allow for specifi c in-vivo therapeutic targeting. Chemotherapy and radiotherapy form the cornerstone of treatment of nasopharyngeal carcinoma. A few molecularly targeting drugs are emerging as clinically active against advanced nasopharyngeal carcinoma.6,7 Use of miRNA signatures in clinical decision-making for prediction of treatment effi cacy and increasingly tailored therapy will be the next step forward, especially for metastatic nasopharyngeal carcinoma for which no phase 3 therapeutic clinical trials have been done so far.

Han Chong Toh, Tam Cam Ha, *Joseph WeeDepartment of Medical Oncology (HCT), Division of Clinical Trials and Epidemiological Sciences (TCH, JW), and Department of Radiation Oncology (JW), National Cancer Centre Singapore, Singapore; and Duke-National University of Singapore Graduate Medical School, Singapore (HCT, TCH, JW)[email protected]


1 Bao H, Gao J, Huang T, Zhou ZM, Zhang B, Xia YF. Relationship between traditional Chinese medicine syndrome diff erentiation and imaging characterization to the radiosensitivity of nasopharyngeal carcinoma. Chin J Cancer 2010; 29: 937–45.

2 Chen S, Lv F, Gao J, et al. HLA class II polymorphisms associated with the physiologic characteristics defi ned by traditional Chinese medicine: linking modern genetics with an ancient medicine. J Altern Complement Med 2007; 13: 231–39.

3 Wang HY, Sun BY, Zhu ZH, et al. Eight-signature classifi er for prediction of nasopharyngeal carcinoma survival. J Clin Oncol 2011; 29: 4516–25.

4 Liu N, Chen NY, Cui RX, et al. Prognostic value of a microRNA signature in nasopharyngeal carcinoma: a microRNA expression analysis. Lancet Oncol 2012; published online May 3. DOI:10.1016/S1470-2045(12)70102-X.

5 Gourzones C, Jimenez AS, Busson P. Profi ling of Epstein-Barr virus-encoded microRNAs in nasopharyngeal carcinoma reveals potential biomarkers and oncomirs. Cancer 2012; 118: 698–710.

6 Chan AT, Hsu MM, Goh BC, et al. Multicenter, phase II study of cetuximab in combination with carboplatin in patients with recurrent or metastatic nasopharyngeal carcinoma. J Clin Oncol 2005; 23: 3568–76.

7 Lim WT, Ng QS, Ivy P, et al. A phase II study of pazopanib in Asian patients with recurrent/metastatic nasopharyngeal carcinoma. Clin Cancer Res 2011; 17: 5481–89.

Polyglutamine diseases and the risk of cancerNeurodegenerative disorders and cancer are two of the most devastating diseases of old age. They are the products of opposing cellular propensities: the former results from increased cell death of distinct neuronal populations in the CNS, and the latter from decreased death of neoplastic cells that eventually progress to become malignant tumours. Although the disorders seem disparate, they might, in fact, be related. As

early as 1954, reduced incidence of cancer was assoc-iated with Parkinson’s disease.1 A meta-analysis of 29 studies reported a reduced aggregate risk of cancer in patients with Parkinson’s disease (0·73, 95% CI 0·63–0·83).2 Similar fi ndings have been reported for patients with Alzheimer’s disease.3 These studies have raised the possibility that neurodegeneration and cancer are two extremes of shared pathways.


Comment


Conventional treatment of prostate cancer is blindfolded; the disease is typically treated on the basis of histological evidence obtained through random sampling of the prostate gland. Focal therapy, however, is changing this.1 With the aid of advanced imaging techniques, such as multiparametric MRI, we are now often able to visualise prostate cancer,2 thus enabling organ-sparing approaches by precise targeting of the cancerous foci as opposed to radical treatment of the entire gland.

Advances in imaging as well as ablative techniques have allowed the development of minimally invasive alternatives. Development of new techniques and refi nement of old techniques has not ceased and we are likely to witness greater integration of these technologies in the management schemes for prostate cancer. The study by Hashim Ahmed and colleagues3 represents the fi rst rigorously designed study of the early outcomes of focal therapy (ie, the application of an organ-sparing approach for a disease that has been traditionally treated with radical, whole-gland therapies).

This study should be considered as an early report, in view of the constant development of both imaging and ablative technology. In fact, in this study ablation was not guided by MRI but by conventional ultrasound whereby the operator had the challenging task of reconstructing the MRI images (and biopsy results) and applying these images to the ultrasound image while accounting for tissue deformation and other factors. With newer technologies, such as MRI and ultrasound fusion platforms this task becomes automated, potentially increasing the accuracy of therapy. Despite these factors, the results support excellent functional outcomes that compare well with those of conventional treatment. This reinforces

the notion that tissue-sparing techniques translate to a reduction in adverse events and thus better preservation of physiological functions that could result in improved quality of life.

What will the future of focal therapy be and how would it refl ect on prostate cancer treatment at large? Although focal therapy remains a developmental approach, evidence is being gathered suggesting that the exploration of such techniques is worthwhile in view of acceptable oncological outcomes paired with excellent preservation of physiological functions, namely urinary and sexual function domains.1,4 Focal therapy needs standardisation, defi nition of consensual treatment schemes and nomenclature, as well as optimisation of patient selection criteria.5 With these issues addressed, focal therapy could potentially become a mainstream therapy for localised prostate cancer that would off er personalised treatment tailored to the patient and his disease characteristics as opposed to conventional, stereotyped, whole-gland treatment.

Furthermore, focal therapy has allowed for imaging of the disease in situ. Modern imaging studies suggest that prostate cancer can not only be seen but also diff erentiated from benign tissue and provide a degree of characterisation of the disease.2,6 This aspect could lead to an entirely new framework for diagnosis of prostate cancer. What if imaging could represent a substitute for random histological sampling of the prostate? What if, upon suspicion of prostate cancer we would undertake an imaging study that would guide the biopsy (similar to most other malignancies)? Would random biopsies be obsolete? Would the same imaging modality be able to diagnose, characterise, and guide treatment of prostate cancer? These questions will need straightforward answers in

Prostate cancer treatment unblinded

Published OnlineApril 17, 2012DOI:10.1016/S1470-2045(12)70136-5


Mau

ro F

erm

arie

llo/S

cienc

e Ph

oto

Libr

ary

3 Heald RJ, Moran BJ, Ryall RD, Sexton R, MacFarlane JK. Rectal cancer: the Basingstoke experience of total mesorectal excision 1978–1997. Arch Surg 1998; 133: 894–99.

4 Quirke P, Durdey P, Dixon MF, Williams NS. Local recurrence of rectal adenocarcinoma due to inadequate surgical resection. Histopathological study of lateral tumour spread and surgical excision. Lancet 1996; 328: 996–99.


6 Taylor FGM, Quirke P, Heald RJ, et al. Preoperative high-resolution magnetic resonance imaging can identify good prognosis stage I, II, and III rectal cancer best managed by surgery alone. Ann Surg 2011; 253: 711–19.

7 Yano H, Moran BJ. The incidence of lateral pelvic side-wall nodal involvement in low rectal cancer may be similar in Japan and the West. Br J Surg 2008; 95: 33–49.

8 Moran BJ, Brown G, Cunningham D, et al. Clarifying the TNM staging of rectal cancer in the context of modern imaging and neoadjuvant therapy: “y” “u” and “p” need “mr” and “ct”. Colorectal Dis 2008; 10: 242–43.

Comment


The main principle of traditional Chinese medicine—which began more than 4000 years ago and is summarised in The Yellow Emperor’s Canon of Medicine—is the notion of syndrome diff erentiation and treatment. This notion aims to analyse, induce, synthesise, judge, and summarise the clinical data of symptoms and signs collected with four diagnostic methods (inspection, listening and smelling, inquiry, and pulse reading and palpation) into a diagnosis of one syndrome. Therapeutic strategies are then decided according to the result of syndrome diff erentiation. Thus, although people living hundreds or thousands of years ago did not have access to next-generation sequencers, they were nevertheless very astute observers of nature and human beings.

A study1 published in 2010 showed that diff erent radiological signs of nasopharyngeal carcinoma (eg, cervical lymph node metastases, skull-base erosions, and radiosensitivity) could be correlated with diff erent syndrome types from traditional Chinese medicine. In another study2 of 706 individuals, Chen and colleagues showed a possible genetic basis for the classifi cation of physical constitution in traditional Chinese medicine. Therefore, syndrome diff erentiation as per traditional Chinese medicine might provide a subtle phenotypic subclassifi cation within an allopathic disease diagnosis, suggesting an underlying genotypic diff erence.

Fast forward to modern allopathic medicine and the TNM system, which has been in existence since the 1940s, and use of classifi cation from anatomical

prognostic factors has withstood the test of time. Such anatomical assessment remains the basis of personalised medicine.

In nasopharyngeal carcinoma, a group from Hong Kong successfully tailored therapy by reducing the radiotherapy dose to the pituitary fossa and neck for patients without cancer involvement of those structures. This group’s next attempt at personalised medicine was confi rmation of the usefulness of addition of concurrent chemotherapy to radiotherapy in two trials. The fi rst trial (NCT00563927) tested the addition of chemotherapy for patients at highest risk of development of distant metastases and the second (NCT00563862) had a component of accelerated radiation in a subset of patients at increased risk for local failure. The same group is presently assessing the use of Epstein-Barr virus (EBV) DNA levels after radiotherapy to identify patients who would benefi t from additional adjuvant treatment (NCT00370890).

Use of molecular signatures has proven to be important for the prediction of outcomes of many cancers. A group from Guangzhou (Guangdong, China) led an international consortium3 to identify an eight signature classifi er for prediction of survival in nasopharyngeal carcinoma. MicroRNAs (miRNAs) make up a large class of small non-protein-coding RNAs that are very infl uential in gene regulation and cellular function. Complex interactions between miRNAs, and their dysregulation, can amplify diverse pathogenic processes that contribute to initiation of tumour growth, proliferation, and metastases.

Ricc

ardo

Cas

siani

-Ingo

ni/S

PL


DOI:10.1016/S1470-2045(12)70115-8

This online publication has been corrected. The corrected

version fi rst appeared at thelancet.com/oncology on

May 28, 2012


Personalised medicine in nasopharyngeal cancer

the near future and the urological and oncological communities have to be alert to this rapidly changing fi eld of prostate cancer management, and specifi cally, focal therapy.

Matvey Tsivian, Michael R Abern, *Thomas J PolascikDivision of Urology, Department of Surgery, and Duke Cancer Institute, Duke University Medical Center, Durham, NC, USA (MT, MRA, TJP)[email protected]


1 de la Rosette J, Ahmed H, Barentsz J, et al. Focal therapy in prostate cancer-report from a consensus panel. J Endourol 2010; 24: 775–80.

2 Yakar D, Debats OA, Bomers JG, et al. Predictive value of MRI in the localization, staging, volume estimation, assessment of aggressiveness, and guidance of radiotherapy and biopsies in prostate cancer. J Magn Reson Imaging 2012; 35: 20–31.

3 Ahmed HU, Hindley RG, Dickinson L, et al. Focal therapy for localised and multifocal prostate cancer: a prospective development study. Lancet Oncol 2012; published online April 17. DOI:10.1016/S1470-2045(12)70121-3.


5 Abern MR, Tsivian M, Polascik TJ. Focal therapy of prostate cancer: evidence-based analysis for modern selection criteria. Curr Urol Rep 2012; published online Feb 2. DOI:10.1007/s11934-012-0241-5.

6 Hoeks CM, Barentsz JO, Hambrock T, et al. Prostate cancer: multiparametric MR imaging for detection, localization, and staging. Radiology 2011; 261: 46–66.

Comment


miRNAs have prognostic signifi cance in various haematological and solid tumours. In The Lancet Oncology, Liu and colleagues4 report a distinct miRNA profi le derived from nasopharyngeal carcinoma tumour tissue, and a fi ve miRNA signature that independently predicts disease-free and overall survival and correlates with TNM staging. Moreover, in combination with TNM staging this miRNA signature improves the prognostic value versus TNM alone in terms of survival. This signature-based scoring could improve oncologists’ ability to identify patients who will likely have poorer outcomes and therefore plan improved therapeutic interventions. Liu and colleagues should be commended for their retrospective analysis of systematically annotated specimens, because biopsy material is usually small in nasopharyngeal carcinoma compared with other cancers in which surgery is defi nitively required. Their study is also the fi rst comprehensive assessment of miRNAs as prognostic biomarkers in this cancer. Such a study potentially allows for increased probing of the biological mechanisms underpinning the role of miRNA in the oncogenesis of this EBV-transformed cancer. Ideally, investigators will eventually be able to replicate and validate this tumour tissue 5-miRNA signature in patients’ serum, which is easier to access and contains miRNA that is as stable as that of tissue. Another report has suggested that serum EBV miRNA of patients with nasopharyngeal carcinoma correlates positively with cellular copy numbers of EBV miRNAs.5 The same group reported that EBV miRNAs were able to inhibit tumour suppressor genes such as PTEN and deregulate pathways including Wnt signalling.

As pointed out by Liu and colleagues, elucidation of the intricate network and roles of miRNA in

development of nasopharyngeal carcinoma could eventually allow for specifi c in-vivo therapeutic targeting. Chemotherapy and radiotherapy form the cornerstone of treatment of nasopharyngeal carcinoma. A few molecularly targeting drugs are emerging as clinically active against advanced nasopharyngeal carcinoma.6,7 Use of miRNA signatures in clinical decision-making for prediction of treatment effi cacy and increasingly tailored therapy will be the next step forward, especially for metastatic nasopharyngeal carcinoma for which no phase 3 therapeutic clinical trials have been done so far.

Han Chong Toh, Tam Cam Ha, *Joseph WeeDepartment of Medical Oncology (HCT), Division of Clinical Trials and Epidemiological Sciences (TCH, JW), and Department of Radiation Oncology (JW), National Cancer Centre Singapore, Singapore; and Duke-National University of Singapore Graduate Medical School, Singapore (HCT, TCH, JW)[email protected]


1 Bao H, Gao J, Huang T, Zhou ZM, Zhang B, Xia YF. Relationship between traditional Chinese medicine syndrome diff erentiation and imaging characterization to the radiosensitivity of nasopharyngeal carcinoma. Chin J Cancer 2010; 29: 937–45.

2 Chen S, Lv F, Gao J, et al. HLA class II polymorphisms associated with the physiologic characteristics defi ned by traditional Chinese medicine: linking modern genetics with an ancient medicine. J Altern Complement Med 2007; 13: 231–39.

3 Wang HY, Sun BY, Zhu ZH, et al. Eight-signature classifi er for prediction of nasopharyngeal carcinoma survival. J Clin Oncol 2011; 29: 4516–25.

4 Liu N, Chen NY, Cui RX, et al. Prognostic value of a microRNA signature in nasopharyngeal carcinoma: a microRNA expression analysis. Lancet Oncol 2012; published online May 3. DOI:10.1016/S1470-2045(12)70102-X.

5 Gourzones C, Jimenez AS, Busson P. Profi ling of Epstein-Barr virus-encoded microRNAs in nasopharyngeal carcinoma reveals potential biomarkers and oncomirs. Cancer 2012; 118: 698–710.


7 Lim WT, Ng QS, Ivy P, et al. A phase II study of pazopanib in Asian patients with recurrent/metastatic nasopharyngeal carcinoma. Clin Cancer Res 2011; 17: 5481–89.

Polyglutamine diseases and the risk of cancerNeurodegenerative disorders and cancer are two of the most devastating diseases of old age. They are the products of opposing cellular propensities: the former results from increased cell death of distinct neuronal populations in the CNS, and the latter from decreased death of neoplastic cells that eventually progress to become malignant tumours. Although the disorders seem disparate, they might, in fact, be related. As

early as 1954, reduced incidence of cancer was assoc-iated with Parkinson’s disease.1 A meta-analysis of 29 studies reported a reduced aggregate risk of cancer in patients with Parkinson’s disease (0·73, 95% CI 0·63–0·83).2 Similar fi ndings have been reported for patients with Alzheimer’s disease.3 These studies have raised the possibility that neurodegeneration and cancer are two extremes of shared pathways.


Comment


In The Lancet Oncology, Ji and colleagues4 report fi ndings that further support an inverse relation be tween neurodegenerative disorders and cancer. Polyglutamine (polyQ) diseases are hereditary, age-dependent, neuro-degenerative disorders that result from expansion of exonic CAG repeats in specifi c but unrelated genes that lead to selective neuronal death.5 Nine polyQ diseases have been identifi ed: Huntington’s disease, spinobulbar muscular atrophy, dentatorubral-pallidoluysian atrophy, and spinocerebellar ataxias 1, 2, 3, 6, 7, and 17. Ji and colleagues used the Swedish national hospital discharge and outpatient registries to identify patients with Huntington’s disease, spinobulbar muscular atrophy, and hereditary ataxias, which were used as a proxy for the six spinocerebellar ataxias. The information was linked with data in the Swedish Cancer Registry to identify cases of cancer in the patients with polyQ diseases and their unaff ected parents. In 1510 patients with Huntington’s disease, 471 with spinobulbar muscular atrophy, and 3425 with hereditary ataxias, the authors report signifi cantly decreased risk of cancer (standardised incidence ratio 0·47, 95% CI 0·38–0·58 for Huntington’s disease, 0·65, 0·45–0·91 for spinobulbar muscular atrophy, and 0·77, 0·70–0·85 for hereditary ataxias).4 The study by Ji and colleagues is the largest so far to assess the relation between cancer and polyQ diseases, and confi rms and extends the fi ndings of previous reports of low incidence of cancer in patients with Huntington’s disease.6,7

These fi ndings of Ji and colleagues, along with those from other studies, support the notion of a delicate balance of common cellular pathways of death and survival operating as part of a continuum, with neurodegeneration at one extreme and cancer at the other. The most well defi ned molecular manifestation of such a continuum is cysteine-aspartic proteases (caspases). Appropriate activation of caspases is important for several processes, including development, immunity, cellular remodelling, wound healing, neurite pruning, learning and memory, and programmed cell death.8 Aberrant activation of caspases is well documented in neurodegenerative diseases and cancer, where increased activation leads to neuronal loss, whereas downregulation results in inappropriate cellular survival and growth, and ultimately leads to cancer.8

The evidence presented by Ji and colleagues4 strongly supports a role for polyQ-tract expansion in resistance to cancer, but some important questions remain unanswered. In cellular and animal models, polyQ length is inversely correlated with age of onset of polyQ disease and degree of toxic eff ects.9,10 If the decreased incidence of cancer in polyQ patients is the result of increased polyQ-related toxic eff ects in precancerous cells, the risk of cancer should be lower in patients with longer polyQ-tract expansion than in those with shorter expansion (fi gure). As the data on polyQ-tract lengths were unavailable for patients assessed by Ji and colleagues (age at diagnosis was used as a surrogate), whether a direct relation between tract length and cancer risk exists remains to be assessed. Also, because the toxic eff ects conferred by expansion of polyQ tracts are thought to refl ect gain-of-function and loss-of-function eff ects,11 the relative contributions of these mechanisms to the reduced risk of cancer requires investigation.

The study strengthens the recognition of broad and systemic eff ects of expansion of polyQ tracts. Although the focus of research in these areas is traditionally on CNS-related pathology, an assessment of alterations that aff ect tissues in both the CNS and peripheral tissues could shed light on key mechanisms that underlie cancer resistance in patients with polyQ disorders. For example, caspase activation might play a part in neuronal death in the CNS, but could also mediate the death of precancerous cells in non-CNS tissues.

The provocative and elegant study by Ji and colleagues4 provides fresh evidence that reinforces a link between

Polyglutamine length

Risk

of d

iseas

e

Risk of neurodegenerationRisk of cancer

Figure: Risk of cancer and neurodegeneration in relation to polyglutamine-tract length

Comment


the seemingly disparate diseases of neurodegeneration and cancer. The study raises the intriguing possibility that identifi cation of the factors that contribute to neurodegeneration could be doubly rewarding, and could potentially provide novel approaches for the treatment of two of the most devastating diseases of old age.

Mahmoud A Pouladi, *Michael R HaydenTranslational Laboratory in Genetic Medicine, Department of Medicine, National University of Singapore, and Agency for Science, Technology and Research (A*STAR), Singapore, Singapore (MAP, MRH); and Centre for Molecular Medicine and Therapeutics, University of British Columbia, and Child and Family Research Institute, 950 West 28th Avenue, Vancouver, BC, Canada V5Z 4H4 (MRH)[email protected]


1 Doshay LJ. Problem situations in the treatment of paralysis agitans. JAMA 1954; 156: 680–84.

2 Bajaj A, Driver JA, Schernhammer ES. Parkinson’s disease and cancer risk: a systematic review and meta-analysis. Cancer Causes Control 2010; 21: 697–707.

3 Roe CM, Fitzpatrick AL, Xiong C, et al. Cancer linked to Alzheimer disease but not vascular dementia. Neurology 2010; 74: 106–12.

4 Ji J, Sundquist K, Sundquist J. Cancer indicence in patients with polyglutamine diseases: a population-based study in Sweden. Lancet Oncol 2012; 13: 642–48.

5 Ashley CT, Warren ST. Trinucleotide repeat expansion and human disease. Ann Rev Genet 1995; 29: 703–28.

6 Sørensen SA, Fenger K. Causes of death in patients with Huntington’s disease and in unaff ected fi rst degree relatives. J Med Genet 1992; 29: 911–14.

7 Sørensen SA, Fenger K, Olsen JH. Signifi cantly lower incidence of cancer among patients with Huntington disease: an apoptotic eff ect of an expanded polyglutamine tract? Cancer 1999; 86: 1342–46.

8 Fuchs Y, Steller H. Programmed cell death in animal development and disease. Cell 2011; 147: 742–58.

9 Stevanin G, Dürr A, Brice A. Clinical and molecular advances in autosomal dominant cerebellar ataxias: from genotype to phenotype and physiopathology. Eur J Hum Gen 2000; 8: 4–18.

10 Andrew SE, Goldberg YP, Kremer B, et al. The relationship between trinucleotide (CAG) repeat length and clinical features of Huntington’s disease. Nat Genet 1993; 4: 398–403.

11 Ross CA. Polyglutamine pathogenesis: emergence of unifying mechanisms for Huntington’s disease and related disorders. Neuron 2002; 35: 819–22.

Correspondence

www.thelancet.com/oncology Vol 13 June 2012 e230

Priorities for cancer prevention Bernard Stewart’s Personal View1 in favour of lifestyle choices to prevent cancer seems to be based on the questionable premise that pollutants should be investigated individually for their ability to cause disease. This approach might work for tobacco smoke and other easily isolated causes with obvious eff ects that lend themselves well to epidemiological study. However, it is not suitable for variable inadvertent exposures to complex mixtures of substances for long periods. Our best understanding is that several carcinogenic exposures accumulate and the overall eff ect is what causes cancer in individuals and populations.

Stewart repeats some of the arguments of Colditz2 and Thun3 after the release of the President’s Cancer Panel report—ie, that people will be diverted from addressing their risky lifestyles by too much public concern about environmental and

occupational exposures. This view implies that people cannot hold two thoughts in their heads at the same time and we cannot as a society try to prevent cancer with several causes. Stewart further cites the President’s Cancer Panel’s steps that individuals can take to reduce exposure to carcinogens,4 but ignores the Panel’s overarching call for a new prevention-oriented chemicals policy.

John Snow is celebrated because he advocated for change in London in the 1850s on the basis of what his research had led him to believe. He is sometimes called the father of epidemiology,5 and he was an early advocate of the precautionary principle—ie, action on the basis of reasonably convincing, although not fully developed evidence. Public health would be well served by the use of this kind of science to obtain solutions. Dozens of known of probable human carcinogens are in use that could be controlled more strictly than they are now or phased out. We do not need to wait for more defi nitive proof

that environmental or occupational exposures have contributed to a specifi c proportion of the overall cancer burden. We declare that we have no confl icts of interest.

Jamie Page, Paul Whaley, Andrew Watterson, *Richard [email protected]

Cancer Prevention and Education Society, Axminster, UK (JP, PW); University of Stirling, Stirling, UK (AW); and University of Massachusetts, Lowell, MA, USA (RC)

1 Stewart BW. Priorities for cancer prevention: lifestyle choices versus unavoidable exposures. Lancet Oncol 2012; 13: e126–33.

2 Fox M. Cancer report energizes activists, not policy. May 9, 2010. http://www.reuters.com/article/2010/05/09/us-cancer-usa-idUSTRE64827I20100509 (accessed April 3, 2012).

3 Sampson D. Cancer and the environment. May 6, 2010. http://acspressroom.wordpress.com/2010/05/06/cancer-and-the-environment/ (accessed April 3, 2012).

4 The President’s Cancer Panel. Reducing environmental cancer risk: what we can do now. April, 2010. http://deainfo.nci.nih.gov/advisory/pcp/annualReports/pcp08-09rpt/PCP_Report_08-09_508.pdf (accessed April 3, 2012).

5 Vachon D. Doctor John Snow blames water pollution for cholera epidemic. 2005. http://www.ph.ucla.edu/epi/snow/fatherofepidemiology.html (accessed April 3, 2012).

Corrections


Correction to Lancet Oncol 2012; 13: 569

Toh HC, Ha TC, Wee J. Personalised medicine in nasopharyngeal cancer. Lancet Oncol 2012; 13: 568–69—In this Comment (published online May 3, 2012), Han Chong Toh and Tam Cam Ha’s names were spelt incorrectly. This correction has been made to the online version as of May 28, 2012, and the printed Article is correct.

News


Canada announced this week that incidences of several major lifestyle cancers have decreased signifi cantly in the past 5 years. “Estimates suggest that nearly 60% of the reduction in cancer mortality between 1990 and 2004 was due to changes in health behaviours and cancer screening practices related to lung, breast, colorectal, and prostate cancers and the Canadian Cancer Statistics Report 2012 highlights that continuing reductions in cancer are largely driven by changes in health behaviours”, comments Erin Strumpf (McGill University, Montreal, QC, Canada).

Unfortunately, these achievements are being eclipsed by the publicity generated by the federal govern-ment’s recent round of health-care budget cuts, some of which might eventually aff ect future cancer sta-tistics. On May 11, 80 doctors protested against cuts announced earlier in May by the Citizenship and Immigration Canada Act, which will prevent refugees in Canada from receiving any primary health-care provision. From the be ginning of July, this vulnerable group will only be able to access health care in an emergency, such as a stroke or heart attack, or if their health poses a risk to others, for example if they have active tuberculosis. “The cuts will undermine any attempts at providing primary health care and that would include

health/nutrition education and cancer screening”, notes Mark Tyndall (University of Ottawa, Ottawa, ON, Canada).

David Stewart (The Ottawa Hospital, Ottawa, ON, Canada) notes that it is very diffi cult to know what eff ect cutting the programme will have. “The approach to refugee claimants is very problematic on several fronts. Many of these people will have limited personal resources, and hence it is diffi cult to understand how they will be able to access any health-care services except in an emergency situation. While it is impossible to predict what impact this would have on later cancer incidence, there are many reasons that this is a bad idea”, he warns. Comments posted on various news articles reporting the health-care cuts for Canadian refugees suggest that the policy is not that unpopular with ordinary Canadians. “There is a lot of misunderstanding; the government is promoting the cuts as a measure to prevent ‘unfounded’ applications for refugee status, but the fact is that providing basic health care to any population is cost eff ective. Treating diabetes, high blood pressure, arthritis and providing health screening not only provides an advantage to the asylum seeker, it saves money in the long term”, stresses Tyndall.

The federal government is responsible for providing health-care services to First Nations, Inuit, and Aboriginal populations, and for providing health insurance to refugee claimants while they are awaiting approval and then for 3 months once a refugee has been approved. “After that they are eligible for public insurance coverage from the province where they live”, notes Strumpf. However, as Tyndall points out, approval, on average, takes 3 years, sometimes much longer. “Some are in limbo for a decade, so the denial of primary health care will have a signifi cant impact”, he says.

More public sympathy was expressed about the abrupt cut in the usual CA$800 000 annual funding to the Pauktuutit Inuit Women of Canada, a strong exponent and provider of health promotion and prevention for native Inuit women in Canada for nearly 30 years. Other recent cuts announced by the Canadian Government include a sweeping reduction in foreign aid, cuts to mental health support for the military, physician pay, which has been cut for key services, including radiology, and a $15 million cut to the federal government’s tobacco control strategy. “We are very concerned about the signifi cant cuts made to Health Canada’s highly successful tobacco control programme”, says Rob Cunningham (Canadian Cancer Society, Toronto, ON, Canada). “This strategy has been working, and these reductions will negatively impact continuing eff orts to reduce smoking rates. The severity of the cuts is disappointing, especially in light of recent positive measures by the federal government, such as banning fl avoured cigarettes and little cigars, and bringing forward larger health warnings on cigarette packages”, he adds.

Whether cuts aff ect cancer outcomes will be largely determined by their eff ect on provision of cancer screening and treatment, or on programmes that target health behaviours such as diet and smoking, Strumpf said to The Lancet Oncology. “The story told by the Canadian Cancer Statistics 2012 report is that population health behaviours may play an even larger role than health-care services in cancer mortality reductions. Therefore, any budget cuts that aff ect the role that Health Canada plays in educating Canadians regarding healthy diets and exercise, and working to reduce smoking rates, are likely to be more problematic”, Strumpf concludes.

Kathryn Senior

Scie

nce

Phot

o Li

brar

y

Cancer screening programmes: under threat?

Canadian cuts overshadow good news on cancer statisticsPublished Online

May 18, 2012 DOI:10.1016/S1470-

2045(12)70221-8

For Canadian Cancer Statistics 2012 see http://www.cancer.ca/Canada-wide/About%20cancer/

Cancer%20statistics.aspx?sc_lang=en

News


Chinese miners at increased risk of cancerA study led by researchers in China has confi rmed the association between inhalation of silica dust and a height ened risk of lung cancer. The study observed 74 040 people working at metal mines and pottery factories in the country. Average follow-up was 33 years. Mortality from all causes was higher in workers exposed to silica dust (993 vs 551 per 100 000 person-years in non-exposed workers). “We saw elevated mortality from lung cancer among exposed workers compared to non-exposed workers”, Weihong Chen (Huazhong University of Science and Technology, Hubei Province, China), co-author of the study, told The Lancet Oncology. He added that the association held true even after adjustment for cigarette smoking (many people in the cohort were smokers). Depending on level of exposure to silica dust, the risk ratio ranged from 1·45 to 1·53.

“It is a huge study—this kind of data is not available anywhere else”, noted Qingyi Wei (MD Anderson Cancer Centre, University of Texas, TX, USA). “It provides the basis of experimental studies in the future [into] how silica may cause lung cancer”. Classifi ed by the International Agency for Research on Cancer as a group 1 carcinogen, silica dust is thrown up whenever one drills into rock. It is associated with various diseases—silicosis, cardiovascular disease, and tuberculosis, for example. People who work in construction, mining, pottery, and stone crushing are vulnerable. In China, this group includes more than 23 million workers.

Occupational health expert Perry Gottesfeld (Occupational Knowledge International, CA, USA) believes that respirable silica is the developing world’s largest work-related health problem. The USA currently mandates a limit roughly similar to that of

China’s, but the US Occupational Safety and Health Administration has proposed for this to be further reduced. “It is a very out-of-date standard, which everyone agrees is not protective for lung cancer”, said Gottesfeld.

Moreover, Chinese safety standard enforcement is poor. Aside from strengthening such capacity, there are simple measures that have a proven ability to alleviate the problem, such as proper ventilation for under-ground mines and respirators for dust-exposed workers. But perhaps foremost is the use of equipment that sprays a fi ne mist of water. Gottesfeld co-authored a study showing that this approach can cut respirable silica by up to 82%. “These exposures can be reduced considerably at little cost; we’ve been able to demonstrate this in India, and it should all be doable in China”, he concluded.

Talha Khan Burki

Helping GPs to diagnose cancer earlierCancer Research UK and the Royal College of General Practitioners (RCGP) have launched an initiative to support general practioners (GPs) to improve early cancer diagnosis rates. The partnership will assemble best-practice models and guidance for GPs, and work with health commissioners on care pathways. Paul Baughan (Scottish Primary Care Cancer Group, Edinburgh, UK) welcomes the appointment of the partnership’s national GP clinical lead, Greg Rubin (Durham University, Durham, UK), an experienced GP and a Professor of General Practice and Primary Care. “Projects such as the National Awareness and Early Diagnosis Initiative have given us a much better understanding of why some people with cancer are diagnosed late and we need to act on these insights”, he comments. Baughan warns, however, that any strategy to detect cancer early

is unlikely to succeed without a clear understanding of general practice. “One of the fi rst objectives of this partnership should be to apply the knowledge around late diagnosis to the realities of general practice in the UK. This will help to ensure that the most appropriate initiatives are progressed fi rst”, he stresses.

“GPs have always accepted that they play an important role in cancer diagnosis, but that role is changing”, notes William Hamilton (Peninsula College of Medicine and Dentistry, Exeter, UK). Until fairly recently, the job of a GP was to spot warning signs and then refer patients on to a specialist for defi nitive diagnosis, as outlined in the National Institute for Health and Clinical Excellence (NICE) guidance issued in 2000 and 2005. GPs were able to undertake physical examinations for breast and testicular cancer and some

tests—eg, prostate-specifi c antigen test for prostate cancer, chest x-rays for lung cancer—but did not have access to the diagnostic armoury now available to a GP. “We can do MRI brain scans, transvaginal ultrasound, blood tests to check for CA-125, and have better access to colonoscopy, so our threshold for investigating the symptomatic patient is reducing”, reports Hamilton, who is clinical lead on an upcoming revision to the NICE guidelines. He also welcomes Rubin’s appointment: “The RCGP has produced a good toolkit for this, co-authored by Rubin, aimed at helping GPs to apply their knowledge, rather than instilling new knowledge. We need to continue providing support for GPs to (gently) lower their threshold for investigating cancer”, he recommends.

Kathryn Senior

Published OnlineApril 27, 2012 DOI:10.1016/S1470-2045(12)70180-8

For more on sentinel signs of disease see Leading Edge Lancet Oncol 2009; 10: 97

For more on variation in GP referrals for patients with cancer see Articles Lancet Oncol 2012; 13: 353–65

Published OnlineApril 27, 2012 DOI:10.1016/S1470-2045(12)70181-X

For more on the silica dust and mortality study see PLoS Med 2012; 9: e1001206.

For Gottesfeld and colleagues’ study see Int J Occup Environ Health 2008; 14: 94–103

John

Gre

im/S

cienc

e Ph

oto

Libr

ary

News


NICE guidance on cabazitaxel for hormone-refractory metastatic prostate cancer previously treated with a docetaxel-containing regimenOn May 11, 2012, the UK National Institute for Health and Clinical Excellence (NICE) published guidance on cabazitaxel in combination with prednisone or prednisolone for the treatment of hormone-refractory metastatic prostate cancer previously treated with a docetaxel-containing regimen.1 Cabazitaxel was not con-sidered a cost-eff ective use of National Health Service (NHS) resources.

Cabazitaxel was appraised under the Single Technology Appraisal process. The manufacturer (Sanofi ) submitted clinical evidence and a health economic model,2 which was critiqued by an independent Evidence Review Group (ERG), based at the School of Health and Related Research (University of Sheffi eld, Sheffi eld, UK).3 An independent Appraisal Committee consisting of NHS health professionals, lay members, industry representatives, health economists and statisticians met twice to develop guidance on cabazitaxel. Clinicians and patient experts attended the fi rst meeting, and the manufacturer attended both meetings.

The clinical evidence was based on one randomised controlled trial (TROPIC) in which patients were randomised to receive either cabazitaxel or mito-xantrone, in addition to prednisone or prednisolone.4 TROPIC enrolled 755 men from 26 countries. The published analysis reported a statistically signifi cant improvement in median overall survival with cabazitaxel compared with mitoxantrone (15·1 months vs 12·7 months; p<0·0001). Cabazitaxel was associated with a statistically signifi cant improvement in median progression-free survival (with progression defi ned as a rise in prostate-specifi c antigen concentration, tumour progression, pain progression, or death; 2·8 months

with cabazitaxel, 1·4 months with mitoxantrone; p<0·0001). The most common adverse events associated with cabazitaxel were neutropenia, its complications, and gastrointestinal toxicity.

The manufacturer submitted a cohort Markov model that included three health states: stable disease, progressive disease, and death. The perspective was that of the NHS and personal social services. For the base-case analysis, the manufacturer used survival data from a post-hoc subgroup of TROPIC consisting of European patients with an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1 who had previously received 225 mg/m2 or more of docetaxel. The manufacturer calculated transition probabilities from stable to progressive disease using progression-free survival, and from stable or progressive disease to death using data for overall survival, both estimated from Kaplan-Meier curves from TROPIC up to the point at which a small number of patients remained in the trial. The manufacturer defi ned this point as the time when no patient had an event over four consecutive cycles of treatment. After this time point, the manufacturer calculated transition probabilities from fi tted parametric curves.

Data for health-related quality of life were not collected in TROPIC. In the model, the manufacturer incorporated a utility value for stable disease from an interim analysis of an ongoing single-arm study collecting EQ-5D data from UK patients receiving cabazitaxel for metastatic prostate cancer. The utility of the progressive disease state was calculated by decreasing the utility value of stable disease by 0·07 (derived from Sullivan

and colleagues5). The manufacturer’s base-case model estimated an incremental cost-eff ectiveness ratio (ICER) of £74 900 per quality-adjusted life-year (QALY) gained.

The ERG considered that the manufacturer’s choice of the base-case population was not appropriate because there was no reason why results from patients recruited at TROPIC’s European centres would diff er from those recruited elsewhere. The ERG considered that the appropriate population for this appraisal should include all patients who received 225 mg/m2 or more of docetaxel and who had an ECOG performance score of 0 or 1. The ERG was not convinced by the manufacturer’s modelling of Kaplan–Meier data, because it was specifi c to the trial population and less generalisable to other populations. The ERG preferred parametric models, which assume an underlying trend and minimise the eff ect of chance observations that could occur in a specifi c clinical trial.

Using survival data specifi c to its preferred population, using parametric curves, and making other minor amendments, the ERG estimated an ICER of £89 476 per QALY gained. The key drivers of the cost-eff ectiveness analysis were the utility values used for stable and progressive disease.

The Committee had concerns about the manufacturer’s base-case population, the methods used for extrapolating survival data, and the utility value used for the stable disease state, which was similar to that observed in the age-matched general population and which the Committee agreed was implausible. The Committee concluded that a diff erence in utility of 0·07 between stable and progressive disease


News


underestimated the diff erence in quality of life because patients with progressive disease generally feel less well and have a worse quality of life than do those with stable disease.

The Committee considered the ICERs in the context of NICE’s supplementary advice for the appraisal of treatments that might give patients with a short life expectancy at least 3 more months of life, and for which the licensed indications aff ect a small proportion of patients.6 The Committee agreed that because people in trials of metastatic prostate cancer randomly assigned to best supportive care have a life expectancy of less than 15 months, and because fewer than 2000 people are estimated to receive second-line chemotherapy, two of the criteria were fulfi lled. In relation to the modelled gain in survival, the Committee noted that it was dependent on the method of extrapolation, which was uncertain. The Committee concluded at its fi rst meeting that, in view of this uncertainty, it could not consider all end-of-life criteria to be met.

On the basis of evidence reviewed at its fi rst meeting, in September, 2011, the Appraisal Committee concluded that cabazitaxel would not be a cost-eff ective use of NHS resources and issued an appraisal consultation document.

During the subsequent consultation period, the manufacturer presented a revised base-case model that included an updated utility value for the stable disease state from a second interim analysis of its ongoing single arm study and that, using the Committee’s preferred population, estimated an ICER of £86 008 per QALY gained. The manufacturer also used a number of diff erent mathematical approaches to extrapolate trial data to model the gain in overall survival, all of which resulted in an extension of life of 3 months or greater.

At its second meeting, in November, 2011, the Appraisal Com-mittee considered all responses to its preliminary guidance7 received from the manufacturer, clinical organisations, and patient groups, as well as additional ex ploratory analyses of overall survival data submitted by the manufacturer. The Committee agreed that a mean improvement of more than 3 months in overall survival had been shown robustly and that all end-of-life criteria were met.

The Committee concluded that of the manufacturer’s proposed methods of extrapolating survival data, the piecewise analysis was the most appropriate, which resulted in an ICER of £87 518 per QALY gained for the Committee’s preferred population. The Committee agreed that changes to the model that would likely further increase the ICER included: a lower utility value for stable disease, a larger diff erence in quality of life between stable and progressive disease, costs for hospital admission in all patients with febrile neutropenia, and fi tting of the curve for mitoxantrone using a piecewise approach.

The Committee agreed that cabazitaxel was an eff ective, life-extending treatment, but that the most plausible ICER exceeded £87 500 per QALY gained. Although the appraisal met the end-of-life criteria, the Committee considered that the additional weight needed to bring the ICER into the range regarded as a cost-eff ective use of NHS resources was too great, and that cabazitaxel could not be recommended as an appropriate use of NHS resources.

The manufacturer appealed against the fi nal appraisal determination.8 The appeal points related to the Committee’s understanding and interpretation of the utility values provided by the manufacturer. The Appeal Panel considered all appeal points at a hearing in March, 2012,

and concluded that the process by which the Committee involves manufacturers at the meetings was followed correctly, and that the Appraisal Committee had not failed to understand the nature of interim data. The Appeal Panel concluded that the recommendations were not aff ected by a diff erence of opinion between the Appraisal Committee and manufacturer and dismissed the appeal on all grounds.

Anwar Jilani, Elisabeth George, Amanda I AdlerNational Institute for Health and Clinical Excellence, MidCity Place, London, UK


1 NICE technology appraisal guidance TA255. Cabazitaxel for hormone-refractory metastatic prostate cancer previously treated with a docetaxel-containing regimen. May, 2012. http://guidance.nice.org.uk/TA255 (accessed May 8, 2012)

2 Manufacturer submissions: cabazitaxel for the second-line treatment of metastatic hormone refractory prostate cancer. June, 2011. http://www.nice.org.uk/nicemedia/live/13237/56549/56549.pdf (accessed Jan 26, 2012).

3 Stevenson M, Lloyd Jones M, Kearns B, Littlewood C, Wong R. Cabazitaxel for the second-line treatment of hormone refractory, metastatic prostate cancer: a single technology appraisal. ScHARR, University of Sheffi eld, 2011. http://guidance.nice.org.uk/TA/Wave23/31/Consultation/EvaluationReport/EvidenceReviewGroupReport (accessed Jan 26, 2012).

4 de Bono JS, Oudard S, Ozguroglu M, et al. Prednisone plus cabazitaxel or mitoxantrone for metastatic castration-resistant prostate cancer progressing after docetaxel treatment: a randomised open-label trial. Lancet 2010; 376: 1147–54.

5 Sullivan PW, Mulani PM, Fishman M, et al. Quality of life fi ndings from a multicenter, multinational, observational study of patients with metastatic hormone-refractory prostate cancer. Qual Life Res 2007; 16: 571–75.

6 Appraising life-extending, end of life treatments. NICE, July, 2009. http://www.nice.org.uk/media/E4A/79/SupplementaryAdviceTACEoL.pdf (accessed Jan 27, 2012).

7 Appraisal consultation document: prostate cancer—cabazitaxel. September, 2011. http://guidance.nice.org.uk/TA/Wave23/31/Consultation/DraftGuidance (accessed Jan 27, 2012).

8 Final appraisal determination: prostate cancer—cabazitaxel. January, 2012. http://www.nice.org.uk/nicemedia/live/13237/57803/57803.pdf (accessed Jan 27, 2012).

News


New treatment options in advanced bladder cancer In a multicentre phase 3 trial of 360 patients with muscle-invasive bladder cancer, British researchers compared radiation therapy alone against radiation and synchronous chemotherapy with fl uorouracil and mitomycin C and found that patients given the combination lived longer than did those on radiation alone. Moreover, patients assigned to combination therapy lived as long as did historical controls who had cystectomy, the gold standard. At 2 years, rates of locoregional disease-free survival were 67% for combination therapy and 54% for patients receiving radiation alone, and at 5 years, overall survival was 48% versus 35%, respectively.

“These results indicate that radiation and chemotherapy with non-platinum drugs produces similar outcomes as radical cystectomy, but without the negative impact on quality of life, and with much less

toxicity than with platinum-based drugs”, said lead author Nicholas James (University of Birmingham, Birmingham, UK).

There were more grade 3 and 4 adverse events in the chemoradiation group (36%) than in the radiation cohort (27%), with gastrointestinal events signifi cantly more common with chemoradiation therapy. However, the combination was less toxic than platinum, which many older patients (who account for the majority of this patient population) cannot tolerate because of its eff ect on the kidneys. These two drugs were chosen because of their fairly low toxicity profi le, extensive data for their use, and their successful use in other cancers such as anal cancer, James said.

Anthony Zietman (Harvard Medical School, Boston, MA, USA), said that this study was so signifi cant that it could potentially change how

advanced bladder cancer is treated. Because of the safety and effi cacy of the combination regimen, radical surgery was ultimately needed in only a minority of patients, making this trial “a very powerful piece of evidence, a proof of principle that this treatment is an excellent alternative to surgery”, Zietman said.

“The combination of two chemotherapeutics worked as radio sensitisers”, said James. His group’s next trial, a phase 1/2 with 40–50 patients, is about to begin, and will add the EGFR-targeting drug cetuximab to radi-ation and chemotherapy. “EGFR is overexpressed in bladder cancer, and works diff erently than chemo-therapy, so we expect it could radiosensitise this cancer in a manner complementary to chemotherapy”, said James.

Vicki Brower

BRCA1 mutations increase risk of prostate cancerThe causes of prostate cancer are little understood, but the observation that breast and prostate cancers cluster within families led to the hypothesis that BRCA1 mutations might confer susceptibility to both diseases. A new study, which sequenced the BRCA1 gene from patients with prostate cancer in the UK, has found that men with deleterious mutations in the gene are almost four times as likely to have the disease compared with the general population—equivalent to an 8·6% cumulative risk by the age of 65 years.

Zsofi a Kote-Jarai (Institute of Cancer Research, London, UK), one of the study’s co-authors, said that while previous research had provided confl icting evidence on the link, this study is “the largest analysis to date of the contribution of BRCA1 mutations to prostate cancer”, adding that their results “confi rmed that although

carriers of germline BRCA1 mutations represent a modest proportion of overall prostate cancer cases, these individuals are at moderately increased risk of developing prostate cancer”.

The investigators sequenced BRCA1 in 886 patients recruited from the UK Genetic Prostate Cancer Study, with the selection enriched for cases with early age of onset. They identifi ed four deleterious mutations, three of which were in patients younger than 65 years at onset. 460 cases were also screened for large-scale rearrangements, al-though none were found in the sample.

Although Kote-Jarai believes that it is not yet feasible to advocate routine BRCA1 screening for patients with early onset prostate cancer, their fi ndings “may have implications for the monitoring and management of male BRCA1 carriers identifi ed through familial breast cancer studies”.

Stephen Chanock (National Cancer Institute, Gaithersburg, MD, USA) told The Lancet Oncology, “not only can this lead to new avenues for understanding some of the key processes that lead to prostate cancer, but it also raises important questions about BRCA1 testing in male family members…BRCA1 testing may be useful in detection of early disease in family members, particularly female relatives of newly diagnosed younger men with prostate cancer and BRCA1 mutations”.

“Like the 8q24 region, it is another example of a genetic region associated with both breast and prostate cancer, which may shed light on common mechanisms that could be exploited in subsequent research and screening”, Chanock added.

Neil Bennet


For the study see Br J Cancer 2012; published online April 19. DOI:10.1038/bjc.2012.146


For the phase 3 study on combination therapy for advanced bladder cancer see New Engl J Med 2012; 366: 1477–88

Davi

d M

ack/

Scie

nce

Phot

o Li

brar

y

News


Published OnlineMay 4, 2012 DOI:10.1016/S1470-2045(12)70201-2

For more on the new imaging technique see Am J Pathol 2012; 180: 1835–42

For more on virtual pathology at the University of Leeds see http://www.virtualpathology.leeds.ac.uk/research/3D/3d_vs.php


For more on the Cancer Drugs Fund see http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_133682

For more on Pharmafocus’ analysis see http://www.inpharm.com/news/172329/government-speed-access-new-cancer-drugs

Faster access to Cancer Drugs Fund Access to England’s Cancer Drugs Fund will be sped up to ensure that patients are off ered drugs within days of applying, the Department of Health has announced.

Introduced by the coalition govern-ment in 2010, the Fund aims to improve patient access to cancer drugs before the expected reform of arrangements for branded drug pricing. The current Pharmaceutical Price Regulation Scheme expires at the end of 2013. However, some doctors and charities have raised concerns that it can take several weeks for patients to access cancer drugs through the Fund.

Pharmafocus magazine recently reported an underspend of around £60 million on the Fund last year by the Strategic Health Authorities (SHAs) accountable for the Fund.

New guidance to the UK National Health Service (NHS) will speed up applications for cancer drugs, which will no longer need to go through the primary care trust (PCT) funding processes. Each SHA should ensure there is a transparent, published process for allocating the funding, and “more challenging timescales” for decision making should be adopted. The present 31-day treatment standard should be seen as an absolute maximum.

The guidance came into force on April 23, 2012, and does not apply retrospectively. It will be kept under review and updated as appropri ate during 2012–13. An audit of the Fund will also be undertaken to provide evidence from drugs funded through the Fund for the benefi t of wider NHS practice.

Andrew Wilson, from the Rarer Cancers Foundation (Canterbury, UK), said, “We welcome the measures aimed at speeding up applications for drugs by bypassing PCT funding processes prior to applying to the Fund. This will help patients to get the fastest possible access to the drugs recommended by their doctors”.

Heather Walker from Cancer Re search UK (London, UK) agreed. “Cancer Research UK welcomes anything that helps patients get the treatments they need faster. But it’s still unclear what will happen to the Fund when we move to value-based pricing in 2014. We would like to see detailed plans from the government to reassure us that cancer patients won’t be left in limbo”, Walker said.

Sanjay Tanday

3D histopathology could help cancer research and treatmentResearchers have developed a new technique for generating high-resolution three-dimensional (3D) representations of tissue samples, with the potential to aid the study of tumour growth and spread.

For the new imaging technique, tissues are sectioned with a micro tome to generate hundreds of slides, which are stained and loaded into digital scanners. Novel 3D histopathological software uses automated virtual slide scanners to reconstruct a digital version of the tissue sample at a cellular resolution level.

Derek Magee (University of Leeds, Leeds, UK), one of the study’s co-authors, told The Lancet Oncology that “the software would be applicable in clinical or biomedical applications where 3D structure is important. For example, determining vascular structure and how close a tumour is to a blood vessel”.

The researchers used the technique to investigate various tissue samples

from both human beings and animals. Among the examples featured in their report were sections from a human liver with a deposit of metastatic colorectal carcinoma, and a cirrhotic liver infected with hepatitis C virus. Sample preparation and staining were labour intensive—taking be-tween 5 and 7·7 h for the published examples—however, the authors suggest that these processes can be automated in the future, minimising the manual input needed.

Peter Morris (University of Notting-ham, Nottingham, UK) said the technique “enhances histopathology as a research tool and has the potential to signifi cantly improve diagnostic procedures”.

“Standard histopathology is based on the examination of two-dimensional (2D) tissue sections from which it is often hard to discern structural relationships—for example, the presence of vessels in multiple sclerosis lesions or tumours—on

which therapeutic strategies might be based. [This paper] describes how standard 2D sections, gener-ated in any histopathology lab, can be automatically co-registered and displayed as 3D images”, he explained.

Commenting on the report, David Hawkes (University College London, London, UK) said, “[The researchers describe] a new method to align, analyse and display complete histology sections of tissue samples in 3D. While many have attempted to do this as a research task, this paper shows a complete pipeline that can be implemented for regular use”.

“The big advantage of this method is that it utilises conventional slicing and staining and so can take advantage of the wealth of information available from these very well developed methods”, Hawkes added.

Neil Bennet

Phil

Boor

man

/Scie

nce

Phot

o Li

brar

y

News



For the RECORD-1 subgroup analysis see Br J Cancer 2012; 106: 1475–80


For more on the Childhood Cancer 2012 conference see http://www.childhoodcancer2012.org.uk/

Second-line everolimus treatment in advanced renal-cell cancerPatients with metastatic renal-cell cancer who discontinued treatment with vascular endothelial growth factor receptor-tyrosine kinase inhi bitors (VEGFR TKIs) because of toxicities were able to tolerate and benefi ted from second-line treat ment with everolimus with no additi onal toxicities, according to a retro spective subgroup analysis of the phase 3 RECORD-1 trial. Risk of disease progression was reduced by 67% for patients taking everolimus, an mTOR inhibitor, compared with those on placebo.

In 58 VEGFR-TKI-intolerant pati-ents (14% of the total number), median progression-free survival was 5·4 months with everolimus and 1·9 months with placebo. For 26 patients who were intolerant to the multitargeted receptor tyrosine kinase sunitinib, median progression-free survival was 5·1 months with everolimus and

2·8 months for placebo. Grade 3 or 4 toxicities seen in everolimus-treated patients were infections, fatigue, and stomatitis, which were similar to those seen in the larger patient pool. Toxicities associated with VEGFR-TKI treatment include hypertension, hand-foot reaction, rash or des-quamation, alopecia, diarrhoea, fatigue, hyponatraemia, and throm-bo cytopenia, and frequently neces-sitate dose interruptions, reductions, or discontinuation of treatment.

“The clinical challenge in metastatic renal-cell cancer is that many patients are intolerant to fi rst-line therapy, and patients need to be treated numerous times for this disease. We found that everolimus, which has a diff erent molecular target and mechanism of action than the VEGFR TKIs, was as eff ective in both patients who were intolerant to VEGFR TKIs and in the larger group”, said study researcher

Robert Figlin (Cedars-Sinai Samuel Oschin Comprehensive Cancer Institute, Los Angeles, CA, USA). The safety profi le of mTOR inhibitors does not generally overlap with that of VEGFR TKIs, and cardiovascular toxicities and hand-foot reactions are not usually seen in everolimus-treated patients, Figlin added.

“The two drug classes may in fact be targeting the same pathway, and may have a similar mechanism of action, which may be why the outcomes in terms of progression-free survival were so similar”, said David Nanus (Weill Cornell Medical College, New York, NY, USA). “But the most important lesson from this study is that everolimus has less toxicity and similar benefi t to VEGFR TKIs in this patient population, who need to be treated repeatedly”, Nanus said.

Vicki Brower

Zeph

yr/S

cienc

e Ph

oto

Libr

ary

Fertility drugs and childhood leukaemia: is there a link?A new study by scientists in France has suggested that fertility drugs could be linked with a heightened risk of childhood leukaemia. The widely publicised study—which involved 2445 children and their mothers—concluded that children conceived after the use of ovarian-stimulating drugs were 2·6 times more likely to develop acute lymphoblastic leukaemia (ALL) and 2·3 times more likely to develop acute myeloid leukaemia.

Presenting the fi ndings at the Childhood Cancer 2012 conference in London, UK, lead author Jeremie Rudant (Center for Research in Epidemiology and Population Health, INSERM, Villejuif, France) noted that this “is the fi rst time that a specifi c increased risk linked to fertility drugs has been found”. Children who were conceived naturally after their mothers had failed to become pregnant for more than a year were found to be

at 50% increased risk of ALL. No heightened risk of leukaemia was associated with in-vitro fertilisation (IVF) or artifi cial insemination.

Simon Fishel (CARE Fertility Group, Nottingham, UK) is sceptical. He points out that patients undergoing IVF also take fertility drugs. “They’re probably more potently used in IVF—for higher doses and for longer periods than you may have if you’re not on IVF”. In which case, how is it possible to posit a link between fertility drugs and leukaemia in one group of patients, but not in another? “It is a major shortcoming and it is very diffi cult to explain it”, Fishel said.

Moreover, the French study is not backed by a publication. “We have to be very circumspect because this hasn’t gone through the peer-review process”, cautions Richard McNally (Newcastle University, Newcastle, UK). Without scrutiny from fellow experts, questions

remain over the robustness of the methods and possible bias in selection of the control group. “The fi ndings would also have to be replicated in a study of greater or similar size”, added McNally. Until all this happens, he does not believe it is possible to draw any conclusions from the research.

The Human Fertilisation and Embryology Authority (HFEA), which regulates the UK’s assisted re pro duc-tion industry, has given permission for researchers at University College London to link its data on treatment cycles to a cancer registry in order to investigate the association between such treatment and site-specifi c risks of cancer incidence. “We ought to be getting more data on this area of research over the next year or so”, the HFEA’s Juliet Tizzard told The Lancet Oncology.

Talha Khan Burki

News



For the new study on panobinostat see J Clin Oncol 2012; published online April 30. DOI:10.1200/JCO.2011.38.1350


For the study on brachytherapy see JAMA 2012; 307: 1827–37.

A new drug for Hodgkin’s lymphomaRelapsed or refractory Hodgkin’s lymphoma could benefi t from panobinostat, according to a new study. Panobinostat is a potent histone deacetylase (HDAC) inhibitor with activity against class I, II, and IV histone deacetylases, which regulate several oncogenic pathways.

About a third of patients with Hodgkin’s lymphoma relapse on standard fi rst-line chemotherapy. Relapsed patients might benefi t from autologous stem-cell transplantation, but again about 50% subsequently have progressive disease. For such relapsed or refractory disease, therapeutic options are still limited.

Anas Younes (MD Anderson Cancer Center, Houston, TX, USA) and colleagues recruited 129 patients with heavily pretreated refractory or relapsed Hodgkin’s lymphoma and gave them panobinostat 40 mg orally three times a week, every week,

in a 3-week cycle. 96 (74%) patients had tumour reductions. Objective response was achieved in 35 (27%) patients, with a median time to response of 2·3 months. Similarly, median duration of response was 6·9 months and median progression-free survival (PFS) was 6·1 months. Prolonged PFS was seen even beyond patients who had objective response. Commenting about the most important fi nding of the study, Younes said, “The fact that the majority of patients had tumour reductions and knowing that HDAC inhibitors exert their antiproliferative activity through multiple mech anisms, this study strongly suggests that HDAC inhibitors can be further developed in combination with other active agents for Hodgkin’s lymphoma”.

Thrombocytopenia was the most common serious adverse event related to panobinostat. TARC (thymus and

activation-regulated chemokine), a chemokine expressed by Hodgkin’s cells and that contributes to “B” symptoms, anaemia and immuno suppression, was reduced early in patients achieving complete or partial response.

Bruce A Chabner (Massachusetts General Hospital, Boston, MA, USA) said, “Panobinostat has shown an impressive result for a heavily treated patient population and it clearly points towards the introduction of this active drug in earlier treatment regimens”.

“It will undoubtedly move into trials for patients failing initial therapy, prior transplant, but ultimately it should be tested in combination with brentuximab (another new Hodgkin’s lymphoma drug) and cytotoxic drugs for fi rst-line therapies”, Chabner added.

Sharan Prakash Sharma

Brachytherapy versus total breast radiationIn a retrospective population-based study of 92 735 American women aged 67 years or older with early-stage breast cancer, brachytherapy following lumpectomy was associ-ated with more compli cations, including subsequent mastectomy and infection, than was total breast radiation. At 5 years, 4·0% of women treated with brachytherapy needed mastectomy versus 2·2% with ra-diation, 16·2% had infectious com-plications versus 10·3%, and 4·5% versus 3·6% had rib fractures, but survival was the same in both groups.

Lead author Benjamin Smith (MD Anderson Cancer Center, Houston, TX, USA) noted that use of brachytherapy has increased from 3·5% of patients in 2003, to 13% in 2007, with at least 50 000 women treated to date in the USA. “One reason for the rapid increase, for patients, is that it lasts 1 week,

whereas total breast radiation is delivered over 5–7 weeks”, said Smith. On the other hand, brachytherapy often costs substantially more than total breast radiation, as it involves a surgeon, he pointed out.

“One glaring problem with this study is that it does not specify what the reason is for the increase in mastectomies”, said Mary Katherine Hayes (New York-Presbyterian/Weill Cornell Medical Center, New York, NY, USA). “Was it due to recurrence, or infection?” There is also no data in the study about the size of tumours, node status, or other features that would make brachytherapy a poor choice for certain patients, “all of which muddies the waters”, Hayes added. “Two sets of guidelines also recommend that younger patients, aged 45–50 years, may receive brachytherapy, not older women, aged 60 years and older”, she said.

“Brachytherapy is not for everyone”, Hayes said. “Physicians must be very careful about patient selection.”

The complication rates seen in both treatment groups were also higher than are generally seen, and for brachytherapy in particular this fi nding might be attributable to the fact that the years in which the study was done represent early use of the technique, Hayes observed. Smith and Hayes both agreed that there is a need for a randomised clinical trial to compare both treatments (such a trial is underway). “In the meantime, there are a variety of opinions as to whether brachytherapy should be used outside a clinical trial in routine practice”, Smith said. Until results are in, patients should be appraised of the risks and benefi ts of both treatments, he added.

Vicki Brower

ISM

/Scie

nce

Phot

o Li

brar

y

News


Low-dose radioiodine ablation for thyroid cancerThe results of two randomised, prospective studies from the UK and France have shown that a low dose of radioiodine (1·1 GBq) is as eff ective as the standard high dose (3·7 GBq) for postoperative ablation after thyroidectomy in the treatment of low-risk thyroid cancer.

Ujjal Mallick (Freeman Hospital, Newcastle upon Tyne, UK), the lead investigator of the UK study, said that “it should now be recommended that suitable patients with thyroid cancer have low-dose radioactive iodine with recombinant human thyrotropin”.

“Importantly, this reduces the chance of developing another cancer later in life and other side-eff ects of treatment. Patients have much better quality of life and the overall societal costs due to sickness absence will be reduced”, Mallick added. The low-dose regimen also means that people can be treated as outpatients, rather than kept in hospital isolation

units for days after receiving radioiodine.

The two trials also evaluated the use of recombinant thyrotropin to stimulate the uptake of radioiodine compared with increasing endogenous thyrotropin through thyroid hormone withdrawal. Although both studies emphasise the benefi ts of using recombinant thyrotropin (such as reduced incidence of adverse events), the authors of an accompanying commentary are careful to point out that both were non-inferiority trials and were not placebo-controlled—so the evidence is inconclusive.

Commenting on the research, Keith Bible (Mayo Clinic, Rochester, MN, USA) told The Lancet Oncology that the two studies “call attention to the fact that most patients with typical low-risk diff erentiated thyroid cancers have excellent overall prognosis, and to the consequent concern that there is risk of overtreatment to their

potential detriment while in parallel imposing signifi cant and potentially unjustifi ed health-care costs”.

“Although the follow-up periods of both studies are, unfortunately, too brief to fully address the posed important questions, they represent important steps in better tailoring therapy to risks and benefi ts in patients recently diagnosed with diff erentiated thyroid cancers”, he continued.

Martin Schlumberger (University of Paris-Sud, Paris, France), lead author of the French study, agreed, saying that the “benefi ts of post-surgical radioiodine administration are still not demonstrated in low-risk patients”, adding that “the next step will be to randomly assign low-risk patients to receive either 1·1 GBq radioiodine following recombinant thyrotropin or no radioiodine at all and monitor them for recurrence”.

Neil Bennet

HPV vaccine dropoff fearThe take up of full human papil-lomavirus (HPV) vaccination among young girls and women in the USA is in steep decline, a study suggests.

Records covering almost 272 000 insured women and girls aged 9 years and older show that the proportion receiving all three injections of the quadrivalent vaccine course has dropped by as much as 63% since the vaccine was approved in 2006.

The steepest decline was noted in girls aged 9–18 years, suggesting a missed opportunity for many in the fi rst generation to benefi t from immunisation, say the researchers.

By 2009, only 21·2% of girls aged 9–12 years who had started the vaccine course had gone on to receive all three doses, compared with 57·5% in 2006. In girls aged 13–18-years, take up of the full course fell from 54·9% to 20·8%.

Study author Jacqueline Hirth (University of Texas Medical Branch, Galveston, TX, USA) pointed out that vaccine initiation increased from 23% to 49%, suggesting that health-care providers were promoting HPV vaccination, but were failing to follow up on second and third doses.

“Our special concern is that the target group of 9-to-18-year olds is where we see the greatest decline.” “Physicians really need to focus on making sure that group complete vaccination”, she said.

The study also showed that paediatricians administered around half of fi rst vaccine doses. Carrie Byington (University of Utah School of Medicine, Salt Lake City, UT, USA) agreed that paediatricians had a vital role here. She also highlighted that the economic recession had led to people delaying preventive care.

“A public health campaign focusing on series completion should be a priority”, she added.

Noel Brewer (University of North Carolina Gillings School of Global Public Health, Chapel Hill, NC, USA) said, “What is more concerning is the low initiation rates seen in the USA when countries like the UK have had tremendous success with school-based programmes.”

What is needed now, says Don Dizon (Brown University, Providence, RI, USA), is a better understanding of why more young girls are taking up the HPV vaccine but not fi nishing the course.

“Providers need to make com-pletion rates for all vaccine series an indicator of quality health care”, Dizon said.

Emma Wilkinson

Published OnlineMay 11, 2012 DOI:10.1016/S1470-2045(12)70219-X

For the UK study see N Engl J Med 2012; 366: 1674–85.

For the French study see N Engl J Med 2012; 366: 1663–73.

For the accompanying commentary see N Engl J Med 2012; 366: 1732–33.


For the study on HPV vaccination see Cancer 2012; published online April 27. DOI:10.1002/cncr.27598

Tim

Ver

non,

Lth

Nhs

Trus

t/Sc

ienc

e Ph

oto

Libr

ary

News


Lenalidomide maintenance for multiple myelomaThree phase 3 randomised controlled trials have shown that patients with multiple myeloma receiving lenalidomide as maintenance therapy had signifi cantly increased progression-free survival (PFS) and, in one study, improved overall survival. In two post-transplant trials, lenalidomide produced median PFS of 39 and 41 months versus 21 and 23 months for placebo, and in a third study in older, transplant-ineligible patients, a PFS of 31 versus 13 months for placebo. “These studies clearly demonstrate that adding lenalidomide as maintenance gives patients another 6–10 months of remission”, said Antonio Palumbo (University of Turin, Turin, Italy), lead investigator of the non-transplant study.

“Taken together, these trials will be practice-changing”, said Saad Usmani (University of Arkansas, Little Rock, AR, USA). “They clearly show that the longer you keep patients on lenalidomide maintenance, regardless of transplant eligibility, the better.”

However, the trials, which were all unmasked early, also showed an increased incidence of second primary cancers. The US Food and Drug Administration issued a warning about increased risk on the basis of these and two other trials, recommending that patients be monitored and risks and benefi ts considered before prescribing.

Researchers’ reaction to the studies ranged from cautious optimism about changing standard of practice to unfettered enthusiasm. But most agreed that they are an important step toward increasing lifespan and improving quality of life for myeloma patients, who in the past survived only a few years after diagnosis.

“We now know that lenalidomide is active at low doses as a single treatment, with an acceptable toxicity profi le”, said Philip McCarthy (Roswell Park Cancer Institute, Buff alo, NY, USA), lead investigator of one post-transplant trial. “We will

have to see with longer follow-up if longer PFS leads to longer overall survival”, he added. However, Ashraf Badros (University of Maryland, Baltimore, MD, USA) and others question the use of PFS as a primary endpoint in maintenance trials, and prefer to see increased overall survival in view of toxicities and high cost before changing standard of care. Paul Richardson (Jerome Lipper Center for Multiple Myeloma, Dana-Farber Cancer Institute, Boston, MA, USA), an investigator on the McCarthy study, disagrees. “With the advent of novel therapies, PFS has rarely failed to show overall survival benefi t with a consistent correlation seen in the relapsed setting, and now in newly diagnosed patients as well”, he said. “In an otherwise incurable disease like myeloma, the use of PFS remains a very important endpoint”, he added.

In McCarthy’s trial, 231 pati ents aged 18–70 years with stable disease received lenalidomide. At 18 months, interim analysis showed that 20% of treated patients versus 44% on placebo had either progressed or died. The trial was unmasked and patients allowed to cross over. At 34 months’ follow-up, time to progression was 46 versus 27 months, and 15% versus 23% of patients had died. More new cancers were diagnosed in patients receiving lenalidomide compared with placebo. Concerning increased second cancers—acute mye-loid leukaemia or myelodysplastic syndrome—McCarthy noted that myeloma patients have increased risk before treatment and slightly higher risk after treatment with alkylating agents, but patients with monoclonal gammopathy of undetermined signi-fi cance, another plasma-cell disorder, who typically receive no treatment, also have increased incidence of these malignancies. “This suggests that these malignancies may be related to microenvironment or stem-cell

abnormalities, not only treatment”, he said.

In a second post-transplant study by Michel Attal (Hôpital Purpan, Toulouse, France), 307 patients younger than 65 years received lenalidomide for two 28-day cycles and then lower dose maintenance, and 307 received placebo. At 30 months, median PFS was 41 versus 23 months. Overall survival at 3 years was similar in both groups: 80% and 84%. 32 second cancers occurred in lenalidomide-treated patients and 12 in the placebo group. “We stopped the study at 2 years because of these cancers”, said Hervé Avet-Loiseau (University of Nantes, Nantes, France). Adverse events, as in other trials, were more common in treated patients, including thromboembolic and hae ma tological problems. “It’s too early to tell if the high cost and toxicities justify long-term use of this drug without increased overall survival”, said Avet-Loiseau.

In Palumbo’s trial, three regimens in older patients not eligible for transplant were compared: induction with melphalan, prednisone, and lenalidomide, followed by main-tenance lenalidomide; mel phalan, prednisone, and lenalidomide, fol-lowed by placebo; and melphalan and prednisone, followed by placebo. PFS in these groups was 31, 14, and 13 months, respectively. “The benefi t was seen more in patients 65–75 years, not older”, Palumbo said. Second tumour incidence at 3 years was 7%, 7%, and 3%, respectively. “These data demonstrate that the risk of second cancers is highest within the fi rst 2 years, and that maintenance should be continued for at least 2 years or until progression”, Richardson said.

To establish whether PFS or overall survival in myeloma is the most relevant endpoint, “we will need to give these studies more time to mature”, said Usmani.

Vicki Brower


For McCarthy and colleagues’ trial of post-transplant lenalidomide maintenance in myeloma see New Engl J Med 2012; published online May 10. DOI:10.1056/NEJMoa1114083

For Attal and colleagues’ trial of post-transplant lenalidomide maintenance in myeloma see New Engl J Med 2012; published online May 10. DOI:10.1056/NEJMoa1114138

For Palumbo and colleagues’ trial of lenalidomide in transplant-ineligible patients with myeloma see New Engl J Med 2012; published online May 10. DOI:10.1056/NEJMoa1112704

For more on secondary cancer risk with lenalidomide see News Lancet Oncol 2012; 13: e54

Tim

Eva

ns/S

cienc

e Ph

oto

Libr

ary

News


mTOR inhibition in hormone-resistant breast cancerAddition of everolimus to tamoxifen in aromatase inhibitor (AI)-resistant metastatic breast cancer could reverse hormone resistance and lead to increased clinical benefi t rate and time to progression, according to a new study. About 30% of oestrogen-receptor-positive metastatic breast cancers become resistant to standard hormone therapy, which is thought to be due to cross-talk between signal transduction pathways including the mammalian target of rapamycin (mTOR) pathway.

Researchers randomly assigned patients with AI-resistant metastatic breast cancer to receive everolimus, an oral inhibitor of mTOR pathway, plus tamoxifen (54 patients) or to tamoxifen alone (57 patients). Clinical benefi t rate at 6 months was 61% in the combination group versus 42% in the tamoxifen alone group. The combination group had increased time

to progression compared with the tamoxifen alone group (8·6 months vs 4·5 months). This increase in time to progression corresponded to a 46% reduction in risk of progression in the combination group.

In an exploratory subgroup analysis, the everolimus benefi t was mostly seen in patients with secondary hormone resistance, defi ned as patients relapsing more than 6 months after stopping adjuvant AIs or responding for more than 6 months to AIs in the metastatic setting. Stephen Johnston (Royal Marsden NHS Foundation Trust, Institute of Cancer Research, London, UK) comments, “The most important aspect of the trial was the stratifi cation into primary or secondary resistance, showing that the main benefi t was in those with acquired AI resistance rather than those with de-novo resistance”.

Observed adverse events in patients on combination therapy were mostly of grade 1 and 2 severity, and the overall incidence of grade 3 and 4 adverse events was similar in both groups.

Lead author Thomas Bachelot (Centre de Recherche en Cancérologie de Lyon, Lyon, France) said their study results are consistent with a recent study (BOLERO-2) that evaluated the addition of everolimus to exemestane in a similar population.

“The similarity of the results of these two trials, with the use of two diff erent hormone therapies (tamoxifen for our study and exemestane for BOLERO-2), strong-ly favours mTOR inhibition as an import ant therapy in acquired hormone resistance in oestrogen-receptor-positive metastatic breast cancer patients”, Bachelot added.

Sharan Prakash Sharma

Avoiding stem cells could spare salivary gland functionThe location of stem cells important in parotid gland recovery after radiotherapy for head and neck cancer was described by Peter van Luijk (University Medical Center Groningen, Netherlands) and colleagues at the 31st conference of the European Society for Radiotherapy and Oncology (Barcelona, Spain, May 9–13, 2012). “The results suggest that recovery of saliva fl ow may be mediated by stem-cell repopulation of the gland. Avoidance of the stem-cell area during targeted radiotherapy may allow better recovery of saliva production after treatment”, observes Christopher Nutting (Royal Marsden Hospital, London, UK).

Damage to the salivary glands during radiotherapy often leads to xerostomia (dry mouth). “This debilitating late side-eff ect aff ects many aspects of the patient’s life, but now we have an insight into the anatomical regions in which the

damage actually occurs”, explains Silke Tribius (University of Hamburg, Germany). A series of experiments using mouse models and human tissue obtained during surgery pinpointed the exact location of the stem-cell component in the major ducts in the parotid gland. “The researchers subsequently devised a mathematical model based on the treatment of 36 patients, confi rming that limiting the radiotherapy dose to major ducts reduced the incidence of parotid dysfunction”, notes Allen Chen (University of California, Davis, CA, USA). “Combining this novel anatomic information with careful radiotherapy planning could signifi cantly reduce long-term xerostomia without compromising, and perhaps even increasing tumour control”, comments David Sher (Rush University Medical Center, Chicago, IL, USA). “This sort of improvement

in therapeutic ratio is the holy grail of modern radiotherapy planning”, he told The Lancet Oncology.

Chen cautions that “further prospective studies are needed to confi rm these provocative fi ndings”. Tribius agrees, adding that “translating results from ‘bench to bedside’ is always challenging and requires confi rmation by prospective controlled trials”. Sher hopes such trials will be pursued: “With younger and healthier patients surviving head and neck cancer, it is vital to research treatments that could improve long-term quality-of-life”. The study could also have implications beyond changing clinical practice in radiotherapy. “It also suggests the possibility of stem-cell treatments for patients with existing dry mouth following previous radiotherapy”, concludes Nutting.

Kathryn Senior


For more on mTOR inhibition in metastatic breast cancer see J Clin Oncol 2012; published online May 7. DOI:10.1200/JCO.2011.39.0708

For more on the BOLERO-2 study see News Lancet Oncol 2012; 13: e10


For more on stem-cell sparing radiotherapy for head and neck cancer see http://www.estro.org/Documents/Press_releases_ESTRO31/van%20Luijk_for%20websites.pdf

Ames

Ste

vens

on/S

cienc

e Ph

oto

Libr

ary


Cancer and Society


Here comes good health! Wellcome Collection, London, UK. Showing until June 3, 2012. http://www.wellcomecollection.org/whats-on/exhibitions/here-comes-good-health.aspx

For more on Change4Life see www.nhs.uk/Change4Life/

For more on Be Clear on Cancer see www.dh.gov.uk/health/tag/be-clear-on-cancer/

Commercial companies and marketing agencies are well acquainted with the power of mass media to shape attitudes and drive behaviours. Modern advertising cam paigns can have a powerful infl uence on the consumer, using an array of psychological nudges to suggest, cajole, and manipulate. Governments and other civic bodies have often sought to harness this power to guide health-related behaviours—to prevent disease and prolong and improve quality of life. Participation in vaccination, cancer screening, and tobacco cessation programmes has been encouraged by media campaigns for years.

In the 1920s and 1930s, residents of Bermondsey in southeast London were recipients of an innovative approach to public health messaging, instigated by the husband-and-wife team of Alfred and Ada Salter, who were Member of Parliament and Mayor of Bermondsey, respectively. The Salters combined modern media and advertising tech-niques to provide their council with information about health practices, thereby improving wellbeing in a time and area where poor health reigned.

The eff orts of this Better than Cure manifesto, inspired by the Salters and realised by Bermondsey’s public health department, can be seen until June 3 at the Wellcome Collection. The exhibition Here comes good health! includes screenings of some of the more than 30 educational public health fi lms that were made and exhibited in Bermondsey and surrounding bor-oughs from 1927 to 1938. Pertinent health messages, regarding issues such as immunisation and cleanliness, were creatively conveyed in dramatic, lecture, or documentary format. The black-and-white shots were screened in public spaces via customised cinemotor vans, using electricity diverted from nearby lampposts; this was an unprecedented approach to targeted health promotion.

Then, as now, content and format were key to a successful health campaign, and these fi lms skillfully use emotional connection to get the message across. Health and clothing is a light-hearted but sincere appeal for Bermondsians to consider how their fashion choices aff ect health. Viewers are encouraged to dress in light-weight clothing that promotes movement and play. The empty bed shows a tragic consequence of not immunising children; here, the desperation of the public health department to combat preventable illness and death is evident. Scare tactics were used to boost uptake of diphtheria vaccination. The campaign, and overall war on infection, led to a 38% mortality reduction over the next 30 years in Bermondsey.

Measuring health outcomes, and the eff ectiveness of media campaigns, resonates with today’s health agenda. The UK Government will expect to see a return on its £75 million investment in the Change4Life campaign, the social media aspect of the Government’s Healthy Weight, Healthy Lives strategy for tackling obesity. The staggering scope of obesity as a health problem, and its direct implications for cancer and other disease prognoses, means that an eff ective campaign could sub-stantially improve population health.

More than a government health initiative, Change4Life is envisaged as a social movement where the agenda and means are provided but the momentum is generated by grass-roots input. Local supporters (schools, businesses, health departments) are invited to download materials for awareness-raising community events. National partners are also involved to deliver joined-up messages to their own customers. For example, the Change4Life website features recipes with ingredients that can be bought at a discount from partner superstores.

When the desired behaviour change is as complex as adopting a more active lifestyle, supportive programmes are needed to supplement the campaign and achieve a population-wide goal. However, when increased uptake for cancer screening is the goal, mass media campaigns can increase awareness quickly and eff ectively promote change. The Be Clear on Cancer campaign, launched by the UK Government in 2011, leaves no media stone unturned to promote awareness and early diagnosis of cancer. National TV, radio, online, press, and bus adverts dispel ignorance and clarify signs and symptoms of lung, breast, bladder, kidney, and oesophagogastric cancers. Public health campaigns help to guide widespread change towards healthier practices, including participation in programmes such as cancer screening that substantially reduce disease burden and mortality. The ever-increasing presence of media in personal lives means growing potential for health messaging to have an impact. Delivery must be timely, accessible, and modern, and the presentation must be engaging to compete in our noisy media world.

Jill Jouret

ExhibitionBe clear on good health

Sout

hwar

k Lo

cal H

istor

y Li

brar

y an

d Ar

chiv

e

Children watching a cinemotor show


Cancer and Society

Cut Poison BurnDirected by Wayne Chesler 2010,

available on DVD March 26, 2012,USA, 87 min

For more on Cut Poison Burn see http://cutpoisonburn.com/

For more about laetrile see http://www.quackwatch.

org/01QuackeryRelatedTopics/Cancer/laetrile.html

For more about Steve McQueen see http://www.nytimes.com/

2005/11/15/health/15essa.html?_r=1

For more about averted cancer deaths see http://onlinelibrary.

wiley.com/doi/10.3322/caac.20121/pdf

On a nice, bright day in Tijuana, Mexico, in May, 2000, a happy and seemingly healthy 4-year-old boy holds a sombrero upside down above his head. He playfully says “There’s something growing on my head, there’s something growing on my head.” In September, 1999, there was something growing inside his head: a medulloblastoma. Thomas Navarro underwent brain surgery to have the tumour removed. His parents James (Jim) and Donna were told that without immediate chemotherapy and radiation the cancer would recur in 9 months. Fearing the eff ects of the standard FDA-approved treatment on their son’s quality of life, despite its curative potential in over 75% of patients, they wanted to seek a non-lethal, non-toxic treatment fi rst. The Navarros went to Mexico to stall for time and to avoid charges fi led against them of medical and child neglect and child abuse.

According to Julian Whitaker, founder of the Whitaker Wellness Institute (Newport Beach, CA, USA), in the documentary Cut Poison Burn, people in the USA have a right to choice in the commercial market except medicine and their “right to choose in cancer is truncated the most”. The terms cut, poison, and burn crudely summarise the generic treatments for cancer: surgery, chemotherapy, and radiation, respectively. Conventional treatments for cancer are presented as being harmful and primitive.

The FDA, American Cancer Society (ACS), and American Medical

Association all declined to be interviewed, so the producers of Cut Poison Burn present a largely one-sided argument for medical freedom with respect to cancer treatment. The FDA, ACS, and National Cancer Institute are portrayed as enemies and bureaucrats in the “war on cancer”. This war declared on Jan 22, 1971, by US President Richard Nixon in his State of the Union address, was the launch of an intensive campaign to fi nd a cure for cancer. The fi lm accuses the aforementioned organisations of stifl ing medical free-dom in terms of treat ment choice and inno vation. For example, Whitaker describes the ACS as “an impediment to a cure for cancer” by making money out of cancer. The cancer industry is presented as being based on a fi nancial agenda rather than a health one. According to Charles Simone, founder of the Simone Protective Cancer Center (Lawrenceville, NJ, USA), “We’re taught to treat in medical school and we’re paid to treat in practice. We don’t get taught prevention and we don’t get paid to prevent”.

The alternative treatment sought after by some of the patients in Cut Poison Burn is antineoplastons at Stanislaw Burzynski’s clinic in Houston, TX, USA. Another sought after altern-ative treatment is laetrile—it was popular in the 1970s and some of its proponents confuse it with amygdalin (vitamin B17), which is found in the pits of some fruits including apricots. Laetrile has not been assessed in clinical trials, is associated with cyanide poisoning, and is illegal in the USA.

In July 1980, the actor and car racer Steve McQueen went to Rosarita Beach, Mexico, for a treatment regimen that included laetrile for his mesothelioma: he died in November, 1980.

Towards the end of the documentary, clips are shown of people applauding President Nixon’s State of the Union address and individuals applauding President Barack Obama’s speech on Feb 24, 2009, “to launch a new eff ort to conquer a disease that has touched the life of nearly every free American, including me, by seeking a cure for cancer in our time”. The message is that nothing has changed in 38 years and is unlikely to do so: since the “war on cancer” began, more than 14·8 million people in the USA have died from the disease. But despite the millions of deaths, according to the ACS nearly 900 000 deaths from cancer were avoided between 1991 and 2007.

The Navarros’ pain is evident when Jim says, “We’ll never be whole” after the death of their son, aged 6 years. Perhaps people who are terminally ill should have medical freedom to choose alternative medicine when the existing treatments are unlikely to prolong their lives. The producers, however, do not address whether and how the effi cacy and safety of alternative treatments could be verifi ed and just tiptoe around how to deal with frauds. Medical freedom has to be tempered with regulations to ensure that vulnerable people will not be exploited by those peddling quackery.

Farhat Yaqub

Cut Poison Burn

Voices from the front: June“See you Friday, doc,” said the well built 30-year-old man with bright eyes and a wry grin as he walked down the hall to his room early Monday morning. He was dressed in sweat pants, a Red Sox shirt, and cap and running shoes. He was starting 4 days of intensive chemotherapy for his sarcoma in the now mostly forgotten pre-ondansetron era. To control his vomiting, he was facing 4 days of sedation from a high-dose metoclopramide infusion with diphenhydramine and lorazepam. He would emerge at the end of the week bleary eyed, unshaven, and stale. Fortunately, his wife reported that he cleaned up nicely.

David M Mastrianni


Cancer and Society

Varia

nce

Film

s

Addiction IncorporatedDirected by Charles Evans Jr, 2011. USA, 101 min.

When US President Barack Obama signed into law the 2009 Tobacco Control Act, which mandated the Food and Drug Administration (FDA) with regulating tobacco products, he commented that the legislation “represents change that’s been decades in the making”. Instrumental to this change was Victor DeNoble, the subject of the smart and sprightly documentary Addiction Incorporated.

The tale kicks off in 1979. With echoes of Mad Men, we see a seductive stewardess sally forth with champagne, cigarette smoke swirling round the fi rst-class cabin. This was a time where there was little stigma attached to smoking. It was the year that Victor DeNoble started working for international tobacco giant Philip Morris. DeNoble was a scientist, the fi rst in his family to attend college (he did so to meet women), hailing from New York. He’s an aff able presence on camera, amusing and entirely at ease. He explains how Philip Morris—who seemed to know his entire history—off ered him a position in their secret research centre to examine the pharmacology of nicotine. The intention, they assured him, was to produce a cigarette which would off er the same jolt as the ones currently in existence but without the concomitant risk of heart attacks and stroke. “To a scientist, that’s a dream”, exclaims DeNoble, “you get to do a job that actually might make a diff erence”.

Addiction Incorporated is essentially a political thriller. It is roughly broken into three parts: the fi rst corresponds to the crime, the second to its exposure, and the third to the aftermath. DeNoble’s research centred on rodents—“we were trying to turn rats into smokers”, he explains. And it proved eminently possible. The fi rst day, the test rat would push the switch to administer a hit of nicotine two to three times; by day 10, they’d be up to 20 times; and

by day 21, 90 times. A series of jaunty cartoons accompany this passage.

But of most interest to Philip Morris was DeNoble’s discovery that acetaldehyde, another constituent of tobacco smoke, acted as an enhancer for nicotine, intensifying the action of nicotine on the brain. Rats would push the switch for nicotine 120 times, they’d push the switch for acetaldehyde 240 times, but if you put the two together, they’d push the switch upwards of 540 times. For a tobacco company, this was the Holy Grail: a means of making a more addictive cigarette.

The early 1980s saw a sharp increase in law suits against tobacco companies. Philip Morris, carefully attuned to changes in the climate, decided to clean up shop. They recognised (correctly) that if word of DeNoble’s work leaked out, it would make it tricky for the company to continue their corrupt claim that cigarettes were not addictive. They shut down the research centre and, crushingly for DeNoble, forced him to withdraw his paper on nicotine enhancers from the journal Psychopharmacology.

The story cuts forward 10 years, shifting its focus to journalism and politics. The American Broadcasting Company led the way with an exposé on the tobacco industry, Philip Morris responded with a US$10 billion lawsuit against the network. The FDA started its own investigation into the industry. But the key event, and the fi lm’s centrepiece, was the exhilarating 1994 Congressional hearing. Spearheaded by Henry Waxman, it called to account the seven chief executive offi cers of America’s largest tobacco companies. Cognisant of DeNoble’s research, but aware that he was unable to testify due to a confi dentiality agreement, a committee member harries the squirming head of Philip Morris into releasing DeNoble from his contract.

He eventually agrees (and small wonder that he had proved reluctant to do so, given that he, and his six fellows, had previously denied under oath that smoking was addictive).

It is diffi cult to overstate the importance of DeNoble’s testimony. The advances of the following decade would have occurred regardless, but perhaps not as rapidly. Soon enough, the tobacco industry would have the Racketeer Infl uenced and Corrupt Organisations (RICO) Act successfully invoked against it—“they had joined the ranks of organised crime” one interviewee says wryly. The public mood turned against the industry, and DeNoble was co-opted into the class-action suit taken against the cigarette companies. The old tactics of intimidation and outspending would no longer work. The companies began to off er settlements.

This is a superb documentary, light-footed and well-organised, it develops a bracing momentum. It adeptly sketches the worlds of science, business, politics and journalism. Today, Victor DeNoble travels the USA educating children on the dangers of addiction (his work, we learn, is funded by the tobacco industry’s settlement with the 50 US states). For this, and for having the courage to blow the whistle on his former employers, he is owed quite a debt.

Talha Khan Burki

FilmsAddiction Incorporated


Cancer and Society

For more on Maggie’s Cancer Caring Centres

see Cancer and Society Lancet Oncol 2008; 9: 521–22

Birmingham Children’s Hospital hosts one of three paediatric liver centres in the UK. If your child falls ill, and it is their liver that is aff ected, you might fi nd yourself facing a lengthy round trip to the nearest specialist unit, perhaps for treatment fi ve times a week. It is expensive, enervating, and disruptive. Of course, hospitals keep rooms in which families can stay, but these fi ll up quickly. Furthermore, a hospital room is an environment that can rapidly become oppressive.

Some families may opt to stay in a local bed-and-breakfast, but a cancer diagnosis exerts a hefty fi nancial toll. However, alternatives do exist. Ronald McDonald House Charities provide “free ‘home from home’ accommo-dation across the UK, enabling families to stay close to their child and maintain a degree of normal family life”. 14 Ronald McDonald houses currently operate in the UK—the one adjacent to Birmingham’s Children Hospital has 60 rooms. The charity was founded in the 1970s in the USA and opened its fi rst UK residence in 1990.

“It is exhausting dealing with a child who’s sick”, explains the charity’s Anne Roberts. “You have to keep the rest of the family following a normal daily routine, there may be siblings to consider who need to go to school, or you may have had a call that your

child has taken a turn for the worse and you have to get to the hospital quickly.” Establishing a home close to the hospital can help alleviate some of these problems. Parents might be able to enrol their other children in local nurseries. There is the opportunity to cook and eat together, and for the stricken child there is the comfort of having their family at hand. The houses—which vary in size from three to fi ve bedrooms to the 60 bedroom house currently being built in Manchester—share communal areas. “It means families can meet other families experiencing the same kind of thing, we fi nd that quite a lot of friendships can be made out of those relationships”, concludes Roberts.

CLIC Sargent’s Home from Home off ers a similar service. Unlike Ronald McDonald houses, Home from Home caters specifi cally for the families of children with cancer. There are nine such homes in the UK, each named after a patient. “It is good for Ryan’s dad and myself”, one parent says of “Alex’s House”. “When Ryan gets some time away from the ward, he loves playing with the other children. No one stares if your child has lost their hair, or has a tube in their chest as we are all in the same boat.”

“We’ve conducted some research into the benefi ts for the families who stay

in the houses and that’s going to be published later in the year”, Roberts told The Lancet Oncology. This research will assess the eff ect of sleep deprivation. Certainly, it seems probable that fam-ilies who stay in a Home from Home or Ronald McDonald house will have improved mental health (and anecdotal evidence points towards this) but it would also be useful to learn whether the residences have any eff ect on treatment outcomes.

The ethos behind Maggie’s Cancer Caring Centres is that a supportive and warm environment away from the hospital can off er patients and their families valuable therapeutic benefi ts. The centres are striking—the one in Nottingham was designed by prominent architect Piers Gough in collaboration with the fashion designer Sir Paul Smith. “In hospitals, patients are recipients of clinical treatment, just by the nature of their relationship with the clinical team, they’re passive—they are having to receive treatment”, explains Laura Lee, CEO of Maggie’s Cancer Caring Centres. Her organisation off ers patients the opportunity to regain their individuality and address issues—emotional, practical, or social—associated with cancer. “The internal architecture communicates that this is not a clinical environment”, adds Lee. “There is no reception desk, people can sit in a quiet corner or go to a course, workshop, or doctor’s lecture or see an advisor”.

It is all part of an overarching aim to help people establish a plan to address the hardships of a cancer diagnosis. Unlike Ronald McDonald’s houses or those of CLIC Sargent’s, Maggie’s Centres are not residential. But in common with the two, they off er a welcome respite from the bleary hours in a hospital ward, and that can only be of benefi t to the patient.

Talha Khan BurkiFree home from home accommodation off ers respite to parents and their families

Russ

Ham

er

CharityHome from Home

www.thelancet.com/oncology Published online April 13, 2012 DOI:10.1016/S1470-2045(12)70116-X 1

Articles

Published OnlineApril 13, 2012 DOI:10.1016/S1470-2045(12)70116-X

Day Treatment Center (TTZ), Interdisciplinary Tumor Center Mannheim (ITM) & III Medical Clinic (Prof R-D Hofh einz MD, D Gencer MD), Clinic for Radiotherapy and Radio-oncology (Prof F Wenz MD), and Clinic and Policlinic for Surgery (Prof S Post MD, Prof P Kienle MD), University Hospital Mannheim, University of Heidelberg, Mannheim, Germany; Clinic for Hematology and Oncology, Caritas Hospital St Theresa, Saarbrücken, Germany (Prof A Matzdorff MD); Clinic for Radio-oncology, University Hospital Tübingen, Tübingen, Germany (S Laechelt MD); Department for Internal Medicine II, University Hospital Schleswig-Holstein, Christian-Albrechts-University, Kiel, Germany (Prof J T Hartmann MD); Oncology Practice, Leer, Germany (L Müller MD); Department for Internal Medicine 1, Westpfalz Hospital, Kaiserslautern, Germany (Prof H Link MD); Department for Interal Medicine I, University Hospital Mainz, Mainz, Germany (Prof M Moehler MD); Hospital Magdeburg, Magdeburg, Germany (E Kettner MD); Clinic for Internal Medicine, Stiftungsklinikum Mittelrhein, Koblenz, Germany (E Fritz MD); Oncology Practice, Mannheim, Germany (U Hieber MD); Clinic for Hematology and Oncology, Katholisches Krankenhaus, Hagen, Germany (H W Lindemann MD); Clinic for Internal Medicine, Hospital Aschersleben-Staßfurt, Aschersleben, Germany (M Grunewald MD); Clinic for Hematology and Oncology, Caritas Hospital, Lebach,

Chemoradiotherapy with capecitabine versus fl uorouracil for locally advanced rectal cancer: a randomised, multicentre, non-inferiority, phase 3 trialRalf-Dieter Hofh einz, Frederik Wenz, Stefan Post, Axel Matzdorff , Stephan Laechelt, Jörg T Hartmann, Lothar Müller, Hartmut Link, Markus Moehler, Erika Kettner, Elisabeth Fritz, Udo Hieber, Hans Walter Lindemann, Martina Grunewald, Stephan Kremers, Christian Constantin, Matthias Hipp, Gernot Hartung, Deniz Gencer, Peter Kienle, Iris Burkholder, Andreas Hochhaus

SummaryBackground Fluorouracil-based chemoradiotherapy is regarded as a standard perioperative treatment in locally advanced rectal cancer. We investigated the effi cacy and safety of substituting fl uorouracil with the oral prodrug capecitabine.

Methods This randomised, open-label, multicentre, non-inferiority, phase 3 trial began in March, 2002, as an adjuvant trial comparing capecitabine-based chemo radiotherapy with fl uorouracil-based chemoradiotherapy, in patients aged 18 years or older with pathological stage II–III locally advanced rectal cancer from 35 German institutions. Patients in the capecitabine group were scheduled to receive two cycles of capecitabine (2500 mg/m² days 1–14, repeated day 22), followed by chemoradiotherapy (50·4 Gy plus capecitabine 1650 mg/m² days 1–38), then three cycles of capecitabine. Patients in the fl uorouracil group received two cycles of bolus fl uorouracil (500 mg/m² days 1–5, repeated day 29), followed by chemoradiotherapy (50·4 Gy plus infusional fl uorouracil 225 mg/m² daily), then two cycles of bolus fl uorouracil. The protocol was amended in March, 2005, to allow a neoadjuvant cohort in which patients in the capecitabine group received chemo radiotherapy (50·4 Gy plus capecitabine 1650 mg/m² daily) followed by radical surgery and fi ve cycles of capecitabine (2500 mg/m² per day for 14 days) and patients in the fl uorouracil group received chemo radiotherapy (50·4 Gy plus infusional fl uorouracil 1000 mg/m² days 1–5 and 29–33) followed by radical surgery and four cycles of bolus fl uorouracil (500 mg/m² for 5 days). Patients were randomly assigned to treatment group in a 1:1 ratio using permuted blocks, with stratifi cation by centre and tumour stage. The primary endpoint was overall survival; analyses were done based on all patients with post-randomisation data. Non-inferiority of capecitabine in terms of 5-year overall survival was tested with a 12·5% margin. This trial is registered with ClinicalTrials.gov, number NCT01500993.

Findings Between March, 2002, and December, 2007, 401 patients were randomly allocated; 392 patients were evaluable (197 in the capecitabine group, 195 in the fl uorouracil group), with a median follow-up of 52 months (IQR 41–72). 5-year overall survival in the capecitabine group was non-inferior to that in the fl uorouracil group (76% [95% CI 67–82] vs 67% [58–74]; p=0·0004; post-hoc test for superiority p=0·05). 3-year disease-free survival was 75% (95% CI 68–81) in the capecitabine group and 67% (59–73) in the fl uorouracil group (p=0·07). Similar numbers of patients had local recurrences in each group (12 [6%] in the capecitabine group vs 14 [7%] in the fl uorouracil group, p=0·67), but fewer patients developed distant metastases in the capecitabine group (37 [19%] vs 54 [28%]; p=0·04). Diarrhoea was the most common adverse event in both groups (any grade: 104 [53%] patients in the capecitabine group vs 85 [44%] in the fl uorouracil group; grade 3–4: 17 [9%] vs four [2%]). Patients in the capecitabine group had more hand-foot skin reactions (62 [31%] any grade, four [2%] grade 3–4 vs three [2%] any grade, no grade 3–4), fatigue (55 [28%] any grade, no grade 3–4 vs 29 [15%], two [1%] grade 3–4), and proctitis (31 [16%] any grade, one [<1%] grade 3–4 vs ten [5%], one [<1%] grade 3–4) than did those in the fl uorouracil group, whereas leucopenia was more frequent with fl uorouracil than with capecitabine (68 [35%] any grade, 16 [8%] grade 3–4 vs 50 [25%] any grade, three [2%] grade 3–4).

Interpretation Capecitabine could replace fl uorouracil in adjuvant or neoadjuvant chemoradiotherapy regimens for patients with locally advanced rectal cancer.

Funding Roche Pharma AG (Grenzach-Wyhlen, Germany).

IntroductionThe combination of optimised surgery (total mesorectal excision [TME]1) and systematic radiotherapy has substantially improved multimodal treatment of rectal cancer;2,3 TME plus short-course radiotherapy yields a 10-year cumulative local recurrence rate of only 5%.3

Fluorouracil in conjunction with neoadjuvant long-term radiotherapy reduced local recurrences in two trials but did not prolong survival.4,5 Fluorouracil is often given as adjuvant treatment of rectal cancer, after resection of the primary tumour and neoadjuvant irradiation.6 In the European Organisation for Research and Treatment of

Articles

2 www.thelancet.com/oncology Published online April 13, 2012 DOI:10.1016/S1470-2045(12)70116-X

Germany (S Kremers MD); Clinic for Hematology and Oncology,

Hospital Lippe, Lemgo, Germany (C Constantin MD);

Clinic and Policlinic for Radiotherapy, University

Hospital Regensburg, Regensburg, Germany

(M Hipp MD); Oncology Practice, Groß-Gerau, Germany

(Prof G Hartung MD); Department of Biostatistics,

German Cancer Research Center (DKFZ) Heidelberg, and STABIL,

Zweibrücken, Germany (I Burkholder PhD); and

Department for Hematology and Oncology, Clinic for

Internal Medicine II, University Hospital Jena, Jena, Germany

(Prof A Hochhaus MD)

Correspondence to:Prof Ralf-Dieter Hofh einz, Day

Treatment Center (TTZ), Interdisciplinary Tumor Center Mannheim (ITM) & III Medical

Clinic, University Hospital Mannheim, University of

Heidelberg, 68167 Mannheim, Germany

ralf.hofh [email protected]

Cancer (EORTC) 22922 trial,5 patients were randomised to bolus fl uorouracil or follow-up after resection of the primary tumour and long-term neoadjuvant radiotherapy. 5-year overall survival was 67·2% in the treatment group and 63·2% in controls (p=0·12); the hazard ratio (HR) for death in the chemotherapy group was 0·85.5 Current evidence favouring the use of fl uorouracil in this setting is limited, and national treatment recommendations refl ect divergent interpretations of the published data. German guidelines consider adjuvant fl uorouracil the standard of care.7

Optimisation of local tumour control has meant that distant metastases now represent the most common type of treatment failure in rectal cancer. Modifi cations of peri-operative fl uorouracil treatment have been investigated in an attempt to improve overall survival and disease-free survival (DFS); however, neither biomodu lation of fl uorouracil by folinic acid or levamisole,8 nor combination with other cytostatic drugs,9,10 have proved superior to bolus fl uorouracil, with the exception of infusional fl uorouracil given during radiotherapy.9

Capecitabine is an oral fl uoropyrimidine derivative that was as eff ective as fl uorouracil plus folinic acid for adjuvant treatment of stage III colon cancer.11 It was also non-inferior to infusional fl uorouracil in combination with oxaliplatin for fi rst-line treatment of metastatic colorectal cancer.12 Several phase 1 and 2 trials have investigated capecitabine as part of combinations for perioperative treatment of rectal cancer,13 but no randomised trial has compared capecitabine with perioperative fl uorouracil in locally advanced disease. Our choice of a non-inferiority trial design was based on the expectation that non-inferiority of capecitabine, given orally on an outpatient basis, would be suffi cient to tip the risk–benefi t ratio in its favour. Here, we report fi nal results of our phase 3 trial comparing capecitabine with fl uorouracil as part of perioperative chemoradiotherapy regimens for locally advanced rectal cancer.

MethodsStudy design and patientsThis was a two-arm, two-cohort, multicentre, random-ised, open-label, non-inferiority, phase 3 trial comparing fl uorouracil with capecitabine for perioperative treatment of patients with locally advanced rectal cancer. Patients were recruited from 35 German institutions between March, 2002, and December, 2007. The protocol was approved by the institutional review boards of all participating centres. All participants provided written informed consent.

The study was initiated to compare fl uorouracil with capecitabine in patients with locally advanced rectal cancer who had undergone TME or partial mesorectal excision (PME) of the primary tumour (adjuvant cohort). In 2004, the German rectal cancer trial6 reported better local control and tolerability with neoadjuvant chemo-radiotherapy; therefore, the study protocol was amended

in March, 2005, to include patients with locally advanced rectal cancer receiving preoperative chemoradiotherapy (neoadjuvant cohort). Recruitment to the adjuvant cohort was continued. The trial steering committee endorsed this amendment on the basis that the German rectal cancer trial showed no diff erence in survival rates or Kaplan-Meier plots of adjuvant versus neoadjuvant groups. Thus, the committee anticipated that the amendment would have no eff ect on the primary endpoint of the present trial, overall survival.

Eligible patients were 18 years or older and had histologically confi rmed adenocarcinoma of the rectum (defi ned as a distal tumour border <16 cm from the anal verge, measured by rigid rectoscopy), with no evidence of distant metastases (identifi ed by abdominal ultra sound or CT scan and chest radiograph). Patients in the adjuvant cohort had to have undergone R0 resection (ie, leaving no residual tumour) for pT3–4 Nany or pTany Npositive non-metastatic rectal cancer. TME was mandatory for tumours in the lower two-thirds of the rectum, with PME being permitted for those in the upper third, provided a distal margin of at least 5 cm without coning was observed. Patients in the neoadjuvant cohort had to have a clinical cT3–4 Nany or cTany Npositive tumour staged by endoscopic ultrasound, provided the lower border of the tumour was 0–16 cm from the anal verge (measured by rigid rectoscopy) and the primary tumour was deemed R0 resectable by TME or PME on the basis of clinical assessment (pelvic CT or MRI were done at the discretion of the local investigators).

Other eligibility criteria were: WHO status 0 or 1; satisfactory liver, renal, and bone-marrow function (leucocytes >3500 cells per μL, platelets >100 000 per μL, haemoglobin >100 g/L); serum bilirubin less than 20 mg/L; and serum creatinine less than 20 mg/L. Exclusion criteria were prior treatment for rectal cancer, prior chemotherapy or immunotherapy, prior pelvic radio therapy, or a history of other malignant disease within the past 5 years, other than successfully treated basal carcinoma of the skin or carcinoma in situ of the uterine cervix. Patients were also excluded if they were participating in another trial, pregnant, breastfeeding, unwilling to use eff ective contraception, or had a medical condition or concomitant illness that might impair protocol compliance.

Randomisation and maskingPatients were randomly allocated (by fax request to the Department of Biostatistics, German Cancer Research Center [Heidelberg, Germany]) in a 1:1 ratio to perioperative treatment with capecitabine or fl uorouracil, using permuted blocks with stratifi cation by centre and clinical or pathological tumour stage (T3–4 N0 vs T1–2 Npositive vs T3–4 Npositive). For each stratifi cation group and participating centre, a list was generated by the data centre using S+ software and used to assign treatments. Local investigators were masked to next assignment in

For more on S+ software see http://spotfi re.tibco.com

Articles


the sequence. The data centre also managed primary and follow-up data using case reports compiled by the participating centres. The study was open-label; patients, treating physicians, and data managers and analysts were not masked to group assignment.

ProceduresPatients randomised to capecitabine were scheduled to receive six cycles of chemotherapy, whereas those in the fl uorouracil group received fi ve cycles. Patients in the adjuvant cohort received two cycles of chemotherapy before starting chemoradiotherapy in week 8, there-after completing chemotherapy with three cycles of capecitabine or two cycles of fl uorouracil (fi gure 1). Patients in the neoadjuvant cohort received chemo-radiotherapy for about 6 weeks. TME or PME was done after 4–6 weeks, followed by fi ve cycles of capecitabine or four cycles of fl uorouracil (fi gure 2).

The total irradiation dose of 50·4 Gy was delivered in conventional fractionation (daily fractions of 1·8 Gy over 5–6 weeks, excluding weekends). Three-dimensional conformal techniques with high-energy photons (6–25 MeV) and belly boards were used. The clinical target volume included the entire macroscopic tumour with a minimum margin of 5 cm, the mesorectum (plus 1·0–1·5 cm margin lateral to the pelvic brim), and the

iliac and presacral lymph nodes up to the L5–S1 junction (or L4–L5 junction in the case of extensive lymph-node involvement).

Capecitabine was given twice daily at a cumulative dose of 2500 mg/m² on days 1–14, and repeated on day 22. The total daily dose was divided into two equal amounts and given roughly 12 h apart and within 30 min after a meal, usually breakfast and dinner. Capecitabine was given at a reduced dose of 1650 mg/m² per day throughout radiotherapy, including weekends. Radio therapy and capecitabine were started on the same day and capecitabine was stopped on the last day of radiotherapy.

Fluorouracil bolus was administered on fi ve consecutive days (days 1–5) and repeated on day 29. Patients in the neoadjuvant cohort received fl uorouracil during radio-therapy according to the CAO/ARO/AIO-94 protocol6 (ie, 1000 mg/m² per day as a continuous infusion on days 1–5 and 29–33). Patients in the adjuvant cohort received 225 mg/m² per day infusional fl uorouracil throughout radiotherapy.9

Vital signs, haematology, and biochemistry were monitored weekly during chemoradiotherapy and before each chemotherapy cycle. The protocol stipulated detailed dose-modifi cation criteria according to toxicity, graded using the National Cancer Institute Common Toxicity Criteria (NCI-CTC) version 2.0.

1 5 9 13 16 20

Radiotherapy 50·4 Gy

Capecitabine 2500 mg/m2 per day (during radiotherapy 1650 mg/m2 per day)

Week

Fluorouracil 500 mg/m2 days 1–5 (during radiotherapy 225 mg/m2 per day)


Capecitabine group

Fluorouracil group

Figure 1: Treatment regimen for the adjuvant cohort

1 5 16 24 28

Week

10 20

Surgery

Surgery

Fluorouracil 1000 mg/m2 days 1–5, days 29–33 during radiotherapy, then 500 mg/m2 days 1–5



Capecitabine 1650 mg/m2 per day during radiotherapy, then 2500 mg/m2 per dayCapecitabine group

Fluorouracil group

Figure 2: Treatment regimen for the neoadjuvant cohort

Articles


TME for tumours of the lower two-thirds of the rectum and PME for the upper third, assuming a 5 cm distal margin without coning, were mandatory for inclusion in the adjuvant cohort and recommended for inclusion in the neoadjuvant cohort, although formal quality assurance in this regard was not implemented. For low-lying tumours, the decision between low anterior resection and abdominoperineal excision was left to the surgeon’s discretion.

Baseline assessments were medical history, clinical examination, complete haematology with diff erential leucocyte count, clinical chemistry including co agulation parameters, tumour markers CEA and CA19-9, electro-cardiogram, abdominal ultrasound, and chest radiograph. Additionally, patients in the neoadjuvant cohort under-went complete colonoscopy, rigid rectoscopy, and endorectal ultrasound. Follow-up, done for 5 years after the start of therapy, included clinical examination, haematology, serum biochemistry, tumour markers, and abdominal ultrasound 3 monthly for the fi rst 2 years, and 6 monthly thereafter, in addition to annual chest radiograph. Regular rectoscopy with endorectal ultra-sound (months 6, 12, 18, 24, 36, 48, and 60 after removal of the primary tumour) and pelvic CT (months 3, 12, and 24) were done to exclude local recurrence.

Statistical analysisThis trial was designed to test the non-inferiority of 5-year overall survival in the capecitabine versus fl uorouracil group. Assuming 57·5% overall survival in the fl uorouracil group, a non-inferiority margin of 12·5%, accrual time of 36 months, follow-up of 48 months, 5% drop-out, 5% type I error, and 80% power, sample size calculation using PASS 2000 yielded a total of at least 372 evaluable patients (ie, 186 per group). The assumption of 57·5% overall survival in the fl uorouracil group was based on interpolation and extrapolation of survival rates reported by two trials investigating adjuvant fl uorouracil-based chemoradiotherapy in rectal cancer.8,9 Data were analysed with SAS (version 9.2) and R (version 2.10.1). All analyses were based on all patients with post-randomisation data.

The primary endpoint of overall survival was calculated from the date of randomisation to the date of death. Non-inferiority of 5-year overall survival was tested with a 12·5% margin, using Greenwood’s variance estimator for the diff erence between two survival proportions at 5 years.14,15 This was the only confi rmatory statistical test. A post-hoc, exploratory, one-sided test of the diff erence in 5-year overall survival between the groups was used to test superiority of the capecitabine group. Kaplan-Meier

116 adjuvant cohort

3 no start 1 delayed wound-healing 1 consent withdrawn 1 other

113 began scheduled treatment

90 completed all scheduled cycles

81 neoadjuvant cohort

4 excluded 2 primary resection 1 sigmoid cancer 1 consent withdrawn

77 began radiochemotherapy

4 not resected 1 pCR 1 PD 2 deaths during CRT

73 resected


115 adjuvant cohort

3 no start 2 delayed wound-healing 1 other


80 neoadjuvant cohort

2 excluded 1 refusal of port implantation 1 consent withdrawn

78 began radiochemotherapy

3 not resected 2 refusals 1 PD


112 began scheduled treatment

197 allocated to capecitabine 195 allocated to fluorouracil

401 patients randomised

9 excluded (no post-randomisation data)

392 patients in full analysis set

75 resected

Figure 3: Trial profi lepCR=pathological complete remission. PD=progressive disease. CRT=chemoradiotherapy.

For more on PASS 2000 software see http://www.ncss.

com

For more on SAS software see http://www.sas.com

Articles


survival estimates were compared using the log-rank test. Cox proportional HRs were calculated for diff erent subgroups (cohort, sex, age, stratum, WHO status, type of surgery, and resection status). The assumption of proportional hazards was assessed using Schoenfeld residuals and by testing a time-dependent covariate defi ned as interaction between treatment group and log survival time.

The secondary endpoint of DFS was calculated from the date of randomisation to the date of disease recurrence (metastasis or local recurrence), development of a second primary cancer (including non-colorectal carcinoma), or death from any cause, whichever occurred fi rst. DFS was analysed using censored failure times with the Kaplan-Meier method, with exploratory two-sided tests of the diff erence in 3-year DFS between groups. Additionally, Cox proportional HR for treatment diff erence in DFS was calculated for the overall population. Other secondary endpoints were local recurrence (pelvic or perineal tumour), distant metastases, and treatment toxicity.

Proportions were compared using a χ² test, and continuous variables were compared using Wilcoxon’s rank-sum test. A p value of 0·05 or lower was considered signifi cant. All statistical tests for secondary endpoints, particularly those concerning safety, were interpreted descriptively and exploratorily and no formal statistical conclusions were drawn. No imputation methods for missing values were applied. A linear association over the ordered categories ypT0 versus ypT1–2 versus ypT3–4 was tested for patients in the neoadjuvant cohort, using an exact Mantel-Haenszel χ² test.

This trial is registered with ClinicalTrials.gov, number NCT01500993.

Role of the funding sourceRoche Pharma AG provided capecitabine and a research grant for the trial, but had no role in the study design, data collection, analysis, or interpretation, writing of this report, or the decision to submit for publication. R-DH, FW, IB, DG, and AH had access to the raw data. The corresponding author had full access to all study data and fi nal responsibility for the decision to submit for publication.

Results401 patients were randomised between March 20, 2002, and Dec 10, 2007 (fi gure 3). Nine patients were excluded because no post-randomisation data were available for analysis. Thus, the full analysis set comprised 392 patients; 197 in the capecitabine group and 195 in the fl uorouracil group (231 in the adjuvant cohort, 161 in the neoadjuvant cohort). Baseline patient and tumour characteristics were well balanced between the two groups (table 1). Men accounted for two-thirds of patients in both groups. Most patients had a cT3 or pT3 tumour, with a slight predominance of T3–4 stages in the

capecitabine group along with somewhat fewer positive nodal stages.

Follow-up was continued until February, 2011. Median follow-up was 52 months (IQR 41–72) and was similar in both groups (51 months [41–75] capecitabine vs 53 months [42–73] fl uorouracil). By the time of the analysis, 93 patients had died (38 in the capecitabine group and 55 in the fl uorouracil group; table 2); 63 of 93 deaths (68%) were due to the underlying cancer (26 [68%] of 38 in the capecitabine group vs 37 [67%] of 55 in the fl uorouracil group; p=0·91). However, there were signifi cantly fewer deaths in the capecitabine group (38 [19%] of 197 vs 55 [28%] of 195; p=0·04), resulting in an absolute reduction of 6% in the risk of disease-related death in the capecitabine group.

Capecitabine (n=197) Fluorouracil (n=195)

Age (years) 65 (30–85) 64 (33–86)

Sex

Male 129 (65%) 131 (67%)

Female 68 (35%) 64 (33%)

WHO status

0 120 (61%) 96 (49%)

1 60 (30%) 78 (40%)

2 3 (2%) 1 (<1%)

Missing data 14 (7%) 20 (10%)

Cohort

Adjuvant 116 (59%) 115 (59%)

Neoadjuvant 81 (41%) 80 (41%)

Tumour category*

T1 or T2 29 (15%) 36 (18%)

T3 150 (76%) 140 (72%)

T4 15 (8%) 14 (7%)

Missing data 3 (2%) 5 (3%)

Nodal category*

Node negative 78 (40%) 69 (35%)

Node positive 112 (57%) 120 (62%)

Missing data 7 (4%) 6 (3%)

Data are n (%) or median (range). *Clinical or pathological category.

Table 1: Baseline characteristics of the study population

Capecitabine (n=197)

Fluorouracil (n=195)

p value*

Site of recurrence

Local 12 (6%) 14 (7%) 0·67*

Distant 37 (19%) 54 (28%) 0·04*

Deaths

Total 38 (19%) 55 (28%) 0·04*

Disease-related 26 (13%) 37 (19%)

Other causes 12 (6%) 15 (8%)

Unknown 0 3 (2%)

Data are cumulative number of events (%). *χ² test.

Table 2: Disease-related events

Articles


5-year overall survival in the capecitabine group was non-inferior to that in the fl uorouracil group (76% [95% CI 67–82] in the capecitabine group vs 67% [58–74] in the fl uorouracil group, non-inferiority p=0·0004; fi gure 4; hazard ratio [HR] for fl uorouracil vs capecitabine 1·5 [95% CI 1·00–2·28]). An exploratory, post-hoc test for superiority in 5-year overall survival was in favour of capecitabine (p=0·05). Better 5-year overall survival with capecitabine than with fl uorouracil was noted in both the adjuvant (81% [95% CI 71–87] vs 71% [60–79]) and the neoadjuvant cohort (66% [46–81]) vs 61% [46–73]).

Capecitabine was associated with improved survival in all subgroups that included a suffi ciently large number of patients (fi gure 5). For example, an HR of 1·70 (1·08–2·68) was seen in the T3–4N+ subgroup (310 patients), suggesting that these patients, who have a higher risk of relapse, derived substantial benefi t from capecitabine. The eff ect of capecitabine relative to fl uorouracil was noted for both cohorts, although it was slightly smaller in the neoadjuvant cohort than in the adjuvant cohort (HR 1·28 [95% CI 0·69–2·37] vs 1·62 [0·92–2·86] in the adjuvant cohort). This diff erence in treatment eff ect between cohorts was not signifi cant. Comparison using a Cox model adjusted for treatment and cohort revealed an HR of 1·5 (0·98–2·24), showing that treatment cohort did not aff ect analysis of the primary endpoint.

The number of patients with a local recurrence was similar between groups (12 [6%] in the capecitabine group vs 14 [7%] in the fl uorouracil group; p=0·67), but fewer patients in the capecitabine group had distant metastasis than did those in the fl uorouracil group (37 [19%] vs 54 [28%]; p=0·04; table 2). DFS was better in the capecitabine group than in the fl uorouracil group (HR 1·4 [95% CI 1·02–2·02]; log-rank p=0·035; fi gure 6). 3-year DFS was higher in the capecitabine group than in the fl uorouracil group (75% [95% CI 68–81] vs 67% [59–73]; p=0·07; table 3). Better 3-year DFS with capecitabine than with fl uorouracil was noted in both the adjuvant (78% [69–85] vs 69% [59–77]) and neoadjuvant cohorts (71% [60–80] vs 63% [51–73]).

In the adjuvant cohort, 90 patients (78%) in the capecitabine group and 92 (80%) in the fl uorouracil group completed their scheduled cycles (table 4). Although a substantial proportion of patients in the neoadjuvant cohort did not continue chemotherapy after resection of the primary tumour, a similar proportion completed chemotherapy in both groups: 37 (46%) in the capecitabine group versus 32 (40%) in the fl uorouracil group (table 4). Of the patients in the neoadjuvant cohort starting postoperative treat ment, 37 (74%) of 50 patients in the capecitabine group and 32 (70%) of 46 patients in the fl uorouracil group completed all postoperative scheduled cycles.

Signifi cant diff erences between groups in the number of patients experiencing a toxic eff ect were noted for fatigue, proctitis, and hand-foot skin reactions, which

Figure 4: Overall survival

Ove

rall

surv

ival

(%)

0

20

40

60

80

100

Time (years)20 5431

195197

FluorouracilCapecitabine

159161

172174

134146

8289

5154

Number at risk

CapecitabineFluorouracil

Test for non-inferiority at 5 years: p=0·0004

Figure 5: Forest plot analysis of overall survival in subgroups (fl uorouracil vs capecitabine)

CohortAdjuvant 231 30/115 20/116 1·62 (0·92–2·86)Neoadjuvant 161 25/80 18/81 1·28 (0·69–2·37)

SexMale 260 37/131 28/129 1·35 (0·83–2·21)Female 132 18/64 10/68 1·78 (0·82–3·90)

Age<50 years 36 1/14 3/22 0·48 (0·05–4·59)50–60 years 103 14/51 9/52 1·71 (0·74–3·95)>60–70 years 150 21/78 16/72 1·13 (0·59–2·16)>70 years 103 19/52 10/51 1·92 (0·89–4·14)

StratumT 3–4 N0 53 2/24 3/29 0·63 (0·11–3·81)T 1–2 N+ 28 5/14 5/14 0·80 (0·23–2·77)T 3–4 N+ 310 48/156 30/154 1·70 (1·08–2·68)

WHO PS0 216 26/96 24/120 1·33 (0·76–2·32)1–2 142 25/79 11/63 1·99 (0·98–4·04)

Resection typeLow anterior 289 36/143 23/146 1·61 (0·96–2·72)Abdominoperineal 79 15/38 12/41 1·48 (0·69–3·18)

Resection statusR0 359 48/177 34/182 1·46 (0·94–2·26)R1–R2 5 2/2 2/3 1·32 (0·18–9·53)

Overall 392 55/195 38/197 1·50 (1·00–2·28)

Patients HR (95% CI)

1010·1Favours fluorouracil Favours capecitabine

Fluorouracil Capecitabine

Articles


were more frequent in the capecitabine group, and leucopenia, which was more frequent in the fl uorouracil group (table 5). Diarrhoea was the most common adverse event (table 5). Rates of diarrhoea during chemotherapy were almost identical between groups (all-grade diarrhoea: 47 [24%] of 197 patients in the capecitabine group vs 43 [22%] of 195 in the fl uorouracil group; p=0·67), but were signifi cantly higher in the capecitabine group during radiochemotherapy (88 [45%] vs 62 [32%]; p=0·009).

Hand-foot skin reactions were reported by 62 (31%) patients receiving capecitabine, but by only three (2%) of those receiving fl uorouracil. In a post-hoc analysis, patients in the capecitabine group who developed any-grade hand-foot skin reactions had better DFS and overall survival than did those who did not develop hand-foot skin reactions (3-year DFS: 83% [95% CI 71–91] vs 71% [63–79], p=0·03; 5-year overall survival: 91% [95% CI 81–96] vs 68% [57–77], p=0·0001). Similar diff erences in survival were noted when comparing patients in the capecitabine group who developed hand-foot skin reactions with the overall study population (for 3-year DFS, p=0·004 and for 5-year overall survival, p<0·0001 vs the remaining population [n=330]).

148 patients in the neoadjuvant cohort underwent surgery: 73 (90%) of 81 patients in the capecitabine group and 75 (94%) of 80 patients in the fl uorouracil group. Unresectable tumour spread made surgery impossible in one patient in the fl uorouracil group; one patient in the capecitabine group had local excision only because of major tumour remission after neoadjuvant therapy (fi gure 3). The mean pretreatment tumour distance from dentate line to lower tumour margin was 5 cm (SD 4) in the capecitabine group and 6 cm (SD 4) in the fl uorouracil group (p=0·04). Resection type and status did not diff er signifi cantly between capecitabine and fl uorouracil groups (low-anterior resection in 53 [73%] of 73 patients in the capecitabine group vs 58 [78%] of 74 in the fl uorouracil group, p=0·42; R0 resection in 69 [96%] of 72 [excluding the patient who underwent local excision] in the capecitabine group vs 68 [92%] of 74 in the fl uorouracil group, p=0·32).

Pathological complete remission (pCR; ypT0N0) was more frequent in the capecitabine group, being achieved by ten (14%) of 73 patients in the capecitabine group versus four (5%) of 74 in the fl uorouracil group (p=0·09). More patients in the capecitabine group had a ypT0–2 tumour (40 [55%] of 73 vs 29 [39%] of 74; p=0·06). A two-sided exact Mantel-Haenszel test for diff erence in the ordered categories ypT0 versus ypT1–2 versus ypT3–4 showed a signifi cant association for lower T categories in the capecitabine group (p=0·03), which suggests improved tumour shrinkage using capecitabine. Information on pretreatment clinical nodal staging was available for 150 patients. About half the patients in each group had a clinically staged node-positive primary tumour (36 [48%] of 75 patients in the capecitabine group

vs 39 [52%] of 75 in the fl uorouracil group; p=0·70). By contrast, patho logical staging after neoadjuvant chemo-radiotherapy revealed more node-negative tumours in the capecitabine group than in the fl uorouracil group

Figure 6: Disease-free survival

20 5431Di

seas

e-fre

e su

rviv

al (%

)Time (years)

195197

133144

161159

108131

7278

4450

0

20

40

60

80

100

FluorouracilCapecitabine

Number at risk

CapecitabineFluorouracil

Two-sided log-rank p=0·035

Capecitabine (n=197) Fluorouracil (n=195)

Disease-free survival, % (95% CI)

3 year 75% (68–81) 67% (59–73)

5 year 68% (60–74) 54% (45–62)

Overall survival, % (95% CI)

3 year 87% (81–91) 83% (77–88)

5 year 76% (67–82) 67% (58–74)

7 year 71% (60–79) 58% (47–67)

Table 3: Kaplan-Meier survival estimates

Capecitabine Fluorouracil

Adjuvant cohort*

1 113 (97%) 112 (97%)

2 110 (95%) 106 (92%)

3 104 (90%) 101 (88%)

4 100 (86%) 96 (84%)

5 98 (85%) 92 (80%)

6 90 (78%) ··

Neoadjuvant cohort†

1 74 (91%) 76 (95%)

2 50 (62%) 46 (58%)

3 46 (57%) 38 (48%)

4 42 (52%) 35 (44%)

5 40 (49%) 32 (40%)

6 37 (46%) ··

Data are n (%). *n=116 for capecitabine, n=115 for fl uorouracil. †n=81 for capecitabine, n=80 for fl uorouracil.

Table 4: Patients receiving scheduled treatment, per cycle

Articles


(51 [71%] of 72 patients in the capecitabine group vs 42 [57%] of 74 in the fl uorouracil group; p=0·08).

DiscussionSix cycles of capecitabine were non-inferior to fi ve cycles of fl uorouracil with regard to 5-year overall survival in patients receiving neoadjuvant or adjuvant chemoradio-therapy for locally advanced rectal cancer. DFS was higher in the capecitabine group than in the fl uorouracil group because fewer distant metastases occurred with capecitabine than with fl uorouracil. A post-hoc explora-tory test of superiority in 5-year overall survival showed a clinically meaningful survival benefi t in favour of capecitabine. This result was substantiated by the Cox proportional HR for the overall population, which favoured capecitabine and had a lower 95% CI boundary of 1·00. In the neoadjuvant cohort, more patients who received capecitabine achieved a pathological complete response or ypT0–2 than did those in the fl uorouracil group; there was also evidence of greater nodal downstaging for patients who received capecitabine than with fl uorouracil. With the exception of a higher rate of gastrointestinal toxicity during radiotherapy and higher rates of hand-foot skin reactions, capecitabine was generally as well tolerated as fl uorouracil.

The present study began as an adjuvant trial in 2002, and was amended to include a neoadjuvant

cohort in 2005, after publication of the German rectal cancer study.6 The German trial showed a signifi cant reduction in local recurrence and improved tolerability with neoadjuvant versus adjuvant chemoradiotherapy, although there was no diff erence in survival. Our steering committee concluded that it was appropriate to add a neoadjuvant cohort to the present study, since there was no indication that such patients would fare better with regard to 5-year overall survival. In our study, an overall benefi t in 5-year overall survival and 3-year DFS was seen for patients who received capecitabine in both adjuvant and neoadjuvant cohorts. Statistical analyses were done to assess the treatment eff ect separately for each cohort. The HR for treatment eff ect (fl uorouracil vs capecitabine) was similar between cohorts, although slightly smaller in the neoadjuvant cohort. Furthermore, the cohort-adjusted overall HR for treatment eff ect diff ered only marginally from the unadjusted overall treatment eff ect. In summary, Cox analyses showed that overall conclusions regarding the primary endpoint of 5-year overall survival were not aff ected by the cohort eff ect, and that a treatment–cohort interaction can be excluded.

Although this study was designed as a non-inferiority trial, the results indicate improved effi cacy with capecitabine. These improvements are not explained by better treatment adherence: the same proportion of patients completed their scheduled cycles in both treatment groups and in both cohorts. Nor can improved effi cacy be explained by improved local control; local recurrence rates were similar in both treatment groups, suggesting good-quality surgery and radiotherapy, although no formal quality control was done. The capecitabine group showed a signifi cant reduction in distant metastases and improved DFS, suggesting greater systemic effi cacy than with bolus fl uorouracil. The X-ACT study11,16 reached a similar conclusion when comparing capecitabine with bolus fl uorouracil plus folinic acid for adjuvant treatment of stage III colon cancer. The study achieved its primary aim of showing at least equivalence in DFS between capecitabine and fl uorouracil plus folinic acid. Superiority analysis showed a non-signifi cant improvement in 3-year DFS with capecitabine (64% vs 61%; p=0·12).11 Long-term follow-up over a median of 6·9 years and preplanned multi variate analyses showed that capecitabine signifi cantly improved DFS (p=0·02) and overall survival (p=0·02) versus fl uorouracil plus folinic acid.16 Thus, the present study and the X-ACT trial support use of capecitabine over bolus fl uorouracil in the adjuvant treatment of colorectal cancer.

Retrospective analyses of phase 2 studies where capecitabine was given in combination with radiotherapy have shown cumulative pCR rates that are similar to infusional fl uorouracil. Sanghera and colleagues17 found similar pCR rates with capecitabine (17%) and infusional fl uorouracil (20%) in a meta-analysis of 71 trials with a total of 4732 patients. In the present trial, we used the German standard infusional fl uorouracil regimen in

Capecitabine (n=197) Fluorouracil (n=195) p value*

Grade 1–2 Grade 3–4 Total† Grade 1–2 Grade 3–4 Total†

Laboratory

Lowered haemoglobin 58 0 62 49 2 52 0·29

Lowered leucocytes 47 3 50 50 16 68 0·04

Lowered platelets 23 0 23 29 1 32 0·18

Raised creatinine 5 0 5 2 0 2 0·26

Raised bilirubin 6 1 8 1 1 2 0·06

Gastrointestinal

Nausea 33 2 36 30 0 32 0·63

Vomiting 11 1 14 8 1 9 0·30

Diarrhoea 83 17 104 76 4 85 0·07

Mucositis 11 1 12 15 2 17 0·32

Stomatitis 8 0 8 11 0 12 0·35

Abdominal pain 19 1 23 11 0 14 0·13

Proctitis 26 1 31 9 1 10 <0·001

Other

Fatigue 50 0 55 27 2 29 0·002

Anorexia 13 0 13 5 1 6 0·10

Alopecia 4 0 4 11 0 11 0·06

Hand-foot skin reaction 56 4 62 3 0 3 <0·001

Radiation dermatitis 22 2 29 32 1 35 0·39

Data are number of patients or p value. *Derived from χ² test comparing total events between treatment groups. †CTC grade was missing in some patients.

Table 5: Toxic eff ects per treatment group, according to National Cancer Institute Common Toxicity Criteria (CTC) version 2.0

Articles


the neoadjuvant cohort: 120 h continuous infusion of fl uorouracil (1000 mg/m² per day) in weeks 1 and 5 of chemoradiotherapy. Capecitabine improved T down-staging and pCR rate (14% vs 5% with fl uorouracil). National Surgical Adjuvant Breast and Bowel Project (NSABP) trial R-0418 compared protracted venous infusional fl uorouracil (225 mg/m² per day) with capecitabine at the same dose as in the present study, for preoperative treatment of rectal cancer. Addition of oxaliplatin to either regimen was investigated using a two-by-two factorial design. Preliminary data showed a slightly better pCR rate with capecitabine with or without oxaliplatin, than with fl uorouracil with or without oxaliplatin (22% [95% CI 19–26] vs 19% [16–22]; p=0·12). The NSABP R-04 study has not yet presented data on pCR rates with capecitabine or fl uorouracil alone. The study had similar rates of sphincter-sparing surgery in both arms, comparable to the neoadjuvant cohort of the present trial; it also noted similar rates of any surgical complication with capecitabine versus fl uorouracil (37% vs 35%). Capecitabine did not increase postoperative morbidity. In the present trial, we noted higher rates of proctitis and diarrhoea with capecitabine during radiotherapy, the latter increasing by 13% compared with fl uorouracil. Similarly, patients who received capecitabine in NSABP R-04 had slightly more symptoms (particularly diarrhoea) than those who received fl uorouracil, as measured with the fl uoropyrimidine-specifi c symptom checklist.19 Never theless, using the Functional Assessment of Cancer Therapy–Colorectal (FACT-C) trial-outcome index, qual ity of life was identical between groups in NSABP R-04, and capecitabine provided greater convenience of care.19

The general level of toxicity observed in the present trial can be regarded as low to moderate in both treat-ment groups. As expected, toxicity patterns diff ered. Capecitabine caused higher gastrointestinal toxicity during radiotherapy, but the rate of diarrhoea during cycles with chemotherapy alone did not diff er between groups. Hand-foot skin reactions and fatigue were more frequent with capecitabine, and leucopenia was more frequent with fl uorouracil. A third of patients receiving capecitabine had hand-foot skin reactions. In a post-hoc analysis, patients in the capecitabine group who developed any-grade hand-foot skin reactions had signifi cantly better 3-year DFS and 5-year overall survival than capecitabine patients with no hand-foot skin reactions, or the remaining study population. None-theless, 3-year DFS for patients who received capecitabine but did not develop hand-foot skin reactions was much the same as for those who received fl uorouracil (71% [95% CI 63–79] vs 67% [59–73]). In this respect, hand-foot skin reactions might be regarded as a positive pharma-codynamic prognostic marker. The same eff ect was reported in the X-ACT trial (panel),16 and in a randomised phase 2 study using capecitabine as part of combination chemotherapy in patients with metastatic colorectal

cancer.20 Assessing individual dose optimisation for capecitabine according to the presence or absence of hand-foot skin reactions might be a worthwhile strategy for future studies.

Current clinical research focuses on improving fl uorouracil or capecitabine-based neoadjuvant18,21,22 or perioperative treatment23 by adding oxaliplatin, as in the ongoing CAO/ARO/AIO-0423 and PETACC-6 (NCT00766155) trials, which have both completed accrual. Until the fi nal results of both studies are reported, capecitabine can be regarded as an eff ective, well tolerated, and convenient alternative to fl uorouracil in patients undergoing adjuvant or neoadjuvant chemoradiotherapy for locally advanced rectal cancer.

ContributorsR-DH, FW, SP, GH, and AH designed the study. R-DH, FW, SP, AM, SL,

JTH, LM, HL, MM, EK, EF, UH, HWL, MG, SK, CC, MH, GH, DG, PK,

and AH were responsible for treatment of study patients and data

collection. R-DH, FW, IB, and AH did the data analysis and interpretation,

and wrote the report. All authors approved the fi nal version.

Confl icts of interestR-DH has consulted for and received honoraria and research grants

from Roche Pharma AG, and has received honoraria from Chugai

Pharma. AM has consulted for and received honoraria and travel grants

from Roche Pharma AG. JTH and MM have received honoraria from

Panel: Research in context

Systematic reviewUntil the early 2000s, adjuvant chemoradiotherapy was considered standard of care for stage II–III rectal cancer in Germany and other European countries. Several trials and strategies, including biomodulation of fl uorouracil, prolongation of adjuvant treatment, and addition of drugs such as semustine, did not improve on results with fl uorouracil treatment.8,9 At the time the present trial was designed, capecitabine was being investigated as an alternative to infusional fl uorouracil in combination regimens in metastatic colorectal cancer.12 No evidence existed as to whether capecitabine could substitute for fl uorouracil in the perioperative treatment of rectal cancer. No systematic review had been done before the start of the present trial, although the steering committee was unaware of other studies on the same question. After our study began, the German rectal cancer study6 showed that neoadjuvant (vs adjuvant) chemoradiotherapy improved local relapse rates and tolerability, with no diff erence in overall survival. Our steering committee decided to include a neoadjuvant treatment cohort, in the absence of evidence that the timing of chemoradiotherapy aff ected overall survival, our primary endpoint.

InterpretationOur data are similar to those from the X-ACT study,11,16 which compared capecitabine with bolus fl uorouracil in stage III colon cancer. X-ACT also showed the non-inferiority of capecitabine, and both trials showed improvement in disease-free survival in the capecitabine group. In the present trial, an association with better disease-free and overall survival with capecitabine was noted for both cohorts, suggesting the absence of a treatment–cohort interaction. Similarly, the NSABP-R04 trial18,19 recently showed that capecitabine can replace infusional fl uorouracil in neoadjuvant chemoradiotherapy for rectal cancer. Our fi ndings reinforce the evidence that capecitabine can replace fl uorouracil in perioperative and palliative treatment of colorectal cancer. Therefore, clinicians might consider using capecitabine instead of fl uorouracil in adjuvant or neoadjuvant chemoradiotherapy for patients with locally advanced rectal cancer.

Articles


12 Cassidy J, Clarke S, Díaz-Rubio E, et al. Randomized phase III study of capecitabine plus oxaliplatin compared with fl uorouracil/folinic acid plus oxaliplatin as fi rst-line therapy for metastatic colorectal cancer. J Clin Oncol 2008; 26: 2006–12.

13 Glynne-Jones R, Dunst J, Sebag-Montefi ore D, et al. The integration of oral capecitabine into chemoradiation regimens for locally advanced rectal cancer: how successful have we been? Ann Oncol 2006; 17: 361–71.

14 Tunes da Silva G, Logan BR, Klein JP. Methods for equivalence and non-inferiority testing. Biol Blood Marrow Transplant 2009; 15 (suppl 1): 120–27.

15 Machin D, Gardner MJ. Calculating confi dence intervals for survival time analyses. BMJ 1988; 296: 1369–71.

16 Twelves C, Scheithauer W, McKendrick J, et al. Capecitabine versus 5 fl uorouracil/folinic acid as adjuvant therapy for stage III colon cancer: fi nal results from the X-ACT trial with analysis by age and preliminary evidence of a pharmacodynamic marker of effi cacy. Ann Oncol published online Sept 6, 2011. DOI:10.1093/annonc/mdr366.

17 Sanghera P, Wong DW, McConkey CC, Geh JI, Hartley A. Chemoradiotherapy for rectal cancer: an updated analysis of factors aff ecting pathological response. Clin Oncol (R Coll Radiol) 2008; 20: 176–83.

18 Roh MS, Yothers GA, O’Connell MJ, et al. The impact of capecitabine and oxaliplatin in the preoperative multimodality treatment in patients with carcinoma of the rectum: NSABP R-04. Proc Am Soc Clin Oncol 2011; 29 (suppl): abstr 3503.

19 Yothers G, Ganz PA, Lopa SH, Ko CY, Wickerham DL, Wolmark N. Patient-reported outcomes (PROs) comparison of 5 FU and capecitabine (cape) with concurrent radiotherapy (RT) for neoadjuvant treatment of rectal cancer: results of NSABP R-04. Proc Am Soc Clin Oncol 2012; 30 (suppl 4): abstr 391.

20 Stintzing S, Fischer von Weikersthal L, Vehling-Kaiser U, et al. Correlation of capecitabine-induced skin toxicity with treatment effi cacy in patients with metastatic colorectal cancer: results from the German AIO KRK-0104 trial. Br J Cancer 2011; 105: 206–11.

21 Gérard JP, Azria D, Gourgou-Bourgade S, et al. Comparison of two neoadjuvant chemoradiotherapy regimens for locally advanced rectal cancer: results of the phase III trial ACCORD 12/0405-Prodige 2. J Clin Oncol 2010; 28: 1638–44.

22 Aschele C, Cionini L, Lonardi S, et al. Primary tumor response to preoperative chemoradiation with or without oxaliplatin in locally advanced rectal cancer: pathologic results of the STAR-01 randomized phase III trial. J Clin Oncol 2011; 29: 2773–80.

23 Roedel C, Becker H, Fietkau R, et al. Preoperative chemoradiotherapy and postoperative chemotherapy with 5-fl uorouracil and oxaliplatin versus 5-fl uorouracil alone in locally advanced rectal cancer: fi rst results of the German CAO/ARO/AIO-04 randomized phase III trial. Proc Am Soc Clin Oncol 2011; 29 (suppl): LBA3505.

Roche Pharma AG. UH has received honoraria from Chugai Pharma

and Roche Pharma AG. AH has consulted for Ariad, Bristol-Myers

Squibb, Novartis, Merck Sharp & Dohme, and Pfi zer, and has received

research grants from Ariad, Bristol-Myers Squibb, Novartis, and Roche

Pharma AG. All other authors declared no confl icts of interest.

AcknowledgmentsWe thank all investigators and study coordinators. We also thank

Lutz Edler, Annette Kopp-Schneider, and Lothar Pilz (Department of

Biostatistics at German Cancer Research Center, Heidelberg, Germany),

and Petra Mura, Renate Kapaun, and Tanja Groh (trial unit at the Third

Department of Internal Medicine, Mannheim University Hospital,

University of Heidelberg, Mannheim, Germany).

References1 Martling A, Holm T, Rutqvist LE, et al. Impact of a surgical training

programme on rectal cancer outcomes in Stockholm. Br J Surg 2005; 92: 225–29.

2 Sebag-Montefi ore D, Stephens RJ, Steele R, et al. Preoperative radiotherapy versus selective postoperative chemoradiotherapy in patients with rectal cancer (MRC CR07 and NCIC-CTG C016): a multicentre, randomised trial. Lancet 2009; 373: 811–20.

3 van Gijn W, Marijnen CA, Nagtegaal ID, et al. Preoperative radiotherapy combined with total mesorectal excision for resectable rectal cancer: 12-year follow-up of the multicentre, randomised controlled TME trial. Lancet Oncol 2011; 12: 575–82.

4 Gérard JP, Conroy T, Bonnetain F, et al. Preoperative radiotherapy with or without concurrent fl uorouracil and leucovorin in T3-4 rectal cancers: results of FFCD 9203. J Clin Oncol 2006; 24: 4620–25.

5 Bosset JF, Collette L, Calais G, et al. Chemotherapy with preoperative radiotherapy in rectal cancer. N Engl J Med 2006; 355: 1114–23.

6 Sauer R, Becker H, Hohenberger W, et al. Preoperative versus postoperative chemoradiotherapy for rectal cancer. N Engl J Med 2004; 351: 1731–40.

7 Schmiegel W, Reinacher-Schick A, Arnold D. Update S3-guideline colorectal cancer 2008. Z Gastroenterol 2008; 46: 799–840 (in German).

8 Tepper J, O’Connell MJ, Niedzwiecki D, et al. Adjuvant therapy in rectal cancer: analysis of stage, sex, and local control—fi nal report of intergroup 0114. J Clin Oncol 2002; 20: 1744–50.

9 O´Connell MJ, Martenson JA, Wieand HS, et al. Improving adjuvant therapy for rectal cancer by combining protracted-infusion fl uorouracil with radiation therapy after curative surgery. N Engl J Med 1994; 331: 502–07.

10 Kalofonos HP, Bamias A, Koutras A, et al. A randomised phase III trial of adjuvant radio-chemotherapy comparing irinotecan, 5FU and leucovorin to 5FU and leucovorin in patients with rectal cancer: a Hellenic Cooperative Oncology Group Study. Eur J Cancer 2008; 44: 1693–700.

11 Twelves C, Wong A, Nowacki MP, et al. Capecitabine as adjuvant treatment for stage III colon cancer. N Engl J Med 2005; 352: 2696–704.


Articles

Adjuvant radiotherapy versus observation alone for patients at risk of lymph-node fi eld relapse after therapeutic lymphadenectomy for melanoma: a randomised trial*Bryan H Burmeister, *Michael A Henderson, Jill Ainslie, Richard Fisher, Juliana Di Iulio, B Mark Smithers, Angela Hong, Kerwin Shannon, Richard A Scolyer, Scott Carruthers, Brendon J Coventry, Scott Babington, Joao Duprat, Harald J Hoekstra, John F Thompson

SummaryBackground The use of radiotherapy after therapeutic lymphadenectomy for patients with melanoma at high risk of further lymph-node fi eld and distant recurrence is controversial. Decisions for radiotherapy in this setting are made on the basis of retrospective, non-randomised studies. We did this randomised trial to assess the eff ect of adjuvant radiotherapy on lymph-node fi eld control in patients who had undergone therapeutic lymphaden ectomy for metastatic melanoma in regional lymph nodes.

Methods This randomised controlled trial included patients from 16 hospitals in Australia, New Zealand, the Netherlands, and Brazil. To be eligible for this trial, patients had to be at high risk of lymph-node fi eld relapse, judged on the basis of number of nodes involved, extranodal spread, and maximum size of involved nodes. After lymphadenectomy, randomisation was done centrally by computer and patients assigned by telephone in a ratio of 1:1 to receive adjuvant radiotherapy of 48 Gy in 20 fractions or observation, with institution, lymph-node fi eld, number of involved nodes, maximum node diameter, and extent of extranodal spread as minimisation factors. Participants, those giving treatment, and those assessing outcomes were not masked to treatment allocation. The primary endpoint was lymph-node fi eld relapse (as a fi rst relapse), analysed for all eligible patients. The study is registered at ClinicalTrials.gov, number NCT00287196. The trial is now closed and follow-up discontinued.

Findings 123 patients were randomly allocated to the adjuvant radiotherapy group and 127 to the observation group between March 20, 2002, and Sept 21, 2007. Two patients withdrew consent and 31 had a major eligibility infringement as decided by the independent data monitoring committee, resulting in 217 eligible for the primary analysis (109 in the adjuvant radiotherapy group and 108 in the observation group). Median follow-up was 40 months (IQR 27–55). Risk of lymph-node fi eld relapse was signifi cantly reduced in the adjuvant radiotherapy group compared with the observation group (20 relapses in the radiotherapy group vs 34 in the observation group, hazard ratio [HR] 0·56, 95% CI 0·32–0·98; p=0·041), but no diff erences were noted for relapse-free survival (70 vs 73 events, HR 0·91, 95% CI 0·65–1·26; p=0·56) or overall survival (59 vs 47 deaths, HR 1·37, 95% CI 0·94–2·01; p=0·12). The most common grade 3 and 4 adverse events were seroma (nine in the radiotherapy group vs 11 in the observation group), radiation dermatitis (19 in the radiotherapy group), and wound infection (three in the radiotherapy group vs seven in the observation group).

Interpretation Adjuvant radiotherapy improves lymph-node fi eld control in patients at high risk of lymph-node fi eld relapse after therapeutic lymphadenectomy for metastatic melanoma. Adjuvant radiotherapy should be discussed with patients at high risk of relapse after lymphadenectomy.

Funding National Health and Medical Research Council of Australia, Cancer Australia, Melanoma Institute Australia, Cancer Council of South Australia.

IntroductionStandard treatment protocols for many solid tumours include adjuvant radiotherapy on the basis of data that show improved local control and, in some situations, improved survival. For primary cutaneous melanoma, the commonest and usually fi rst site of recurrence after defi nitive excision of the primary tumour is in the draining lymph-node fi eld. After therapeutic lymph adenectomy for isolated lymph-node fi eld relapse, patients with substantial disease burden in the regional lymph-node fi eld have a high risk of recurrence,1–3 which can cause morbidity including pain, ulceration, malodour, lymph oedema, and impaired function, particularly in the leg. Further

lymph-node fi eld relapse is predicted by extranodal spread of melanoma, increased number of tumour-positive lymph nodes, and increasing size of involved nodes.4–6 The use of adjuvant radiotherapy after lymphadenectomy to reduce the risk of further relapse is controversial. In many centres it is recommended in patients considered to be at high risk of further relapse; however, evidence from randomised studies is scarce. In 1993, the Radiation Therapy Oncology Group initiated a randomised trial of post-operative adjuvant radiotherapy (RTOG 93.02), but the trial was halted after failure to accrue suffi cient patients, with no results reported. Some retrospective single institution studies and several reviews suggest that


See Online/CommentDOI:10.1016/S1470-2045(12)70165-1

*Both authors contributed equally

Princess Alexandra Hospital, University of Queensland, Brisbane, Australia (Prof B H Burmeister MD, B M Smithers FRACS); Peter MacCallum Cancer Centre (Prof M A Henderson MD, J Ainslie FRANZCR), Centre for Biostatistics and Clinical Trials (R Fisher PhD), University of Melbourne, Melbourne, Australia; Centre for Biostatistics and Clinical Trials, Melbourne, Australia (J Di Iulio PhD); Melanoma Institute of Australia, University of Sydney, Sydney, Australia (A Hong PhD, K Shannon FRACS, Prof R A Scolyer MD, Prof J F Thompson MD); Royal Adelaide Hospital, University of Adelaide, Adelaide, Australia (S Carruthers FRANZCR, B J Coventry FRACS); Christchurch Hospital, Christchurch, New Zealand (S Babington FRANZCR); Hospital Do Cancer, Sao Paulo, Brazil (Prof J Duprat PhD); and University Medical Center Groningen, Groningen, Netherlands (Prof H J Hoekstra MD)

Correspondence to:Prof Bryan H Burmeister, Division of Cancer Services, Princess Alexandra Hospital, Woolloongabba, Brisbane 4102, [email protected]

Articles


radiotherapy after therapeutic lymphadenectomy im-proves lymph-node fi eld control; however, the eff ect of radiotherapy on survival is much less clear.7–19

In 1996, the Trans-Tasman Radiation Oncology Group (TROG) initiated a phase 2 trial20 (TROG 96.06) of adjuvant radiotherapy for melanoma patients with completely resected lymph-node disease judged to be at substantial risk of further melanoma relapse. Regional control was good, with few toxic eff ects.20,21

With the same radiotherapy regimen as TROG 96.06, we did a randomised trial to compare adjuvant radiotherapy with observation alone in patients at high risk of lymph-node fi eld relapse who had undergone therapeutic lymphadenectomy for metastatic melanoma in regional lymph nodes. The study was started in 2002, by the Australian and New Zealand Melanoma Trials Group and TROG (ANZMTG 01.02/TROG 02.01).

MethodsPatientsPatients from 16 hospitals in Australia, four in New Zealand, one in the Netherlands, and one in Brazil were

considered for inclusion. Patients were eligible if they had palpable metastatic lymph-node fi eld disease; had a complete cervical, axillary, or inguinal lymphadenectomy; were at high risk of further lymph-node fi eld relapse; had an ECOG performance status of 0 or 1; were aged 18 years or older; had a life expectancy in the absence of melanoma of 2 years or more; were staged (by CT scan of lymph node fi eld, chest, abdomen, or pelvis, and CT or MRI of brain); had a serum lactate dehydrogenase (LDH) concentration of less than 1·5-times the upper limit of normal; had normal full blood count and biochemistry; and had given informed consent. High risk of further lymph-node fi eld relapse was defi ned as any one of the following factors: one or more involved parotid nodes, two or more involved cervical or axillary nodes, or three or more involved inguinal nodes; presence of extranodal tumour spread; or the maximum diameter of the largest metastatic lymph node was 3 cm or more (for a cervical node) or 4 cm or more (for an axillary or inguinal node). The ethics committees of all participating centres approved the study. Before random isation the involved lymph-node fi eld (including the nodal basin and the tissues to be targeted for radiotherapy) was photographed, with protocol-defi ned anatomical boundaries drawn on the skin. The photograph assisted defi nition of subsequent lymph-node fi eld relapse versus non-lymph node fi eld relapse (local, in transit, or distant).

Patients were ineligible if they had a concurrent or previous history of local, in transit, or distant relapse; had impalpable (including detected by sentinel-node biopsy) lymph-node fi eld relapse; or had had cancer previously (unless diagnosed more than 5 years before with an estimated risk of recurrence of less than 10%). PET scanning was permitted in centres where it was routine practice. Adjuvant systemic therapy, including interferon, was permitted although the protocol stipu lated that no cytotoxic drugs be given during, or close to radiotherapy.

Randomisation and maskingWe randomly assigned eligible patients centrally with a computer program (with our own algorithm, constructed in Microsoft Access) in a 1:1 ratio to receive adjuvant radiotherapy or observation by a telephone call to the trial centre. Randomisation was done with minimisation using a random component, with balancing factors of institution, lymph-node fi eld (parotid and cervical, axilla, or groin), number of involved nodes (≤three or >three), involved node maximum diameter (≤4 cm or >4 cm), and extent of extranodal spread (none, limited, or extensive).

ProceduresAdjuvant radiotherapy was started within 12 weeks of lymphadenectomy. The dose was 48 Gy given in 20 fractions over 4 weeks. Treatment was given over a maximum of 30 days, by diff erent techniques for each nodal site. The planned treatment volume had prespecifi ed boundaries that included the dissected

Adjuvant radiotherapy group (n=17)

Observation group (n=24)

Previous local or in-transit relapse 5 8

Non-palpable nodes 5 7

Local or in-transit relapse at randomisation

1 5

Two node fi elds involved 3 1

Distant metastases at randomisation 2 1

Incompletely resected nodal disease 1 0

Previous nodal surgery 0 1

No melanoma in lymph nodes 0 1

Table 1: Major eligibility infringements

250 patients randomly assigned

127 allocated to observation 1 received radiotherapy 126 did not receive radiotherapy

123 allocated to receive adjuvant radiotherapy 115 received radiotherapy 8 did not receive radiotherapy

122 analysed by intention to treat 126 analysed by intention to treat

109 eligible, analysed per protocol 108 eligible, analysed per protocol

1 withdrew consent

18 had major eligibility infringements

1 withdrew consent

13 had major eligibility infringements

Figure 1: Trial profi le

Articles


lymph-node fi eld and lymphadenectomy scar. For the axillary and ilio-inguinal lymph-node fi elds multiple photon fi elds were recommended, but in the head and neck region both photon and electron plans were allowed. Maximum doses were stipulated for the spinal cord (40 Gy), brachial plexus (45 Gy), and femoral neck (40 Gy). In patients who were assigned to observation and later developed isolated lymph-node fi eld relapse, resection with postoperative radiotherapy was permitted.

Patients were followed up once every 3 months for 2 years, then every 6 months until 5 years, and annually thereafter. We assessed surgical morbidity, quality of life, and limb circumference at baseline and at each visit. Patients assigned to radiotherapy were assessed for acute radiation toxic eff ects with the Radiation Therapy Oncology Group criteria22 at 2 and 6 weeks after completion of radiotherapy, and at the fi rst 3 month follow-up visit. A full blood cell count, serum urea and

Intention-to-treat population Eligible population





Sex

Male 92 (75%) 92 (73%) 83 (76%) 81 (75%)

Female 30 (25%) 34 (27%) 26 (24%) 27 (25%)

Age at randomisation (years)

Median (IQR; range) 58 (46–69; 22–80) 56 (45–68; 22–87) 59 (47–69; 22–80) 56 (45–68; 25–87)

ECOG performance status

0 88 (72%) 90 (71%) 77 (71%) 74 (69%)

1 34 (28%) 36 (29%) 32 (29%) 34 (31%)

Lymph-node fi eld

Head and neck 31 (25%) 34 (27%) 29 (27%) 30 (28%)

Axilla 52 (43%) 52 (41%) 45 (41%) 44 (41%)

Groin 39 (32%) 40 (32%) 35 (32%) 34 (31%)

Node maximum diameter

≤4 cm 75 (61%) 80 (63%) 64 (59%) 66 (61%)

>4 cm 47 (39%) 46 (37%) 45 (41%) 42 (39%)

Number of positive nodes

Head and neck

Mean (SD) 1·90 (1·60) 3·32 (5·43) 1·93 (1·65) 3·53 (5·75)

Median (IQR; range) 1 (1–2; 1–9) 2 (1–3; 1–31) 1 (1–2; 1–9) 2 (1–3; 1–31)

Axilla

Mean (SD) 3·79 (4·38) 4·23 (7·07) 3·96 (4·69) 4·55 (7·57)

Median (IQR; range) 3 (1–4; 1–25) 2 (1–4; 1–47) 3 (1–5; 1–25) 2 (1–4; 1–47)

Groin

Mean (SD) 3·54 (2·26) 3·35 (2·25) 3·66 (2·35) 3·18 (2·19)

Median (IQR; range) 3 (2–4; 1–12) 3 (2–4·25; 1–10) 3 (2–4·5; 1–12) 2·5 (2–4; 1–10)

Extent of extranodal spread

None 61 (50%) 62 (49%) 51 (47%) 56 (52%)

Limited 37 (30%) 44 (35%) 36 (33%) 35 (32%)

Extensive* 24 (20%) 20 (16%) 22 (20%) 17 (16%)

Primary site

Head and neck 19 (16%) 22 (17%) 18 (17%) 20 (19%)

Trunk 37 (30%) 27 (21%) 30 (28%) 22 (20%)

Arm 17 (14%) 22 (17%) 15 (14%) 16 (15%)

Leg 29 (24%) 30 (24%) 26 (24%) 25 (23%)

Unknown 20 (16%) 25 (20%) 20 (18%) 25 (23%)

Use of interferon

None 118 (97%) 120 (95%) 105 (96%) 103 (95%)

Planned or started at randomisation 4 (3%) 6 (5%) 4 (4%) 5 (5%)

*Matted nodes, multiple node extracapsular spread, or multiple deposits of lymphovascular space invasion or multiple deposits of tumour in fat or connective tissue unrelated to nodes or vessels.

Table 2: Baseline characteristics of the intention-to-treat and eligible populations

Articles


creatinine concentrations, liver function tests, and CT scans of the head, chest, abdomen, pelvis, and involved lymph-node fi eld were done by the treating clinician annually.

The trial was overseen by an independent data monitoring committee, which consisted of a surgeon, a radiation oncologist, and a statistician. Quality assess-ments of the trial data were done throughout the trial. These reviews included eligibility criteria, pathology reports, operation reports, photographs of the lymph-node fi eld taken before randomisation (inadequate

photographs were retaken), informed consent forms, primary melanoma details, lymph-node fi eld relapse (using original and relapse photographs), other sites of relapse (local, in transit, and distant), and verifi cation of date and cause of death.

The primary endpoint was lymph-node fi eld relapse (as fi rst relapse). Secondary objectives were measurement of acute toxic eff ects, relapse-free survival, and overall survival. Late toxic eff ects, quality of life, and long-term survival will be assessed in later reports.

Statistical analysisThe trial was designed to have a target sample size of 270 patients, but included a planned reassessment of the sample size on the basis of the lymph-node fi eld relapse rate in both groups. This reassessment (overseen by two independent experienced clinical researchers) resulted in a decrease of the target sample size because of a slower accrual rate than was expected and because the lymph-node fi eld relapse rate for both groups combined was higher than expected. Consequently, we amended the protocol twice: the fi rst introduced an interim analysis for effi cacy based on the primary endpoint. The interim α was 0·048, based on the proportion of the total number of lymph-node fi eld relapses and the O’Brien-Fleming spending function. After the interim analysis, the data monitoring committee recommended continuation of the trial to full accrual. The second amendment occurred after the identifi cation of eligibility infringements and permitted an increase in the sample size accordingly.

The fi nal target sample size was 250 patients, on the basis of the assumption that 40% of patients in the observation group would have a lymph-node fi eld relapse according to an exponential distribution and that 70% of these relapses would occur within 1 year. The trial was designed to have 80% power to detect a 20% diff erence in the proportion of patients with lymph-node fi eld relapse beween groups (20% in the radiotherapy group vs 40% in the observation alone group), corresponding to a hazard ratio (HR) of 0·437. 48 lymph-node fi eld relapses were needed to ensure adequate power. We initially planned to do effi cacy analyses in the intention-to-treat population. However, during the course of the trial, the independent data monitoring committee identifi ed several eligibility infringements. On the advice of the committee, a formal independent, external review was done for all patients for whom eligibility concerns existed. The eligibility review panel consisted of a radiation oncologist and a surgical oncologist who were masked to the patient’s institution, assignment, and outcome. This review iden tifi ed 41 major eligibility infringements in 31 patients (table 1). The committee recommended that these patients be excluded from the primary analysis. Thus, two study populations were analysed; the intention-to-treat population, and the eligible population.

We used the Kaplan-Meier method to estimate curves for time to lymph-node fi eld relapse, overall survival,

0

10

20

30

40

50

60

70

80

90

100

Prop

ortio

n fre

e of l

ymph

-nod

e fie

ld re

laps

e (%

)

Time since randomisation (years)

0

10

20

30

40

50

60

70

80

90

100

Prop

ortio

n fre

e of l

ymph

-nod

e fie

ld re

laps

e (%

)

5421 30 6

5421 30 6 Number at risk Adjuvant radiotherapy 109 59 39 20 8 5 Observation 108 59 31 20 6 4

Number at risk Adjuvant radiotherapy 122 66 44 24 10 6 Observation 126 68 36 22 7 4

B

A

Adjuvant radiotherapy groupObservation group

HR 0·56; 95% CI 0·32–0·98; p=0·041

HR 0·47; 95% CI 0·28–0·81; p=0·005

Figure 2: Kaplan-Meier curves of time to lymph-node fi eld relapse by treatment groupIn the eligible population (A) and the intention-to-treat population (B). Hazard ratios are for adjuvant radiotherapy versus observation.

Articles


and relapse-free survival. The start date for all time-to-event outcomes was the date of randomisation. Time to lymph-node fi eld relapse was defi ned as time from start date to lymph-node fi eld relapse, either alone or concurrent with relapse at any other site. Time to lymph-node relapse was censored by relapse in other sites, the cutoff date, and loss to follow-up. Overall survival was defi ned as time from start date to death from any cause and relapse-free survival was defi ned as relapse at any site or death without previous relapse. Both measures were censored by the cutoff date and loss to follow-up. We used a cutoff date for follow-up of Nov 15, 2008, and all living patients were followed up to this date; any follow-up after this date was ignored to minimise reporting bias. Potential follow-up time was defi ned for each patient as time from entry to the trial to the cutoff date, or to the last follow-up date for patients lost to follow-up. We analysed time-to-event with unadjusted and stratifi ed log-rank tests and Cox regression methods, including testing model assumptions. We did two-sample comparisons with unadjusted and stratifi ed exact log-rank tests based on hypergeometric probabilities. We analysed patterns of failure with a competing risks method, presented with cumulative incidence curves and compared with Gray’s test.

To assess the eff ects of potential prognostic factors in the eligible population for lymph-node fi eld relapse and overall survival we did a multivariable analysis including lymph-node fi eld site (head and neck, axilla, or groin), status of primary lesion (known or unknown primary), extranodal spread (none, limited, or extensive; coded as 0, 1, or 2 and analysed as a continuous variable), and number of positive nodes (one, two or three, or four or more; coded as 1, 2, or 3 and analysed as a continuous variable).

All p values are two sided. All analyses were done with the R statistical package (version 2.14.0). This study is registered with ClinicalTrials.gov, number NCT00287196.

Role of the funding sourceThe sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to submit for publication.

ResultsFrom March 20, 2002, to Sept 21, 2007, 250 patients from 16 institutions (six did not accrue any patients) in Australia (222 patients), New Zealand (14 patients), the Netherlands (12 patients), and Brazil (two patients) were randomly assigned. The accrual rate (average 3·7 patients per month) was generally uniform throughout the accrual period. Two patients (one from each group) withdrew consent soon after randomisation and were excluded from all analyses (fi gure 1). The intention-to-treat population thus consisted of 248 patients, while the eligible population consisted of 217 patients.

Median potential follow-up in the intention-to-treat population was 40 months (IQR 27–55) and, in patients not lost to follow-up, ranged from 14 to 80 months. Two patients in the observation group were lost to follow-up, after 13 and 21 months. Table 2 shows baseline char-acteristics, including patient demographics, for the eli gible and intention-to-treat populations. Char acteristics were much the same in both groups and in both analysis populations. Median time from primary melanoma

Cum

ulat

ive

incid

ence

(%)

Cum

ulat

ive

incid

ence

(%)

Time since randomisation (years)21 30

40

30

20

10

0

50

40

30

20

10

0

B

AAdjuvant radiotherapy, lymph-node field onlyObservation, lymph-node field onlyAdjuvant radiotherapy, distant onlyObservation, distant onlyAdjuvant radiotherapy, lymph-node field and distantObservation, lymph-node field and distant


Figure 3: Cumulative incidence curves of relapse(A) Curves of competing events: isolated lymph-node fi eld relapse, isolated distant relapse, and concurrent lymph node and distant relapse in each group in the 217 patients in the eligible population. Additionally (data not shown), nine patients had isolated local or in-transit disease and one patient died without relapse (local or in-transit disease concurrent with lymph-node fi eld and distant relapse are included in presented curves). (B) Curves for lymph-node fi eld relapse, with or without concurrent relapse in other sites, as a fi rst relapse, in each group in the eligible population.

For the statistical package see http://www.R-project.org/

Articles


diagnosis to diagnosis of isolated lymph-node fi eld relapse in the eligible population was 16 months (IQR 5–41).

Of the 122 patients due to to receive adjuvant radio-therapy in the intention-to-treat population, seven did not (four refused and three developed pro gressive disease before starting radiotherapy). Of the remaining 115, 109 (95%) received the prescribed dose of 48 Gy, fi ve (4%) received less than 48 Gy, and one (<1%) received more than 48 Gy. 99 patients (86%) received their treatment without interruption and 16 (14%) had interruptions of 1–11 days because of machine malfunc tion, public holidays, personal circumstances, concur rent interferon administration, or unplanned hospital admissions for other disorders. 54 patients had their compliance to the radiotherapy protocol reviewed; compliance (defi ned as complete adherence to the radiotherapy protocol) was 79%.

In the eligible population, 54 of 217 patients had lymph-node fi eld relapse as a fi rst relapse. Fewer patients in the radiotherapy group (20 of 109) relapsed than did those in the observation group (34 of 108; HR 0·56, 95% CI 0·32–0·98; p=0·041; fi gure 2A). Results were similar for the analysis adjusted for lymph-node fi eld (head and neck, axilla, or groin; HR 0·56, 95% CI 0·33–0·96; p=0·041). Higher risk of lymph-node fi eld relapse in the observation group than in the radiotherapy group was consistent across all nodal sites (data not shown). The risk of lymph-node fi eld relapse in the intention-to-treat population was comparable with that in the eligible population (HR 0·47, 95% CI 0·28–0·81; p=0·005; fi gure 2B). Of 23 eligible patients in the observation group who had an isolated lymph-node fi eld relapse as a fi rst relapse, 18 received salvage therapy and seven of these patients have died. Survival from fi rst relapse in all eligible patients who relapsed at any site was 44% (95% CI 36–53) at 1 year, 24% (17–34) at 2 years,

and 18% (11–27) at 3 years; median survival was 10 months (IQR 7–13).

Of the 217 eligible patients, 142 relapsed with melanoma (69 in the radiotherapy group, 73 in the observation group) and one assigned to the radiotherapy group died without having relapsed. Figure 3A shows cumulative incidence curves for diff erent fi rst relapse sites in the eligible population, and sites of relapse are shown in table 3. At 3 years, the cumulative incidence of lymph node relapse was 19% (95% CI 11–27) in the radiotherapy group and 31% (20–40) in the observation group. Relapse-free survival did not diff er between in the radiotherapy group compared with the observation group (HR 0·91, 95% CI 0·65–1·26; p=0·56; fi gure 4A). Median recurrence-free survival was 15 months (95% CI 11–27) in the radiotherapy group and 14 months (95% CI 9–23) in the observation group. The results for relapse-free survival were similar in the intention-to-treat population (HR 0·90, 95% CI 0·66–1·22; p=0·53); 163 of the 248 patients had a relapse at any site (n=161; table 3) or death without relapse (two patients in the radiotherapy group). Time to distant relapse (as a fi rst relapse) for the eligible patients did not diff er signifi cantly between treatment groups (HR 1·06, 95% CI 0·72–1·57; p=0·77). Time to any distant relapse was almost identical for the two groups (HR 1·00, 0·71–1·41; p=0·8). Results were much the same for the intention-to-treat population (data not shown).

106 eligible patients died by the data cutoff , 59 in the radiotherapy group and 47 in the observation group (including three deaths that were not from melanoma). This diff erence was not statistically signifi cant (HR 1·37, 95% CI 0·94–2·01; p=0·12; fi gure 4B). Median survival was 32 months (95% CI 20 to not yet reached) for patients in the radiotherapy group and 47 months (95% CI 30 to not yet reached) for patients in the observation group. 120 patients in the intention-to-treat population died, 66 in the radiotherapy group and 55 in the observation group, all but two (one in each group) of melanoma; the diff erence was not statistically signifi cant (HR 1·35, 95% CI 0·94–1·92; p=0·12).

The most common early surgery-related adverse events were seroma formation in the groin or axilla, infection, nerve damage, wound necrosis, and local pain (table 4). For patients in the radiotherapy group, the most commonly early reported toxic eff ects were radiation dermatitis, and pain (table 4).

In a multivariable analysis of potential prognostic factors, extranodal spread (none vs limited vs extensive) was the only independent risk factor for lymph-node fi eld relapse (HR 1·77 per degree of spread, 95% CI 1·26–2·49; p=0·001 [HR lymph-node fi eld relapse for adjuvant radiotherapy vs observation 0·53, 95% CI 0·30–0·92; p=0·025]). Seven of 51 patients (14%) in the radiotherapy group with no extranodal spread had lymph-node fi eld relapse, compared with six of 36 (17%) with limited extranodal spread, and seven of 22 (32%)

Intention-to-treat population

Eligible population





Lymph-node fi eld only 8 26 8 23

Local or in-transit only 7 3 7 2

Lymph-node fi eld and local or in-transit 0 2 0 0

Distant only 49 39 42 37

Lymph-node fi eld and distant 11 11 11 9

Distant and local or in-transit 0 1 0 0

Lymph-node fi eld and distant and local or in-transit

1 3 1 2

Total relapsed 76 85 69 73

No relapse 46 41 40 35

Of the 161 patients in the intention-to-treat population who had a fi rst relapse, 132 relapsed in one site only (64 in the radiotherapy group vs 68 in the observation group) and 29 relapsed in multiple sites (two sites, 11 vs 14; or all three sites, 1 vs 3). Of the 142 patients in the eligible population who had a fi rst relapse, 119 relapsed in one site only (57 vs 62) and 23 relapsed in multiple sites (two sites, 11 vs 9; or all three sites, 1 vs 2).

Table 3: Sites of fi rst relapse

Articles


with extensive extranodal spread. Extranodal spread (HR 1·66 per degree of spread, 95% CI 1·30–2·13, p=0·0001) and number of positive nodes involved (one vs two to three vs four or more; HR 1·45 per level, 95% CI 1·13–1·86, p=0·004) were independently predictive of poor overall survival. Treatment group, adjusting for these two factors, was not signifi cantly related to survival (HR 1·36, 95% CI 0·92–2·00; p=0·12).

DiscussionThe results of our study suggest that the use of adjuvant radiotherapy after complete surgical resection for isolated lymph-node fi eld melanoma relapse in patients at high risk of relapse substantially reduced the risk of further lymph-node fi eld relapse, although there was no eff ect on overall survival. As used in this trial, adjuvant radiotherapy was associated with acceptable early toxic eff ects.

The fi rst study to investigate the role of adjuvant radiotherapy in melanoma patients with resected lymph-node fi eld recurrences was a randomised study of 56 patients done in 1978.23 The duration of follow-up was short, the radiation therapy used was suboptimum by current standards, and the rate of lymph-node fi eld relapse was not reported. The investigators attributed the trend towards improved survival in patients receiving radiotherapy to an imbalance of prognostic factors between the two groups (age and the number of lymph nodes involved), leading them to conclude that radiotherapy had no eff ect on outcome. Since this report, several groups have published retrospective reviews which, in most but not all cases, indicate improved lymph-node fi eld control with adjuvant radiotherapy compared with surgery alone.10–21 Results of several systematic reviews have also shown that adjuvant radiotherapy reduces lymph-node fi eld relapse.24,25 How ever, other reports indicate no signifi cant improvement in lymph-node fi eld control for radiotherapy after lymphadenectomy (panel).26,27

The evidence supporting a survival benefi t for patients receiving adjuvant radiotherapy is even more controversial and confl icting.17 Although many major melanoma treatment centres around the world off er adjuvant radiotherapy for selected patients, many others are still cautious because of the absence of a clear survival benefi t in patients with a very high risk of succumbing to the disease and concerns about the possibility of long-term morbidity associated with radiotherapy. Lymph oedema is the most common long-term problem, aff ecting many patients receiving adjuvant radiotherapy after axillary or inguinal lymphadenectomy.10,11,13,21 No comprehensive, prospective assessments of long-term radiotherapy-associated complications or eff ects of radiotherapy on quality of life exist. Our study closed in November, 2011, and analyses of long-term toxic eff ects and quality-of-life outcomes are not yet complete.

Concerns about long-term morbidity caused by radio-therapy restricts the use of adjuvant radiotherapy in most centres to patients at high risk of lymph-node fi eld relapse after lymphadenectomy. These risks are well documented; they are principally related to the number and maximum size of tumour-involved nodes, and the presence of extranodal tumour extension.4–6 Our results show that radiotherapy aff ects the timing and pattern of fi rst relapse and reduces the risk of lymph-node fi eld relapse. Overall survival was not signifi cantly diff erent between groups.

0

10

20

30

40

50

60

70

80

90

100

Rela

pse-

free

surv

ival

(%)

0

10

20

30

40

50

60

70

80

90

100

Ove

rall

surv

ival

(%)

Time since randomisation (years)421 30

421 30

B

A

Number at risk Adjuvant radiotherapy 109 59 39 20 8 Observation 108 59 31 20 6

Number at risk Adjuvant radiotherapy 109 89 54 35 18 Observation 109 84 49 28 11


HR 0·91; 95% CI 0·65–1·26; p=0·56

HR 1·37; 95% CI 0·94–2·01; p=0·12

Figure 4: Survival by group in the eligible populationRelapse-free survival (A) and overall survival (B). Hazard ratios are for adjuvant radiotherapy versus observation.

Articles


Some centres use hypofractionated regimens for melanoma rather than traditional standard fractionation schedules, largely on the basis of radiobiological data from in-vitro studies, which suggested that melanoma was relatively radioresistant compared with other types of cancer, but might respond to higher dose-per-fraction schedules.28 However, more recent radio biological and clinical studies suggest that most melanomas are radiosensitive.29 Hypofractionated regimens have not been directly compared with standard fraction regimens in the adjuvant setting. Hypo fractionation regimens might be associated with increased risk of long-term tissue eff ects, notably lymphoedema.10,11 Improved selection of patients for adjuvant systemic therapies or adjuvant radiotherapy is clearly a priority. High standard uptake values in FDG-PET and raised serum S100 concentration are useful predictors of outcome and potential of metastasis in operable stage 3 melanoma.30 Several new targeted therapies—eg, ipilimumab and verumafenib—are highly eff ective in patients with

advanced melanoma. These agents will likely be investigated in adjuvant studies of patients at high risk of locoregional and distant recurrence.

A limitation of our study was the high number of ineligible patients, which reduced the power of the study. Our future studies are centred on further exploring the role of radiotherapy in the preoperative setting in which modern imaging modalities are able to identify high risk patients before surgery.

ContributorsBHB, MAH, and JFT designed the study, developed the protocol,

recruited patients, and drafted the Article. JA designed the study,

developed the protocol, and recruited patients. RF designed the study,

developed the protocol, analysed data, and drafted the Article. JDI

coordinated the trial and collected data, was involved in quality

assurance, and drafted the Article. BMS, KS, SC, BJC, SB, JD, and HJH

were involved in patient accrual. AH was involved in patient accrual and

developed the protocol. RAS was involved in pathology review.

Confl icts of interestWe declare that we have no confl icts of interest.

AcknowledgmentsWe thank all participating centres for their contribution to this study.

For a list of collaborators, see appendix. The study was funded by the

National Health and Medical Research Council of Australia, Cancer

Australia, Melanoma Institute Australia, and the Cancer Council of

South Australia.

References1 Byers RM. The role of modifi ed neck dissection in the treatment of

cutaneous melanoma of the head and neck. Arch Surg 1986; 121: 1338–41.

2 Balch CM, Soong SJ, Murad TM, Ingalls AL, Maddox WA. A multifactorial analysis of melanoma III: prognostic factors in melanoma patients with lymph node metastases. Ann Surg 1983; 193: 377–88.

3 Karakousis CP, Seddig MK, Moore R. Prognostic value of lymph node dissection in malignant melanoma. Arch Surg 1980; 115: 719–22.

4 Burmeister BH, Smithers BM, Poulsen M, et al. Radiation therapy for nodal disease in malignant melanoma. World J Surg 1995; 19: 369–71.

5 O’Brien CJ, Petersen-Schaefer K, Stevens GN, et al. Adjuvant radiotherapy following neck dissection and parotidectomy for metastatic melanoma. Head Neck 1997; 19: 589–94.

6 Corry J, Smith JG, Bishop M, Ainslie J. Nodal radiation therapy for metastatic melanoma. Int J Radiat Oncol Biol Phys 1999; 44: 1065–69.

7 Stevens G, Thompson JF, Firth I, et al. Locally advanced melanoma: results of postoperative hypofractionated radiation therapy. Cancer 2000; 88: 88–94.

8 Ballo MT, Strom EA, Zagars GK, et al. Adjuvant irradiation for axillary metastases from malignant melanoma. Int J Radiat Oncol Biol Phys 2002; 52: 964–72.

9 Ballo MT, Bonnen MD, Garden AS, et al. Adjuvant irradiation for cervical lymph node metastases from melanoma. Cancer 2003; 97: 1789–96.

10 Ballo MT, Zagars GK, Gershenwald JE, et al. A critical assessment of adjuvant radiotherapy for inguinal lymph node metastases from melanoma. Ann Surg Oncol 2004; 11: 1079–84.

11 Ballo MT, Ross MI, Cormier JN, et al. Combined modality therapy for patients with regional nodal metastases from melanoma. Int J Radiat Oncol Biol Phys 2006; 64: 106–13.

12 Hamming-Vrieze O, Balm AJ, Heemsbergen WD, et al. Regional control of melanoma neck node metastasis after selective neck dissection with or without adjuvant radiotherapy. Arch Otolaryngol Head Neck Surg 2009; 135: 795–800.

13 Conill C, Valduvieco I, Domingo-Domenech J, et al. Loco-regional control after postoperative radiotherapy for patients with regional nodal metastases from melanoma. Clin Transl Oncol 2009; 11: 688–93.


Systematic reviewTo assess the evidence supporting the use of adjuvant radiotherapy in melanoma, we searched Cochrane, PubMed, and Medline with the terms “melanoma, regional recurrence, lymph-node fi eld recurrence and melanoma”, for studies in English published between 1970, and 2001. We did not identify any completed randomised trials with the primary endpoint of lymph-node fi eld relapse. Data from retrospective studies and one prospective study suggest reduction in the risk of lymph-node fi eld relapse by about 50%. Three reviews published after 2001 support this assessment.

ImplicationsThis report confi rms that adjuvant radiotherapy reduces the risk of further lymph-node fi eld relapse after lymphadenectomy in patients at high risk of relapse, although it had no signifi cant eff ect on overall survival. Early toxic eff ects related to radiotherapy were infrequent and minor. If the intention of treatment is to reduce the risk of regional recurrence, adjuvant radiotherapy is a treatment option that should be discussed with patients at high risk of lymph-node fi eld relapse after lymphadenectomy.

Head and neck Axilla Ilio-inguinal

Adjuvant radiotherapy group

Observation group


Observation group


Observation group

Related to surgery*

Seroma 0 0 5 4 4 7

Wound infection 1 0 0 1 2 6

Nerve damage 1 1 0 0 0 0

Wound necrosis 0 0 0 0 0 1

Pain 0 0 0 0 0 1

Related to radiation therapy†

Dermatitis 3 ·· 10 ·· 6 ··

Pain 0 ·· 2 ·· 0 ··

Based on the common toxicity criteria version 2.0.22 *At registration. †2 weeks after radiotherapy.

Table 4: Early adverse events (grade 3 and 4)

Articles


14 Agrawal S, Kane JM, Guadagnolo BA, et al. The benefi ts of adjuvant radiation therapy after therapeutic lymphadenectomy for clinically advanced, high-risk, lymph node metastatic melanoma. Cancer 2009; 115: 5836–44.

15 Strojan P, Jancar B, Cemazar M, et al. Melanoma metastases to neck nodes: role of adjuvant radiation. Int J Radiat Oncol Biol Phys 2010; 77: 1039–45.

16 Bibault JE, Dewas S, Mirabel X, et al. Adjuvant radiation therapy in metastatic lymph nodes from melanoma. Radiat Oncol 2011; 6: 12.

17 Burmeister BH, Smithers BM, Burmeister E, et al. A prospective phase II study of adjuvant postoperative radiation therapy following nodal surgery in malignant melanoma—Trans Tasman Radiation Oncology Group (TROG) Study 96.06. Radiother Oncol 2006; 81: 136–42.

18 Burmeister BH, Smithers BM, Davis S, et al. Radiation therapy following nodal surgery for melanoma: an analysis of late toxicity. ANZ J Surg 2002; 72: 344–48.

19 Cox JD, Stetz J, Pajak TF. Toxicity criteria of the Radiation Therapy Oncology Group (RTOG) and the European Organization for Research and Treatment of Cancer (EORTC). Int J Rad Oncol Biol 1995; 31: 1341–46.

20 Creagan ET, Cupps RE, Ivins JC, et al. Adjuvant radiation therapy for regional nodal metastases from malignant melanoma: a randomized prospective study. Cancer 1978; 42: 2206–10.

21 Guadagnolo BA, Zagars GK. Adjuvant radiation therapy for high-risk nodal metastases from cutaneous melanoma. Lancet Oncol 2009; 10: 409–16.

22 Bastiaannet E, Beukema JC, Hoekstra HJ, et al. Radiation therapy following lymph node dissection in melanoma patients: treatment, outcome and complications. Cancer Treat Rev 2005; 31: 18–26.

23 Shen P, Wanek LA, Morton DL. Is adjuvant radiotherapy necessary after positive lymph node dissection in head and neck melanomas? Ann Surg Oncol 2000; 7: 554–59.

24 Fuhrmann D, Lippold A, Borrosch F, et al. Should adjuvant radiotherapy be recommended following resection of regional lymph node metastases of malignant melanomas? Br J Dermatol 2001; 144: 66–70.

25 Miller EJ, Daly JM, Synnestfedt M, et al. Loco-regional relapse in melanoma. Surg Oncol 1992; 1: 333–40.

26 Gadd MA, Coit DG. Recurrence patterns and outcomes in 1019 patients undergoing axillary or inguinal lymphadenectomy for melanoma. Arch Surg 1992; 127: 1412–16.

27 Calabro A, Singletary SE, Balch CM. Patterns of relapse in 1001 consecutive patients with melanoma nodal metastases. Arch Surg 1989; 124: 1051–55.

28 Overgaard J. The role of radiotherapy in recurrent and metastatic malignant melanoma: a clinical radiobiological study. Int J Rad Oncol Biol Phys 1986; 12: 867–72.

29 Stevens G, McKay MJ. Dispelling the myths surrounding radiation therapy for treatment of cutaneous melanoma. Lancet Oncol 2006; 7: 575–83.

30 Kruijff S, Bastiaannet E, Speijers MJ, et al. The value of pre-operative S-100B and SUV in clinically stage III melanoma patients undergoing therapeutic lymph node dissection. Eur J Surg Oncol 2011; 37: 225–32.


Articles



Colorectal Surgery Division, National Cancer Center Hospital, Tokyo, Japan (S Fujita MD, T Akasu MD, Y Moriya MD); JCOG Data Center, Multi-institutional Clinical Trial Support Center, National Cancer Center, Tokyo, Japan (J Mizusawa MSc); Department of Surgery, National Cancer Center Hospital East, Kashiwa, Japan (N Saito MD); Department of Surgery, Shizuoka Cancer Center, Shizuoka, Japan (Y Kinugasa MD); Department of Surgery, Aichi Cancer Center Hospital, Nagoya, Japan (Y Kanemitsu MD); Department of Surgery, Osaka Medical Center and Cardiovascular Diseases, Osaka, Japan (M Ohue MD); Department of Surgery, Yokohama City University Medical Center, Yokohama, Japan (S Fujii MD); Department of Surgery, Kanagawa Cancer Center, Yokohama, Japan (M Shiozawa MD); Department of Surgery, Kyoto Medical Center, Kyoto, Japan (T Yamaguchi MD)

Correspondence to:Dr Shin Fujita, Colorectal Surgery Division 1-1, Tukiji 5-chome, Chuo-ku, Tokyo, [email protected]

Postoperative morbidity and mortality after mesorectal excision with and without lateral lymph node dissection for clinical stage II or stage III lower rectal cancer (JCOG0212): results from a multicentre, randomised controlled, non-inferiority trial Shin Fujita, Takayuki Akasu, Junki Mizusawa, Norio Saito, Yusuke Kinugasa, Yukihide Kanemitsu, Masayuki Ohue, Shoichi Fujii, Manabu Shiozawa, Takashi Yamaguchi, Yoshihiro Moriya, on behalf of the Colorectal Cancer Study Group of Japan Clinical Oncology Group

SummaryBackground Mesorectal excision is the international standard surgical procedure for lower rectal cancer. However, lateral pelvic lymph node metastasis occasionally occurs in patients with clinical stage II or stage III rectal cancer, and therefore mesorectal excision with lateral lymph node dissection is the standard procedure in Japan. We did a randomised controlled trial to confi rm that the results of mesorectal excision alone are not inferior to those of mesorectal excision with lateral lymph node dissection.

Methods This study was undertaken at 33 major hospitals in Japan. Eligibility criteria included histologically proven rectal cancer of clinical stage II or stage III, with the main lesion located in the rectum with the lower margin below the peritoneal refl ection, and no lateral pelvic lymph node enlargement. After surgeons had confi rmed macroscopic R0 resection by mesorectal excision, patients were intraoperatively randomised to mesorectal excision alone or with lateral lymph node dissection. The groups were balanced by a minimisation method according to clinical N staging (N0 or N1, 2), sex, and institution. Allocated procedure was not masked to investigators or patients. This study is now in the follow-up stage. The primary endpoint is relapse-free survival and will be reported after the primary analysis planned for 2015. Here, we compare operation time, blood loss, postoperative morbidity (grade 3 or 4), and hospital mortality between the two groups. Analysis was by intention-to-treat. This trial is registered with ClinicalTrials.gov, number NCT00190541.

Findings 351 patients were randomly assigned to mesoretcal excision with lateral lymph node dissection and 350 to mesorectal excision alone, between June 11, 2003, and Aug 6, 2010. One patient in the mesorectal excision alone group underwent lateral lymph node dissection, but was analysed in their assigned group. Operation time was signifi cantly longer in the mesorectal excision with lateral lymph node dissection group (median 360 min, IQR 296–429) than in the meso rectal excision alone group (254 min, 210–307, p<0·0001). Blood loss was signifi cantly higher in the mesorectal excision with lateral lymph node dissection group (576 mL, IQR 352–900) than in the mesorectal excision alone group (337 mL, 170–566; p<0·0001). 26 (7%) patients in the mesorectal excision with lateral lymph node dissection group had lateral pelvic lymph node metastasis. Grade 3–4 postoperative complications occurred in 76 (22%) patients in the mesorectal excision with lateral lymph node dissection group and 56 (16%) patients in the mesorectal excision alone group. The most common grade 3 or 4 postoperative complication was anastomotic leakage (18 [6%] patients in the mesorectal excision with lateral lymph node dissection group vs 13 [5%] in the mesorectal excision alone group; p=0·46). One patient in the mesorectal excision with lateral lymph node dissection group died of anastomotic leakage followed by sepsis.

Interpretation Mesorectal excision with lateral lymph node dissection required a signifi cantly longer operation time and resulted in signifi cantly greater blood loss than mesorectal excision alone. The primary analysis will help to show whether or not mesorectal excision alone is non-inferior to mesorectal excision with lateral lymph node dissection.

Funding National Cancer Center, Ministry of Health, Labour and Welfare of Japan.

IntroductionTotal mesorectal excision or mesorectal excision, in which at least a clear margin of 4 cm of the attached mesorectum distal to the tumour is resected, is the international standard surgical procedure for rectal cancer because it has a lower rate of associated local recurrence and higher rate of patient survival than conventional surgery.1–3

However, metastasis to lateral pelvic lymph nodes occasionally occurs in patients with clinical stage II or stage III lower rectal cancer, the lower margin of which is located at or below the peritoneal refl ection.

The incidence of lateral pelvic lymph node metastasis from lower rectal cancer is about 15%, and mesorectal excision with lateral lymph node dissection has been the

Articles


standard procedure for patients with lower rectal cancer in Japan4–6 since it was introduced in the 1970s. Pelvic autonomic nerve-sparing lateral lymph node dissection has been developed and refi ned since in the mid-1980s.7 If metastatic lymph node metastases are not dissected, local or systemic recurrence can develop.8,9 However, the incidence of local recurrence in patients with rectal cancer who undergo total mesorectal excision or mesorectal excision without lateral lymph node dissection at major hospitals in Europe and North America is reported to be less than 10%.10–13 Although this incidence is much the same as the rate for patients undergoing standard treatment in major hospitals in Japan,4–6 comparison is diffi cult because of diff erences in the backgrounds of patients.

The diffi culty of comparison between diff erent proced-ures in distinct populations prompted us to assess the survival benefi t, local control, operative complications, and sexual and urinary function of patients with rectal cancer undergoing mesorectal excision alone or with lateral lymph node dissection in a randomised controlled trial in major hospitals in Japan. The study aims to determine whether or not mesorectal excision alone is non-inferior to mesorectal excision with lateral lymph node dissection in terms of effi cacy. The primary analysis is planned for 2015, and this study is now in the follow-up stage. In this report, we present the data obtained so far for operation time, blood loss, and postoperative morbidity (grade 3 or 4) and mortality. Further analyses of urinary and sexual function are underway and will be reported at a later date.

MethodsStudy design and participantsPreoperative inclusion criteria were histologically confi rmed adenocarcinoma of clinical stage II or III (as determined by digital rectal examination, CT or MRI, and endoscopy); main lesion of tumour located in the rectum, with the lower tumour margin below peritoneal refl ection; no extramesorectal lymph node enlargement (ie, lymph nodes with a short-axis diameter of less than 10 mm shown by CT scan or MRI is not regarded as lymph node enlargement); and no invasion to other organs. Eligible patients were aged between 20 and 75 years with performance status 0 or 1 and no history of chemotherapy, pelvic surgery, or radiation. Intraoperative inclusion criteria were completed mesorectal excision, confi rmation that the main lesion of the tumour was located in the rectum, with the lower tumour margin below peritoneal refl ection, and macroscopic R0 (ie, no residual tumour) after the mesorectal excision. Exclusion criteria were synchronous or metachronous (within 5 years) malignancies other than carcinoma in situ or mucosal carcinoma, pregnancy or breastfeeding in women, or a psychological disorder or severe mental illness. Patients undergoing treatment with systemic steroids, or with a history of myocardial infarction or unstable angina pectoris within 6 months, or with severe pulmonary emphysema or

pulmonary fi brosis were also excluded. The attending physician had the fi nal decision for exclusion.

Clinical stage was based on the results of digital rectal examination, imaging (CT or MRI), and endoscopy. Clinical stage I rectal tumours and tumours in which the lower margin was located above the peritoneal refl ection were not included, because the incidence of lateral pelvic lymph node metastasis in such cases is very low. If lateral pelvic lymph node enlargement was detected by CT or MRI with 5 mm thick sections and the short-axis diameter of the nodes exceeded 10 mm, which is the minimum measurable size in such sections, patients were not included in this study and underwent mesorectal excision with lateral lymph node dissection.

Only surgeons specialising in both procedures from 33 Japanese institutions (listed in the appendix) participated in the study. We obtained written informed consent from all patients before surgery and the protocol was approved by institutional review boards.

Randomisation and maskingRandomisation and data handling were done by the JCOG Data Center. After surgeons had confi rmed macroscopic R0 resection (ie, no residual tumour) by mesorectal excision and macroscopic absence of lymph node metastasis in the lateral pelvic lymph area, patients were randomised intraoperatively to mesorectal excision alone or with lateral lymph node dissection by phone call to the JCOG Data Center. The groups were balanced by a minimisation method with biased-coin assignment according to clinical N staging by imaging (CT or MRI) and surgical exploration (N0 or N1, 2), sex, and institution. Allocated procedure was not masked to investigators or patients.

ProceduresMesorectal excision was done by open surgery in accordance with reported methods.1 Under direct vision with sharp dissection, the rectum was mobilised keeping the plane around the mesorectum, and the attached mesorectum with at least a 4 cm clearance margin distal to the tumour was resected. If the length of the attached mesorectum distal to the tumour was less than 4 cm, the mesorectum was totally resected. The inferior mesenteric artery was ligated at its root. If the blood supply to the distal colon was deemed inadequate as a result of this procedure, preservation of the left colonic artery after lymph node dissection at its root was allowed.

Lateral lymph node dissection was done in accordance with reported methods.4,5,14 Lateral pelvic lymph nodes include the common iliac node, internal iliac node, external iliac node, obturator node, and middle sacral node. Because metastasis to the external iliac node and middle sacral node in the patients eligible for this study without clinical lateral pelvic lymph node metastasis is rare,15 dissection of those nodes was not deemed necessary. The other lateral pelvic lymph nodes in the fatty and

See Online for appendix

Articles


connective tissues outside the pelvic plexus, around the common, internal, and oburator fossa were dissected after mesorectal excision (fi gure 1). All the autonomic nerves were preserved because lymph node metastasis around these nerves is rare in patients without clinical lateral pelvic lymph node metastasis.

For surgical quality control and assurance, intra operative photographs were taken. In the mesorectal excision alone group, fi ve photos were taken: the site of inferior mesenteric artery ligation, the preserved right and left hypogastric nerves, and the anterior and posterior sides of the resected specimen. In the mesorectal excision with lateral lymph node dissection group, 11 photos were taken: the site of inferior mesenteric artery ligation, the preserved right and left hypogastric nerves, the right and left internal iliac artery, the right and left obturator fossa, the anterior and posterior sides of the resected specimen, and the right and left dissected fatty and connective tissues in the lateral

pelvic lymph node area. These photographs were assessed and scored by the committee for quality control and assessment of surgery, and the surgical procedure was discussed and assured according to the score at meetings held twice a year.

Adjuvant chemotherapy with the Roswell Park regimen of intravenous fl uorouracil (500 mg/m²) and l-leucovorin (250 mg/m²) was given to patients with pathological stage III tumours in both groups. Patients who were stage II did not receive adjuvant chemotherapy.16 This regimen con-sisted of three courses of six doses of weekly chemo therapy followed by a 2-week rest. Adjuvant radiotherapy was not used.

Operative methods and pathology results were recorded according to the Japanese Classifi cation of Colon and Rectal Carcinoma (sixth edition)17 and TNM classifi cation (fi fth edition).18 The primary endpoint was relapse-free survival, and the secondary endpoints were overall survival, local recurrence-free survival, incidence of adverse events, incidence of major adverse events, operation time, blood loss, and incidence of sexual and urinary dysfunction. Operation time, blood loss, and all post operative mor-bidities during hospital stay were recorded prospectively on case report forms. Post operative morbidity was described according to the National Cancer Institute-Common Toxicity Criteria version 2·0. Hospital mortality was defi ned as postoperative death from any cause within 30 days.

Statistical analysisWe originally estimated that 5-year relapse-free survival after mesorectal excision with lateral lymph node dissection and mesorectal excision alone would be 65%, and the initial sample size was 600 patients, which was determined with one-sided alpha of 0·05, a power of 0·75, and a non-inferiority margin for a hazard ratio (HR) of 1·34. However, we calculated the 5-year relapse-free survival for all randomised patients 5 years after the start of registration, and recorded that it was about 75%. Therefore, the sample size was increased to 700 patients to maintain the required statistical power. Planned accrual and

Figure 1: Lateral lymph node dissection(A) The obturator fossa after lateral lymph node dissection, with the dissected fatty and connective tissues (right side). (B) Dissected fatty and connective tissues including lymph nodes.

B

AExternal iliac vessels

Internal iliac artery

Obturator nerve

Ureter

Figure 2: Trial profi leWe did not collect data for the number of eligible patients before enrolment. ME=mesorectal excision. LLND=lateral lymph node dissection.

701 patients enrolled and randomly assigned intraoperatively after ME

351 included in safety analysis 350 included in safety analysis

350 assigned ME alone 348 underwent ME alone 1 underwent ME and liver resection 1 underwent ME with LLND

351 assigned ME with LLND 350 underwent ME with LLND 1 underwent ME with LLND and liver resection

Articles


follow-up were 7 years and 5 years, respectively. Incidences of operative morbidity and mortality were expressed as the number of cases divided by the total number of registered patients. Diff erences in proportions between groups were

assessed with Fisher’s exact test. Diff erences in operation time and blood loss were compared with the Wilcoxon rank sum test. All p values were two-sided, and statistical analysis was done with SAS version 9·1. The data presented in this paper were as of June 12, 2011. Analysis was by intention-to-treat. This trial is registered with ClinicalTrials.gov, number NCT00190541, and UMIN-CTR, number C000000034.

Role of the funding sourceThe funding sources had no role in the design of the study, collection, analysis, interpretation of the data, writing of the report, or in the decision to submit for publication. The corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to submit the report for publication.

Results701 patients were randomly assigned to the mesorectal excision alone group (n=350) or the mesorectal excision with lateral lymph node dissection group (n=351) between June 11, 2003, and Aug 6, 2010 (fi gure 2). All but three patients received the allocated surgery. Liver metastasis was identifi ed after randomisation in one patient in each group and they underwent hepatic resection after rectal cancer surgery. Lateral lymph node metastasis was strongly suspected after randomisation in one patient allocated to the mesorectal excision alone group and the patient underwent lateral lymph node dissection. These three patients were eligible and included in this analysis. Two patients assigned to the mesorectal excision with lateral lymph node dissection group were found to have clinical stage I disease, despite being reported as clinical stage II or III at enrolment. Two other patients assigned to the same group had synchronous multiple cancers. Three patients (one in the mesorectal excision with lateral lymph node dissection group and two in the mesorectal excision alone group) were judged to have residual tumours before randomisation. We included these seven patients in this analysis, but their data will be excluded from the fi nal survival analysis.

Table 1 shows the characteristics of all patients. Low anterior resection was done in 568 (81%) of 701 patients. Mesorectal excision with lateral lymph node dissection required a signifi cantly longer operation time and resulted in signifi cantly greater blood loss than did mesorectal excision alone (table 2). Of the 26 patients in the mesorectal excision with lateral lymph node dissection group who had lateral pelvic lymph node metastasis, 11 (42%) were clinical stage II and 15 (58%) were clinical stage III. 19 (73%) had pathological mesorectal lymph node metastasis and seven (27%) had no pathological mesorectal lymph node metastasis. Although more common in the mesorectal excision with lateral lymph node dissection group than with mesorectal excision alone, diff erences between groups in grade 3 and 4 postoperative complications were not signifi cant (table 3). Anastomotic leakage of all grades,

ME with LLND (n=351) ME (n=350)

Sex

Male 236 (67%) 236 (67%)

Female 115 (33%) 114 (33%)

Age (years)

Median (IQR) 61 (54–67) 62 (55–68)

Clinical stage

II 188 (54%) 197 (56%)

III 163 (46%) 153 (44%)

Tumour location*

Ra 81 (23%) 80 (23%)

Rb 270 (77%) 270 (77%)

Tumour distance from anal verge (cm)†

Median (IQR) 5·0 (4·0–6·0) 5·0 (3·7–6·0)

ME=mesorectal excision. LLND=lateral lymph node dissection.*Ra=tumour centre located above the peritoneal refl ection, Rb=tumour centre located below the peritoneal refl ection. †Data for fi ve patients are missing.

Table 1: Characteristics of patients

ME with LLND (n=351) ME (n=350) p value*

Type of surgery ··

Low anterior resection 284 (81%) 284 (81%)

Abdominoperineal resection 66 (19%) 64 (18%)

Hartmann’s procedure 1 (<1%) 2 (<1%)

Time (min)

Median (IQR) 360 (296–429) 254 (210–307) <0·0001

Blood loss (mL)

Median (IQR) 576 (352–900) 337 (170–566) <0·0001

Lateral lymph node metastasis

Number (%) 26 (7%) ·· ··

ME=mesorectal excision. LLND=lateral lymph node dissection. *Wilcoxon rank sum test, two-sided.

Table 2: Operative details

ME with LLND (n=351) ME (n=350) p value*

Any grade 3–4 complication† 76 (22%) 56 (16%) 0·07

Anastomotic leakage‡ 18 (6%) 13 (5%) 0·46

Urinary retention 18 (5%) 10 (3%) 0·18

Infection with normal absolute neutrophil count

16 (5%) 17 (5%) 0·86

Haemorrhage with surgery 13 (4%) 5 (1%) 0·09

Wound infection 10 (3%) 8 (2%) 0·81

Pelvic abscess 6 (2%) 2 (<1%) 0·29

Bowel obstruction 4 (1%) 3 (<1%) 1·00

Other§ 12 (3%) 9 (3%) 0·66

ME=mesorectal excision. LLND=lateral lymph node dissection. *Fisher’s exact test, two-sided. †National Cancer Institute-Common Toxicity Criteria Version 2·0. ‡Denominator is patients with anastomosis (ME with LLND=284, ME=284). §Other=fever, melaena, fi stula, thrombosis, urinary frequency.

Table 3: Grade 3–4 postoperative morbidity

Articles


which is the major complication after low anterior resection, occurred in 37 (13%) of 284 patients in the meso-rectal excision alone group and 32 (11%) of 284 patients in the mesorectal excision with lateral lymph node dissection group (p=0·61). One patient in the mesorectal excision with lateral lymph node dissection group died of anastomotic leakage followed by sepsis. All other patients recovered from surgery and were discharged from hospital.

DiscussionAs expected, mesorectal excision with lateral lymph node dissection required a signifi cantly longer operation time and resulted in signifi cantly greater blood loss than did mesorectal excision alone. Although the incidence of grade 3 or grade 4 complications was higher in the mesorectal excision with lateral lymph node dissection group than in the mesorectal excision alone group, these diff erences were not signifi cant.

In previous reports, the mean diff erence in intraoperative blood loss between surgical procedures with and without lateral lymph node dissection was more than 500 mL.19–22 Blood loss might have been less in our study because none of the eligible patients had clinical evidence of lateral pelvic lymph node metastasis. In these patients, lateral lymph node dissection is easier than it is in those with clinical evidence of such metastasis. Also, because expertise with the lateral lymph node procedure is improving, blood loss might have been minimised compared with earlier studies.

The median operation time needed for mesorectal ex-icison with lateral lymph node dissection was longer than that for mesorectal excision alone. This result is attributable to the time needed for lateral lymph node dissection, which is a meticulous procedure, and confi rms previous results with regard to the diff erence in operation time.20–22

The incidence of all grade 3 or 4 postoperative complications, apart from infection with a normal absolute neutrophil count, was higher in the mesorectal excision with lateral lymph node dissection group than in the mesorectal excision alone group, but diff erences were not signifi cant. Results of a previous meta-analysis19 comparing extended lymphadenectomy including lateral lymph node dissection and conventional surgery for rectal cancer showed that the incidence of perioperative morbidity was higher for extended lymphadenectomy than for conventional surgery. However, one of the major complications, anastomotic leakage of all grades, showed no diff erence in incidence between the groups. Although we did not collect data for defunctioning stoma, the incidences of anastomotic leakage of all grades in patients who underwent low anterior resection in the mesorectal excision with lateral lymph node dissection group and mesorectal excision alone group were much the same, which suggests that lateral lymph node dissection was not a highly invasive surgical procedure.

Only one patient died from sepsis after anastomotic leakage. The reported mortality after mesorectal excision for rectal cancer surgery in Europe and North America is

1–3%,11–13,23 and that after mesorectal excision with lateral lymph node dissection in Japan is 1%,19 which is in line with our results (panel). The low mortality in our study can be attributed to several factors. Only surgeons specialising in both mesorectal excision and lateral lymph node dissection participated in this trial. Second, only patients who were judged to be capable of tolerating lateral lymph node dissection were selected and only high-volume centres for cancer treatment were allowed to enrol patients by the Colorectal Cancer Study Group.

Neoadjuvant chemoradiotherapy for rectal cancer is used worldwide. However, patients undergoing such treatment were not included and adjuvant radiotherapy was not used in our study for two reasons. First, the eff ectiveness and safety of adjuvant or neoadjuvant chemoradiotherapy for rectal cancer had not been clearly shown when we designed the protocol of this study. Second, adjuvant radiotherapy is not commonly used in Japan because of the lower local recurrence rate and better prognosis for patients in Japan than for those in Europe and North America.

Kim and colleagues8 showed that lateral pelvic lymph node metastasis is a major cause of local recurrence of rectal cancer. With serial sections from human fetuses and three-dimensional reconstruction, Kusters and colleagues24 showed that tumour recurrence might arise from lateral pelvic lymph nodes. However, other reports from Europe and North America have not supported these results. Syk and colleagues25 examined the pattern of local recurrence after total mesorectal excision and concluded that lateral pelvic lymph node metastases are not a major cause of local recurrence. The results of a Dutch trial of total mesorectal excision showed that the rate of lateral site


Systematic reviewTotal mesorectal excision or mesorectal excision is the international standard surgical procedure for lower rectal cancer.1 However, lateral pelvic lymph node metastasis occasionally occurs in patients with clinical stage II or stage III rectal cancer, and therefore mesorectal excision with lateral lymph node dissection is the standard procedure in Japan. When metastatic lateral pelvic lymph nodes are not dissected, the patients can have local or systemic recurrence. Although we did not do a systematic search of published work before starting this trial, the reported incidence of local recurrence in rectal cancer patients undergoing mesorectal excision without lateral lymph node dissection at major hospitals in Europe and North America is less than 10%,10–13 which is much the same as the incidence in patients who undergo mesorectal excision with lateral lymph node dissection at major hospitals in Japan.4–6 Therefore, we did a randomised controlled trial to determine whether mesorectal excision alone is non-inferior to mesorectal excision with lateral lymph node dissection.

Interpretation7% of the patients with lower rectal cancer without lateral pelvic lymph node enlargement had lateral pelvic lymph node metastasis. Mesorectal excision with lateral lymph node dissection required a signifi cantly longer operation time and resulted in signifi cantly greater blood loss than mesorectal excision alone. The primary analysis will help to determine whether or not mesorectal excision alone is non-inferior to mesorectal excision with lateral lymph node dissection.

Articles


recurrence was only 3% in patients with lower rectal cancer, being much the same as results for patients who underwent lateral lymph node dissection at the National Cancer Center, Tokyo.26 Analysis of the pattern of local recurrence in our study is very important, and should give a reliable indication of the incidence of lateral pelvic lymph node metastasis. The incidence of such metastasis was 7%, which was lower than the 15% reported in a retrospective multicentre study in Japan,6 because only patients who had no clinical evidence of lateral pelvic lymph node enlargement were eligible for our study. This result shows that even in patients without clinically evident lateral pelvic lymph node metastasis, such metastasis is sometimes present pathologically.

Our patient population was defi ned as being lateral pelvic lymph node negative by CT or MRI. Nonetheless, the 7% of patients in the mesorectal excision with lateral lymph node dissection group were found to have lateral pelvic lymph node metastasis after lymph node dissection. Therefore, a similar proportion of patients undergoing mesorectal excision alone probably have such metastasis. If all patients with lateral pelvic lymph node metastasis have local or systemic recurrence, then the relapse rate will be about 7% higher in patients who undergo mesorectal excision alone than in those who also have lateral lymph node dissection. If the results for the primary analysis planned for 2015 show that the upper confi dence limit of the HR is less than 1·34, which corresponds to an 8% diff erence in 5-year relapse-free survival between the groups, then the non-inferiority of mesorectal exicision alone will be confi rmed in terms of outcome. If not, mesorectal excision with lateral lymph node dissection should be considered the standard surgical procedure for lower rectal cancer.

ContributorsSFujita, TA, NS, and YM contributed to study design. SFujita, TA, NS, YKi,

YKa, MO, SFujii, MS, TY, and YM contributed to data collection, data

analysis, and interpretation. JM contributed to statistical analyses. All the

authors contributed to writing or review of the report and approved the

fi nal version.


AcknowledgmentsWe thank the members of the JCOG Data Center and JCOG Operations

Offi ce for their support in preparing the report (Kenichi Nakamura,

Hiroshi Katayama, Atsuo Takashima), data management (Harumi Kaba,

Ayaka Nakano), and overseeing of study management (Haruhiko Fukuda).

This study was supported in part by the National Cancer Center Research

and Development Fund (23-A-16, 23-A-19) and Grants-in-Aid for Cancer

Research (14S-4, 17S-3, 17S-5, 20S-3, 20S-6), Health Sciences Research

Grants for Medical Frontier Strategy Research, Health and Labour Sciences

Research Grants for Clinical Research for Evidence-based Medicine and

Health, and a Labour Sciences Research Grant for Clinical Cancer Research

from the Ministry of Health, Labour and Welfare of Japan.

References1 Heald RJ, Husband EM, Ryall RD. The mesorectum in rectal cancer

surgery—the clue to pelvic recurrence? Br J Surg 1982; 69: 613–16.

2 MacFarlane JK, Ryall RD, Heald RJ. Mesorectal excision for rectal cancer. Lancet 1993; 341: 457–60.

3 Nelson H, Petrelli N, Carlin A, et al. Guidelines 2000 for colon and rectal cancer surgery. J Natl Cancer Inst 2001; 93: 583–96.

4 Moriya Y, Sugihara K, Akasu T, Fujita S. Importance of extended lymphadenectomy with lateral node dissection for advanced lower rectal cancer. World J Surg 1997; 21: 728–32.

5 Takahashi T, Ueno M, Azekura K, Ohta H. Lateral node dissection and total mesorectal excision for rectal cancer. Dis Colon Rectum 2000; 43: S59–68.

6 Sugihara K, Kobayashi H, Kato T, et al. Indication and benefi t of pelvic sidewall dissection for rectal cancer. Dis Colon Rectum 2006; 49: 1663–72.

7 Moriya Y, Sugihara K, Akasu T, Fujita S. Nerve-sparing surgery with lateral node dissection for advanced lower rectal cancer. Eur J Cancer 1995; 31A: 1229–32.

8 Kim TH, Jeong SY, Choi DH, et al. Lateral lymph node metastasis is a major cause of locoregional recurrence in rectal cancer treated with preoperative chemoradiotherapy and curative resection. Ann Surg Oncol 2008; 15: 729–37.

9 MERCURY Study Group. Relevance of magnetic resonance imaging-detected pelvic sidewall lymph node involvement in rectal cancer. Br J Surg 2011; 98: 1798–804.

10 Enker WE, Thaler HT, Cranor ML, Polyak T. Total mesorectal excision in the operative treatment of carcinoma of the rectum. J Am Coll Surg 1995; 181: 335–46.

11 Heald RJ, Moran BJ, Ryall RD, Sexton R, MacFarlane JK. Rectal cancer: the Basingstoke experience of total mesorectal excision, 1978–1997. Arch Surg 1998; 133: 894–99.

12 Lopez-Kostner F, Lavery IC, Hool GR, Rybicki LA, Fazio VW. Total mesorectal excision is not necessary for cancers of the upper rectum. Surgery 1998; 124: 612–17.

13 Zaheer S, Pemberton JH, Farouk R, Dozois RR, Wolff BG, Ilstrup D. Surgical treatment of adenocarcinoma of the rectum. Ann Surg 1998; 227: 800–11.

14 Akasu T, Moriya Y. Abdominopelvic lymphadenectomy with autonomic preservation for carcinoma of the rectum: Japanese experience. In: Wanebo HJ, ed, Surgery for gastrointestinal cancer: a multidisciplinary approach. Philadelphia: Lippincott-Raven, 1997: 667–80.

15 Kobayashi H, Mochizuki H, Kato T, et al. Outcomes of surgery alone for lower rectal cancer with and without pelvic sidewall dissection. Dis Colon Rectum 2009; 52: 567–76.

16 Haller DG, Catalano PJ, Macdonald JS, et al. Phase III study of fl uorouracil, leucovorin, and levamisole in high-risk stage II and III colon cancer: fi nal report of Intergroup 0089. J Clin Oncol 2005; 23: 8671–78.

17 Japanese Society of Cancer of the Colon and Rectum. General rules for clinical and pathological studies on cancer of the colon, rectum and anus, 6th edn. Tokyo: Kanehara, 1998.

18 Sobin LH, Wittekind C. TNM classifi cation of malignant tumours, 5th edn. New York: Wiley-Liss, 1997.

19 Georgiou P, Tan E, Gouvas N, et al. Extended lymphadenectomy versus conventional surgery for rectal cancer: a meta-analysis. Lancet Oncol 2009; 10: 1053–62.

20 Hojo K, Sawada T, Moriya Y. An analysis of survival and voiding, sexual function after wide iliopelvic lymphadenectomy in patients with carcinoma of the rectum, compared with conventional lymphadenectomy. Dis Colon Rectum 1989; 32: 128–33.

21 Nagawa H, Muto T, Sunouchi K, et al. Randomized, controlled trial of lateral node dissection vs. nerve-preserving resection in patients with rectal cancer after preoperative radiotherapy. Dis Colon Rectum 2001; 44: 1274–80.

22 Fujita S, Yamamoto S, Akasu T, Moriya Y. Lateral pelvic lymph node dissection for advanced lower rectal cancer. Br J Surg 2003; 90: 1580–85.


24 Kusters M, Wallner C, Lange MM, et al. Origin of presacral local recurrence after rectal cancer treatment. Br J Surg 2010; 97: 1582–87.

25 Syk E, Torkzad MR, Blomqvist L, Ljungqvist O, Glimelius B. Radiological fi ndings do not support lateral residual tumour as a major cause of local recurrence of rectal cancer. Br J Surg 2006; 93: 113–19.

26 Kusters M, Beets GL, van de Velde CJ, et al. A comparison between the treatment of low rectal cancer in Japan and the Netherlands, focusing on the patterns of local recurrence. Ann Surg 2009; 249: 229–35.


Articles

Lancet Oncol 2012; 13: 633–41

Published OnlineMay 3, 2012DOI:10.1016/S1470-2045(12)70102-X

See Comment page 568

*Authors contributed equally

†Senior authors contributed equally to this work

Department of Radiation Oncology (N Liu MD, R-X Cui MD, W-F Li MD, Y Sun PhD, M Chen MD, N Jiang MD, L Chen MD, Prof J Ma MD), Department of Pathology (Prof J-P Yun PhD, J Zeng PhD), Imaging Diagnosis and Interventional Center (L-Z Liu MD, Prof L Li PhD), State Key Laboratory of Oncology in South China (R-R Wei PhD, M-Y Zhang MD, B-J Huang MD, Q-M He MD, Prof H-Y Wang PhD), and the National Clinical Study Center for Anticancer Drugs (Y Guo PhD), Sun Yat-sen University Cancer Center, Guangzhou, China; Department of Radiation Oncology, State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, Sichuan University, Chengdu, China (Prof N-Y Chen PhD, Y Li MD); and Department of Clinical Oncology, Queen Elizabeth Hospital, Hong Kong, China (Prof W C S Cho PhD)

Correspondence to:Prof Jun Ma, Department of Radiation Oncology, State Key Laboratory of Oncology in South China, Sun Yat-sen University Cancer Center, 651 Dongfeng Road East, Guangzhou 510060, [email protected]

Prognostic value of a microRNA signature in nasopharyngeal carcinoma: a microRNA expression analysisNa Liu*, Nian-Yong Chen*, Rui-Xue Cui*, Wen-Fei Li, Yan Li, Rong-Rong Wei, Mei-Yin Zhang, Ying Sun, Bi-Jun Huang, Mo Chen, Qing-Mei He, Ning Jiang, Lei Chen, William C S Cho, Jing-Ping Yun, Jing Zeng, Li-Zhi Liu, Li Li, Ying Guo, Hui-Yun Wang†, Jun Ma†

SummaryBackground MicroRNAs (miRNAs) can be used as prognostic biomarkers in many types of cancer. We aimed to identify miRNAs that were prognostic in patients with nasopharyngeal carcinoma.

Methods We retrospectively analysed miRNA expression profi les in 312 paraffi n-embedded specimens of nasopharyngeal carcinoma from Sun Yat-sen University Cancer Center (Guangzhou, China) and 18 specimens of non-cancer nasopharyngitis. Using an 873 probe microarray, we assessed associations between miRNA signatures and clinical outcome in a randomly selected 156 samples (training set) and validated fi ndings in the remaining 156 samples (internal validation set). We confi rmed the miRNAs signature using quantitative RT-PCR analysis in 156 samples from a second randomisation of the 312 samples, and validated the miRNA signature in 153 samples from the West China Hospital of Sichuan University in Chengdu, China (independent set). We used the Kaplan-Meier method and log-rank tests to estimate correlations of the miRNA signature with disease-free survival (DFS), distant metastasis-free survival (DMFS), and overall survival.

Findings 41 miRNAs were diff erentially expressed between nasopharyngeal carcinoma and non-cancer nasopharyngitis tissues. A signature of fi ve miRNAs, each signifi cantly associated with DFS, was identifi ed in the training set. We calculated a risk score from the signature and classifi ed patients as high risk or low risk. Compared with patients with low-risk scores, patients with high risk scores in the training set had shorter DFS (hazard ratio [HR] 2·73, 95% CI 1·46–5·11; p=0·0019), DMFS (3·48, 1·57–7·75; p=0·0020), and overall survival (2·48, 1·24–4·96; p=0·010). We noted equivalent fi ndings in the internal validation set for DFS (2·47, 1·32–4·61; p=0·0052), DMFS (2·28, 1·09–4·80; p=0·030), and overall survival (2·87, 1·38–5·96; p=0·0051) and in the independent set for DFS (3·16, 1·65–6·04; p=0·0011), DMFS (2·39, 1·05–5·42; p=0·037), and overall survival (3·07, 1·34–7·01; p=0·0082). The fi ve-miRNA signature was an independent prognostic factor. A combination of this signature and TNM stage had better prognostic value than did TNM stage alone in the training set (area under receiver operating characteristics 0·68 [95% CI 0·60–0·76] vs 0·60 [0·52–0·67]; p=0·013), the internal validation set (0·70 [0·61–0·78] vs 0·61 [0·54–0·68]; p=0·012), and the independent set (0·70 [0·62–0·78] vs 0·63 [0·56–0·69]; p=0·032).

Interpretation Identifi cation of patients with the fi ve-miRNA signature might add prognostic value to the TNM staging system and inform treatment decisions for patients at high risk of progression.

Funding Science Foundation of Chinese Ministry of Health, National Natural Science Foundation of China, Pearl River Scholar Funded Scheme, Guangdong Key Scientifi c and Technological Innovation Program, Guangdong Natural Science Foundation, Fundamental Research Funds for the Central Universities.

IntroductionAccording to the International Agency for Research on Cancer, 84 400 incident cases of nasopharyngeal carcinoma and 51 600 disease-related deaths occurred in 2008.1 The highest prevalence (20–50 cases per 100 000 individuals) of this cancer occurred in south China whereas the lowest prevalence (0·5 cases per 100 000 individuals) was reported in white populations in Europe and the USA.2

The TNM staging system provides a useful benchmark for prognostic defi nition and establishment of treatment strategy. According to the US National Comprehensive Cancer Network guidelines, patients with early-stage nasopharyngeal carcinoma should be treated with radiotherapy, whereas patients with advanced disease should receive chemo radiotherapy. However, large

variations in the clinical outcomes are reported in patients with the same stage and similar treatment regimens.3 These fi ndings suggest that the present staging system is inadequate for defi nition of prognosis and does not refl ect the biological heterogeneity of cancer because it classifi es the extent of disease chiefl y on the basis of anatomical information. With the rapid advance in understanding of the molecular biology of nasopharyngeal carcinoma, various biomarkers that associate with prognosis or effi cacy of treatment have been reported, including Epstein-Barr virus (EBV) DNA, lactate dehydrogenase, and VEGF.4–6 Nevertheless, more studies are required to confi rm their clinical roles and new prognostic biomarkers are needed for this disease.

MicroRNAs (miRNAs) are a class of small non-coding RNAs that regulate various biological processes

Articles


post-transcriptionally and are dysregulated in most cancer types.7,8 miRNAs have been reported in the development and progression of many cancers9–15 and are potential biomarkers for cancer diagnosis, prognosis, and person alised therapy.16 miRNAs were reported to be expressed aberrantly in nasopharyngeal carcinoma tissues compared with normal epithelial tissue, and promote an aggressive tumour phenotype by changing the expression of their mRNA targets.17,18 The clinical signifi cance of miRNAs in nasopharyngeal carcinoma, however, has not been established.

We aimed to examine the miRNA expression profi les in tissue samples of nasopharyngeal carcinoma to explore their clinical signifi cance in disease develop-ment and progression, and provide information for personalised therapy.

MethodsClinical specimens and study designWe obtained 465 pathologically proven and non-distant-metastatic paraffi n-embedded nasopharyngeal carcinoma speci mens for this study. 312 specimens were obtained from the Sun Yat-sen University Cancer Center (Guangzhou, China) between Jan 16, 2003, and Feb 25, 2006, and 153 were collected from the West China Hospital of Sichuan University (Chengdu, China) between April 1, 2002, and May 22, 2008. We also acquired 18 paraffi n-embedded non-cancer nasopharyngitis biopsy samples from patients who were originally suspected to have nasopharyngeal carcinoma from the Sun Yat-sen University Cancer Center between January, 2003, and May, 2011. All samples were pathologically reassessed by two pathologists (J-PY and JZ), and the percentage of tumour cells was 70% or more in all nasopharyngeal carcinoma specimens. None of the patients had received radiotherapy or chemotherapy before biopsy sampling. This study was approved by the institutional ethical review boards of both hospitals, and written informed consent was obtained from all patients.

All MRI and CT scans were reassessed separately by two radiologists (L-ZL and LL), and any disagreements were resolved by consensus. Clinical staging was reclassifi ed according to the criteria of the American Joint Committee on Cancer staging manual (seventh edition). All patients were treated with conventional two-dimensional radio therapy and 159 (50%) of 317 patients with stage III–IV disease also received concurrent platinum-based chemotherapy.19 Median follow-up was 62·1 months (IQR 47·7–71·5).

We established miRNA expression profi les of the 312 nasopharyngeal carcinoma samples and 18 non-cancer nasopharyngitis samples with microarray ana lysis. According to Simon’s proposal on class prediction based on the microarray experiment,20 we randomly divided the 312 specimens from Sun Yat-sen University into a training set of 156 samples and a validation set of 156 samples on the basis of a computer-generated

allocation sequence. We aimed to identify a clinically signifi cant prognostic miRNA signature from the training set and tested it in the internal validation set.

After this internal validation, we did a second random-isation for the 312 specimens from Sun Yat-sen University (two sets of 156) and aimed to replicate the miRNA signature in quantitative RT-PCR analysis in one set. This analysis was designed to assess repeatability of the microarray fi ndings in a cost-eff ective setting.

To assess whether any miRNA signature identifi ed from the quantitative RT-PCR analysis had the same or similar prognostic value in diff erent populations, we used the 153 samples from Sichuan University as an independent set to externally validate the results.

ProceduresWe isolated total RNA as described previously.21 RNA was extracted with an acid phenol–chloroform extraction method followed by ethanol precipitation. We used a microarray containing 873 miRNA probes and undertook miRNA probe design, RNA labelling, and microarray hybridisation as described elsewhere.22,23 Briefl y, total RNA (2·5 μg) was labelled with pCp-DY647 (Dharmacon, Lafayette, CO, USA). After hybridisation, we scanned arrays with the LuxScan 10K Microarray Scanner (CapitalBio, Beijing, China), and analysed images with GenePix Pro 6.0 software (Axon Instruments, Foster City, CA, USA).

For quantitative RT-PCR, total RNA (2 μg) was reversely transcribed with Bulge-Loop miRNA-specifi c reverse transcription-primers (RiboBio, Guangzhou, China) and Moloney murine leukaemia virus reverse transcriptase (Promega, Madison, WI, USA). Quantitative PCR reac-tions were done with Platinum SYBR Green qPCR SuperMix-UDG reagents (Invitrogen, Carlsbad, CA, USA) and Bulge-Loop primers (RiboBio) on the PRISM 7900HT system (Applied Biosystems, Carlsbad, CA, USA) with small nuclear RNA U6 as the normalisation control. We did all assays in triplicate. Reactions without RNA template or reverse transcriptase were used as negative controls. We calculated relative expression of miRNAs with the ΔΔCT equation, in which the expression pattern in normal tissue is subtracted from that identifi ed in nasopharangeal cancer.

The microarray data has been deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus (GSE32960). For data processing, we fi rst did background subtraction and quantile normalisation, and then excluded miRNAs from the analysis when they were not detectable in more than 80% of the samples and when they had a less than 1·5-fold change in either direction from the median value in less than 10% of the samples.10,12–15 After fi ltration, 515 (59%) of 873 miRNAs were included for further analysis.

We used Biometric Research Branch-ArrayTools version 3.8.0 for microarray data analysis.24 We did class comparison with unpaired t tests to identify

For the Gene Expression Omnibus see http://www.ncbi.

nlm.nih.gov/geo/

Articles


diff erentially expressed miRNAs between 312 samples of nasopharyngeal carcinoma and 18 samples of non-cancer nasopharyngitis, with a threshold of 2·5 for the fold expression change, and diff erences in miRNA expression were regarded as signifi cant if p values were less than 0·001. We also did a global test to establish whether the expression profi les diff ered between the nasopharyngeal carcinoma and non-cancer naso-pharyngitis groups by permuting the labels of the arrays corresponding to each group. For every permu tation, the p values were recomputed, and the number of signifi cant diff erences in miRNA expression (p<0·001) level were noted. The proportion of permutations that resulted in at least as many signifi cant diff erences in miRNA expression as the actual data was the signifi cance level of the global test. We did hierarchical clustering analysis with the average linkage method and centred Pearson’s correlation coeffi cients with MEV version 4.2.25

To identify which of the diff erentially expressed miRNAs were signifi cantly associated with disease-free survival (DFS) in the training set, we did univariate Cox regression analysis with BRB-ArrayTools, and identifi ed signifi cant diff erences (p<0·05) in miRNA expression. miRNAs that were associated signifi cantly with DFS were selected to construct an miRNA signature with the risk-score method11,26 and were used for survival analysis.

Statistical analysisOur study had a power of 80% to detect a hazard ratio (HR) for treatment failure of 2·44 based on an assumed 5-year DFS rate of 58% in the high-risk group and 80% in the low-risk group with a two-sided log-rank test at a signifi cance level of 0·05. We anticipated that 45 events were needed from 156 patients in the training set (78 per risk score group).27

Our primary outcome of interest was DFS. Secondary outcomes included distant metastasis-free survival (DMFS) and overall survival. We calculated DFS from treatment to the date of the fi rst relapse at any site or death from any cause, whichever occurred fi rst, DMFS to the fi rst distant relapse, and overall survival to death from any cause. We assessed the relation between clinical characteristics and miRNA expression with Student’s t test, χ² test, or Fisher’s exact test. We analysed correlation between the microarray data and the RT-PCR data with the Spearman correlation. We used the Kaplan-Meier method and the log-rank test to estimate DFS, DMFS, and overall survival, and calculated HRs with the adjusted multivariate Cox regression analysis.

We did a multivariate Cox regression analysis using a backward stepwise approach to test if the signature was an independent prognostic factor of DFS, DMFS, and overall survival. miRNA signature, age, sex, staging and patho-logical type, radiotherapy, concurrent chemotherapy and EBV seromarkers were used as covariates. We did a stratifi ed analysis to test whether the miRNA signature was associated with survival independent of stage and

used receiver operating characteristics (ROC) curves to compare the sensitivity and specifi city for the prediction of survival by the miRNA signature, TNM stage, age, and sex. A prognostic score model was constructed combining the miRNA signature and TNM stage to compare its prognostic validity with the TNM stage-alone model and the miRNA signature-alone model with a ROC analysis (see appendix p 1 for details).28 All statistical analyses were done with Stata version 10.0 with two-tailed tests, and signifi cance was defi ned as p values of less than 0·05.

Role of the funding sourceThe sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The lead authors (JM and H-YW) had full access to all the data in the study and had fi nal responsibility for the decision to submit for publication.

Results465 nasopharyngeal carcinoma specimens were obtained for this study. Table 1 lists the clinical charac teristics of patients with nasopharyngeal carcinoma in the training, internal validation, and independent sets. On microarray analysis, 41 miRNAs were diff erentially expressed between 312 samples of nasopharyngeal carcinoma from the combined training and internal validation sets and 18 non-cancer nasopharyngitis samples (fold change ≥2·5, false discovery rate 0; appendix pp 2–3). For this 41 miRNA profi le, clustering analysis displayed a perfect separation between the 312 nasopharyngeal carcinoma samples and the 18 non-cancer nasopharyngitis samples (appendix p 13).

In the training set, we identifi ed fi ve miRNAs that were signifi cantly associated with DFS from the 41 diff erentially expressed miRNAs (appendix 4–5). In the fi ve miRNAs, expression levels of four miRNAs (miR-142-3p, miR-29c, miR-26a, and miR-30e) were positively associated with DFS, and the expression level of the other one (miR-93) was inversely associated with DFS (appendix p 6).

To validate the microarray data, we did quantitative RT-PCR on the fi ve identifi ed miRNAs in a set of 156 nasopharyngeal carcinoma tissues from the second randomisation and 18 non-cancer nasopha ryngitis tissues. Expression of these miRNAs measured by RT-PCR was notably diff erent between nasopharyn geal carcinoma and non-cancer nasopharyngitis tissues (appendix p 13) and were signifi cantly correlated with their microarray data (appendix p 14).

We derived a formula to calculate the risk score for every patient from the expression values of the fi ve miRNAs, weighted by regression coeffi cient:11,26

Risk score = (0·487 × expression value of miR-93) – (0·395 × expression value of miR-142-3p) – (0·367 × expression value of miR-29c) – (0·355 × expression value of miR-26a) – (0·804 × expression value of miR-30e)


Articles


Training set (n=156) Validation set (n =156) Independent set (n=153)

High-risk group (n=78)

Low-risk group (n=78)

p value* High-risk group (n=81)


p value* High-risk group (n=75)


p value*

Age, years 47·46 (10·75) 47·21 (10·59) 0·74† 46·84 (12·00) 45·79 (10·80) 0·27† 47·40 (9·86) 47·85 (12·03) 0·80†

Sex, male 60 (77%) 59 (76%) 0·85 60 (74%) 54 (72%) 0·77 59 (79%) 53 (68%) 0·14

WHO pathological type 0·84 0·58 0·51

Undiff erentiated non-keratinising 75 (96%) 76 (97%) 78 (96%) 72 (96%) 73 (97%) 76 (97%)

Diff erentiated non-keratinising 2 (3%) 1 (1%) 3 (4%) 2 (3%) 2 (3%) 1 (1%)

Keratinising squamous cell 1 (1%) 1 (1%) 0 1 (1%) 0 1 (1%)

VCA-IgA 0·22 0·37 0·76

<1:80 14 (18%) 8 (10%) 10 (12%) 14 (19%) 15 (20%) 13 (17%)

1:80–1:320 45 (58%) 43 (55%) 46 (57%) 44 (59%) 39 (52%) 45 (58%)

≥1:640 19 (24%) 27 (35%) 25 (31%) 17 (23%) 21 (28%) 20 (26%)

EA-IgA 0·88 0·50 0·65

<1:10 19 (24%) 20 (26%) 18 (22%) 19 (25%) 21 (28%) 18 (23%)

1:10–1:20 28 (36%) 25 (32%) 22 (27%) 25 (33%) 23 (31%) 29 (37%)

≥1:40 31 (40%) 33 (42%) 41 (51%) 31 (41%) 31 (41%) 31 (40%)

T stage 0·26 0·64 0·58

T1 15 (19%) 18 (23%) 15 (19%) 18 (24%) 18 (24%) 18 (23%)

T2 21 (27%) 26 (33%) 20 (25%) 22 (29%) 15 (20%) 23 (29%)

T3 16 (21%) 19 (24%) 20 (25%) 16 (21%) 21 (28%) 18 (23%)

T4 26 (33%) 15 (19%) 26 (32%) 19 (25%) 21 (28%) 19 (24%)

N stage 0·95 0·45 0·50

N0 10 (13%) 10 (13%) 11 (14%) 13 (17%) 13 (17%) 8 (10%)

N1 37 (47%) 39 (50%) 40 (49%) 32 (43%) 36 (48%) 38 (49%)

N2 18 (23%) 15 (19%) 17 (21%) 22 (29%) 18 (24%) 25 (32%)

N3 13 (17%) 14 (18%) 13 (16%) 8 (11%) 8 (11%) 7 (9%)

TNM stage 0·17 0·66 0·72

I 2 (3%) 4 (5%) 3 (4%) 3 (4%) 7 (9%) 8 (10%)

II 21 (27%) 24 (31%) 20 (25%) 21 (28%) 15 (20%) 20 (26%)

III 18 (23%) 26 (33%) 22 (27%) 25 (33%) 27 (36%) 29 (37%)

IV 37 (47%) 24 (31%) 36 (44%) 26 (35%) 26 (35%) 21 (27%)

Radiotherapy period interruptions 0·63 0·84 0·91

0 days 41 (53%) 38 (49%) 43 (53%) 41 (55%) 33 (44%) 35 (45%)

≥1 day 37 (47%) 40 (51%) 38 (47%) 34 (45%) 42 (56%) 43 (55%)

Radiotherapy boosting

No 66 (85%) 68 (87%) 0·65 69 (85%) 65 (87%) 0·79 70 (93%) 75 (96%) 0·49

Yes 12 (15%) 10 (13%) 12 (15%) 10 (13%) 5 (7%) 3 (4%)

Chemotherapy 0·42 0·47 0·63

Yes 36 (46%) 31 (40%) 37 (46%) 30 (40%) 39 (52%) 37 (47%)

No 42 (54%) 47 (60%) 44 (54%) 45 (60%) 36 (48%) 41 (53%)

Disease-free survival‡

Relapses or deaths‡ 33 (42%) 14 (18%) 34 (42%) 14 (19%) 31 (41%) 14 (18%)

5 year 59% 82% 0·0011§ 59% 81% 0·0026§ 57% 83% 0·0012§

Distant metastasis-free survival

Distant metastases 25 (32%) 8 (10%) 23 (28%) 10 (13%) 17 (23%) 9 (12%)

5-year 68% 91% 0·0010§ 72% 87% 0·023§ 76% 88% 0·046§

Overall survival

Deaths 25 (32%) 12 (15%) 27 (33%) 10 (13%) 20 (27%) 8 (10%)

5-year 71% 87% 0·0068§ 72% 87% 0·0029§ 72% 92% 0·0058§

Data are mean (SD) or n (%), unless otherwise stated. VCA=viral capsid antigen. IgA=immunoglobulin A. EA=early antigen. *χ² test or Fisher’s exact test. †Student’s t test. ‡Disease-free survival from treatment to the date of the fi rst relapse at any site or death from any cause, whichever occurred fi rst. §Log-rank test.

Table 1: Clinical characteristics of patients according to the microRNA signature in the training, validation, and independent sets

Articles


With this risk score formula, patients in the training set were divided into high-risk or low-risk groups with the median risk score (–13·67) as the cutoff . Distribution of clinical characteristics did not vary signifi cantly between the high-risk and low-risk groups (table 1). Compared with patients with low risk scores, patients with high risk scores had shorter DFS, DMFS, and overall survival (fi gure 1).

To confi rm that these fi ve miRNAs were essential for the fi ve-miRNA signature, we constructed fi ve-minus-one miRNA signatures by deleting each miRNA in turn, and compared the survival of these four miRNA signatures with the original fi ve-miRNA signature with log-rank test.11 Unlike the fi ve-miRNA signature, none

of the fi ve-minus-one miRNA signatures was consistently associated with DFS, DMFS, and overall survival in the training set, the internal validation set, and the independent set (appendix p 7).

Patients in the internal validation set were classifi ed into high-risk and low-risk groups with the same cutoff as used in the training set. As expected, patients in the validation set with high risk-scores had shorter DFS, DMFS, and overall survival than did those with low risk scores (fi gure 1). Clinical characteristics did not diff er between the two groups (table 1).

We also assessed the distribution of risk scores, survival statuses, and miRNA expression in patients from the combined training and validation sets. Patients with low

Dise

ase-

free

surv

ival

1·00

0·75

0·25

0·50

0 Dist

ant m

etas

tasis

-fre

e su

rviv

al 1·00

0·75

0·25

0·50

0O

vera

ll su

rviv

al

1·00

0·75

0·25

0·50

0

Dise

ase-

free

surv

ival

1·00

0·75

0·25

0·50

0 Dist

ant m

etas

tasis

-fre

e su

rviv

al 1·00

0·75

0·25

0·50

0

Ove

rall

surv

ival

1·00

0·75

0·25

0·50

0

Dise

ase-

free

surv

ival

1·00

0·75

0·25

0·50

0 Dist

ant m

etas

tasis

-fre

e su

rviv

al 1·00

0·75

0·25

0·50

0

Ove

rall

surv

ival

1·00

0·75

0·25

0·50

0

Time (months)96722412 480 846036

Time (months)96722412 480 846036

Time (months)96722412 480 846036

96722412 480 846036 96722412 480 846036 96722412 480 846036

96722412 480 846036 96722412 480 846036 96722412 480 846036Number at risk

Low-risk 78 71 68 63 61 60 28 6 0 High-risk 78 67 53 46 44 41 7 4 0

78 73 72 68 64 63 29 6 0 78 69 56 52 51 44 7 4 0

78 77 72 68 65 63 29 6 0 78 75 64 57 54 49 7 4 0

Number at riskLow-risk 75 68 62 59 59 58 23 9 0

High-risk 81 71 59 52 49 45 12 4 0

75 69 64 62 61 61 25 9 0 81 72 62 56 55 49 13 5 0

75 72 67 66 65 62 25 9 0 81 78 68 63 59 54 14 5 0

Number at riskLow-risk 78 75 73 68 56 40 19 15 8

High-risk 75 64 58 52 33 19 10 6 2

78 75 74 70 57 41 19 15 8 75 68 63 58 38 22 13 9 4

78 76 75 74 60 43 21 16 8 75 73 70 64 45 25 15 11 5

D

A

G

E

B

H

F

C

I

Low-riskHigh-risk

HR 2·47 (95% CI 1·32–4·61)p=0·0052

HR 2·73 (95% CI 1·46–5·11)p=0·0019

HR 3·16 (95% CI 1·65–6·04)p=0·0011

HR 2·28 (95% CI 1·09–4·80)p=0·030

HR 3·48 (95% CI 1·57–7·75)p=0·0020

HR 2·39 (95% CI 1·05–5·42)p=0·037

HR 2·87 (95% CI 1·38–5·96)p=0·0051

HR 2·48 (95% CI 1·24–4·96)p=0·010

HR 3·07 (95% CI 1·34–7·01)p=0·0082

Figure 1: Kaplan-Meier curves of disease-free survival (DFS), distant metastasis-free survival (DMFS), and overall survival according to expression of the fi ve-microRNA (miRNA) signature in patients with nasopharyngeal carcinomaDFS (A), DMFS (B), and overall survival (C) in 156 patients in the training set. DFS (D), DMFS (E), and overall survival (F) in 156 patients in the internal validation set. DFS (G), DMFS (H), and overall survival (I) in 153 patients in the independent set. We calculated hazard ratios (HRs) and p values with an adjusted multivariate Cox regression analysis, including miRNA signature (high risk vs low risk), sex, age (<45 years vs ≥45 years), TNM stage (stage III–IV vs I–II), WHO pathological type (undiff erentiated non-keratinising vs diff erentiated non-keratinising vs keratinising squamous cell), radiotherapy period interruptions (0 days vs ≥1 day), radiotherapy boosting (no vs yes), concurrent chemotherapy (yes vs no), viral capsid antigen immunoglobulin A (<1:80 vs 1:80–1:320 vs ≥1:640), and early antigen immunoglobulin A (<1:10 vs 1:10–1:20 vs ≥1:40) as covariates for each analysis.

Articles


risk scores generally had higher expression levels of protective miRNAs and fewer died compared with those with high risk scores (appendix p 15). When this group was subdivided into the training (appendix p 16) and internal validation (appendix p 17) groups, we noted equivalent results.

In addition, with the same method used in the training set, we developed a new formula for the RT-PCR data from the 156 randomly selected nasopharyngeal carcin-oma tissues from the second randomisation. We calculated the following new risk score formula:

Risk score = (0·212 × expression value of miR-93) – (0·154 × expression value of miR-142-3p) – (0·183 × expression value of miR-29c) – (0·116 × expression value of miR-26a) – (0·141 × expression value of miR-30e)

Patients were divided into high-risk or low-risk groups with the median risk score of 3·97 as a cutoff point, and similar prognostic value was found in these 156 randomly selected patients (appendix p 18–20). We only randomly selected 156 samples from the second randomisation used for RT-PCR analyses.

To assess whether the miRNA signature had the same or similar prognostic value in diff erent populations, we applied the same formula and cutoff point developed from the RT-PCR data in the independent set (153 samples from West China Hospital of Sichuan University); these patients were divided into high-risk and low-risk groups. Patients with high risk scores had

shorter DFS, DMFS, and overall survival than did those with low risk scores (fi gure 1). Clinical characteristics did not diff er between the two groups (table 1).

In multivariate Cox regression analysis, miRNA signature and TNM stage were independent prognostic factors for DFS, DMFS, and overall survival in the combined training and validation sets (table 2). Much the same results were also noted in the training, internal validation, and independent sets (appendix p 8).

We did stratifi ed analyses of patients with stage II, III, and IV disease from the combined training and internal validation sets. Stage I patients were excluded from the analysis, because no relapse or metastasis had occurred. Compared with patients with low risk scores, patients with stage II, III, and IV disease and high risk scores had shorter DFS (stage II p=0·090, stage III p=0·015, and stage IV p=0·0040), DMFS (p=0·14, p=0·056, and p=0·030), and OS (p=0·015, p=0·075, and p=0·010; appendix pp 21–27). Appendix p 9 provides additional TNM stage subgroup analysis. We also did an exploratory subset analysis by concurrent chemotherapy, which indicated that patients with stage III–IV disease and low risk scores had a favourable response to concurrent chemotherapy but patients of the same stages with high risk scores did not benefi t from concurrent chemotherapy (appendix pp 30–35).

In ROC analysis to compare sensitivity and specifi city of survival prediction, the fi ve-miRNA signature showed a better prediction of survival than did age and sex with regard to DFS, DMFS, and overall survival in the train-ing and internal validation sets (fi gure 2); however, in the independent set a signifi cant diff erence in these measures was not observed between age and sex with the fi ve-miRNA signature. When compared with TNM stage, the predictive ability of the fi ve-miRNA signature was much the same (all p>0·05; fi gure 2).

To develop a more sensitive predictive tool, we constructed a prognostic score model combining two independent prognostic factors, fi ve-miRNA signature and TNM stage, based on the training set (appendix pp 10–11), and com pared its prognostic validity with the TNM stage alone and fi ve-miRNA signature alone models. Combination of the fi ve-miRNA signature and TNM stage had a better prognostic value than did TNM stage alone in the training set, which were corroborated in the validation and independent sets (fi gure 3).

DiscussionPrognostic assessment is crucial for formation of appropriate treatment choices. In routine clinical practice, the TNM staging system is the key prognostic determinant for patients with nasopharyngeal carcin-oma. However, large variations in the clinical outcomes of patients with the same cancer stage have been reported, suggesting that the present staging system is not adequate for prognosis. We developed a fi ve-miRNA signature that was predictive of survival of patients with

Hazard ratio (95% CI)

p value

Disease-free survival

miRNA signature (high risk vs low risk) 2·64 (1·70–4·11) <0·0001

TNM stage (III–IV vs I–II) 2·48 (1·45–4·25) 0·0013

Sex (male vs female) 1·85 (1·08–3·17) 0·025

Distant metastasis-free survival


TNM stage (III–IV vs I–II) 4·09 (1·87–8·97) <0·0001

Overall survival


TNM stage (III–IV vs I–II) 3·13 (1·61–6·10) 0·0009

Sex (male vs female) 2·09 (1·10–3·97) 0·024

We calculated hazard ratios and p values with an adjusted multivariate Cox proportional hazards regression model, including miRNA signature (high risk vs low risk), sex, age (<45 years vs ≥45 years), TNM stage (stage III–IV vs I–II), WHO pathological type (undiff erentiated non-keratinising vs diff erentiated non-keratinising vs keratinising squamous cell), radiotherapy period interruptions (0 days vs ≥1 day), radiotherapy boosting (no vs yes), concurrent chemotherapy (yes vs no), viral capsid antigen immunoglobulin A (<1:80 vs 1:80–1:320 vs ≥1:640), and early antigen immunoglobulin A (<1:10 vs 1:10–1:20 vs ≥1:40) as covariates. We selected variables with the backward stepwise approach. Only variables that were signifi cantly associated with survival are presented.

Table 2: Multivariable Cox regression analysis of microRNA (miRNA) signature expression and survival in the combined training and validation set

Articles


nasopharyngeal carcinoma independent of TNM stage, supporting the prognostic value of the signature and allowing clinicians to potentially identify candidates for aggressive therapy to improve treatment outcomes. In addition, ROC analysis suggested that the fi ve-miRNA signature has a similar survival predictive ability to TNM stage. TNM staging is key to assessment of prognosis and establishment of treatment strategy but is done mainly on the basis of anatomical information; conversely, the miRNA signature could show the biological characteristics of nasopharyngeal carcinoma and provide diff erent information from TNM staging. Identifi cation of several miRNAs in patients with quantitative RT-PCR might be a straightforward and clinically applicable procedure. In our study, combination of miRNA signature and TNM stage had a

better prognostic value than did TNM staging alone, suggesting that the miRNA signature reinforced the prognostic ability of TNM stage. Ultimately, patients classifi ed with the same TNM stage naso pharyngeal carcinoma might be able to be stratifi ed into diff erent disease recurrence-risk groups on the basis of the fi ve-miRNA signature, and thus treated with systemic approaches of diff erent intensities to improve outcomes. Such stratifi cation could lead to a more personalised treatment for patients with nasopharyngeal carcinoma (panel).

About a quarter of patients from European and US populations have keratinising nasopharyngeal carcinoma, whereas more than 95% of patients in China have non-keratinising disease, and the prognosis for every histological subtype is diff erent.29 Therefore, our study

Figure 2: Comparisons of the sensitivity and specifi city for prediction of survival by the fi ve-microRNA signature, TNM stage, age, and sex in patients with nasopharyngeal carcinomaReceiver operating characteristics (ROC) curves for DFS (A), DMFS (B), and overall survival (C) in the training set. ROC curves for DFS (D), DMFS (E), and overall survival (F) in the internal validation set. ROC curves for DFS (G), DMFS (H), and OS (I) in the independent set. p values show the area under the ROC (AUROC) of fi ve-microRNA signature versus the AUROC of TNM stage, age, or sex. DFS=disease free survival. DMFS=distant metastasis-free survival. OS=overall survival.

AUROC 95% CI p valuemicroRNA 0·70 0·61–0·79 TNM stage 0·62 0·54–0·71 0·20Sex 0·55 0·48–0·62 0·0045Age 0·54 0·45–0·62 0·014

Sens

itivi

ty

1·00

0·75

0·25

0·50

0

Sens

itivi

ty

1·00

0·75

0·25

0·50

0

Sens

itivi

ty

1·00

0·75

0·25

0·50

0

1–specificity0·500·25 1·000 0·75

1–specificity0·500·25 1·000 0·75

1–specificity0·500·25 1·000 0·75

microRNATNM stageSexAgeReference









A Training: DFS B Training: DMFS C Training: OS

D Internal validation: DFS E Internal validation: DMFS F Internal validation: OS

G Independent: DFS H Independent: DMFS I Independent: OS

Articles


has limitations in terms of generalisability because all specimens were obtained from patients in China (where nasopharyngeal carcinoma is prevalent). Add itional sets of independent samples from non-Asian patients will be needed to confi rm our fi ndings before this fi ve-miRNA signature can be used in the clinic globally. Furthermore, the mechanism behind the prognostic value of the fi ve-miRNA signature in nasopharyngeal carcinoma is unclear, and further studies on these miRNAs might

provide more information for better understanding the roles of miRNAs in the development and progression of nasopharyngeal carcinoma. Additionally, 873-probe microarray was used in the study. Because the discoveries of miRNAs are fast, we cannot profi le every miRNA that has been identifi ed.

At present, the standard treatment for locoregionally advanced nasopharyngeal carcinoma is concurrent chemoradiotherapy with adjuvant or induction chemo-therapy, and 30% of patients die due to distant metastasis.30 This pattern of failure suggests that a subgroup of patients does not benefi t from present chemoradiotherapy strategies. Thus, the accurate iden-tifi cation of subgroups of patients will improve the prognostic system and lead to more personalised therapy. Patients with high-risk scores had poorer survival than did those with low-risk scores, and thus present chemoradiotherapy strategies are not suffi cient for such patients. Further studies are needed to identify more intensive systemic approaches to improve the treatment outcomes of patients with high-risk scores. Promisings results have been seen in the palliative setting with taxanes and gemcitabine, and with molecularly targeted drugs.31,32 Therefore, we suggest that patients with high-risk scores should be treated with new combinations of more tolerable drugs as an adjunct to potentiate the effi cacy of systemic control. Furthermore, miRNAs are candidate therapeutic targets, and inhibition or over-expression of specifi c miRNAs might have therapeutic implications.33 The function and mechanism of the fi ve miRNAs in nasopharyngeal carcinoma pathogenesis have not been shown thoroughly, and further inves-tigation into their functions may provide additional targets and strategies for treatment.

To our knowledge, this study possesses the largest sample size used to date to analyse the potential

Sens

itivi

ty

1·00

0·75

0·25

0·50

0

1–specificity0·500·25 1·000 0·75

1–specificity0·500·25 1·000 0·75

1–specificity0·500·25 1·000 0·75

A B CTNM stage and microRNATNM stagemicroRNAReference

AUROC 95% CI p valueTNM stage 0·68 0·60–0·76 andmicroRNA TNM stage 0·60 0·52–0·67 0·013microRNA 0·64 0·56–0·73 0·21



Figure 3: Comparisons of the sensitivity and specifi city for the prediction of survival by the combined fi ve-microRNA (miRNA) signature and TNM stage model, the TNM stage alone model, and the fi ve-miRNA signature alone modelReceiver operating characteristics (ROC) curves in the training set (A), internal validation set (B), and independent set (C). p values show the area under the ROC (AUROC) of the combined fi ve-miRNA signature and TNM stage model versus AUROCs of the TNM stage alone model or the fi ve-miRNA signature alone model.


Systematic reviewThe TNM staging system is inadequate for defi nition of prognosis, and biomarkers could add prognostic value to the staging system. miRNAs are dysregulated in most tumour types,7,8 and the prognostic value of these miRNAs has been shown for chronic lymphocytic leukaemia,9 lung cancer,10,11 pancreatic cancer,12 colon adenocarcinoma,13 liver cancer,14 and gastric cancer.15 We systemically searched PubMed for reports published between Jan 1, 1993, and Jan 10, 2011, with the terms “nasopharyngeal carcinoma”, “microRNA”, and “prognosis”. No publications on the relation between miRNA expression profi les and prognosis of nasopharyngeal carcinoma were identifi ed. Only a few reports showed that miRNAs were expressed aberrantly in nasopharyngeal carcinoma tissues and promote an aggressive tumour phenotype by changing the expression of their mRNA targets.17,18 These studies showed that miRNAs had important roles in the development and progression of nasopharyngeal carcinoma and could serve as novel prognostic biomarkers. Therefore, we did a study to assess the prognostic value of miRNAs for patients with nasopharyngeal carcinoma.

InterpretationOur study assessed a large number of patients to analyse the potential prognostic value of miRNAs in nasopharyngeal carcinoma, and used three sets of patients for internal and external validation. We noted systematic changes in miRNA expression in nasopharyngeal carcinoma tissues, and a fi ve-miRNA signature was associated with survival independent of stage. The fi ve-miRNA signature could add prognostic value to TNM staging. These results suggest that the fi ve-miRNA signature might be a useful biomarker for prognosis and help development of personalised therapy.

Articles


prognostic value of miRNAs in nasopharyngeal carcin-oma. Our results suggest that the fi ve-miRNA signature may be a useful biomarker for prognosis, leading to increasingly personalised therapy. We acknowledge that prospective, large-scale, multicentre studies are necessary to confi rm our results before this miRNA signature can be applied in the clinic, but the role of these miRNAs in nasopharyngeal carcinoma warrants further study.

ContributorsJM and H-YW designed the study. NL, N-YC, R-XC, W-FL, YL, R-RW,

M-YZ, J-PY, JZ, L-ZL, and LL collected the data. JM, H-YW, NL, N-YC,

YS, B-JH, MC, Q-MH, NJ, LC, YG, WCSC did the data analysis and

interpretation. NL, JM, H-YW wrote the report. WCSC, N-YC, R-XC,

W-FL, YL, R-RW, M-YZ, YS, B-JH, MC, Q-MH, NJ, LC, J-PY, JZ, L-ZL,

LL, and YG revised the report. NL, YG, and R-RW did the statistical

analysis. All authors reviewed the report and approved the fi nal version.


AcknowledgmentsWe thank Kwok-wai Lo (Department of Anatomical and Cellular Pathology,

Prince of Wales Hospital, The Chinese University of Hong Kong, Hong

Kong, China) for helpful discussion and editorial assistance. This work

was supported by grants from the Science Foundation of the Key Hospital

Clinical Program of the Ministry of Health, China (2010-178), the National

Natural Science Foundation of China (81071835), the Guangdong Province

Universities and Colleges Pearl River Scholar Funded Scheme (2011,

2050205), the Key Scientifi c and Technological Innovation Program for

Universities of Guangdong Province (No.cxzd1005), the Guangdong

Natural Science Foundation (10151008901000099), and the Fundamental

Research Funds for the Central Universities (10ykjc08).

References1 Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global

cancer statistics. CA Cancer J Clin 2011; 61: 69–90.

2 Cuardo MP, Edwards BK, Shin HR, et al. Cancer incidence in fi ve continents. Vol IX. Lyon, France: International Agency for Research on Cancer, 2007.

3 Mao YP, Xie FY, Liu LZ, et al. Re-evaluation of 6th edition of AJCC staging system for nasopharyngeal carcinoma and proposed improvement based on magnetic resonance image. Int J Radiat Oncol Biol Phys 2009; 73: 1326–34.

4 Leung SF, Zee B, Ma BB, et al. Plasma Epstein-Barr viral deoxyribonucleic acid quantitation complements tumor-node-metastasis staging prognostication in nasopharyngeal carcinoma. J Clin Oncol 2006; 24: 5414–18.

5 Turen S, Ozyar E, Altundag K, Gullu I, Atahan IL. Serum lactate dehydrogenase level is a prognostic factor in patients with locoregionally advanced nasopharyngeal carcinoma treated with chemoradiotherapy. Cancer Invest 2007; 25: 315–21.

6 Lv X, Xiang YQ, Cao SM, et al. Prospective validation of the prognostic value of elevated serum vascular endothelial growth factor in patients with nasopharyngeal carcinoma: more distant metastases and shorter overall survival after treatment. Head Neck 2011; 33: 780–85.

7 Lu J, Getz G, Miska EA, et al. MicroRNA expression profi les classify human cancers. Nature 2005; 435: 834–38.

8 Calin GA, Croce CM. MicroRNA signatures in human cancers. Nat Rev Cancer 2006; 6: 857–66.

9 Calin GA, Ferracin M, Cimmino A, et al. A MicroRNA signature associated with prognosis and progression in chronic lymphocytic leukemia. N Engl J Med 2005; 353: 1793–801.

10 Yanaihara N, Caplen N, Bowman E, et al. Unique microRNA molecular profi les in lung cancer diagnosis and prognosis. Cancer Cell 2006; 9: 189–98.

11 Yu SL, Chen HY, Chang GC, et al. MicroRNA signature predicts survival and relapse in lung cancer. Cancer Cell 2008; 13: 48–57.

12 Bloomston M, Frankel WL, Petrocca F, et al. MicroRNA expression patterns to diff erentiate pancreatic adenocarcinoma from normal pancreas and chronic pancreatitis. JAMA 2007; 297: 1901–08.

13 Schetter AJ, Leung SY, Sohn JJ, et al. MicroRNA expression profi les associated with prognosis and therapeutic outcome in colon adenocarcinoma. JAMA 2008; 299: 425–36.

14 Ji J, Shi J, Budhu A, et al. MicroRNA expression, survival, and response to interferon in liver cancer. N Engl J Med 2009; 361: 1437–47.

15 Ueda T, Volinia S, Okumura H, et al. Relation between microRNA expression and progression and prognosis of gastric cancer: a microRNA expression analysis. Lancet Oncol 2010; 11: 136–46.

16 Cho WC. MicroRNAs: potential biomarkers for cancer diagnosis, prognosis and targets for therapy. Int J Biochem Cell Biol 2010; 42: 1273–81.

17 Sengupta S, den Boon JA, Chen IH, et al. MicroRNA 29c is down-regulated in nasopharyngeal carcinomas, up-regulating mRNAs encoding extracellular matrix proteins. Proc Natl Acad Sci USA 2008; 105: 5874–78.

18 Chen HC, Chen GH, Chen YH, et al. MicroRNA deregulation and pathway alterations in nasopharyngeal carcinoma. Br J Cancer 2009; 100: 1002–11.

19 Chen Y, Liu MZ, Liang SB, et al. Preliminary results of a prospective randomized trial comparing concurrent chemoradiotherapy plus adjuvant chemotherapy with radiotherapy alone in patients with locoregionally advanced nasopharyngeal carcinoma in endemic regions of China. Int J Radiat Oncol Biol Phys 2008; 71: 1356–64.

20 Simon R, Radmacher MD, Dobbin K, McShane LM. Pitfalls in the use of DNA microarray data for diagnostic and prognostic classifi cation. J Natl Cancer Inst 2003; 95: 14–18.

21 Korbler T, Grskovic M, Dominis M, Antica M. A simple method for RNA isolation from formalin-fi xed and paraffi n-embedded lymphatic tissues. Exp Mol Pathol 2003; 74: 336–40.

22 Wang HY, Luo M, Tereshchenko IV, et al. A genotyping system capable of simultaneously analyzing >1000 single nucleotide polymorphisms in a haploid genome. Genomen Res 2005; 15: 276–83.

23 Wang H, Ach RA, Curry B. Direct and sensitive miRNA profi ling from low-input total RNA. RNA 2007; 13: 151–59.

24 Biometric Research Branch-ArrayTools. http://linus.nci.nih.gov/BRB-ArrayTools.html (accessed Dec 1, 2009).

25 MeV: MultiExperiment Viewer. http://www.tm4.org/mev.html (accessed Aug 1, 2008).

26 Lossos IS, Czerwinski DK, Alizadeh AA, et al. Prediction of survival in diff use large-B-cell lymphoma based on the expression of six genes. N Engl J Med 2004; 350: 1828–37.

27 Chow SC, Shao J, Wang H. Sample size calculations in clinical research. New York, USA: Marcel Dekker, 2003.

28 Yang H-I, Yuen M-F, Chan HL-Y, et al, for the REACH-B Working Group. Risk estimation for hepatocellular carcinoma in chronic hepatitis B (REACH-B): development and validation of a predictive score. Lancet Oncol 2011; 12: 568–74.

29 Marks JE, Philips JL, Menck HR, et al. The national cancer data base report on the relationship of race and national origin to the histology of race and national origin to the histology of nasopharyngeal carcinoma. Cancer 1998; 83: 582–88.

30 Hui EP, Leung SF, Au JS, et al. Lung metastasis alone in nasopharyngeal carcinoma: a relatively favorable prognostic group. A study by the Hong Kong Nasopharyngeal Carcinoma Study Group. Cancer 2004; 101: 300–06.

31 Leong SS, Wee J, Tay MH, et al. Paclitaxel, carboplatin, and gemcitabine in metastatic nasopharyngeal carcinomas. Cancer 2005; 103: 569–75.


33 Cho WC. MicroRNAs in cancer—from research to therapy. Biochim Biophys Acta 2010; 1805: 209–17.


Review

Lancet Oncol 2012; 13: e240–48

Breast Cancer Translational Research Laboratory, Faculty of Medicine, Université Libre de Bruxelles, Institut Jules Bordet, Brussels, Belgium (D Fumagalli MD, S Michiels PhD, C Sotiriou MD, S Loi MD); Division of Medical Oncology and Hematology, Princess Margaret Hospital, Department of Medicine, University of Toronto, Toronto, ON, Canada (P L Bedard MD); Division of Hematology-Oncology, Department of Internal Medicine TTUHSC-Paul L Foster School of Medicine, El Paso, TX, USA (Z Nahleh MD); Albert Einstein College of Medicine, Montefi ore Medical Center, Bronx, NY, USA (Prof J A Sparano MD); Siteman Cancer Center, Washington University in St Louis, St Louis, MO, USA (Prof M Ellis MD); Department of Radiology and Biomedical Imaging, University of California San Francisco, San Francisco, CA, USA (Prof N Hylton PhD); National Cancer Institute, Cancer Therapy Evaluation Program, Division of Cancer Treatment and Diagnosis, Bethesda, MD, USA (J A Zujewski MD); Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY, USA (Prof C Hudis MD); Department of Surgery and Radiology, University of California San Francisco, San Francisco, CA , USA (Prof L Esserman MD); and Breast International Group, Institut Jules Bordet, Brussels, Belgium (Prof M Piccart MD)

Correspondence to:Prof Martine Piccart, Breast International Group, Institut Jules Bordet, Boulevard de Waterloo 121, 1000 Brussels, [email protected]

A common language in neoadjuvant breast cancer clinical trials: proposals for standard defi nitions and endpointsDebora Fumagalli, Philippe L Bedard, Zeina Nahleh, Stefan Michiels, Christos Sotiriou, Sherene Loi, Joseph A Sparano, Matthew Ellis, Nola Hylton, Jo Anne Zujewski, Cliff ord Hudis, Laura Esserman, Martine Piccart, on behalf of the BIG-NABCG collaboration

The neoadjuvant setting provides a unique opportunity to study the eff ect of systemic treatments on breast cancer biology and to identify clinically useful prognostic and predictive biomarkers. Discrepancies and incon sistencies in the use of defi nitions and endpoint assessments in this setting confound the analysis and interpretation of results across clinical trials and hinder research progress. This Review represents a joint eff ort of the Breast International Group and the National Cancer Institute-sponsored North American Breast Cancer Group to provide clinicians and researchers with a series of standardised defi nitions and endpoints that could be implemented in future neoadjuvant clinical trials. Defi nitions of the setting of interest and of survival endpoints are recommended, together with proposals for standard assessment of the response to treatment, use of functional and molecular imaging endpoints, and characterisation and selection of the population to treat. We expect that implementation of these recommendations will improve the conduct, reporting, and eff ectiveness of clinical trials and fully exploit the clinical and scientifi c potential of the neoadjuvant setting in breast cancer.

IntroductionNeoadjuvant chemotherapy has been used for treatment of breast cancer since the 1970s, when it was shown to induce tumour response and to facilitate local control through subsequent surgery and radiation therapy in patients with locally advanced and infl ammatory breast cancer.1 The role of neoadjuvant treatment has evolved since this time (fi gure).

Results of a fi rst generation of randomised clinical trials showed not only that neoadjuvant chemotherapy could safely downstage operable tumours so that con servation surgery could be considered as an alternative to radical mastectomy but also that the long-term rates of distant tumour control were comparable with those obtained with standard adjuvant systemic therapy.2–4 Furthermore, the neoadjuvant setting off ered the oppor tunity to assess the prognostic factor pathological complete response (pCR)—ie, the absence of residual invasive disease after neoadjuvant chemotherapy in the breast, lymph nodes, or both. Patients with pCR had a superior long-term outcome when compared with patients with residual tumour at the time of surgery, suggesting that pCR could be a marker for long-term eff ects of systemic treatment on dis-seminated tumour cells.

A second generation of neoadjuvant clinical trials focused on addition of new agents and alternative treatment schedules.5–7 Although addition of taxanes generally led to higher pCR rates, a clinically meaningful improvement in long-term outcomes was not shown consistently, leading many investigators to conclude that early improvements in pCR rates cannot yet act as surrogate endpoints. However, most neoadjuvant trials undertaken so far have enrolled unselected populations of patients. More recently, the genomic complexity of breast cancer has started to be appreciated, with several subtypes with specifi c molecular profi les, clinical behaviours, and response to treatment being recog-nised.8–12 A remarkable increased rate of pCR has been

recorded with use of the monoclonal antibody trastuzumab to treat the population of patients over-expressing its target, the human epidermal growth factor receptor 2 (HER2).13,14 Similarly, endocrine therapy is now used to raise the breast conservation rate in post meno-pausal women with oestrogen receptor (ER)-positive breast cancer, a subtype less sensitive to chemo therapy than ER-negative disease.15,16

These considerations highlight the importance of designing trials in which the molecular characteristics of tumours are used to select the most eff ective neoadjuvant treatment. However, the lack of consensus on standard defi nitions and endpoint assessments in this setting still confounds the analysis and interpretation of results across clinical trials and diminishes the power to draw robust conclusions.

This Review represents an attempt by the Breast International Group (BIG) and the National Cancer Institute-sponsored North American Breast Cancer Group (NABCG) to provide a series of standard defi nitions and endpoints to be implemented in neo adjuvant clinical trials in breast cancer, to standardise their conduct and reporting. By summarising the crucial issues and making recommendations, we hope that this Review will provoke debate on more contentious matters so that consensus can ultimately be reached or new studies initiated.

Panel 1 summarises all proposed defi nitions and endpoint assessments. Our recommendations can be deemed applicable to all breast cancer histological types, although their validity for favourable histological subtypes—such as tubular, colloid, and adenoid cystic—needs to be investigated further.17,18

Proposal for the standard defi nition of the setting of interestCurrently, the terms neoadjuvant and presurgical are used interchangeably, and their proper defi nition is still confused. We recommend that, in clinical trials,

e241 www.thelancet.com/oncology Vol 13 June 2012

Review

neoadjuvant should be used when referring to a treatment given before surgery, with therapeutic intent—ie, to downstage the disease or to determine the effi cacy of a regimen against early breast cancer; and presurgical should be used when referring to an intervention undertaken before surgery, without thera-peutic intent but with the main aim to investigate the biological or pharmacodynamic eff ect of a compound on breast cancer.

At present, presurgical trials are mostly referred to as biological window trials—ie, those in which a short course (usually 1–3 weeks) of a compound is administered before surgery to test its short-term biological or pharmacodynamic activity on an in-vivo target rather than a conventional therapeutic effi cacy endpoint.19,20 In the future, other designs of presurgical trials could possibly be developed.

Proposal for the standard evaluation of the response to treatmentPathological complete responseThe extent of residual tumour after neoadjuvant treatment can be regarded as an intermediate endpoint for breast cancer relapse and survival, and patients who achieve pCR after neoadjuvant treatment benefi t from substantially longer disease-free and overall survival when compared with patients with residual disease.2 Currently, terminology used to assess the pathological response to neoadjuvant treatment and to defi ne pCR varies between clinical trials (panel 2), and caution should be exercised when comparing results across trials that used diff erent defi nitions.2,6,21–24 Reaching a consensus on a standard defi nition of pCR would allow for consistent interpretation of clinical trial results and permit cross-trial comparisons.

Retrospective analysis of a database including 2302 pa tients with breast cancer treated with neoadjuvant chemotherapy at MD Anderson Cancer Center indicated no signifi cant diff erence in disease-free and overall survival between patients with pCR and residual ductal

carcinoma in situ (DCIS).25 Other studies have shown that when no residual invasive cancer was detected in the breast, the number of involved axillary lymph nodes was inversely correlated with survival and, conversely, patients who converted to node-negative status after treatment had excellent survival, even if residual disease was present in the breast.26

Taking into account these observations, we recommend that in neoadjuvant clinical trials, pCR is defi ned as the absence of residual invasive cancer within both the breast and lymph nodes. In view of the importance of adequate gross evaluation, sampling, and reporting of the post-treatment specimen to defi ne tumour response, for examination of breast specimens at surgery we recommend investigators follow the guidelines of the College of American Pathologists.27

American Joint Committee on Cancer stage and residual (proliferative) cancer burden for assessment of residual diseaseIn the neoadjuvant setting, the stage—initial clinical or fi nal pathological—that is most meaningful in terms of prognosis and further treatment decisions is currently unclear. This area needs to be investigated further. Inter-national guidelines recommend use of the American Joint Committee on Cancer (AJCC) tumour-nodes-metastasis (TNM) system to defi ne the clinical and pathological stage of disease,28,29 and reports indicate that prognosis for patients treated with neoadjuvant chemo therapy was best determined using the fi nal pathological stage defi ned by the 2003 AJCC TNM system, when compared with other systems.30

Evaluation of residual disease after neoadjuvant treat-ment is a tricky task. Moreover, even if several reports have shown that the extent of residual disease after neoadjuvant treatment is associated with relapse and survival,2–7,21–24 at present, no consensus exists on how to best evaluate residual disease and what meaning to attribute to near-pCR responses.

In an attempt to better defi ne the extent of residual disease, investigators at the MD Anderson Cancer Center have defi ned a composite risk of residual (proliferative) cancer burden (RCB) that combines relevant pathological characteristics of primary tumour dimension, cellularity of the tumour bed, and axillary nodal burden with prognostic importance. Through their work, RCB was shown to be strongly prognostic, as both a continuous and a categorical variable.31–33 RCB could in fact be divided into four classes (RCB-0 [pCR], RCB-I [near-PCR], RCB-II [moderate residual disease], and RCB-III [extensive residual disease]) that were reported to add prognostic power when compared with post-treatment pathological AJCC stage for stage II/III tumours, although they did not add signifi cant prognostic information for stage I tumours.31 Therefore, RCB could provide an accurate way to characterise and defi ne the extent of residual disease after neoadjuvant chemotherapy.

1970 1980 1990 2000

Locally advanced

↑Local control

Yes

Early

↑Breastconservation

Yes

Early

↑Survival

No

Early

↑Treatment tailoring

Probably

↑Treatment tailoring

Probably

↑Breast conservation

Yes

Treatment of unfit elderly women

Yes

Chemotherapy

Stage of disease

Goal of neoadjuvant therapy

Outcome achieved

Endocrine therapy

Goal of neoadjuvant therapy

Outcome achieved

Figure: Evolution of neoadjuvant treatment for breast cancer


Review

We recommend that in neoadjuvant trials using chemotherapy, besides using the AJCC defi nition for stage, RCB should be defi ned at surgery. Some variables used to defi ne RCB are not captured routinely in synoptic reports; however, they can be easily obtained from pathological review, and clear guidelines about how to describe and sample the tumour bed and how to describe and report microscopic fi ndings are available.34

For future clinical trials, it could be useful to set up working groups in which local pathologists engaged in defi nition of RCB are trained and sustained by in-person

or web-based meetings (or both). This training would also help in standardising the procedure across insti-tutions worldwide and in increasing its applicability.

Ki67Determining the best schedule to administer neoadjuvant hormone treatment and how to subsequently assess pathological response on the surgical specimen are still matters for discussion. With currently used schedules, pCR is only very rarely recorded after neoadjuvant hormone treatment (about 1% of cases),16 and many

Proposal for the standard defi nition of the setting of interestNeoadjuvant: should be used when referring to a treatment given before surgery, with therapeutic intent—ie, to downstage the disease or to determine the effi cacy of a regimen against early breast cancer.

Presurgical: should be used when referring to an intervention undertaken before surgery, without therapeutic intent but with the main aim to investigate the biological or pharmacodynamic eff ect of a compound on breast cancer.

Proposal for the standard evaluation of the response to treatmentpCR: we recommend defi ning pCR as the absence of residual invasive cancer within both the breast and lymph nodes.

AJCC stage and RCB: in neoadjuvant trials using chemotherapy, besides using the AJCC defi nition for stage, RCB should be defi ned at surgery.

Ki67: for patients receiving neoadjuvant endocrine treatment in the context of clinical trials, we recommend assessment of Ki67 on baseline biopsy samples, on biopsy specimens collected during treatment, and on surgical specimens and that its score be recorded as an absolute value for research purposes.

PEPI: we recommend assessment of the PEPI score 12–16 weeks after treatment in neoadjuvant trials using endocrine therapy, for research purposes.

Proposal for the standard defi nition of survival endpointsWe recommend that in neoadjuvant clinical trials, survival endpoints should be defi ned using standardised defi nitions for effi cacy endpoints and events proposed for the adjuvant setting in the STEEP system, taking into consideration that the starting point of the time-to-event variables should generally be as follows:• The date of the fi rst course of treatment in patients treated

with therapeutic intent—ie, in the neoadjuvant setting.• The date of surgery for all other cases—ie, in presurgical or

biological window trials.

Proposal for the use of functional and molecular imaging as endpointsMRI: we recommend restricting use of MRI to clinical trials for research purposes and to perform it at the following timepoints:

• Before the start of chemotherapy.• At an early timepoint—eg, after the fi rst cycle of treatment.• In the event progressive disease is clinically suspected.• After the fi nal course of chemotherapy to assess residual

disease.• If treatment with two non-crossresistant cytotoxic regimens

is foreseen, before switching to the second regimen to assess the response after treatment with the fi rst regimen.

FDG-PET: we recommend using FDG-PET in neoadjuvant clinical trials for research purposes only.

Ultrasonography: we recommend using ultrasonography in neoadjuvant clinical trials if MRI is not available, or in selected circumstances, such as the staging of nodal disease. It could also be used for placement of clips to mark the original biopsy site.

Mammography: we recommend using mammography in neoadjuvant clinical trials before the start of treatment only if other imaging modalities are not available and after treatment only for clip/tumour localisation for defi nitive surgical resection.

Proposal for the standard characterisation of the population to treatIn neoadjuvant clinical trials, we recommend that breast cancer subtypes should be defi ned with immunohistochemical assays done according to approved ASCO/CAP guidelines, possibly in a central laboratory, with their defi nitions as follows:• Luminal A subtype: ER-positive and/or PgR-positive,

HER2-negative, low proliferative (Ki67 <14%)*.• Luminal B, HER2-negative subtype: ER-positive and/or

PgR-positive, HER2-negative, high proliferative (Ki67 ≥14%)*.• Luminal B, HER2-positive subtype: ER-positive and/or

PgR-positive, HER2-positive, high proliferative (any Ki67)*.• HER2-positive subtype: HER2-positive, ER-negative, and

PgR-negative.• Triple negative subtype: ER-negative, PgR-negative, and

HER2-negative.

AJCC=American Joint Committee on Cancer. ASCO=American Society of Clinical Oncology. BIG=Breast International Group. CAP=College of American Pathologists. ER=oestrogen receptor. FDG=fl uorodeoxyglucose. NABCG=North American Breast Cancer Group. pCR=pathological complete response. PEPI=Preoperative Endocrine Prognostic Index. PgR=progesterone receptor. RCB=residual (proliferative) cancer burden. STEEP=standardised defi nitions for effi cacy endpoints. *The defi nition of high and low proliferative luminal tumours, and the optimum cutoff to be used for Ki67, might change in the future, and therefore we recommend to record absolute values.

Panel 1: BIG and NABCG proposals for standard defi nitions and endpoints in neoadjuvant breast cancer clinical trials


Review

patients with residual disease have favourable long-term outcome. Therefore, alternative biological markers of response would be more useful for patients receiving such treatment. Tumour immunostaining for Ki67, a biomarker that measures proliferation, seems promising in this respect.15,16,35–38

The Immediate Preoperative Arimidex, tamoxifen, or Combined with Tamoxifen (IMPACT) study, in which a comparison was done of neoadjuvant anastrozole versus tamoxifen versus the combination of the two,15,35–37 and the P024 study of neoadjuvant letrozole versus

tamoxifen16,38 are two neoadjuvant endocrine trials that have evaluated retrospectively use of Ki67 as a primary endpoint. Follow-up data of patients enrolled in IMPACT indicated that the Ki67 level measured 2 weeks after treatment was a better predictor of relapse-free survival than pretreatment levels.39 These fi ndings suggest that Ki67 could be used as an intermediate endpoint, able to provide evidence of therapeutic eff ectiveness during endocrine treatment, and that serial Ki67 analysis from baseline to 2 weeks after treatment provides an index of residual risk following short-term endocrine therapy.

The implications of changes in Ki67 staining during neoadjuvant chemotherapy are less robust although, in a recent study, Ki67 levels in the residual tumour in patients without pCR were associated strongly with outcome.40 This fi nding suggests that measurement of Ki67 could be used to identify patients for trials of additional adjuvant treatment after neoadjuvant chemotherapy.

One drawback of Ki67 as a biomarker is the lack of agreement about the analytical method, interpretation or scoring method, and data handling strategy to use when determining its value, but guidelines addressing this issue have recently been published.41 Moreover, currently insuffi cient evidence is available to establish a standard cutoff for the value of Ki67 as a response endpoint. Nevertheless, acknowledging the potential of this marker, we recommend that for patients receiving neoadjuvant endocrine treatment in the context of clinical trials, Ki67 should be assessed on baseline biopsy samples, on biopsy specimens collected during treatment, and on surgical specimens, and its score should be recorded as an absolute value for research purposes.

Preoperative Endocrine Prognostic IndexKi67 data and other clinical and pathological variables were recently reanalysed in the P024 study to evaluate their ability to predict long-term outcome in patients treated with endocrine therapy.42 The Ki67 index, patho-logical tumour size, nodal status, and ER status in tumours after 4 months of neoadjuvant endocrine therapy were shown to be independently associated with relapse-free and overall survival. The Preoperative Endocrine Prognostic Index (PEPI; table), derived from the combination of these factors and validated subsequently with an independent dataset from the IMPACT trial, identifi ed a group of patients at very low-risk of recurrence who might be spared additional chemotherapy on the basis of favourable prognosis alone. Of note, the authors have suggested that patients with high PEPI scores are those who should receive chemotherapy, since their tumours seem to be relatively endocrine resistant.

On the basis of these data, for the time being, we recom mend assessment of the PEPI score 12–16 weeks after treatment in neoadjuvant clinical trials using endocrine therapy for research purposes. These recom-mendations could evolve as our understanding of neo-adjuvant endocrine therapy grows.

Panel 2: Examples of response classifi cation systems used in neoadjuvant breast cancer clinical trials

Milan Cancer Institute2

• Pathological complete response: no invasive tumour in the breast• Good partial response: >50% reduction in bidimensional tumour size• Minor or non-response: <50% reduction in bidimensional tumour size

Chevallier21

• Grade 1: disappearance of all tumour either on macroscopic or microscopic assessment• Grade 2: presence of in-situ carcinoma in the breast, no invasive tumour, and no

tumour found in the lymph nodes• Grade 3: presence of invasive carcinoma with stromal alteration (eg, sclerosis or

fi brosis)• Grade 4: no or few modifi cations of the tumour appearance

Sataloff 22

Primary site:• T-A: total or near therapeutic eff ect• T-B: subjectively >50% therapeutic eff ect, but less than total or near total• T-C: <50% therapeutic eff ect, but eff ect evident• T-D: no therapeutic eff ect

Axillary lymph nodes:• N-A: evidence of therapeutic eff ect, no metastatic disease• N-B: no lymph-node metastasis or therapeutic eff ect• N-C: evidence of therapeutic eff ect, but lymph-node metastasis still present• N-D: viable metastatic disease, no therapeutic eff ect

National Surgical Adjuvant Breast and Bowel Project (NSABP) classifi cation6

• Pathological complete response: no invasive tumour in the breast

Ogston23 (Miller and Payne grading system)Primary site response:• Grade 1: no change or some alteration to individual malignant cells but no reduction

in overall cellularity• Grade 2: a minor loss of tumour cells but overall cellularity still high, up to 30%• Grade 3: between an estimated 30% and 90% reduction in tumour cells• Grade 4: a signifi cant disappearance of tumour cells such that only small clusters or

widely dispersed individual cells remain; more than 90% loss of tumour cells• Grade 5: no malignant cells identifi able in sections from the site of the tumour; only

vascular fi broelastotic stroma remains often containing macrophages; ductal carcinoma in situ might be present

MD Anderson Cancer Center classifi cation24

• Pathological complete response: no histopathological evidence of any residual invasive cancer cells in breast or axillary lymph nodes


Review

Proposal for the standard defi nition of survival endpointsConsistency in reporting of survival endpoints is lacking, mostly in relation to inclusion and exclusion of events in defi nitions of disease-related endpoints. This issue has already been addressed with respect to breast cancer clinical trials in the adjuvant setting, for which a panel of experts convened to formulate recommendations for standard defi nitions for effi cacy endpoints and events (referred to as the STEEP system).43

An issue of particular concern in the neoadjuvant setting is how to defi ne the starting point for assessment of time-to-event data. For instance, in neoadjuvant trials, disease-free survival has been defi ned inconsistently either from the date of study entry or from the date of surgery. This issue is complicated further because some patients do not undergo surgery after treatment, aff ecting assessment of events defi ned from the date of surgical excision.

We recommend that in neoadjuvant trials, survival endpoints should be defi ned using STEEP system proposals for the adjuvant setting. This recommendation takes into consideration that the starting point of these events should generally be either the date of the fi rst course of treatment in patients treated with therapeutic intent—ie, in the neoadjuvant setting—or the date of surgery for all other cases (ie, presurgical or biological window trials). Being aware that investigators might need to modify a standard endpoint for a trial-specifi c reason, we recommend that the name of the modifi ed survival endpoint should clearly refl ect the endpoint being measured.

Proposal for the use of functional and molecular imaging endpointsUse of functional and molecular imaging in the neoadjuvant setting could enable assessment of tumour biology in vivo during treatment, to measure early response and predict tumour behaviour, so that patients who are unlikely to respond to further therapy could be spared the toxic eff ects of additional ineff ective treatments. Standard imaging techniques such as mammography and ultrasonography are based on anatomical features and physical coordinates (ie, changes in tumour size) whereas newer modalities such as MRI and PET can also image functional and molecular tissue properties (eg, perfusion, metabolism, and marker expression).44 This aspect is very important, since anatomical measurement alone might be insuffi cient for detection of treatment response. Neoadjuvant systemic treatment induces cancer cell death by promotion of apoptosis and cell necrosis, with reduction of tumour volume and tumour shrinkage. However, in some cases, either the connective tissue component of the tumour might persist or destroyed cancer cells can be replaced by a hyaline amorphous scar, resulting in misinter pretation of the residual

tumour mass. Moreover, molecularly targeted agents—used increasingly in the neoadjuvant setting—can produce modest tumour shrinkage, which could be diffi cult to measure with traditional imaging technologies.

MRIThe role of MRI in the neoadjuvant setting is controversial with respect to both assessment of disease extent and response to treatment. In studies comparing diff erent imaging modalities, MRI was superior to clinical examination, mammography, and breast ultrasound for detection and staging of invasive breast cancer.45,46 Moreover, MRI measurements after neo-adjuvant treatment are more highly correlated with patho logical tumour size than mammography or ultra-sound.47,48 Several studies have reported that changes during treatment of MRI-defi ned variables, such as tumour apparent diff usion coeffi cient and tumour volume, could predict treatment response and survival.49,50 Findings suggest that the value of MRI could be of particular importance for some breast cancer subtypes.51,52

Despite this promise, MRI has some drawbacks. False-positive fi ndings on breast MRI can arise after neoadjuvant chemotherapy, and surgical decisions should not be based solely on MRI fi ndings. Furthermore, MRI could overestimate the extent of residual disease whereas in other instances it is unable to detect small residual tumour foci that can persist after neoadjuvant chemotherapy.53–55 Other issues, such as the need for a dedicated breast coil, for an expert skilled breast imaging team, and for standardisation of technical procedures across multi-institutional trials, should also not be neglected. Moreover, insuffi cient evidence is available to support routine use of MRI to assess the response to

Score

Maximum tumour diameter

≤5 cm 0

>5 cm 3

Lymph-node status

Negative 0

Positive 3

Ki67 immunostaining*

0–2·7% 0

>2·7–19·7% 1

>19·7–53·1% 2

>53·1% 3

Oestrogen receptor immunostaining†

0–2 3

3–8 0

PEPI=Preoperative Endocrine Prognostic Index.*% of positive cells. †Allred staining (scoring category).

Table: PEPI scoring system


Review

neoadjuvant endocrine therapy, and prospective data do not exist to show that MRI improves clinical outcome. On balance, however, we judge MRI to be the most promising research imaging method to investigate in the neoadjuvant setting at present.

We recommend to restrict MRI use to clinical trials for research purposes and to perform it at several timepoints: before the start of chemotherapy; at an early timepoint (eg, after the fi rst cycle of treatment); in the event progressive disease is clinically suspected; after the fi nal course of chemotherapy to assess residual disease; and if treatment with two non-crossresistant cytotoxic regimens is foreseen, before switching to the second regimen to measure response after treatment with the fi rst regimen.

Fluorodeoxyglucose PETResults available on use of fl uorodeoxyglucose (FDG) PET in the neoadjuvant setting are contradictory. Some trials have shown that metabolic information obtained from FDG-PET provides a reliable marker of tumour viability and treatment response, being associated with response to neoadjuvant chemotherapy at an early stage, after one or two cycles of neoadjuvant chemotherapy, and accurately visualising lymph-node metastases.56,57 How ever, other reports suggest superiority of MRI for monitoring the eff ect of neoadjuvant chemotherapy in advanced breast cancer and better diagnostic performance of ultrasonography and contrast-enhanced CT for detection of axillary lymph-node metastasis.58,59

Current guidelines do not support use of FDG-PET or FDG-PET with CT for staging of breast cancer because of the high false-negative rate for detection of lesions that are small (<1 cm) or low grade, the relatively low sensitivity for detection of axillary nodal metastases, the low previous probability for detection of metastatic disease at presentation, and the high rate of false-positive scans.29 We recommend that FDG-PET be used in neoadjuvant clinical trials for research purposes only. We believe that metabolic information from FDG-PET cannot be used reliably to assess residual disease, largely because of the high rate of false-positive fi ndings.

UltrasonographyUltrasonography tends to overestimate residual tumour volume and, compared with mammography and MRI, it has the highest rate of false-positive fi ndings and low specifi city.60 Some reports show poor agreement between tumour size measurements by ultra sono graphy and histopathological analysis, whereas the methodology seems to be especially accurate in nodal staging.60,61

We recommend using ultrasonography in neo adjuvant clinical trials either if MRI is not available or in selected circumstances, such as the staging of nodal disease. Ultrasonography could also be used for place ment of clips to mark the original biopsy site.

MammographySimilar to ultrasonography, mammography tends to overestimate residual tumour volume after neoadjuvant chemotherapy.60 Some reports show poor agreement between tumour size measurements by mammography and histopathological analysis.60 Moreover, specifi city of mammography is low and prediction of pathological outcome is poor, especially when calcifi cations are present.62 We recommend using mammography in neo-adjuvant clinical trials before the start of treatment only if other imaging modalities are not available and after treatment only for clip or tumour localisation for defi nitive surgical resection.

Proposal for the standard characterisation of the population to treatEvidence suggests that breast cancer can be classifi ed into at least four intrinsic molecular subtypes with diff erent prognoses and responses to treatment: the basal-like subtype (predominantly ER-negative, proges-terone receptor (PgR)-negative, and HER2-negative); the HER2-enriched subtype (with overexpression of HER2 and several other genes located on the HER2 amplicon); and two ER-positive subtypes, luminal A and luminal B, characterised by low and high proliferation levels, respectively.8–12 Currently, the prognostic and predictive role of these intrinsic subtypes beyond available clinical and pathological variables is still a matter for research. However, future neoadjuvant trials should take into consideration the biological and clinical diff erences of these diseases by either selection of patients according to subtypes or, at least, designation with enough statistical power to allow subtype stratifi cation or adjustment.

To limit discrepancies across clinical trials and to facilitate data analysis and interpretation, defi nition of breast cancer subtypes needs to be standardised. Gene expression profi les obtained with microarray could be used for this purpose,63–65 but several issues still limit their application in routine clinical practice.66

Immunohistochemical assays undertaken in a high-quality laboratory could provide information similar to that of a gene-expression based classifi er.67 An immuno-histochemical assay to assess expression of ER, HER2, and Ki67 was able to distinguish luminal A type from luminal B in an ER-positive population of breast cancers, with prognostic value.68

Although substantial intralaboratory and inter -laboratory variation of immunohistochemical assess-ment is acknow ledged, guidelines are available to standardise these assays, in particular for ER, HER2, and Ki67.41,69,70 For the time being, in neoadjuvant clinical trials, we recommend that breast cancer subtypes should be defi ned with immuno histochemical assays done according to guidelines approved by the American Society of Clinical Oncology and the College of American Pathologists, possibly in a central laboratory. Subtypes


Review

are defi ned as luminal A (ER-positive and/or PgR-positive, HER2-negative, low proliferative [Ki67 <14%]); luminal B, HER2-negative (ER-positive and/or PgR-positive, HER2-negative, high proliferative [Ki67 ≥14%]); luminal B, HER2-positive (ER-positive and/or PgR-positive, HER2-positive, high proliferative [any Ki67]); HER2-positive (HER2-positive, ER-negative, and PgR-negative); and triple negative (ER-negative, PgR-negative, and HER2-negative). The defi nition of high and low proliferative luminal tumours, and the optimum cutoff to be used for Ki67, might change in the future, and therefore we recommend to record absolute values.

Future perspectives and conclusionsWe have attempted to standardise some of the defi nitions and endpoints currently used in neoadjuvant clinical trials undertaken worldwide. However, our recom-mendations might evolve on the basis of emerging evidence from ongoing and future research.

Because the neoadjuvant setting can be used for both therapeutic and research purposes, one can envisage a future in which neoadjuvant treatment could be recommended to all patients eligible for adjuvant chemo therapy on the basis of their clinical and molecu-lar characteristics. In this vision, use of molecular methods could be implemented for several reasons. Gene expression signatures under investigation in the adjuvant setting could be used as prognostic or predictive methods to select patients most likely to respond to neoadjuvant treatment or to identify patients who might benefi t from additional adjuvant treatment,71 to defi ne patients’ intrinsic molecular subtype,65 or as more objective ways to ascertain variables such as tumour grade. Even if some reports suggest that these methods

could help in the neoadjuvant setting,72–75 we believe that, at the moment, not enough evidence exists to support their integration into clinical practice.

Further insights into understanding the mechanism of action of new drugs and identifi cation of predictive and prognostic biomarkers could instead derive from implementation of biological window or other presurgical trials. Waiting for further confi rmation, we hope that adoption of these guidelines in the design of neoadjuvant breast cancer trials will improve conduct and reporting of studies and, in turn, increase quality of fi ndings and promote research in this area.

ContributorsAll authors contributed their ideas for the report. DF and MP wrote the

report. All authors were given the opportunity to edit the report and all

approved the fi nal version.

Confl icts of interestMP and CS have patent interests in the Gene Expression Grade Index

(GGI). ME has patent and licensing interests in PAM50. The other

authors declare that they have no confl icts of interest.

AcknowledgmentsWe thank Carolyn Straehle for editorial assistance. The BIG and NABCG

collaboration is supported by a Breast Cancer Research Foundation grant.

References1 De Lena M, Varini M, Zucali R, et al. Multimodal treatment for

locally advanced breast cancer: result of chemotherapy-radiotherapy versus chemotherapy-surgery. Cancer Clinical Trials 1981; 4: 229–36.

2 Bonadonna G, Valagussa P, Brambilla C, et al. Primary chemotherapy in operable breast cancer: eight year experience of the Milan Cancer Institute. J Clin Oncol 1998; 16: 93–100.

3 Wolmark N, Wang J, Mamounas E, Bryant J, Fisher B. Preoperative chemotherapy in patients with operable breast cancer: nine-year results from National Surgical Adjuvant Breast and Bowel Project B-18. J Natl Cancer Inst Monogr 2001; 30: 96–102.

4 van der Hage JA, van de Velde CJ, Julien JP, Tubiana-Hulin M, Vandervelden C, Duchateau L. Preoperative chemotherapy in primary operable breast cancer: results from the European Organization for Research and Treatment of Cancer trial 10902. J Clin Oncol 2001; 19: 4224–37.

5 Smith IC, Heys SD, Hutcheon AW, et al. Neoadjuvant chemotherapy in breast cancer: signifi cantly enhanced response with docetaxel. J Clin Oncol 2002; 20: 1456–66.

6 Bear HD, Anderson S, Smith RE, et al. Sequential preoperative or postoperative docetaxel added to preoperative doxorubicin plus cyclophosphamide for operable breast cancer: National Surgical Adjuvant Breast and Bowel Project Protocol B-27. J Clin Oncol 2006; 24: 2019–27.

7 von Minckwitz G, Raab G, Caputo A, et al. Doxorubicin with cyclophosphamide followed by docetaxel every 21 days compared with doxorubicin and docetaxel every 14 days as preoperative treatment in operable breast cancer: the GEPARDUO study of the German Breast Group. J Clin Oncol 2005; 23: 2676–85.

8 Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature 2000; 406: 747–52.

9 Sorlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA 2001; 98: 10869–74.

10 Sorlie T, Tibshirani R, Parker J, et al. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci USA 2003; 100: 8418–23.

11 Sotiriou C, Neo SY, McShane LM, et al. Breast cancer classifi cation and prognosis based on gene expression profi les from a population-based study. Proc Natl Acad Sci USA 2003; 100: 10393–98.

12 Rouzier R, Perou CM, Symmans WF, et al. Breast cancer molecular subtypes respond diff erently to preoperative chemotherapy. Clin Cancer Res 2005; 11: 5678–85.

Search strategy and selection criteria

References for this Review were either provided by the authors or identifi ed on PubMed with the search terms “breast cancer” and one of “neoadjuvant”, “presurgical”, “pathologic complete response”, “residual disease”, “TNM”, “residual cancer burden”, “Ki67”, “Preoperative Endocrine Prognostic Index”, “survival endpoints”, “magnetic resonance”, “FDG-PET”, “ultrasonography”, “mammography”, “molecular subtypes”, or “gene expression signatures”. Only papers published in English between Jan 1, 1980, and Nov 30, 2011, in peer-reviewed journals were considered for inclusion, and they were selected according to their importance. To reach consensus among this panel of experts, a full-day meeting was held for extensive discussion to identify defi nitions and endpoints for which discrepancies and inconsistencies existed in breast cancer neoadjuvant clinical trials. A manuscript with proposals for standard defi nitions and endpoint assessments was then drafted, circulated, and revised over a period of several months among all the members of the panel until consensus was reached.


Review

13 Untch M, Rezai M, Loibl S, et al. Neoadjuvant treatment with trastuzumab in HER2-positive breast cancer: results from the GeparQuattro study. J Clin Oncol 2010; 28: 2024–31.

14 Gianni L, Eiermann W, Semiglazov V, et al. Neoadjuvant chemotherapy with trastuzumab followed by adjuvant trastuzumab versus neoadjuvant chemotherapy alone, in patients with HER2-positive locally advanced breast cancer (the NOAH trial): a randomised controlled superiority trial with a parallel HER2-negative cohort. Lancet 2010; 375: 377–84.

15 Smith IE, Dowsett M, Ebbs SR, et al. Neoadjuvant treatment of postmenopausal breast cancer with anastrozole, tamoxifen, or both in combination: the Immediate Preoperative Anastrozole, Tamoxifen, or Combined with Tamoxifen (IMPACT) multicenter double-blind randomized trial. J Clin Oncol 2005; 23: 5108–16.

16 Ellis MJ, Coop A, Singh B, et al. Letrozole is more eff ective neo-adjuvant endocrine therapy than tamoxifen for ErbB-1- and/or ErbB-2-positive, estrogen receptor-positive breast cancer: evidence from a phase III randomized trial. J Clin Oncol 2001; 19: 3808–16.

17 Weigelt B, Geyer FC, Reis-Filho JS. Histological types of breast cancer: how special are they? Mol Oncol 2010; 4: 192–208.

18 Mook S, Schmidt MK, Rutgers EJ, et al. Calibration and discriminatory accuracy of prognosis calculation for breast cancer with the online Adjuvant! program: a hospital-based retrospective cohort study. Lancet Oncol 2009; 10: 1070–76.

19 Guix M, Granja Nde M, Meszoely I, et al. Short preoperative treatment with erlotinib inhibits tumor cell proliferation in hormone receptor-positive breast cancers. J Clin Oncol 2008; 26: 897–906.

20 Martin L-A, Davies GLS, Weigel MT, et al. Pre-surgical study of the biological eff ects of the selective cyclo-oxygenase-2 inhibitor celecoxib in patients with primary breast cancer. Breast Cancer Res Treat 2010; 123: 829–36.

21 Chevallier B, Roche H, Olivier JP, Chollet P, Hurteloup P. Infl ammatory breast cancer: pilot study of intensive induction chemotherapy (FEC-HD) results in a high histologic response rate. Am J Clin Oncol 1993; 16: 223–28.

22 Sataloff DM, Mason BA, Prestipino AJ, Seinige UL, Lieber CP, Baloch Z. Pathologic response to induction chemotherapy in locally advanced carcinoma of the breast: a determinant of outcome. J Am Coll Surg 1995; 180: 297–306.

23 Ogston KN, Miller ID, Payne S, et al. A new histological grading system to assess response of breast cancers to primary chemotherapy: prognostic signifi cance and survival. Breast 2003; 12: 320–27.

24 Green MC, Buzdar AU, Smith T, et al. Weekly paclitaxel improves pathologic complete remission in operable breast cancer when compared with paclitaxel once every 3 weeks. J Clin Oncol 2005; 23: 5983–92.

25 Mazouni C, Peintinger F, Wan-Kau S, et al. Residual ductal carcinoma in situ in patients with complete eradication of invasive breast cancer after neoadjuvant chemotherapy does not adversely aff ect patient outcome. J Clin Oncol 2007; 25: 2650–55.

26 Rouzier R, Extra JM, Klijanienko J, et al. Incidence and prognostic signifi cance of complete axillary downstaging after primary chemotherapy in breast cancer patients with T1 to T3 tumors and cytologically proven axillary metastatic lymph nodes. J Clin Oncol 2002; 20: 1304–10.

27 Lester SC, Bose S, Chen Y-Y, et al. Protocol for the examination of specimens from patients with invasive carcinoma of the breast. Arch Pathol Lab Med 2009; 133: 1515–38.

28 Anon. Breast. In: Edge SB, Byrd DR, Compton CC, et al, eds. AJCC Cancer Staging Manual, 7th edn. New York: Springer, 2010: 347–76.

29 National Cancer Institute. Breast cancer treatment (PDQ). Nov 21, 2011. http://www.cancer.gov/cancertopics/pdq/treatment/breast/healthprofessional/page3 (accessed Dec 15, 2011).

30 Carey L, Metzger R, Dees EC, et al. American Joint Committee on Cancer tumor-node-metastasis stage after neoadjuvant chemotherapy and breast cancer outcome. J Natl Cancer Inst 2005; 97: 1137–42.

31 Symmans WF, Peintinger F, Hatzis C, et al. Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy. J Clin Oncol 2007; 25: 4414–22.

32 Pusztai L. Preoperative systemic chemotherapy and pathologic assessment of response. Pathol Oncol Res 2008; 14: 169–71.

33 Nahlah Z, Sivasubramaniam D, Dhaliwal S, Sundarajan V, Komrokji R. Residual cancer burden in locally advanced breast cancer: a superior tool. Curr Oncol 2008; 15: 271–78.

34 MD Anderson Cancer Center. Residual cancer burden calculator. 2011. http://www3.mdanderson.org/app/medcalc/index.cfm?pagename=jsconvert3 (accessed Dec 16, 2011).

35 Dowsett M, Ebbs SR, Dixon JM, et al. Biomarker changes during neoadjuvant anastrozole, tamoxifen, or the combination: infl uence of hormonal status and HER-2 in breast cancer—a study from the IMPACT trialists. J Clin Oncol 2005; 23: 2477–92.

36 Dowsett M, Smith IE, Ebbs SR, et al. Short-term changes in Ki-67 during neoadjuvant treatment of primary breast cancer with anastrozole or tamoxifen alone or combined correlate with recurrence-free survival. Clin Cancer Res 2005; 11: 951–58s.

37 Dowsett M, Smith IE, Ebbs SR, et al. Proliferation and apoptosis as markers of benefi t in neoadjuvant endocrine therapy of breast cancer. Clin Cancer Res 2006; 12: 1024–30s.

38 Ellis MJ, Coop A, Singh B, et al. Letrozole inhibits tumor proliferation more eff ectively than tamoxifen independent of HER1/2 expression status. Cancer Res 2003; 63: 6523–31.

39 Dowsett M, Smith IE, Ebbs SR, et al. Prognostic value of Ki67 expression after short-term presurgical endocrine therapy for primary breast cancer. J Natl Cancer Inst 2007; 99: 167–70.

40 Jones RL, Salter J, A’Hern R, et al. The prognostic signifi cance of Ki67 before and after neoadjuvant chemotherapy in breast cancer. Breast Cancer Res Treat 2009; 116: 53–68.

41 Dowsett M, Nielsen TO, A’Hern R. Assessment of ki67 in breast cancer: recommendations from the international ki67 in breast cancer working group. J Natl Cancer Inst 2011; 103: 1656–64.

42 Ellis MJ, Tao Y, Luo J, et al. Outcome prediction for estrogen receptor-positive breast cancer based on postneoadjuvant endocrine therapy tumor characteristics. J Natl Cancer Inst 2008; 100: 1380–88.

43 Hudis CA, Barlow WE, Costantino JP, et al. Proposal for standardized defi nitions for effi cacy end points in adjuvant breast cancer trials: the STEEP system. J Clin Oncol 2007; 25: 2127–32.

44 Oude Munnink TH, Nagengast WB, Brouwers AH, et al. Molecular imaging of breast cancer. Breast 2009; 18: S66–73.

45 Esserman L, Hylton N, Yassa L, Barclay J, Frankel S, Sickles E. Utility of magnetic resonance imaging in the management of breast cancer: evidence for improved preoperative staging. J Clin Oncol 1999; 17: 110–19.

46 Van Goethem M, Schelfout K, Dijckmans L, et al. MR mammography in the pre-operative staging of breast cancer in patients with dense breast tissue: comparison with mammography and ultrasound. Eur Radiol 2004; 14: 809–16.

47 Londero V, Bazzocchi M, Del Frate C, et al. Locally advanced breast cancer: comparison of mammography, sonography and MR imaging in evaluation of residual disease in women receiving neoadjuvant chemotherapy. Eur Radiol 2004; 14: 1371–79.

48 Yeh E, Slanetz P, Kopans DB, et al. Prospective comparison of mammography, sonography, and MRI in patients undergoing neoadjuvant chemotherapy for palpable breast cancer. AJR Am J Roentgenol 2005; 184: 868–77.

49 Fangberget A, Nilsen LB, Hole K, et al. Neoadjuvant chemotherapy in breast cancer-response evaluation and prediction of response to treatment using dynamic contrast-enhanced and diff usion-weighted MR imaging. Eur Radiol 2011; 21: 1188–99.

50 Li XR, Cheng LQ, Liu M, et al. DW-MRI ADC values can predict treatment response in patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy. Med Oncol 2011; published online Feb 1. DOI:10.1007/s12032-011-9842-y.

51 Loo CE, Straver ME, Rodenhuis S, et al. Magnetic resonance imaging response monitoring of breast cancer during neoadjuvant chemotherapy: relevance of breast cancer subtype. J Clin Oncol 2011; 29: 660–66.

52 Kawashima H, Inokuchi M, Furukawa H, Kitamura S. Triple-negative breast cancer are the imaging fi ndings diff erent between responders and nonresponders to neoadjuvant chemotherapy? Acad Radiol 2011; 18: 963–69.

53 Rosen EL, Blackwell KL, Baker JA, et al. Accuracy of MRI in the detection of residual breast cancer after neoadjuvant chemotherapy. AJR Am J Roentgenol 2003; 181: 1275–82.

54 Partridge SC, Gibbs JE, Lu Y, Esserman LJ, Sudilovsky D, Hylton NM. Accuracy of MR imaging for revealing residual breast cancer in patients who have undergone neoadjuvant chemotherapy. AJR Am J Roentgenol 2002; 179: 1193–99.


Review

55 Kim HJ, Im YH, Han BK, et al. Accuracy of MRI for estimating residual tumor size after neoadjuvant chemotherapy in locally advanced breast cancer: relation to response patterns on MRI. Acta Oncol 2007; 46: 996–1003.

56 Duch J, Fuster D, Muñoz M, et al. 18F-FDG PET/CT for early prediction of response to neoadjuvant chemotherapy in breast cancer. Eur J Nucl Med Mol Imaging 2009; 36: 1551–57.

57 Straver ME, Aukema TS, Olmos RA, et al. Feasibility of FDG PET/CT to monitor the response of axillary lymph node metastases to neoadjuvant chemotherapy in breast cancer patients. Eur J Nucl Med Mol Imaging 2010; 37: 1069–76.

58 Choi JH, Lim HI, Lee SK, et al. The role of PET CT to evaluate the response to neoadjuvant chemotherapy in advanced breast cancer: comparison with ultrasonography and magnetic resonance imaging. J Surg Oncol 2010; 102: 392–97.

59 Monzawa S, Adachi S, Suzuki K, et al. Diagnostic performance of fl uorodeoxyglucose-positron emission tomography/computed tomography of breast cancer in detecting axillary lymph node metastasis: comparison with ultrasonography and contrast-enhanced CT. Ann Nucl Med 2009; 23: 855–61.

60 Yeh E, Slanetz P, Kopans DB, et al. Prospective comparison of mammography, sonography, and MRI in patients undergoing neoadjuvant chemotherapy for palpable breast cancer. AJR Am J Roentgenol 2005; 184: 868–77.

61 Khan A, Sabel MS, Nees A, et al. Comprehensive axillary evaluation in neoadjuvant chemotherapy patients with ultrasonography and sentinel lymph node biopsy. Ann Surg Oncol 2005; 12: 697–704.

62 Tardivon AA, Ollivier L, El Khoury C, Thibault F. Monitoring therapeutic effi cacy in breast carcinomas. Eur Radiol 2006; 16: 2549–58.

63 Sorlie T, Tibshirani R, Parker J, et al. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci USA 2003; 100: 8418–23.

64 Hu Z, Fan C, Oh DS, et al. The molecular portraits of breast tumors are conserved across microarray platforms. BMC Genomics 2006; 7: 96.

65 Parker JS, Mullins M, Cheang MC, et al. Supervised risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol 2009; 27: 1160–67.

66 Weigelt B, Mackay A, A’hern R, et al. Breast cancer molecular profi ling with single sample predictors: a retrospective analysis. Lancet Oncol 2010; 11: 339–49.

67 Cuzick J, Dowsett M, Wale C, et al. Prognostic value of a combined ER, PgR, Ki67, Her2 immunohistochemical (IHC4) score and comparison with the GHI recurrence score: results from TransATAC. Cancer Res 2009; 69: 74 (abstr).

68 Cheang MC, Chia SK, Voduc D, et al. Ki67 index, HER2 status, and prognosis of patients with luminal B breast cancer. J Natl Cancer Inst 2009; 101: 736–50.

69 Hammond ME, Hayes DF, Dowsett M, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. J Clin Oncol 2010; 25: 2784–95.

70 Wolff AC, Hammond ME, Schwartz JN, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 2007; 25: 118–45.

71 Sotiriou C, Pusztai L. Gene-expression signatures in breast cancer. N Engl J Med 2009; 360: 790–800.

72 Gianni L, Zambetti M, Clark K, et al. Gene expression profi les in paraffi n-embedded core biopsy tissue predict response to chemotherapy in women with locally advanced breast cancer. J Clin Oncol 2005; 23: 7265–77.

73 Straver ME, Glas AM, Hannemann J, et al. The 70-gene signature as a response predictor for neoadjuvant chemotherapy in breast cancer. Breast Cancer Res Treat 2010; 119: 551–58.

74 Liedtke C, Hatzis C, Symmans WF, et al. Genomic grade index is associated with response to chemotherapy in patients with breast cancer. J Clin Oncol 2009; 27: 3185–91.

75 Ellis MJ, Suman VJ, Hoog J, et al. Randomized phase II neoadjuvant comparison between letrozole, anastrozole, and exemestane for postmenopausal women with estrogen receptor-rich stage 2 to 3 breast cancer: clinical and biomarker outcomes and predictive value of the baseline PAM50-based intrinsic subtype—ACOSOG Z1031. J Clin Oncol 2011; 29: 2342–49.


Articles

Lancet Oncol 2012; 13: 622–32

Published OnlineApril 17, 2012

DOI:10.1016/S1470-2045(12)70121-3


Division of Surgery and Interventional Science,

University College London, London, UK (H U Ahmed MRCS,

L Dickinson MRCS, R Scott RN, Prof M Emberton FRCS); Department of Urology

(H U Ahmed, L Dickinson, M Sahu MRCS), Department of

Histopathology (A Freeman FRCPath), and Department of Radiology

(A P Kirkham FRCR, C Allen FRCR), University

College London Hospitals NHS Foundation Trust, London, UK;

Department of Urology, Hampshire Hospitals NHS

Foundation Trust, Basingstoke, UK (R G Hindley FRCS); Clinical

Eff ectiveness Unit, Royal College of Surgeons of England,

London, UK (Prof J Van der Meulen PhD); and

Health Services Research Unit, The London School of Hygiene

and Tropical Medicine, London, UK (Prof J Van der Meulen)

Correspondence to:Dr Hashim Uddin Ahmed,

Division of Surgery and Interventional Sciences,

University College London, 67 Riding House Street,

London, W1P 7NN, [email protected]

Focal therapy for localised unifocal and multifocal prostate cancer: a prospective development studyHashim U Ahmed, Richard G Hindley, Louise Dickinson, Alex Freeman, Alex P Kirkham, Mahua Sahu, Rebecca Scott, Clare Allen, Jan Van der Meulen, Mark Emberton

SummaryBackground Radical whole-gland therapy can lead to signifi cant genitourinary and rectal side-eff ects for men with localised prostate cancer. We report on whether selective focal ablation of unifocal and multifocal cancer lesions can reduce this treatment burden.

Methods Men aged 45–80 years were eligible for this prospective development study if they had low-risk to high-risk localised prostate cancer (prostate specifi c antigen [PSA] ≤15 ng/mL, Gleason score ≤4 + 3, stage ≤T2), with no previous androgen deprivation or treatment for prostate cancer, and who could safely undergo multiparametric MRI and have a general anaesthetic. Patients received focal therapy using high-intensity focused ultrasound, delivered to all known cancer lesions, with a margin of normal tissue, identifi ed on multiparametric MRI, template prostate-mapping biopsies, or both. Primary endpoints were adverse events (serious and otherwise) and urinary symptoms and erectile function assessed using patient questionnaires. Analyses were done on a per-protocol basis. This study is registered with ClinicalTrials.gov, number NCT00561314.

Findings 42 men were recruited between June 27, 2007, and June 30, 2010; one man died from an unrelated cause (pneumonia) 3 months after treatment and was excluded from analyses. After treatment, one man was admitted to hospital for acute urinary retention, and another had stricture interventions requiring hospital admission. Nine men (22%, 95% CI 11–38) had self-resolving, mild to moderate, intermittent dysuria (median duration 5·0 days [IQR 2·5–18·5]). Urinary debris occurred in 14 men (34%, 95% CI 20–51), with a median duration of 14·5 days (IQR 6·0–16·5). Urinary tract infection was noted in seven men (17%, 95% CI 7–32). Median overall International Index of Erectile Function-15 (IIEF-15) scores were similar at baseline and at 12 months (p=0·060), as were median IIEF-15 scores for intercourse satisfaction (p=0·454), sexual desire (p=0·644), and overall satisfaction (p=0·257). Signifi cant deteriorations between baseline and 12 months were noted for IIEF-15 erectile (p=0·042) and orgasmic function (p=0·003). Of 35 men with good baseline function, 31 (89%, 95% CI 73–97) had erections suffi cient for penetration 12 months after focal therapy. Median UCLA Expanded Prostate Cancer Index Composite (EPIC) urinary incontinence scores were similar at baseline as and 12 months (p=0·045). There was an improvement in lower urinary tract symptoms, assessed by International Prostate Symptom Score (IPSS), between baseline and 12 months (p=0·026), but the IPSS-quality of life score showed no diff erence between baseline and 12 months (p=0·655). All 38 men with no baseline urinary incontinence were leak-free and pad-free by 9 months. All 40 men pad-free at baseline were pad-free by 3 months and maintained pad-free continence at 12 months. No signifi cant diff erence was reported in median Trial Outcomes Index scores between baseline and 12 months (p=0·113) but signifi cant improvement was shown in median Functional Assessment of Cancer Therapy (FACT)-Prostate (p=0·045) and median FACT-General scores (p=0·041). No histological evidence of cancer was identifi ed in 30 of 39 men biopsied at 6 months (77%, 95% CI 61–89); 36 (92%, 79–98) were free of clinically signifi cant cancer. After retreatment in four men, 39 of 41 (95%, 95% CI 83–99) had no evidence of disease on multiparametric MRI at 12 months.

Interpretation Focal therapy of individual prostate cancer lesions, whether multifocal or unifocal, leads to a low rate of genitourinary side-eff ects and an encouraging rate of early absence of clinically signifi cant prostate cancer.

Funding Medical Research Council (UK), Pelican Cancer Foundation, and St Peters Trust.

IntroductionThe management of localised prostate cancer remains controversial because the systematic over-diagnosis that accompanies the current diagnostic pathway results in over-treatment.1 At present, radical whole-gland surgery or radiotherapy can result in substantial side-eff ects that are a consequence of damage to surrounding structures. These include urinary incontinence (5–20%), erectile dysfunction (30–70%), and bowel toxicity (5–10%).2,3

Technological refi ne ments do not seem to have reduced the burden of harm.4,5

Apart from active surveillance for low-risk disease, few strategies are available to address the burden of treatment-related side-eff ects in other risk categories. One strategy that has shown promise relates to managing prostate cancer in the same manner as most other solid organ malignancies—by focusing the therapy to the cancer lesion, injury to the

Articles


bladder, rectum, and neurovascular bundles could be reduced.6,7

We have previously assessed hemi-ablation of patients with localised unilateral prostate cancer,8 which included treatment of the entire half of the prostate associated with cancer. Regardless of the grade, volume, or location of cancer that aff ected half of the prostate, the entire side was ablated. This strategy is the most straight forward to undertake, standardise, and train others to do, but is limited because only one in fi ve men have true unilateral disease on template biopsies. Furthermore, hemi-ablation might represent overtreatment since a low-volume, low-grade lesion would be treated in the same manner as a high-volume, high-grade cancer. In this study, we postulated that selective focal ablation of unifocal and multifocal cancer lesions with a margin of normal tissue could reduce genitourinary and rectal side-eff ects for men with localised prostate cancer.

MethodsStudy design and patientsWe undertook a two-centre, prospective development study, as defi ned by the IDEAL (Idea, Development, Exploration, Assessment, and Long-term) guidelines for assessing innovation in surgery.9 Men could enter into the study if they had localised prostate cancer on multiparametric MRI and transperineal template-prostate-mapping biopsies.10

We included men with low-risk to high-risk disease (prostate-specifi c antigen [PSA] ≤15 ng/mL, Gleason score ≤4+3, stage ≤T2), aged 45–80 years with a life expectancy of 5 years or more, a prostate volume of 40 mL or less or maximum anterior–posterior length of 40 mm or less who had undergone multiparametric MRI and transperineal template (5 mm spaced) biopsies in the 6 months before recruitment. We excluded men who had androgen suppression within the previous 6 months, previous radiation therapy or chemotherapy for prostate cancer, latex allergies, previous rectal surgery preventing insertion of transrectal probe, intraprostatic calcifi cations making high-intensity focused ultrasound (HIFU) of focal areas of cancer diffi cult, previous transurethral resection of the prostate or laser prostatectomy in 5 years before recruitment, previous HIFU, cryosurgery, or thermal or microwave therapy to the prostate at any point before recruitment. Men who were not fi t for general anaesthesia or regional anaes thesia as assessed by a consultant anaesthetist, or were unable to have MRI scanning (eg, severe claustrophobia, permanent cardiac pacemaker, metallic implant likely to contribute signifi cant artifact to images) were also excluded. All men gave written informed consent.

Our trial was approved by the University College London Hospitals Local Research Ethics Committee A, UK, which is under the auspices of the National Research Ethics Service. The study was independently audited by hospital research and development offi cials. Additionally,

the protocol was anonymously peer-reviewed by the National Cancer Research Network (NCRN), UK, and the Medical Research Council, UK.

ProceduresTo locate areas of cancer, multiparametric MRI was done at 1·5 T magnetic fi eld strength with pelvic phased-array coils. Sequences included T2-weighting, dynamic gadolinium contrast-enhance ment and diff usion-weighting. Template-prostate-mapping biopsies were done under general or spinal anaesthesia with the prostate sampled at 5 mm intervals. Two biopsies were taken at the same grid coordinate if the prostate was longer than the standard length of a biopsy core. All biopsies were reported by a single uropathologist.

Men underwent focal ablation with a transrectal HIFU device (Sonablate 500; Focus Surgery, Indianapolis, IN, USA). Ultrasonic waves were generated with a cylindrical piezoelectric ceramic transducer and then focused with a spherical plate onto a target area determined by the focal length of the transducer. The sound waves were transmitted to the tissues by a coupling mechanism from a transducer placed either extracorporeally or transrectally. Transmission of sound waves transrectally was achieved by placing the probe in a condom fi lled with chilled circulating degassed water. The dimensions of the target area were determined by the focal length of the transducer, the applied frequency, the intensity of the applied power (W/cm²), and the duration of the pulse. The lesion produced was pseudoellipsoid in shape and referred to as the focal zone. Its long axis lies at right angles to the transducer and is greatest in length towards the transducer.

Tissue destruction is produced by thermal, mechanical, and cavitation eff ects to produce a clearly demarcated region of coagulative necrosis surrounded by normal tissue on microscopic examination. Thermal energy comes from absorption of mechanical energy. Adequate cell destruction can be produced by short exposure (1 s) to temperatures of 60°C or more, which has therefore been adopted as the minimum target temperature. In practice, this temperature is easily attained with temperatures of 80°C or more recorded during HIFU therapy. Cooling due to tissue perfusion in the focal zone is not a problem because the rate of heating is greater than that of cooling when the exposure time is within a window of 3 s. The mechanical eff ects of HIFU are more complex and involve shear forces, torque, and streaming. These forces result in destruction by both physical and thermal means. Cavitation results from gas (bubble) formation within cells due to heat and mechanical energy deposition causing bubbles to oscillate.

All patients had sterile urine on culture before treatment. If culture was positive for infection, men were treated with antibiotics and their treatment rescheduled; prophy lactic intravenous gentamicin was given to all men at the time of general anaesthetic. A suprapubic

For the study protocol see https://www.ucl.ac.uk/focal/therapy/hifu/focal/Focal-HIFU-Protocol

Articles


catheter was inserted before HIFU under the same anaesthetic.

We standardised the process of focal therapy by setting three broad guidelines. First, a maximum of 60% of the prostate could be ablated. Second, the edge of the ablation zone had to be at least 10 mm from a neurovascular bundle. The ablation zone had to be at least 5 mm from both neurovascular bundles if disease was bilateral. Third, untreated areas could not have any histological evidence of prostate cancer; high-grade prostate intraepithelial neoplasia and atypical small acinar proliferation were permitted. The operator made judgments as to the location and boundaries of the cancer lesions for treatment planning on the basis of the information from both multiparametric MRI (when a lesion was visible) and template-prostate-mapping biopsies. The areas positive for cancer were treated with at least a 3–5 mm margin around the lesion (2–4 HIFU pulses). In areas where a discrepancy between multi parametric MRI and template-prostate-mapping biopsies was identifi ed, histopathological fi ndings took prece dence. Some cancer lesions were quite close and were therefore included in the same area of treatment. As a result, more than two lesions could be treated as long as there were only two areas of treatment. Designation of individual lesions was usually straightforward but when positive biopsies were close together, lesions were labelled separately if there was at least one intervening normal biopsy.

After ablation, the suprapubic catheter was placed on free drainage into a urinary leg-bag for 1–2 days and urethral voiding was encouraged thereafter by closure of a valve attached to the catheter. Because many men travelled a long distance, the timing of catheter removal was delayed to coincide with the fi rst trial visit after the operation at 10–14 days (for an early MRI) even if urethral voiding was restored earlier. All men were given ciprofl oxacin and oral analgesia (co-dydramol) for 7 days. A contrast-enhanced MRI was done 10–14 days after focal HIFU to confi rm the area of ablation, as shown by a confl uent perfusion defi cit.

Follow-up consisted of clinic visits at 1 month, 3 months, 6 months, 9 months, and 12 months for PSA measurement and adverse event reporting. Men fi lled in validated questionnaires at each clinic visit. Questionnaires included the International Prostate Symptom Score (IPSS), International Index of Erectile Function-15 (IIEF-15), UCLA-Expanded Prostate Cancer Index Composite (EPIC) urinary incontinence scale, and the Functional Assessment of Cancer Therapy-Prostate (FACT-P) score, which includes FACT-General (FACT-G) scores and the Trials Outcome Index.11–15 Phosphodiesterase-5 inhibitors were permitted at any timepoint during follow-up. At 6 months, another multiparametric MRI followed by targeted biopsies of the treated areas were scheduled with a minimum requirement for sampling every 1 mL of residual tissue with one core. Our justifi cation for one biopsy for every

mL of residual tissue refl ects the biopsy density of the original template biopsies before focal HIFU (which was about 1 mL for every biopsy). Retreatment with a further focal HIFU was permitted if biopsies were positive. A further multi parametic MRI scan was done at 12 months. As the purpose of the 6-month biopsies was to determine whether the ablation was successful, our ethics com mittee did not permit sampling of untreated tissue due to the requirement for another general anaesthetic. However, biopsies of the untreated tissue were permitted if a new, potentially malignant lesion was seen on multiparametric MRI.

The primary outcomes were feasibility, patient acceptability, and side-eff ect profi le of focal HIFU. Feasibility and acceptability were reported with rates and description of adverse events, serious and otherwise. Urinary symptoms and erectile function were assessed with patient questionnaires. The IIEF-15 was used to report the proportion of men capable of having erectile function suffi cient for penetration at 12 months as well as total IIEF-15 score and domain scores on erectile function, intercourse satisfaction, orgasmic function, sexual desire, and overall satisfaction. Urinary function was assessed with IPSS, the IPSS quality-of-life questionnaire, and the UCLA EPIC continence questionnaire. The continence questionnaire included total scores as well as the proportion of patients who were pad-free, or leak-free and pad-free at 12 months. Quality of life was measured with FACT-P with summary measures of the Trial Outcome Index score, FACT-P score, and FACT-G score.

Secondary outcomes were histological and imaging measures of cancer control. A composite outcome in the form of trifecta status was assessed at 12 months. For this status to be met, we defi ned trifecta as leak-free and pad-free continence, erections suffi cient for penetration, and no evidence of clinically signifi cant disease at 12 months multiparametric MRI16,17 in those men with normal baseline genitourinary function.

Statistical analysisSince the primary objective of the study was to determine the side-eff ect profi le of focal ablation, the sample size was powered on a common event, namely erectile dysfunction. We estimated that focal ablation therapy would lead to an absolute rate of 15% erectile dysfunction (insuffi cient for penetrative sex) at 12 months. The sample size calculation was based on a comparison with a known rate of 40% erectile dysfunction,18 which usually occurs when HIFU is applied to the whole prostate.19 Therefore, with an α level of 0·05 and power of 90% (1–β), the sample size required was at least 33 men with good baseline function. We adjusted the sample size to allow for the estimated rate of 25% of men having poor baseline erectile function in the general population, and therefore aimed to recruit 42 men in total.

Validated questionnaires were analysed with standard methods. Missing values for patient-reported outcome

Articles


variables (between 2% and 10% missing values for individual questions) were imputed with a fully con-ditional specifi cation method and logistic regression model (categorical data). The imputation was based on the observation of values missing completely at random. We classed the variables for which missing values were to be imputed (questionnaires) as categorical. Therefore, logistic regression was used with available values of the same variables at diff erent timepoints either side of the imputed value as predictors. Imputed values were rounded up to next integer values. Maximum and minimum values were set according to extremes of questionnaire item scales.

Categorical outcomes were reported as point estimates with binomial 95% CIs to demonstrate level of precision. Wilcoxon signed rank test (two-tailed) was used to assess diff erences between continuous variables that were not normally distributed (PSA and questionnaire scores) measured at baseline and at the 12-month follow-up visit. Changes over time were reported with box-and-whisker plots. Subgroup analyses were hypothesis-generating and with small numbers in each subgroup, it was deemed inappropriate to run statistical tests of signifi -cance in such comparisons. p values of 0·05 or less were deemed signifi cant. All statistical tests were done with SPSS (version 17.0). This study is registered with ClinicalTrials.gov, number NCT00561314.

Role of the funding sourceThe sponsors of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. HUA, LD, RS, JvdM, and ME had access to the raw data. The corresponding author had full access to all the data and had fi nal responsibility for the decision to submit for publication.

Results42 men were recruited between June 27, 2007, and June 30, 2010. One man had an unrelated death 3 months after focal therapy. He had baseline retro-peritoneal fi brosis, hypertension, and a single kidney. He had an uneventful recovery after HIFU and had no respiratory symptoms immediately before treatment, or at 2 weeks and 6 weeks after treatment during the formal trial visits. The cause was independently verifi ed by local physicians as respiratory failure following pneumonia. As a result, 41 men were included in analyses; 30 (73%) had intermediate and high-risk disease (table 1).20 Three men had baseline mild stress urinary incontinence but required no pads; 35 had good baseline sexual function with erections suffi cient for penetration. Of 41 men, 20 (49%) had unilateral, one-area ablation, 15 (37%) had bilateral, two-area ablation, and six (15%) had midline, one-area treatment (fi gure 1). Baseline characteristics of these three groups are shown in the appendix.

30 (73%, 95% CI 57–86) men had a hospital stay of less than 24 h. All 41 were able to void urine through the

Patients (N=41)

Age (years) 63 (58·0–66·0)

Serum PSA (ng/mL) 6·6 (5·4–7·7)

Reason for PSA test and biopsy

PSA screening (patient request) 31 (76%)

Lower urinary tract symptoms* 10 (24%)

Prostate volume (mL) 35 (29·0–45·5)

PSA density (ng/mL per mL prostate) 0·18 (0·14–0·22)

Initial biopsy

TRUS biopsy 35 (85%)

TPM biopsy 6 (15%)

Gleason (TRUS-guided biopsy)

3+3 24 (59%)

3+4 7 (17%)

4+3 5 (12%)

No TRUS biopsy 5 (12%)

Gleason (TPM biopsies)

3+3 13 (32%)

3+4 24 (59%)

4+3 4 (10%)

Clinical stage

T1c 37 (90%)

T2a 4 (10%)

TRUS guided biopsies

Total cores 10·0 (8·0–12·0)

Total positive cores 2·0 (1·0–3·0)

Percent positive cores 11·0 (6·3–33·8)

TPM biopsies

Total cores 46 (35·5–65·5)

Total positive cores† 5 (3·0–9·0)

Positive cores (%) 9·4% (4·6–18·5)

Core density (biopsies/mL) 1·4 (0·9–1·9)

Number of lesions on TPM

One 21

Two 17

Three 3

Disease distribution

Unifocal

Unilateral 15 (37%)

Bilateral (midline lesion) 6 (15%)

Multifocal

Unilateral 5 (12%)

Bilateral 15 (37%)

NCCN risk category20

Low 11 (27%)

Intermediate 26 (63%)

High 4 (10%)

Data are median (IQR), number, or number (%). PSA=prostate-specifi c antigen. TRUS=transrectal ultrasound. TPM=template prostate mapping. NCCN=National Comprehensive Cancer Network. *These men were opportunistically screened with a PSA test when they presented with symptoms of lower urinary tract infection, rather than part of a formal request (by their physician or by the men themselves) for a PSA test. †A high number of positive cores were retrieved, despite only a maximum of three lesions, because large dominant lesions were sampled several times.

Table 1: Baseline characteristics


Articles


urethra on the fi rst day after the operation with the suprapubic catheter clamped. About a fi fth had self-resolving, mild to moderate, intermittent dysuria (table 2). One man had acute retention of urine 3 days after removal of the suprapubic catheter, which required a urethral catheter for 5 days. In terms of rectal adverse events, one man had diarrhoea and mucus discharge 2 weeks after focal HIFU while the suprapubic catheter was still in situ. He had multiparametric MRI and urethrogram within 4 weeks, which showed extravasation of urine outside the prostate; although the serosal layer of the rectum was aff ected by treatment no defi nitive rectourethral fi stula was seen. However, as a precau-tionary measure, he was managed conservatively with a suprapubic catheter and quinolone antibiotics, serial

2 monthly multiparametric MRI, and a further repeat urethrogram until the extravasation had resolved by 6 months. At that point, he developed a stricture requiring endoscopic dilatation; a further two limited endoscopic prostate-tissue resections to a prostate that had a total volume of 80 mL were done when voiding did not return to normal.

Overall IIEF-15 scores initially decreased, indicating diminished erectile function but showed a gradual return to baseline by 12 months (p=0·060; fi gure 2A). IIEF-15 domain scores in intercourse satisfaction (p=0·454), sexual desire (p=0·644), and overall satisfaction (p=0·257) all showed decreased scores at 1 month and 3 months, but there was no signifi cant diff erence between baseline and 12 months (fi gure 2 C, F). IIEF-15 erectile and orgasmic domains showed signifi cant deteriorations from baseline to 12 months (p=0·042 and p=0·003, respectively; fi gure 2 B, D). Of 35 men with erectile function satisfactory for penetration before treatment, 31 (89%, 95% CI 73–97) described erections suffi cient for penetration at 12 months. No formal programme of penile rehabilitation was available, but all men were off ered phosphodiesterase-5 inhibitors (eg, sildenafi l, tadalafi l) if needed; 14 of the 31 men required phospho-diesterase-5 inhibitors.

In a post-hoc analysis to explore whether type of ablation made any diff erence to erectile dys function, we assessed the following factors. Of those men who had erections suffi cient for penetration at baseline, 28 of 31 (90%, 95% CI 74–98) of those who had unilateral nerve-sparing ablation and four of four (100%, 40–100) who had bilateral nerve-sparing ablation had erections suffi cient for penetration at 12 months. Although only six patients received midline nerve-sparing ablation, we assessed the hypothesis that bilateral ablation might result in higher erectile dysfunction rates than unilateral ablation. Of those men who had erections suffi cient for penetration at baseline, 17 of 18 (94%, 95% CI 73–100) who had unilateral

Figure 1: Schematic diagrams of the types of focal therapyUnilateral one-area ablation (A). Bilateral two-area ablation with preservation of one neurovascular bundle (B). Midline one-area ablation allows preservation of both neurovascular bundles (C). Large red areas represent dominant cancers (index lesions) whereas small green areas represent small low-grade secondary lesions. Grey transparent boxes represent ablation zones on the high-intensity focused ultrasound device.

Neurovascular bundles

BA C

Value

Total anaesthetic time (min) 135·0 (115·0–150·0)

Procedure time (Suprapubic catheter+focal HIFU; min) 105·0 (87·0–125·0)

Total hospitalisation time (admission to discharge; h) 12·0 (10·0–27·0)

Discharge time (end procedure to discharge; h) 6·0 (5·0–18·0)

Time with suprapubic catheter (days)*† 8·5 (8·0–15·0)

Dysuria (negative urine culture) 9/41 (22%, 11–38)

Duration of dysuria (days) 5 (2·5–18·5)

Intermittent haematuria (start of stream only) 16/41 (39%, 24–56)

Duration of intermittent haematuria (days) 15 (10·3–15·0)

Urinary debris 14/41 (34%, 20–51)

Duration of urinary debris (days) 14·5 (6·0–16·5)

Urinary tract infection (positive urine culture) 7/41 (17%, 7–32)

Acute retention of urine 1/41 (2%, 0–13)

Data are median (IQR) or number of patients aff ected/N (%, 95% CI). HIFU=high-intensity focused ultrasound. *Suprapubic catheter was usually removed at the same time as the postoperative early contrast MRI for convenience to reduce visits for men who travelled far to the study centre. †The man who had diarrhoea and mucus discharge and had a suprapubic catheter for 6 months was excluded from these descriptive values.

Table 2: Perioperative outcomes in men undergoing focal high-intensity focused ultrasound for unifocal and multifocal localised prostate cancer

Articles


ablation and 15 of 17 (88%, 95% CI 64–99) who had bilateral or midline ablation had erections suffi cient for penetration at 12 months.

UCLA EPIC urinary incontinence scores showed an initial deterioration in continence function but had returned to a similar value as at baseline by 12 months

Figure 2: Sexual function after focal high-intensity focused ultrasound for unifocal and multifocal localised prostate cancer, measured with the International Index of Erectile Function-15 (IIEF-15) questionnaireTwo-tailed p values were reported for Wilcoxon signed ranks test comparing baseline and 12-month median scores. Box and whiskers plots indicate median with IQR (boxes), and range (whiskers). Dots are outliers. Median baseline versus 12-month scores: total IIEF-15 (57·5 [IQR 30·0–67·0] vs 47·0 [29·5–63·3], p=0·060; A); IIEF-15 erectile-function domain (24·0 [13·0–29·0] vs 21·0 [10·3–27·3], p=0·042; B); IIEF-15 intercourse-satisfaction domain (9·0 [0·0–12·0] vs 8·0 [0·0–11·0], p=0·454; C); IIEF-15 orgasmic-function domain (10·0 [6·5–10·0] vs 7·0 [5·0–8·5], p=0·003; D); IIEF-15 sexual-desire domain (7·0 [5·0–8·5] vs 7·0 [5·0–8·0], p=0·644; E); IIEF-15 overall-satisfaction domain (7·5 [4·0–9·0] vs 8·0 [6·0–9·0], p=0·257; F).

80

60

40

20

0

–20

N=41 N=41 N=41 N=41 N=41 N=41 40

30

20

10

0

–10

–20

N=41 N=41 N=41 N=41 N=41 N=41

15

10

5

0

–5

N=41 N=41 N=41 N=41 N=41 N=41 15

10

5

0

–5

N=41 N=41 N=41 N=41 N=41 N=41

10

8

6

4

2

0

N=41 N=41 N=41 N=41 N=41 N=4112

9

6

3

0

N=41 N=41 N=41 N=41 N=41 N=41

Ove

rall

IIEF

scor

e

IIEF

erec

tile

func

tion

scor

e

IIEF

inte

rcou

rse

satis

fact

ion

scor

e

IIEF

orga

smic

func

tion

scor

e

IIEF

sexu

al d

esire

scor

e

IIEF

over

all s

atisf

actio

n sc

ore

Time (months)

1231 90 6

Time (months)

1231 90 6

(n=2)(n=2)

(n=3)

(n=1) (n=1)

(n=1)

(n=1)

BA

DC

FE

Articles


(p=0·045; fi gure 3). IPSS values increased initially although by 12 months showed signifi cantly lower values, suggesting improvement in lower urinary-tract symptoms from baseline (p=0·026; fi gure 4). No signifi cant diff erence was seen in IPSS-quality-of-life score between baseline and 12 months (p=0·655; fi gure 4). Of 38 men with no urinary leak at baseline all (100%, 95% CI 91–100) were leak-free and pad-free by 9 months; 40 men pad-free at baseline were pad-free again by 3 months and maintained pad-free continence at 12 month (100%, 91–100; one man did not report on this parameter at 12 months and was excluded).

Signifi cant deterioration of health-related quality of life was shown between baseline and 12 months on the total FACT-P and total FACT-G scores (p=0·045 and p=0·041, respectively; fi gure 5). No signifi cant diff erence was seen between baseline and 12 months in the Trial Outcomes Index between baseline and 12 months (p=0·113; fi gure 5).

Compared with baseline, a signifi cant decrease in PSA levels was reported at 12 months (p<0·0001; fi gure 6). The time to nadir was not calculated because the PSA changes showed a pattern of ongoing small decreases in PSA up to trial end at 12 months.

One man refused to undergo biopsy at 6 months because of his concern over the eff ect of further biopsies on sexual function. The man with diarrhoea and rectal mucus discharge was excluded from biopsy because of risk of infection and promotion of fi stula formation. Both had multiparametric MRI at 6 months, which showed no evidence of disease. The second man had transurethral resection of obstructing prostate tissue at 7 months, which was histologically confi rmed as benign. Of the

39 men biopsied, nine (23%, 95% CI 11–39) had evidence of cancer while three (8%, 2–21) had evidence of clinically signifi cant cancer (Epstein criteria21,22—Gleason >3 + 3, >2 cores positive, >2 mm cancer involve ment; table 3). All biopsies done after focal HIFU had one or more of the following features: necrosis, fi brosis, or giant-cell reaction. Presence of these features showed that treated areas were accurately targeted. Multiparametric MRI at 6 months showed signs of residual cancer in the treated areas in nine men; seven of whom had cancer confi rmed on biopsy. Two men with negative multiparametric MRI but positive biopsies both had clinically insignifi cant disease. No areas of residual cancer were identifi ed in untreated areas on multiparametric MRI so untreated areas were not biopsied.

Of those men with positive biopsies at 6 months, fi ve chose to undergo active surveillance and four had

Figure 3: Continence function after focal high-intensity focused ultrasound for unifocal and multifocal localised prostate cancer, measured with UCLA Expanded Prostate Cancer Index Composite (EPIC) incontinence questionnaireTwo-tailed p values were reported for Wilcoxon signed ranks test comparing baseline and 12-month median scores. Median baseline vs 12 month scores: 100 (IQR 86·0–100·0) vs 100 (92·5–100·0, p=0·045). Box and whiskers plots show median with IQR (boxes) and range (whiskers). Dots are outliers.

Time (months)1231 90 6

(n=1)

(n=1)

(n=2)

(n=3)

(n=1)

(n=2) (n=2)

(n=2)

(n=2)

100

80

60

40

EPIC

urin

ary d

omai

n

Figure 4: Urinary function after focal high-intensity focused ultrasound for unifocal and multifocal localised prostate cancer, measured with International Prostate Symptom Score (IPSS) questionnaireTwo-tailed p values were reported for Wilcoxon signed ranks test comparing baseline and 12-month median scores. Box and whiskers plots show median with IQR (boxes) and range (whiskers). Dots are outliers. Median baseline versus 12-month International Prostate Symptom Score (IPSS): 8·0 (IQR 5·5–13·0) vs 7·0 (3·0–12·0, p=0·026; A). Median baseline versus 12-month IPSS-quality of life: 1·0 (0·0–2·0) vs 1·0 (1·0–1·0, p=0·655; B).

5

4

3

2

1

0

–1

N=41 N=41 N=41 N=41 N=41 N=41

30

25

20

15

10

5

0

N=41 N=41 N=41 N=41 N=41 N=41


(n=2)

(n=1)

(n=1)(n=3)(n=1)

(n=1)

(n=3)

(n=1)

(n=4)

B

A

IPSS

IPSS

qua

lity o

f life

Articles


retreatment. PSA levels in the retreatment group changed from a median of 3·9 ng/mL (IQR 3·7–4·5) at 6 months (before retreatment) to 3·9 ng/mL (3·7–4·1) at

9 months and 2·9 ng/mL (1·8–3·6) at 12 months. None of the four men undergoing repeat focal therapy consented to further biopsies but all had multiparametric MRI at trial exit showing no evidence of clinically signifi cant disease at 12 months. No man required adjuvant radiotherapy, prostate cancer surgery, or androgen deprivation therapy during the trial duration.

Figure 6: PSA levels after focal high-intensity focused ultrasound for unifocal and multifocal localised prostate cancerWilcoxon signed ranks test comparing median prostate-specifi c antigen levels at baseline (6·6 [IQR 5·4–7·7] ng/mL) and at 12 months (1·9 [0·8–3·3] ng/mL; two-tailed p<0·0001). Box and whiskers plots show median with IQR (boxes) and range (whiskers). Dots are outliers.


20

15

10

5

0PS

A (n

g/m

L)

(n=1)

(n=1)

(n=1)

(n=1)

(n=1)

N=36 N=36 N=36 N=36 N=36 N=36

Figure 5: Quality-of-life outcomes after focal high-intensity focused ultrasound for unifocal and multifocal localised prostate cancer, measured with the Functional Assessment of Cancer Therapy (FACT) questionnaireTwo-tailed p values were reported for Wilcoxon signed ranks test comparing baseline and 12-month median scores. Box and whiskers plots show median with IQR (boxes) and range (whiskers). Dots are outliers. Median baseline versus 12-month Trials Outcome Index score: 94·0 (IQR 89·0–97·3) vs 97·5 (91·0–101·0, p=0·113; A). Median baseline versus 12-month Functional Assessment of Cancer Therapy (FACT)-Prostate score: 138·5 (133·0–147·0) vs 145·3 (137·0–152·0, p=0·045; B). FACT-general score: 96·0 (91·0–102·3) vs 102·0 (96·0–105·0, p=0·041; C).

120

100

80

60

40

20

150

125

100

75

50

120

100

80

60

40


FACT

-P Tr

ial O

utco

me

Inde

x sc

ore

FACT

-P sc

ore

FACT

-G sc

ore

N=41 N=41 N=41 N=41 N=41 N=41

N=41 N=41 N=41 N=41 N=41 N=41

N=41 N=41 N=41 N=41 N=41 N=41

B

A

C

(n=1)

(n=1)

(n=1)

(n=1)

(n=1)(n=1)

(n=1)

(n=1)

(n=1)

(n=1)(n=1)

(n=1)

(n=1)(n=1)

(n=1)(n=1)

(n=1)

(n=1)

(n=1)

(n=1)

(n=1)

(n=1) (n=1)(n=1)

(n=1)

(n=1)

(n=1)

(n=1)

Value

Number of cores taken 6·0 (5·0–7·0)

Absence of any cancer 30/39 (77%, 61–89)

Positive biopsy outcomes*

Positive biopsies† 9/39 (23%, 11–39)

Maximum cancer core length (in positive cores), mm

1·0 (1·0–3·5)

Gleason (N)

3+3 6

3+4 3

Absence of clinically signifi cant disease‡ 36/39 (92%, 79–98)

Other histological fi ndings

Prostatic acini 21/39 (54%, 37–70)

Atrophy 25/39 (64%, 47–79)

Fibrosis 35/39 (90%, 76–97)

Giant-cell reaction 4/39 (10%, 3–24)

Necrosis 15/39 (38%, 23–55)

Data are median (IQR) or number of patients/N (%, 95% CI). *Two men were not biopsied because of suprapubic catheter in situ in one patient and refusal by another. †Five men opted for surveillance. Of these, four had 1 mm of Gleason 3+3 and one had 2 mm of Gleason 3+4. Four men opted for retreatment. Of these, two had clinically signifi cant cancer (5 mm and 6 mm of Gleason 3+4) and two had no more than 1 mm of Gleason 3+3. ‡As defi ned by Epstein criteria: Gleason >3+3, >2 cores positive, ≥3 mm cancer involvement.

Table 3: Histological outcomes at 6 months in men undergoing focal high-intensity focused ultrasound for unifocal and multifocal localised prostate cancer

Articles


Of the 31 men with good baseline function, 26 (84%, 95% CI 66–95) achieved the trifecta status of having leak-free and pad-free continence, erections suffi cient for intercourse, and no evidence of clinically signifi cant disease on multipara metric MRI at 12 months (fi gure 7).

DiscussionTo our knowledge, our study is the fi rst to assess targeting of individual known cancer areas, with a margin of normal tissue, in men with multifocal as well as unifocal prostate cancer across all cancer-risk categories. Focal therapy of discrete areas of cancer, whether unifocal or multifocal, is feasible, safe, and can be delivered in an ambulatory care

setting. Early self-resolving lower urinary-tract symptoms were common. However, the strategy was well tolerated in the genitourinary functional domains. Almost 90% of men reported having erections satisfactory for intercourse at 12 months, and all were continent.

There was a signifi cant decrease in PSA levels from baseline to 12 months, with concentrations of serum PSA continuing to decline months after the initial treatment. This prevents an estimate of time to PSA nadir. Conclusions from this fi nding should not be made, although if this decline in serum PSA were to be reproduced, one possible reason for it might be the presentation of antigens to the immune response leading to a secondary immune response against the remaining prostate tissue.23

There are several limitations to our study. First, the number of participants included in the study was small. As a prospective development study representing stage 2b of the IDEAL guidelines for evaluating novel surgical strategies,24 we designed our trial to primarily assess side-eff ects. Second, characterisation of disease with template-prostate-mapping biopsies before focal therapy was diff erent to the 6-month verifi cation biopsy, because of research ethics committee stipulations to limit patient burden. We have recently started recruiting patients to a multicentre, phase 2 trial that will address some of the legitimate concerns regarding the natural history of untreated prostate tissue. The trial will assess focal HIFU applied to clinically signifi cant areas of prostate cancer identifi ed on entry by multiparametric MRI and template prostate-mapping biopsies25,26 followed by further multi-parametric MRI and template prostate-mapping biopsies applied to treated and untreated tissue at 3 years (NCT01194648). Third, our focal therapeutic strategy included ablation of normal surrounding tissue, which might contribute to the adverse events reported in our study. Destruction of some normal tissue is necessary to incorporate an adequate margin but because of the nature of the HIFU therapy, our margins are likely to be larger than they need to be. Other ablative modalities could serve to reduce the margins of ablated normal tissue. Furthermore, image-registration of preoperative MRI to treatment delivery could further help to reduce destruction of normal tissue by allowing the clinician to more accurately defi ne the boundaries of the target lesion.

Many retrospective case series have reported encouraging short-term functional and cancer-control outcomes of men treated in a focal manner with HIFU and cryotherapy.27–32 A prospective feasibility trial has reported on the use of focal interstitial laser therapy in a small cohort of 12 men with very low-risk unifocal disease.33 We have previously reported8 the outcomes from a prospective development study of 20 men with unilateral disease undergoing ablation of an entire prostate lobe using HIFU. 18 (90%) were leak-free and pad-free continent while 19 (95%) were pad-free after

Figure 7: Trifecta rate after focal high-intensity focused ultrasound for unifocal and multifocal localised prostate cancerPatient-reported trifecta outcomes were reported with validated questionnaires. Data are number of patients (%, 95% CI). *Derived using UCLA EPIC urinary domain questions: “over the past 4 weeks how often do you leak urine?” and “over the past 4 weeks how many pads or adult diapers per day did you usually use to control leakage?”. †Before imputation of data, 36 of 36 men were leak-free and pad-free. ‡Before imputation of data, 38 of 38 men were pad-free. §Proportion of men scoring ≥2 on question 2 of IIEF-15: “over the past 4 weeks when you had erections with sexual stimulation, how often were your erections hard enough for penetration?”. ¶Before imputation of data, 29 (88%) of 33 men had erections suffi cient for penetration. ||Phosphodiesterase-5 inhibitors (tadalafi l, sildenafi l, or vardenafi l; percentage calculated with denominator as those achieving erections suffi cient for intercourse).

Prop

ortio

n (%

)

100

80

60

40

20

0

Time (months)

12931 60

Functional outcomeContinence* Pad-free and leak free 38 26 36 37 38 38 (68%, 51–83) (95%,82–99) (97%, 86–100) (100%, 91–100) (100%, 91–100)† Pad-free 40 33 40 40 40 40 (83%, 67–93) (100%, 91–100) (100%, 91–100) (100%, 91–100) (100%, 91–100)‡Erections 35 19 29 29 31 31 satisfactory for (54%, 37–71) (83%, 66–93) (83%, 66–93) (89%, 73–97) (89%, 73–97)¶penetration§Use of 3/35 (9%) 4/19 14/29 17/29 17/31 14/31 phosphodiesterase- (21%, 6–46) (48%, 29–67) (59%, 39–76) (55%, 36–73) (45%, 27–64)5 inhibitors||

39/41* (95%, 95% CI 83–99)

26/31* (84%, 95% CI 66–95)

9/39* (23%, 95% CI 11–39)

3/39 (8%, 95% CI 2–21)

Pad-free continenceLeak-free/pad-free continenceErections sufficient for penetrationNo evidence of diseaseTrifecta (pad-free, leak-free, and erectile function satisfatory for penetration, and no evidence of clinically significant disease)Any cancer on biopsyClinically significant cancer on biopsy(≥3 mm and/or Gleason >3+3)

Articles


hemi-ablation; 17 (89%) of 19 achieved trifecta status at 12 months.8

The current study allowed user or operator deter-mination of ablative zones within the prostate on an individual basis provided our standardised method of focal therapy was followed. However, the treatments invariably followed one of three patterns of ablation (fi gure 1), by contrast with our previous study,8 whereby the treatment method was fi xed and standardised by mandating therapy to the entire half of the prostate associated with cancer (hemi-ablation) regardless of

individual lesion grade, volume, or location and proximity to a neurovascular bundle.

While our current study is not directly comparable to previous studies of focal therapy (panel), it continues to support the proposition that tissue preservation leads to functional preservation. The histological outcomes in this study were slightly worse than those reported for hemi-ablation, perhaps because of the requirements for increased pre cision of focal ablation—individual areas of cancer were targeted as opposed to a standardised hemi-ablation—with a resulting reduction in margin around the cancer. Image-registration software to accurately fuse, in real-time, pretreatment location data to intraoperative ultrasound images could improve histological outcomes.34 Another reasonable explanation might relate to physical limitations of the ablative technology. Other ablative therapies as well as brachy therapy35 and image-guided radiosurgery platforms are able to treat discrete volumes of tissue and might have diff ering outcomes. New platforms such as irreversible electroporation and photodynamic therapy are theor etically more tissue-specifi c and could allow neurovascular bundle preservation even if the ablative zone is close to the prostate capsule; tissue specifi city of these techniques has not yet been assessed.

In conclusion, focal therapy of individual prostate-cancer lesions, regardless of whether they are multifocal or unifocal, leads to a low rate of genitourinary side-eff ects and an encouraging rate of early freedom from clinically signifi cant prostate cancer. If the functional outcomes that we report are reproduced in larger studies and coupled with acceptable rates of cancer control in the medium to long term, focal therapy could off er a strategy by which the burden of treatment-related side-eff ects are addressed for a substantial proportion of men with localised prostate cancer. The design and execution of comparative-eff ectiveness research assessing long-term cancer control needs to be prioritised, especially at a pace than can match the potential for informal diff usion.

ContributorsME and HUA conceived the study. ME supervised the project and is

guarantor of the data. HUA, LD, RGH, RS, and MS recruited, treated,

and followed up patients. HUA and JVdM analysed the data. HUA wrote

the fi rst draft of the paper. AF provided histological expertise for the

project and reported all biopsies. APK and CA undertook MRIs and

provided imaging expertise. All authors contributed to the drafting and

editing of the manuscript and approved the fi nal version.

Confl icts of interestME and HUA received funding from USHIFU, GlaxoSmithKline, and

Advanced Medical Diagnostics for clinical trials. ME and HUA are paid

consultants to Steba Biotech and have received funding from USHIFU,

Focused Surgery, Misonix (manufacturers and distributors of the

Sonablate 500 high-intensity focused ultrasound device), and Oncura

and GE Healthcare for medical consultancy and travel to conferences.

All other authors declared no confl icts of interest.

AcknowledgmentsThis study was funded by the UK Medical Research Council, the Pelican

Cancer Foundation, and St Peters Trust, and sponsored by University

College London Hospitals NHS Foundation Trust. We thank the men


Systematic reviewIn 2007, we did a systematic review7 of reports on Medline and PubMed databases using the terms “focal therapy” and ”prostate cancer” and/or “high intensity focused ultrasound” and/or “cryotherapy/cryoablation/cryosurgery”. We also searched for all prostate cancer clinical trials in the UK National Cancer Research Network portfolio, ClinicalTrials.gov, and ISRCTN trial registries. We only identifi ed one prospective, commercially funded, phase 1 trial assessing photodynamic therapy in progress at the time. All other data were from two retrospective case series that were poorly reported. Subsequent to this review, we started a health technology assessment of focal therapy, following the phased approach described by the Medical Research Council complex interventions guidelines and subsequently formalised in the IDEAL guidelines for assessing surgical procedures.9

InterpretationOur study showed that the rate of genitourinary side-eff ects associated with focal therapy is low, coupled with an encouraging rate of early absence of clinically signifi cant prostate cancer. These fi ndings reaffi rm two other prospective development studies8,33 in which focal using high-intensity focused ultrasound and photothermal therapy was used.

Focal therapy could hold promise in mitigating the harms that result from current therapeutic strategies. Prioritisation and support of a pragmatic, randomised, clinical trial comparing focal therapy with whole-gland treatments is urgently needed. Such a trial should be done before informal diff usion and dissemination of focal therapy. Any randomised controlled trial should be pragmatic in nature and adaptive in execution, so that actual clinical practice is refl ected and new technological developments can be incorporated as they occur. Furthermore, since the natural history of prostate cancer is long, timelines based on metastases and mortality, which would require a trial at least 15 years in duration, might not be feasible or warranted. Therefore, endpoints that are clinically meaningful in the medium term are needed so that fi ndings are delivered effi ciently and in a timely fashion to change practice. Subsequent linkage to national electronic registries will ensure that robust cancer-control outcomes are still reported at a later date.

Articles


who agreed to participate in this trial, and Jane Coe and Victor Abu for

their invaluable help and support during the duration of the study. HUA

and ME receive funding for other research projects from the Wellcome

Trust, National Institute of Health Research-Health Technology

Assessment programme, the US National Institute of Health—National

Cancer Institute, Prostate Action, and Prostate Cancer Research Centre. ME receives funding in part from the UK National Institute of Health

Research University College London Hospitals/University College

London Comprehensive Biomedical Research Centre.

References1 Djulbegovic M, Beyth RJ, Neuberger MM, et al. Screening for

prostate cancer: systematic review and meta-analysis of randomised controlled trials. BMJ 2010; 341: c4543.

2 Sanda MG, Dunn RL, Michalski J, et al. Quality of life and satisfaction with outcome among prostate-cancer survivors. N Engl J Med 2008; 358: 1250–61.

3 Wilt TJ, MacDonald R, Rutks I, et al. Systematic review: comparative eff ectiveness and harms of treatments for clinically localized prostate cancer. Ann Intern Med 2008; 148: 435–48.

4 Barry MJ, Gallagher PM, Skinner JS, Fowler FJ Jr. Adverse eff ects of robotic-assisted laparoscopic versus open retropubic radical prostatectomy among a nationwide random sample of medicare-age men. J Clin Oncol 2012; 30: 513–18.

5 Nilsson S, Norlén BJ, Widmark A. A systematic overview of radiation therapy eff ects in prostate cancer. Acta Oncol 2004; 43: 316–81.


7 Ahmed HU, Pendse D, Illing R, Allen C, van der Meulen JH, Emberton M. Will focal therapy become a standard of care for men with localized prostate cancer? Nat Clin Pract Oncol 2007; 4: 632–42.

8 Ahmed HU, Freeman A, Kirkham A, et al. Focal therapy for localized prostate cancer: a phase I/II trial. J Urol 2011; 185: 1246–54.

9 McCulloch P, Altman DG, Campbell WB, et al. No surgical innovation without evaluation: the IDEAL recommendations. Lancet 2009; 374: 1105–12.

10 Barzell WE, Melamed MR. Appropriate patient selection in the focal treatment of prostate cancer: the role of transperineal 3-dimensional pathologic mapping of the prostate—a 4-year experience. Urology 2007; 70 (suppl 6): 27–35.

11 Litwin MS, Hays RD, Fink A, et al. The UCLA Prostate Cancer Index: development, reliability, and validity of a health-related quality of life measure. Med Care 1998; 36: 1002–12.

12 Esper P, Mo F, Chodak G, et al. Measuring quality of life in men with prostate cancer using the Functional Assessment of Cancer Therapy-Prostate (FACT-P) instrument. Urology 1997; 50: 920–28.

13 Rosen RC, Cappelleri JC, Smith MD, et al. Development and evaluation of an abridged, 5-item version of the international index of erectile function (IIEF-5) as a diagnostic tool for erectile dysfunction. Int J Impot Res 1999; 11: 319–26.

14 Rosen RC, Riley A, Wagner G, et al. The international index of erectile function (IIEF) a multidimensional scale for assessment of erectile dysfunction. Urology 1997; 49: 822–30.

15 Webster K, Cella D, Yost K. The Functional Assessment of Chronic Illness Therapy (FACIT) measurement system: properties, applications, and interpretation. Health Qual Life Outcomes 2003; 1: 79.

16 Punwani S, Emberton M, Walkden M, et al. Prostatic cancer surveillance following whole-gland high-intensity focused ultrasound: comparison of MRI and prostate-specifi c antigen for detection of residual or recurrent disease. Br J Radiol 2012; published online Jan 17. DOI:10.1259/bjr/61380797.

17 Rouvière O, Girouin N, Glas L, et al. Prostate cancer transrectal HIFU ablation: detection of local recurrences using T2-weighted and dynamic contrast-enhanced MRI. Eur Radiol 2010; 20: 48–55.

18 Machin D, Campbell M, Fayers P, et al. Sample size tables for clinical studies, 2nd edn. London: Blackwell Science, 1997: 21–22.

19 Ahmed HU, Zacharakis E, Dudderidge T, et al. High-intensity-focused ultrasound in the treatment of primary prostate cancer: the fi rst UK series. Br J Cancer 2009; 101: 19–26.

20 Mohler J, Bahnson RR, Boston B, et al. NCCN clinical practice guidelines in oncology: prostate cancer. J Natl Compr Canc Netw 2010; 8: 162–200.

21 Epstein JI, Walsh PC, Carmichael M, Brendler CB. Pathologic and clinical fi ndings to predict tumor extent of nonpalpable (stage T1c) prostate cancer. JAMA 1994; 271: 368–74.

22 Epstein JI. Prognostic signifi cance of tumor volume in radical prostatectomy and needle biopsy specimens. J Urol 2011; 186: 790–97.

23 Finley DS, Pouliot F, Shuch B, et al. Ultrasound-based combination therapy: potential in urologic cancer. Expert Rev Anticancer Ther 2011; 11: 107–13.

24 Medical Research Council Health Services and Public Health Research Board. A framework for development and evaluation of RCTs for complex interventions to improve health. London: Medical Research Council, 2000.

25 Ahmed HU. The index lesion and the origin of prostate cancer. N Engl J Med 2009; 361: 1704–06.

26 Ahmed HU, Hu Y, Carter T, et al. Characterizing clinically signifi cant prostate cancer using template prostate mapping biopsy. J Urol 2011; 186: 458–64.

27 Onik G, Vaughan D, Lotenfoe R, et al. The “male lumpectomy”: focal therapy for prostate cancer using cryoablation results in 48 patients with at least 2-year follow-up. Urol Oncol 2008; 26: 500–05.

28 Lambert EH, Bolte K, Masson P, et al. Focal cryosurgery: encouraging health outcomes for unifocal prostate cancer. Urology 2007; 69: 1117–20.

29 Muto S, Yoshii T, Saito K, et al. Focal therapy with high-intensity-focused ultrasound in the treatment of localized prostate cancer. Jpn J Clin Oncol 2008; 38: 192–99.

30 Ward JF, Jones JS. Focal cryotherapy for localized prostate cancer: a report from the national Cryo On-Line Database (COLD) Registry. BJU Int 2011; published online Oct 28. DOI:10.1111/j.1464-410X.2011.10578.x.

31 El Fegoun AB, Barret E, Prapotnich D, et al. Focal therapy with high-intensity focused ultrasound for prostate cancer in the elderly. A feasibility study with 10 years follow-up. Int Braz J Urol 2011; 37: 213–19.

32 Lambert EH, Bolte K, Masson P, et al. Focal cryosurgery: encouraging health outcomes for unifocal prostate cancer. Urology 2007; 69: 1117–20.

33 Lindner U, Weersink RA, Haider MA, et al. Image guided photothermal focal therapy for localized prostate cancer: phase I trial. J Urol 2009; 182: 1371–77.

34 Hu Y, Ahmed HU, Taylor Z, et al. MR to ultrasound registration for image-guided prostate interventions. Med Image Anal 2012; 16: 687–703.

35 Langley S, Ahmed HU, Al-Qaisieh B, et al. Report of a consensus meeting on focal low dose rate brachytherapy for prostate cancer. BJU Int 2012; 109 (suppl 1): 7–16.

www.thelancet.com/oncology Published online April 12, 2012 DOI:10.1016/S1470-2045(12)70132-8 1

Articles


Centre for Primary Health Care Research, Lund University and Region Skåne, Lund, Sweden (J Ji MD, Prof K Sundquist MD, Prof J Sundquist MD); and Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, CA, USA (Prof J Sundquist)

Correspondence to:Dr Jianguang Ji, Centre for Primary Health Care Research, Lund University, CRC, Building 28, Floor 11, Entrance 72, Skåne University Hospital, 205 02 Malmö, [email protected]

Cancer incidence in patients with polyglutamine diseases: a population-based study in SwedenJianguang Ji, Kristina Sundquist, Jan Sundquist

SummaryBackground Polyglutamine (polyQ) diseases are characterised by the expansion of CAG triplet repeats in specifi c genes. The accumulated encoded proteins aff ect the transcription of numerous transcription factors. We investigated whether polyQ diseases reduce the risk of cancer development.

Methods Data on patients with the polyQ diseases Huntington’s disease (HD), spinobulbar muscular atrophy (SBMA), and hereditary ataxia (HA) in Sweden were linked to the Swedish Cancer Registry. We calculated standardised incidence ratios for cancers at specifi c sites or of specifi c types and the risks were compared with those in the general population. We also analysed risks in the unaff ected parents of patients.

Findings In the period January, 1969, to December, 2008, we identifi ed 1510 patients with HD, 471 with SBMA, and 3425 with HA. Cancer was diagnosed in 91 (6·0%) HD patients, 34 (7·2%) SBMA patients, and 421 (12·3%) HA patients. The standardised incidence ratios were 0·47 (95% CI 0·38–0·58), 0·65 (0·45–0·91), and 0·77 (0·70–0·85), respectively. Before diagnosis of polyQ disease, the risk of cancer was even lower. Cancer incidence and risk in the unaff ected parents of patients with polyQ diseases were similar to those in the general population.

Interpretation The consistently decreased incidence of cancer in patients with polyQ diseases suggests that a common mechanism protects against the development of cancer. This feature could be related to the polyQ-tract expansion seen in these diseases. Further studies are warranted to investigate the underlying mechanisms linking cancer and polyQ diseases.

Funding Swedish Cancer Society, Swedish Council for Working Life and Social Research.

IntroductionThe polyglutamine (polyQ) diseases belong to a group of rare neurodegenerative disorders that are characterised by the expansion of CAG triplet repeats in specifi c genes.1–4 Nine diseases have been identifi ed: Huntington’s disease (HD), spinobulbar muscular atrophy (SBMA), dentatorubral and pallidoluysian atrophy, and spino cerebellar ataxia types 1, 2, 3, 6, 7, and 17. The genes underlying these disorders share no homology aside from expansion of the polyQ tracts.5–7 All patients with polyQ diseases present with progressive degeneration of a population of neurons in the CNS that are involved in motor control. Notably, HD is charac terised by chorea, impaired judgment and cognition, and loss of coordination, SBMA by progressive muscular atrophy and loss of motor neurons, and spinocerebellar ataxias by progressive coordination impairment that aff ects the legs and hands, and speech.

Common pathogenic mechanisms have been sug-gested for the polyQ diseases. PolyQ-containing proteins accumulate in the cytoplasm, nucleus, or both, and interact with numerous transcription factors,1–4 which impedes the activity of many genes. Eventually, accumulation leads to dysfunction of the cellular machinery and to cell death. In patients with HD, muta-tions lead to expanded polyQ tracts in the HTT gene, which causes upregulation of nuclear concentrations

and transcriptional activity of tumour suppressor P53.8–10 The incidence of cancer is reported to be decreased in patients with HD.11 Other neuropsychiatric and neurodegenerative disorders, such as Parkinson’s disease, Alzheimer’s disease, and schizophrenia, are also associated with decreased risk of cancer.12,13 The fi ndings might, however, have been aff ected by negative surveillance bias, medication, change in diet after diagnosis of the neurological disorder, or a combination of these factors. One example of the negative surveillance bias is clinicians neglecting to fully assess symptoms on the basis of physical complaints in patients with HD or other neuropsychiatric and neurodegenerative disorders, which can lead to delayed or reduced diagnosis of cancer, or both.

In this study we analysed data from Swedish nationwide registries to estimate the incidence rates for all cancers and for cancers by specifi c site or type in patients with a diagnosis of HD, SBMA, or hereditary ataxia (HA). We used the latter as a proxy for spinocerebellar ataxia types 1, 2, 3, 6, 7, and 17 because of the lack of specifi c diagnostic codes for these diseases. Most types of HA are inherited in an autosomal dominant manner; only a few, such as Friedreich’s ataxia, are autosomal recessive. Spino cerebellar ataxia types 1, 2, 3, 6, 7, and 17 account for over 60% of all cases of HA.14 This study is, to the best of our knowledge, the largest so far on this subject.

Articles

2 www.thelancet.com/oncology Published online April 12, 2012 DOI:10.1016/S1470-2045(12)70132-8

MethodsStudy populationThe study was approved by the ethics committee at Lund University, Lund, Sweden. We used the Swedish Hospital Discharge Register and Outpatient Registry to identify individuals treated in hospitals for HD, SBMA, and HA between Jan 1, 1969, and Dec 31, 2008. The Swedish Hospital Discharge Register was founded in 1964–65 by the National Board of Health and Welfare, and has had complete nationwide coverage since 1987. It is part of a research database maintained at the Centre for Primary Health Care Research, Lund University. Outpatient data have been included in the database from 2000 onwards. We used codes in the various revisions of the International Classifi cation of Diseases (ICD) that covered the study period to identify cases of each disease. ICD-8 covered the years 1969–86, ICD-9 1987–96, and ICD-10 1997 onwards. For HD, the codes were 3310 (in ICD-8), 3334 (ICD-9), and G10 and F022 (ICD-10), for SBMA they were 3352 (ICD-9) and G121 (ICD-10), and for HA they were 332 (ICD-8), 334 (ICD-9), and G11 (ICD-10). Dentatorubral and pallidoluysian atrophy was not included in the present study because of a lack of specifi c diagnostic codes. Information on identifi ed patients was linked to the Swedish Multi-Generation Register to identify unaff ected parents.

To fi nd all incident cases of cancer during the study period, information on patients and unaff ected parents was linked to data in the Swedish Cancer Registry, which was started in 1958 and is maintained by the National Board of Health and Welfare. Clinicians and pathologists

or cytologists must report all newly diagnosed cancers to the registry15 and, therefore, it has almost 100% coverage. We retrieved diagnoses of malignant and benign tumours made during the study period (ICD-7 codes 140–209, as used in the Swedish Cancer Registry). Because the numbers of cancers were low for some sites in the body, we grouped them for this study as cancers of the digestive system (with colorectal cancers listed separately), lung, breast, female genital organs, prostate, urinary system, haemopoietic and lymphoid tissues, and other sites. We investigated the incidence of cancer before the diagnosis of polyQ diseases to control for the confounding eff ects, such as negative surveillance bias and medication. We assessed the risks of developing other diseases as control conditions, and analysed the risk of cancer according to age at diagnosis of HD and SBMA. Cancer risk in patients with HA was not analysed by age at diagnosis because it involves six diseases with pathogenic CAG repeats of diff erent lengths (range 21–55) and, therefore, the age of onset varies widely. We investigated the incidence of cancer among unaff ected parents to determine whether the biological connection between polyQ diseases and cancer is due to the polyQ tract itself or to environmental, genetic, or other factors shared by family members.

Additional linkages to identify individual-level data were made to the Swedish National Population and Housing Census (for socioeconomic status, available for every 10th year since 1960), the Cause of Death Register (for date of death), and the Emigration Registry (for date of emigration). Patients were found by use of their individual national identifi cation numbers, which are assigned to residents of Sweden for their lifetimes. In the study data, these numbers were replaced by serial numbers to preserve anonymity.

Statistical analysisPerson-years of risk were calculated from the date of birth, date of immigration, or Jan 1, 1965 (whichever was latest), until diagnosis of cancer, death, emigration, or the end of the study period (whichever came fi rst) for polyQ patients and for the general population. We calculated ratios of the observed and expected numbers of cases by indirect standardisation (standard ised incidence ratios):16,17

where O=Σoj is the total number of cancer cases in the study group of patients with polyQ diseases; E* is calculated by the multiplication of the stratum-specifi c person-years (nj) in patients with polyQ diseases with stratum-specifi c standard incidence rates (λ*j) obtained from the reference group (all individuals without a diagnosis of HD, SBMA, or HA); oj represents the observed number of cases in the study group contributing

SIR =Σj = 1oj

J

Σj = 1nj λj*J =

OE *

Huntington’s disease (n=1510)

Spinobulbar muscular atrophy (n=471)

Hereditary ataxia (n=3425)

Sex

Male 732 (48·5%) 296 (62·8%) 1902 (55·5%)

Female 777 (51·5%) 175 (37·2%) 1523 (44·5%)

Age (years)

<40 283 (18·7%) 204 (43·3%) 1075 (31·4%)

40–49 297 (19·7%) 38 (8·1%) 342 (10·0%)

50–59 381 (25·2%) 84 (17·8%) 534 (15·6%))

≥60 549 (36·4%) 145 (30·8%) 1474 (43·0%)

Median 54 (43–65) 47 (18 –65) 56 (30–70)

Time period

1969–79 287 (19·0%) 0 357 (10·4%)

1980–89 298 (19·7%) 60 (12·7%) 833 (24·3%)

1990–99 271 (17·9%) 96 (20·4%) 928 (27·1%)

2000–08 654 (43·3%) 315 (66·9%) 1307 (38·2%)

Deaths 879 (58·2%) 134 (28·5%) 1536 (44·8%)

Median age at death (years) 64 (53–74) 73 (63–81) 75 (64–82)

Cancer cases 91 (6·0%) 34 (7·2%) 421 (12·3%)

Median duration of survival (years) 5 (1–9) 4 (2–8) 5 (2–10)

Data are number (%) or median (IQR).

Table 1: Characteristics of patients with Huntington’s disease, spinobulbar muscular atrophy, and hereditary ataxia in Sweden

Articles


to the jth stratum; and j represents the strata, defi ned by the cross-classifi cation of variables, including age (5-year age groups), sex, socioeconomic status (farmer, self-employed, manual worker, white-collar worker, professional, and other), time period (5-year periods), and region of residence (large cities, southern Sweden, and northern Sweden).18 We assumed a Poisson distribution for calculation of 95% CIs.18 All analyses were done with SAS (version 9.1).

Role of the funding sourceThe sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had fi nal responsibility for the decision to submit for publication.

ResultsWe identifi ed 1510 patients with HD, 471 with SBMA, and 3425 with HA during the study period (table 1). Similar numbers of male and female patients had HD and HA, but more male patients had SBMA. The median ages at diagnosis were 54 (IQR 43–65), 47 (18–65), and 56 (30–70) years, respectively, for HD, SBMA, and HA, and the median ages at death were 64 (53–74), 73 (63–81), and 75 (64–82) years, respectively. The median survival for patients with polyQ diseases who were also diagnosed as having cancer was 5 (range 2–10) years, which did not

diff er signifi cantly from that in cancer patients without polyQ diseases, in whom the median survial was 4 years (IQR 0–6).

91 (6·0%) individuals with HD developed cancer during the study period, as did 34 (7·2%) patients with SBMA and 421 (12·3%) patients with HA. The overall standardised incidence ratios showed sig nifi cantly reduced risks of cancer in all polyQ disease groups (HD, 0·47, 95% CI 0·38–0·58; SBMA, 0·65, 0·45–0·91; and HA, 0·77, 0·70–0·85; table 2). The overall incidence and risk of cancer were even lower before diagnosis of polyQ

HD SBMA HA

Overall risk Before diagnosis of polyQ disease



Observed cancer cases

SIR (95% CI) Observed cancer cases





SIR (95% CI)

Digestive system 19 0·49 (0·29–0·73)*

6 0·26 (0·09–0·51)*

5 0·46 (0·14–0·94)*

1 0·13 (0·00–0·50)*

73 0·60 (0·47–0·75)*

17 0·22 (0·13–0·34)*

Colorectum 14 0·70 (0·38–1·11)

5 0·43 (0·14–0·90)*

4 0·71 (0·18–1·57)

1 0·25 (0·00–1·00)

47 0·76 (0·56–0·99)*

12 0·32 (0·17–0·53)*

Lung 6 0·43 (0·16–0·95)*

2 0·25 (0·02–0·93)*

0 ·· 0 ·· 17 0·39 (0·23–0·63)*

1 0·04 (0·00–0·21)*

Breast 16 0·54 (0·31–0·88)*

12 0·64 (0·33–1·12)

4 0·79 (0·21–2·04)

2 0·55 (0·05–2·02)

47 0·79 (0·58–1·05)

25 0·64 (0·41–0·94)*

Female genital organs 7 0·40 (0·16–0·75)*

6 0·50 (0·18–0·98)*

1 0·36 (0·00–1·43)

1 0·47 (0·00–1·82)

23 0·65 (0·41–0·95)*

13 0·52 (0·28–0·84)*

Prostate 12 0·52 (0·27–0·91)*

6 0·51 (0·18–1·11)

5 0·52 (0·16–1·23)

2 0·34 (0·03–1·27)

64 0·72 (0·55–0·91)*

23 0·50 (0·32–0·76)*

Urinary system 9 0·67 (0·30–1·17)

4 0·49 (0·13–1·09)

3 0·70 (0·13–1·71)

2 0·65 (0·06–1·85)

38 0·87 (0·62–1·17)

19 0·70 (0·42–1·05)

Haemopoietic and lymphoid tissues

4 0·27 (0·07–0·61)*

2 0·22 (0·02–0·63)*

2 0·47 (0·04–1·34)

0 .. 38 0·91 (0·64–1·22)

16 0·61 (0·35–0·94)*

Other sites 18 0·43 (0·26–0·65)*

9 0·34 (0·15–0·59)*

14 1·20 (0·65–1·90)

7 0·82 (0·33–1·54)

121 1·08 (0·89–1·28)

68 0·96 (0·75–1·20)

All 91 0·47 (0·38–0·58)*

47 0·40 (0·30–0·54)*

34 0·65 (0·45–0·91)*

15 0·41 (0·23–0·67)*

421 0·77 (0·70–0·85)*

181 0·54 (0·46–0·62)*

HD=Huntington’s disease. SBMA=spinobulbar muscular atrophy. HA=hereditary ataxia. polyQ=polyglutamine. SIR=standardised incidence ratio. *Signifi cant (95% CI does not include 1).

Table 2: Cancer risk in patients with Huntington’s disease, spinobulbar muscular atrophy, and hereditary ataxia in Sweden


SIR (95% CI)

Digestive system* 0 ··

Lung 0 ··

Breast 6 1·69 (0·61–3·71)

Female genital organs 3 1·63 (0·31–4·00)

Other female genital 0 ··

Prostate 2 1·07 (0·1–3·95)

Urinary system 1 0·89 (0·00–3·50)

Haemopoietic and lymphoid tissues 1 0·61 (0·00–2·39)

Other sites 8 1·62 (0·69–2·94)

All 21 1·09 (0·67–1·67)

SIR=standardised incidence ratio. *Includes colorectal cancer.

Table 3: Cancer risk in patients with Friedreich’s ataxia

Articles


diseases, and almost all were decreased when cancers were assessed by specifi c sites or types (table 2). Some standardised incidence ratios were not signifi cant, probably because of small numbers of cancer cases.

To minimise confounding eff ects of recessive forms of HA, such as Friedreich’s ataxia and ataxia telangiectasia, which typically occur during infancy, we excluded from the analysis patients who had been diagnosed as having HA before age 20 years. The risk of cancer in the remaining 409 patients was not aff ected (standardised incidence ratio 0·75, 95% CI 0·68–0·83). The overall risk of cancer in patients with Friedreich’s ataxia was not notably diff erent from that in the general population (standardised incidence ratio 1·09, 95% CI 0·67–1·67; table 3) and no signifi cant standardised incidence ratios were noted for any specifi c cancer site or type (table 3).

The risks of patients with polyQ diseases developing stroke, diabetes mellitus, or chronic obstructive pulmonary disease were generally similar to those in the general population (appendix). The risk of developing

benign tumours was lower than that in the general population, but only that in patients with HD was signifi cant (0·77, 95% CI 0·58–0·99; appendix).

The decrease in risk of cancer was greater in patients younger than 60 years at diagnosis of HD and SBMA than in those aged 60 years or older (table 4), but the diff erences between age groups were not signifi cant. Among patients younger than 40 years at diagnosis of polyQ disease, four HD patients developed cancer (standardised incidence ratio 0·73, 95% CI 0·19–1·61; one astrocytoma and one ependymoma in the brain, one melanoma, and one thyroid gland tumour), whereas no patients aged younger than 40 years at diagnosis of SBMA developed cancer during their lifetime.

The overall risk of cancer in the unaff ected parents of patients with polyQ disorders was similar to that in the general population (table 5). Incidence was signifi cantly decreased for cancers of the digestive system in the unaff ected parents of patients with HD and for breast cancer in those of patients with HA.

HD SBMA

Age <60 years Age ≥60 years Age <60 years Age ≥60 years





SIR (95% CI)

Digestive system* 5 0·61 (0·19–1·25) 14 0·47 (0·25–0·74)† 1 0·60 (0·00–2·33) 4 0·43 (0·11–0·96)†

Lung 3 0·75 (0·14–2·23) 3 0·30 (0·06–0·90)† 0 ·· 0 ··

Breast 8 0·68 (0·29–1·34) 8 0·45 (0·19–0·89)† 2 0·95 (0·09–3·50) 2 0·67 (0·06–2·48)

Female genital organs 2 0·31 (0·03–0·88)† 5 0·45 (0·14–0·93)† 0 0·00 (1·06–1·06) 1 0·54 (0·00–2·13)

Prostate 2 0·48 (0·04–1·75) 10 0·53 (0·25–0·98)† 1 0·72 (0·00–4·15) 4 0·49 (0·13–1·26)

Urinary system 2 0·58 (0·05–1·65) 7 0·70 (0·28–1·31) 1 1·45 (0·00–5·68) 2 0·55 (0·05–1·59)

Haemopoietic and lymphoid tissues 1 0·20 (0·00–0·80)† 3 0·31 (0·06–0·76)† 0 ·· 2 0·65 (0·06–1·86)

Other sites 11 0·67 (0·33–1·12) 7 0·27 (0·11–0·51)† 4 1·13 (0·29–2·51) 10 1·22 (0·58–2·10)

All 34 0·57 (0·39–0·80)† 57 0·43 (0·32–0·55)† 9 0·74 (0·33–1·41) 25 0·62 (0·40–0·92)†

HD=Huntington’s disease. SBMA=spinobulbar muscular atrophy. SIR=standardised incidence ratio. *Includes colorectal cancer. †Signifi cant (95% CI does not include 1).

Table 4: Cancer risk in patients with Huntington’s disease and spinobulbar muscular atrophy by age at diagnosis

HD SBMA HA




SIR (95% CI)

Digestive system 30 0·66 (0·44–0·91)* 15 0·72 (0·40–1·13) 130 0·99 (0·83–1·17)

Colorectum 13 0·57 (0·30–0·92)* 11 1·03 (0·51–1·72) 64 0·97 (0·74–1·22)

Lung 16 1·09 (0·62–1·78) 10 1·50 (0·71–2·77) 45 1·08 (0·79–1·45)

Breast 23 0·91 (0·57–1·36) 9 0·74 (0·34–1·42) 56 0·75 (0·57–0·98)*

Female genital organs 14 0·87 (0·47–1·38) 6 0·85 (0·31–1·66) 48 1·06 (0·78–1·38)

Prostate 30 1·20 (0·81–1·71) 14 1·00 (0·55–1·69) 89 1·16 (0·93–1·43)

Urinary system 14 0·96 (0·52–1·52) 11 1·56 (0·78–2·63) 54 1·27 (0·95–1·63)

Haemopoietic and lymphoid tissues 15 1·04 (0·58–1·63) 3 0·41 (0·08–1·01) 39 0·90 (0·64–1·20)

Other sites 43 0·98 (0·86–1·11) 21 0·96 (0·79–1·15) 109 0·99 (0·92–1·07)

All 185 0·95 (0·82–1·10) 89 0·94 (0·76–1·16) 570 1·00 (0·92–1·09)

HD=Huntington’s disease. SBMA=spinobulbar muscular atrophy. HA=hereditary ataxia. SIR=standardised incidence ratio. *Signifi cant (95% CI does not include 1).

Table 5: Cancer risk in unaff ected parents of patients with Huntington’s disease, spinobulbar muscular atrophy, and hereditary ataxia in Sweden


Articles


DiscussionIn this population-based study, we found that the incidence of cancer was signifi cantly lower among patients with polyQ diseases than in the general population after we accounted for confounding factors, such as age, sex, time period, geographic region, and socioeconomic status. Furthermore, as the decrease in cancer risk was even more pronounced before the diagnosis of polyQ diseases, some of the causal factors predisposing to these disorders might be protective against the development of cancer. Cancer risk in the unaff ected parents of patients with polyQ diseases was similar to that in the general population.

A possible common protective mechanism against cancer development across polyQ diseases could be the expansion of polyQ tracts. The absence of a reduction in risk in the unaff ected parents of patients with polyQ diseases suggests that familial factors, such as other shared genetic and environmental factors, are unlikely to contribute notably to the protective eff ect. Additionally, owing to the strongly reduced risk of developing cancer before the diagnosis of polyQ diseases, individual factors—eg, medication, diet change, and low detection rates of cancer because of the poor prognosis in patients with polyQ diseases—could be excluded. Unhealthy behaviours that increase the risk of cancer, such as cigarette smoking and high alcohol consumption, are more frequently seen in presymp tomatic HD patients than in the general population.19 This fi nding supports the hypothesis of a relation between decreased risk of cancer and expansion of polyQ tracts, because otherwise the risk would be expected to remain high after the onset of HD. Furthermore, the genes involved in diff erent polyQ diseases have little in common, but the diseases have similar clinical features and similar protective eff ects on cancer, which suggest that their biological features are not related to specifi c proteins. Previous data from studies in mice suggest that a long polyQ tract can by itself be toxic to cells.20,21 Cancer risk in patients with Friedreich’s ataxia, a recessive HA disease not related to polyQ-tract expansion, did not diff er from that in the general population. Again, this fi nding supports the theory that expansion of polyQ tracts is protective against cancer.

The overall risk of cancer was lower in HD patients than in patients with other polyQ diseases (standardised incidence ratio 0·47 for HD vs 0·65 for SBMA and 0·77 for HA), possibly because of the higher mortality and earlier age at death in HD patients (table 1). However, the overall standardised incidence ratios for cancer before the diagnosis of polyQ diseases were similar (0·40 in HD, 0·41 in SBMA, and 0·54 in HA), which suggests the protective eff ects against cancer are similar in all three groups.

The hypothesis that the length of CAG is negatively associated with risk of cancer in patients with polyQ diseases seems reasonable. Unfortunately, data on CAG-repeat length are not available in our nationwide

databases. Therefore, we used age at diagnosis as a surrogate in an attempt to investigate a relation with risk of developing cancer.22–24 We could not fi nd a negative association between age at diagnosis and cancer risk; patients aged 60 years or older at the time of diagnosis of a polyQ disease had slightly lower risks of cancer than did patients who were younger at diagnosis, but this diff erence was not signifi cant. One possible explanation for this fi nding is that age at diagnosis in hospital does not fully represent age at onset because only the most severe cases are treated in hospital. In addition, CAG-repeat length can explain only half of the variation in age at onset for patients with HD,22–24 and even less in those with SBMA.25–27 The age at onset of HD diff ers by up to 7 years in monozygotic twins.28 Marked diff erences in clinical symptoms and behav ioural abilities have also been reported in monozygotic twins,29 which suggest that environmental factors are highly associated with age at onset.28 Alcohol abuse and cigarette smoking have been reported to be associated with early onset of HD.19,30 Further studies are needed to directly assess the predictive values of CAG-repeat length and other possible factors, such as alcohol consumption and smoking.

The mechanisms behind the protective eff ects against cancer are unclear and further research is warranted (panel). The tumour-suppressor P53 protein is highly expressed in patients with HD,8,9 which could partly explain the decreased risk of cancer. However, no association has been reported between P53 expression and other polyQ diseases, and mutations in P53 are found in less than 50% of human tumours,31,32 which suggests that other mechanisms, such as apoptosis caused by expanded polyQ tracts or altered neurotransmitter release, are involved.11,33 A toxic gain-of-function mutation has


Systematic reviewWe searched PubMed for articles published in English from Jan 1, 1970, to March 20, 2011, with the MeSH terms “Huntington’s disease”, “spinobulbar muscular atrophy”, “dentatorubral and pallidoluysian atrophy”, “spinocerebellar ataxias”, “polyQ diseases”, “cancer”, “incidence”, and “cancer mortality” in diff erent combinations. We found that Huntington’s disease has been associated with a decreased incidence of cancer,11 but the associations with other polyglutamine (polyQ) diseases are not reported. Other neuropsychiatric and neurodegenerative disorders, such as Parkinson’s disease, Alzheimer’s disease, and schizophrenia, have been associated with decreased risk of cancer. 12,13 Therefore, we undertook a large, population-based study to compare the risk of cancer in patients with polyQ diseases with that in the general population.

InterpretationThe incidence and risk of cancer, overall and by specifi c site or type of cancer, were signifi cantly lower in patients with polyQ diseases than in the general population, but not in unaff ected parents. Additionally, the decrease in risk was even greater before the diagnosis of all polyQ diseases. Thus, confounding eff ects, such as negative surveillance bias, medication, or change in diet after diagnosis, could be excluded. Our fi ndings suggest a common mechanism in patients with polyQ diseases that protects against the development of cancer, and expansion of the polyQ tracts is likely.

Articles


been proposed to be involved in polyQ diseases,8,34 although the lack of common genes across disorders suggests a greater role for the polyQ-tract expansion itself. Additionally, the risk of benign tumours was decreased in patients with polyQ diseases. Therefore, the protective eff ect against cancer does not simply involve suppression of metastasis.35 On the basis of the similar protective eff ects across the polyQ diseases studied, these disorders could be reconsidered as a set of hereditary cancer-suppressive diseases.

Previous studies have found decreased risks of cancers in patients with Parkinson’s disease, Alzheimer’s disease, and schizophrenia,12,13 and specifi c genes have been reported to account partly for these associations. For example, the tumour-suppressor gene PARK2 is mutated in autosomal recessive Parkinson’s disease. Genome-wide association studies in patients with polyQ diseases might uncover mechanisms underlying the protective eff ects against cancer. Research on epigenetic modi-fi cations, microRNA alterations, or both, could also provide evidence regarding the pathogenesis of both polyQ diseases and cancer. Whether the protective eff ects against cancer in polyQ diseases and other neurode-generative disorders involve similar pathways should also be investigated.

Strengths of this study include the analysis of a large number of patients from a national population, which enabled adequate numbers of those with rare diseases to be studied. The use of a nationwide database that includes the entire population avoided selection bias. Other strengths are the prospective design and the completeness of the follow-up.

A limitation of the study is that patients with spinocerebellar ataxias 1, 2, 3, 6, 7, and 17 could not be identifi ed in our database because specifi c diagnostic codes were not provided, although we used HA as a proxy. The spinocerebellar ataxias of interest account for over 60% of HA cases,14 and the risk of cancer was not aff ected when we excluded potential recessive HA. Another limitation is that outpatient-clinic data were only available from 2000, and the Swedish Hospital Discharge Registry has only had complete nationwide coverage since 1987. Some individuals with polyQ diseases might, therefore, have been missed and some with pathogenic CAG-repeat mutations could have been included in the reference group. Nevertheless, as polyQ diseases are chronic neurodegenerative diseases, we believe that there would be few aff ected individuals who had not been treated as outpatients at least once between 2000 and 2008, and, therefore, any bias due to missing data is probably conservative. The decreased risk of colon and breast cancers in unaff ected parents of patients with HD and HA could be related to the multiple comparisons undertaken in this study. Thus, we cannot exclude the possibility that these fi ndings were due to chance. Additionally, the diagnosis of polyQ diseases could have been more diffi cult in the parental generation than in the

off spring generation because the genetic tools might have been unavailable before death. Some people with pathogenic CAG-repeat mutations might, therefore, have been incorrectly classifi ed as being unaff ected. In turn, this misclassifi cation might have lowered the incidence of cancer. Another limitation is that the ICD coding system was changed during the study period, which could have led to inconsistency between some records, especially around the times when new revisions were published. However, the overall accuracy of diagnosis from the Swedish Hospital Discharge Registry was suggested in a review to be 85–95%,36 and any such bias is likely to be non-diff erential. Finally, age at admission to hospital or at the time of hospital diagnosis of outpatients does not necessarily represent age at onset.

In summary, in this large study, we found that incidence of cancer, overall and for specifi c sites or types, was signifi cantly lower in patients with polyQ diseases than in the general population. Cancer incidence in the unaff ected parents of patients with polyQ diseases was similar to that in the general population. These fi ndings suggest that expansion of polyQ tracts in polyQ diseases contributes to a protective eff ect against the development of cancer. Future studies should investigate the specifi c biological mechanisms underlying the reduced cancer risk in patients with polyQ diseases.

ContributorsJJ conceived and designed the study and analysed and interpreted the

data. KS and JS obtained and interpreted the data. JJ, KS, and JS drafted

the report and all authors revised drafts and approved the fi nal version.


AcknowledgmentsWe thank Stephen Gilliver, Centre for Primary Health Care Research,

Malmö, Sweden, for his valuable comments on the text.

References1 Paulson HL, Bonini NM, Roth KA. Polyglutamine disease and

neuronal cell death. Proc Natl Acad Sci USA 2000; 97: 12957–58.

2 Gatchel JR, Zoghbi HY. Diseases of unstable repeat expansion: mechanisms and common principles. Nat Rev Genet 2005; 6: 743–55.

3 Ross CA. Polyglutamine pathogenesis: emergence of unifying mechanisms for Huntington’s disease and related disorders. Neuron 2002; 35: 819–22.

4 Shao J, Diamond MI. Polyglutamine diseases: emerging concepts in pathogenesis and therapy. Hum Mol Genet 2007; 16: R115–23.

5 Walker FO. Huntington’s disease. Lancet 2007; 369: 218–28.

6 Klockgether T. Update on degenerative ataxias. Curr Opin Neurol 2011; 24: 339–45.

7 Durr A. Autosomal dominant cerebellar ataxias: polyglutamine expansions and beyond. Lancet Neurol 2010; 9: 885–94.

8 Nucifora FC Jr, Sasaki M, Peters MF, et al. Interference by huntingtin and atrophin-1 with cbp-mediated transcription leading to cellular toxicity. Science 2001; 291: 2423–28.

9 Steff an JS, Kazantsev A, Spasic-Boskovic O, et al. The Huntington’s disease protein interacts with p53 and CREB-binding protein and represses transcription. Proc Natl Acad Sci USA 2000; 97: 6763–68.

10 Bae BI, Xu H, Igarashi S, et al. p53 mediates cellular dysfunction and behavioral abnormalities in Huntington’s disease. Neuron 2005; 47: 29–41.

11 Sorensen SA, Fenger K, Olsen JH. Signifi cantly lower incidence of cancer among patients with Huntington disease: an apoptotic eff ect of an expanded polyglutamine tract? Cancer 1999; 86: 1342–46.

Articles


12 Tabarés-Seisdedos R, Dumont N, Baudot A, et al. No paradox, no progress: inverse cancer comorbidity in people with other complex diseases. Lancet Oncol 2011; 12: 604–08.

13 Plun-Favreau H, Lewis PA, Hardy J, Martins LM, Wood NW. Cancer and neurodegeneration: between the devil and the deep blue sea. PLoS Genet 2010; 6: e1001257.

14 O’Sullivan Smith C, Michelson SJ, Bennett RL, Bird TD. Spinocerebellar ataxia: making an informed choice about genetic testing. http://depts.washington.edu/neurogen/downloads/ataxia.pdf (accessed March 22, 2012).

15 National Board of Health and Welfare. Cancer incidence in Sweden 2007. http://www.socialstyrelsen.se/Lists/Artikelkatalog/Attachments/8715/2008-125-16_200812516_rev2.pdf (accessed March 22, 2012).

16 Rothman KJ, Greenland S. Modern epidemiology, 2nd edn. Philadelphia, PA: Lippincott-Raven, 1998.

17 Breslow NE, Day NE, eds. Statistical methods in cancer research: volume II—the design and analysis of cohort studies. Lyon: International Agency for Research on Cancer, 1987.

18 Esteève J, Benhamou E, Raymond L, eds. Statistical methods in cancer research: volume IV—descriptive epidemiology. Lyon: International Agency for Research on Cancer, 1994.

19 Ehret JC, Day PS, Wiegand R, Wojcieszek J, Chambers RA. Huntington disease as a dual diagnosis disorder: data from the National Research Roster for Huntington disease patients and families. Drug Alcohol Depend 2007; 86: 283–86.

20 Martindale D, Hackam A, Wieczorek A, et al. Length of huntingtin and its polyglutamine tract infl uences localization and frequency of intracellular aggregates. Nat Genet 1998; 18: 150–54.

21 Kim YJ, Yi Y, Sapp E, et al. Caspase 3-cleaved N-terminal fragments of wild-type and mutant huntingtin are present in normal and Huntington’s disease brains, associate with membranes, and undergo calpain-dependent proteolysis. Proc Natl Acad Sci USA 2001; 98: 12784–89.

22 Snell RG, MacMillan JC, Cheadle JP, et al. Relationship between trinucleotide repeat expansion and phenotypic variation in Huntington’s disease. Nat Genet 1993; 4: 393–97.

23 Andrew SE, Goldberg YP, Kremer B, et al. The relationship between trinucleotide (CAG) repeat length and clinical features of Huntington’s disease. Nat Genet 1993; 4: 398–403.

24 Duyao M, Ambrose C, Myers R, et al. Trinucleotide repeat length instability and age of onset in Huntington’s disease. Nat Genet 1993; 4: 387–92.

25 MacLean HE, Warne GL, Zajac JD. Spinal and bulbar muscular atrophy: androgen receptor dysfunction caused by a trinucleotide repeat expansion. J Neurol Sci 1996; 135: 149–57.

26 MacLean HE, Choi WT, Rekaris G, Warne GL, Zajac JD. Abnormal androgen receptor binding affi nity in subjects with Kennedy’s disease (spinal and bulbar muscular atrophy). J Clin Endocrinol Metab 1995; 80: 508–16.

27 Biancalana V, Serville F, Pommier J, Julien J, Hanauer A, Mandel JL. Moderate instability of the trinucleotide repeat in spino bulbar muscular atrophy. Hum Mol Genet 1992; 1: 255–58.

28 Friedman JH, Trieschmann ME, Myers RH, Fernandez HH. Monozygotic twins discordant for Huntington disease after 7 years. Arch Neurol 2005; 62: 995–97.

29 Georgiou N, Bradshaw JL, Chiu E, Tudor A, O’Gorman L, Phillips JG. Diff erential clinical and motor control function in a pair of monozygotic twins with Huntington’s disease. Mov Disord 1999; 14: 320–25.

30 Byars JA, Beglinger LJ, Moser DJ, Gonzalez-Alegre P, Nopoulos P. Substance abuse may be a risk factor for earlier onset of Huntington disease. J Neurol 2012; published online Jan 25. DOI:10.1007/s00415-012-6415-8.

31 Muller PA, Vousden KH, Norman JC. p53 and its mutants in tumor cell migration and invasion. J Cell Biol 2011; 192: 209–18.

32 Levine AJ, Oren M. The fi rst 30 years of p53: growing ever more complex. Nat Rev Cancer 2009; 9: 749–58.

33 Schuller HM. Is cancer triggered by altered signalling of nicotinic acetylcholine receptors? Nat Rev Cancer 2009; 9: 195–205.

34 Lieberman AP, Harmison G, Strand AD, Olson JM, Fischbeck KH. Altered transcriptional regulation in cells expressing the expanded polyglutamine androgen receptor. Hum Mol Genet 2002; 11: 1967–76.

35 Lazebnik Y. What are the hallmarks of cancer? Nat Rev Cancer 2010; 10: 232–33.

36 Ludvigsson JF, Andersson E, Ekbon A, et al. External review and validation of the Swedish national inpatient register. BMC Public Health 2011; 11: 450.


Articles

Global burden of cancers attributable to infections in 2008: a review and synthetic analysisCatherine de Martel, Jacques Ferlay, Silvia Franceschi, Jérôme Vignat, Freddie Bray, David Forman, Martyn Plummer

SummaryBackground Infections with certain viruses, bacteria, and parasites have been identifi ed as strong risk factors for specifi c cancers. An update of their respective contribution to the global burden of cancer is warranted.

Methods We considered infectious agents classifi ed as carcinogenic to humans by the International Agency for Research on Cancer. We calculated their population attributable fraction worldwide and in eight geographical regions, using statistics on estimated cancer incidence in 2008. When associations were very strong, calculations were based on the prevalence of infection in cancer cases rather than in the general population. Estimates of infection prevalence and relative risk were extracted from published data.

Findings Of the 12·7 million new cancer cases that occurred in 2008, the population attributable fraction (PAF) for infectious agents was 16·1%, meaning that around 2 million new cancer cases were attributable to infections. This fraction was higher in less developed countries (22·9%) than in more developed countries (7·4%), and varied from 3·3% in Australia and New Zealand to 32·7% in sub-Saharan Africa. Helicobacter pylori, hepatitis B and C viruses, and human papillomaviruses were responsible for 1·9 million cases, mainly gastric, liver, and cervix uteri cancers. In women, cervix uteri cancer accounted for about half of the infection-related burden of cancer; in men, liver and gastric cancers accounted for more than 80%. Around 30% of infection-attributable cases occur in people younger than 50 years.

Interpretation Around 2 million cancer cases each year are caused by infectious agents. Application of existing public health methods for infection prevention, such as vaccination, safer injection practice, or antimicrobial treatments, could have a substantial eff ect on the future burden of cancer worldwide.

Funding Fondation Innovations en Infectiologie (FINOVI) and the Bill & Melinda Gates Foundation (BMGF).

IntroductionInfection is recognised as a major cause of cancer worldwide. Prevention and treatment of infectious agents have already had a substantial eff ect on cancer prevention.1 A useful statistic to quantify this eff ect is the population attributable fraction (PAF), defi ned as the proportion of new cancer cases in a specifi c population that would have been prevented by a hypothetical intervention on a specifi c exposure. For infectious agents classifi ed as carcinogenic to humans,2 we calcu lated the PAF worldwide and in eight regions, using GLOBOCAN statistics on estimated cancer incidence in 2008.3 Similar calculations have been done for cancer incidence data from 19904 and 2002.5 In this report, we substantially revised the methods to reduce uncertainties and biases resulting from lack of data on population-specifi c and age-specifi c infection prevalence. We also discuss a framework for cal-culating global attributable fractions that might be applied to other causes of cancer. Some physical or chemical carcinogens act synergistically with in fec-tious agents to cause cancers; in these cases, the attributable fractions can add to more than 100%. We report the attributable fractions of infectious agents but do not report the contribution of any non-infectious cofactor.

MethodsInfectious agentsIn February, 2009, an expert working group reviewed infectious agents that have been classifi ed as carcinogenic to humans by the International Agency for Research on Cancer (IARC) Monographs programme.6 Panel 1 shows these agents and their associated cancers, namely Helicobacter pylori (H pylori), hepatitis B virus (HBV), hepatitis C virus (HCV), Opisthorchis viverrini, Clonorchis sinensis, human papillomavirus (HPV), Epstein-Barr virus (EBV), human T-cell lymphotropic virus type 1 (HTLV-1), human herpes virus type 8 (HHV-8; also known as Kaposi’s sarcoma herpes virus), and Schistosoma haematobium. Other cancer sites and infectious agents for which the evidence of carcinogenicity is weaker, or that have not been evaluated by the IARC Monographs programme, are not considered. Since HIV causes cancer through immunosuppression, thereby enhancing the carcinogenic action of other viruses, it is considered a cofactor and a PAF is not separately calculated.2 The appendix includes a short review of each infectious agent and its associated cancer sites.

Geographical areasGlobal estimates of the number of cancer cases caused by infections were calculated for the eight geographical regions shown in fi gure 1; these are based on UN



International Agency for Research on Cancer, Lyon, France (C de Martel MD, J Ferlay ME, S Franceschi MD, J Vignat MSc, F Bray PhD, D Forman PhD, M Plummer PhD)

Correspondence to:Dr Martyn Plummer, International Agency for Research on Cancer, 150 cours Albert Thomas, 69372 Lyon cedex 08, [email protected]


Articles


geographical regions and were used by the GLOBOCAN 2008 project. Additional estimates were calculated for China, India, Japan, and Australia and New Zealand because of their large population sizes within their respective regions. For some infections, calculations were restricted to the corresponding endemic areas. Countries were classifi ed by development status using UN defi nitions.7 All countries in Europe and North America (as shown in fi gure 1), as well as Australia, New Zealand, and Japan, were considered more developed; all other countries were considered less developed.

Source of cancer incidence dataEstimates of the number of new cancer cases in 2008 were obtained from the GLOBOCAN 2008 report,8 which provided age-specifi c and sex-specifi c incidence estimates for 27 cancers in 184 countries. Incidence estimates were directly available for seven of the infection-associated cancers considered in the present report (cancer of the liver, cervix, nasopharynx, bladder, Hodgkin’s lymphoma, non-Hodgkin lymphoma, and Kaposi’s sarcoma in sub-Saharan Africa). Some specifi c cancer subsites were included as part of broader categories in GLOBOCAN 2008, so incidence data were not readily available. Anal cancer (International Classi-fi cation of Diseases code C21) was included in the colorectal cancer category (C18–C21). Oropharyngeal sites including tonsils and base of tongue (C01, C09, and C10) were included in the lip, oral cavity (C00–08), and other pharynx (C09, C10, C12–14) categories. Vulval, vaginal, and penile carcinoma, and Kaposi’s sarcoma outside sub-Saharan Africa, were included in the category of other and unspecifi ed cancers. For these cancers, we estimated the number of cases by multiply ing the estimated number of cases in the broader GLOBOCAN 2008 category by the proportion of cancer subsites or subtypes from cancer registry data. These proportions were stratifi ed by region, age, and sex, and, in general, they were derived from the same registry data used for the GLOBOCAN 2008 estimates. However, when these

subtype proportions were based on micro scopically verifi ed cases (Burkitt’s lymphoma, gastric lymphoma, and adult T-cell leukaemia) or specifi c anatomic location (non-cardia gastric cancers), only cancer registries with the highest standard of information were used.9,10

Attributable-risk calculationThe number of new cancer cases attributable to each infection was calculated by multiplying incidence esti-mates by PAF. PAF is an estimate of the proportion of cases of a disease that could theoretically be avoided if exposure to a specifi c risk factor was modifi ed or removed. This estimate relies on strong causal assumptions and a simplifi ed statistical model. PAF combines the magnitude of eff ect of a risk factor with its distribution in the population. For a binary exposure, which is either present or absent, PAF can be calculated as

where p is the prevalence of exposure in the general population and R is the relative risk associated with exposure. PAF can also be calculated retrospectively using the prevalence of cases (pC ):

11

The quantity (R-1)/R is also known as the attributable risk in the exposed (ARE). As R increases, ARE increases to a limit of 1, so that all cases among the exposed are attributed to the exposure. At this limit, PAF is equal to the prevalence in cases:

We used this simplifi ed formula when mechanistic knowledge strongly suggests that the presence of infec-tion in a cancer is suffi cient to infer that infection caused the cancer. When more than one infectious agent was associated with the same cancer site, we assumed that the infections do not interact, but represent alternate pathways to cancer. Thus, PAFs for diff erent infections are additive.

Sources of infection prevalence and relative-risk dataSources of data used for PAF calculations are summarised in table 1 and discussed in the appendix. For all cancers other than cholangiocarcinomas, we used retrospective calculation of PAF based on prevalence of infection in cancer cases (formula 2 or 3) rather than prevalence of infection in the population (formula 1). Large population-based surveys of infec tion prevalence are not always available and might not be representative of the population from which the cancer cases were

Panel 1: Major cancer sites associated with group 1 infectious agents*

• Stomach: Helicobacter pylori• Liver: Hepatitis B virus, hepatitis C virus (HCV), Opisthorchis viverrini, Clonorchis sinensis• Cervix uteri: Human papillomavirus (HPV) with or without HIV• Anogenital (penile, vulva, vagina, anus): HPV with or without HIV• Nasopharynx: Epstein-Barr virus (EBV)• Oropharynx: HPV with or without tobacco or alcohol consumption• Kaposi’s sarcoma: Human herpes virus type 8 with or without HIV• Non-Hodgkin lymphoma: H pylori, EBV with or without HIV, HCV, human T-cell

lymphotropic virus type 1• Hodgkin’s lymphoma: EBV with or without HIV• Bladder: Schistosoma haematobium

*Classifi ed as carcinogenic to humans in International Agency for Research on Cancer Monograph 100B.2

=PAF (formula 1)p(R–1)

1+p(R–1)

=PAF (formula 2)(R–1)

Rpc

=PAF (formula 3)pc

Articles


derived—they can often be focused on young and healthy subgroups such as blood donors, pregnant women, or military recruits. Population prevalence surveys might also be done for diff erent age groups over diff erent time periods, leading to diffi culty calculating PAF for infections that show strong secular trends in prevalence such as H pylori, or that generally accumulate with age like HCV. Conversely, infection prevalence in cancer cases is usually straightforward to measure using molecular methods and can be assumed to directly represent the current population at risk.

For some infections (H pylori, HPV at anogenital sites, HHV-8, and HTLV-1), a global prevalence estimate was used because the literature showed no evidence of regional variations for prevalence of the infection in cases. For cancers that showed strong evidence of heterogeneity across countries, regional estimates of infection prevalence were calculated using one of two methods: geographical pooling or risk-based pooling. For geographical pooling, a weighted average of infection prevalence estimates for all countries in a region was calculated, where the weights were the product of the sample size of available studies of prevalence in cases and the number of incident cancer cases in each country given by GLOBOCAN 2008. This weighting scheme favours prevalence estimates from countries that make a larger contribution to the global cancer burden, while respecting the fact that larger prevalence surveys provide more statistical infor-mation. In geographical pooling, the underlying assumption is that countries within the same region have the same infection prevalence. For HBV and HCV in liver cancer, where geographical coverage of infection prevalence data was insuffi cient for some regions, risk-based pooling was used. The underlying assumption of risk-based pooling is that countries with

similar incidence rates for a specifi c infection-associated cancer have similar infection prevalence. Countries were stratifi ed by development status and a logistic regression model was fi tted with infection prevalence as outcome and cancer incidence as predictor. Fitted values from this model were used to impute country-specifi c prevalence estimates for all countries, and these estimates were pooled by region and weighted by the number of incident cancers, as with geographical pooling.

Relative risk estimates for infection were extracted from studies reviewed in IARC Monograph 100B2 or in formal meta-analyses. The association between an infection and its related cancer was assumed to be constant worldwide, and so a single relative risk esti-mate was used in the calculation of each PAF. SAS version 9.2 was used to compile the data and create the descriptive tables. R version 2.14.0 was used for specifi c calculations, such as risk-based pooling, and to create the fi gures. In the tables, estimated numbers of cases have been rounded to two signifi cant fi gures to avoid spurious precision. In some cases, this creates small discrepancies with the displayed totals and percentages, which are based on the data before rounding.

Role of the funding sourceThe sponsors had no role in the study design, collection, analysis and interpretation of data, or writing of the report. CDM and MP had full access to all the data in the study and fi nal responsibility for the decision to submit for publication.

ResultsTable 2 shows the estimated number of cancer cases attributed to infection in 2008, in less developed and more developed regions. Of the estimated 12·7 million

Sub-Saharan AfricaNorth Africa and west AsiaCentral AsiaEast AsiaSouth AmericaNorth AmericaEuropeOceania

Figure 1: Regions used to derive global estimates of the number of cancer cases caused by infection

Articles


new cancers worldwide, around 2 million were attributable to infec tions, of which 1·6 million (80%) occurred in less developed regions. HBV, HCV, HPV, and H pylori were together responsible for 1·9 million cases worldwide. Figure 2 shows the contribution of these infectious agents to cancer burden in less developed and more developed regions.

Table 3 shows a breakdown by geographical region of the number of new cancer cases and the number attributable to infection. Overall, 16·1% of cancer cases in 2008 were attributable to infection. The proportion was higher in less developed countries (22·9%) than in more developed countries (7·4%). The

attributable fraction varied greatly between regions, from 3·3% in Australia and New Zealand to 32·7% in sub-Saharan Africa.

Table 4 shows a more detailed breakdown of attrib-utable cancer cases, according to sex, age group, and development status of the country. Cervical cancer accounted for half of the attributable cases in women. The burden of gastric cancer and liver cancer was much higher in men than in women. The total number of cases attributable to infection was much the same in men and women. This similarity between sexes was noted across age groups, apart from in individuals younger than 40 years, where women had a higher burden of

Relative risk (RR) estimate Prevalence of infectious agent in cancer cases

Types of studies used for RR estimation

Laboratory method used for measurement of exposure

RR Geographical area Prevalence in cases (%)

Comments and strength of data

Helicobacter pylori

Non-cardia gastric cancer† (C16.1–16.9)

Cohort with >10 years follow-up

ELISA in serum 5·9 World 90% Data based on a pooled analysis of major prospective studies, all using ELISA serology.12 Strong data

Non-Hodgkin lymphoma of gastric location† (MALT and DLBC) (C82–85, C96)

Cohort and case–control

ELISA in serum 7·2 World 86% Consensus that nearly 100% of gastric MALT lymphomas are H pylori-related. Strong dataNo consensus on DLBC. Sparse data

Hepatitis viruses

Liver cancer (C22) Case–control (>10 cases)

HBV: HBsAg in serumHCV: ELISA in serum (fi rst generation excluded)

2317

Sub-Saharan AfricaNorth Africa and west AsiaSouth-central AsiaIndiaEast AsiaChinaJapanSouth AmericaNorth AmericaEuropeOceania

84%82%80%79%86%86%87%82%42%48%44%

For countries with no data, prevalence in cases imputed by liver cancer incidence and more developed or less developed statusData based on two large meta-analyses13 and original data.14 Strong data

Non-Hodgkin lymphomas (C82–85, C96)


ELISA in serum (fi rst generation excluded)

2·5 Southern EuropeJapan and KoreaOther more developed regionsLess developed regions

18%5%

10%17%

Data based on one meta-analysis.15 Limited data

Human papillomavirus (HPV; high-risk types)

Cervix uteri carcinoma (C53)

Case–control PCR in tumour tissue or cells

>100 World 100% High-risk HPV types are considered a necessary cause of cervical cancer. Strong data

Penile carcinoma† (C60) Case–control PCR in tumour tissue NR‡ World 50% Assumption is that detection of high-risk HPV DNA in tumour tissue signifi es cancer attributable to HPVData based on one meta-analysis.16 Limited data

Anal carcinoma† (C21) Case–control PCR in tumour tissue NR‡ World 88% Same assumption as for penile cancerData based on one meta-analysis.17 Strong data

Vulvar carcinoma† (C51) Case–control PCR in tumour tissue NR‡ World 43% Same assumption as for penile cancerData based on one meta-analysis.17 Limited data

Vaginal carcinoma† (C52) Case–control PCR in tumour tissue NR‡ World 70% Same assumption as for penile cancerData based on one meta-analysis.17 Limited data

Oropharynx† including tonsils and base of tongue (C01, C09–C10)

Case–control PCR in tumour tissue with HPV E6 or E7 expression

NR‡ North AmericaNorthern and western EuropeEastern EuropeSouthern EuropeAustraliaJapanRest of world

56%39%38%17%45%52%13%

Few prevalence studies available for less developed regionsSame assumption as for penile cancer, except for the diffi culty separating strong eff ect of tobacco and alcohol. Limited data

(Continues on next page)

Articles


infection-related cancer, on account of cervical cancer, in less developed and more developed regions (fi gure 3).

DiscussionThe analysis described in this report and in the appendix shows that infection is an important contributor to the global cancer burden, with 16·1% of cancers diagnosed in 2008 being attributable to infections, although the contribution due to infection varies widely from region to region. The estimated burden of cancer in 2008 attributable to infec tions is an update of previous

estimates for 20025 and 1990.4 Our estimates for 2008 are slightly lower than those for 2002, for global burden of cancer (16·1% vs 17·8%) and by development status (7·4% vs 7·7% for more developed regions; 22·9% vs 26·5% for less developed regions). Overall, the results are remarkably similar, in view of the change in methodology to incorporate retrospective calculation of PAF based on prevalence of infection in cancer cases. For the four main infections altogether, the relative contribution of HPV to cancer burden is similar in more developed and less developed areas. The contribution of

Relative risk (RR) estimate Prevalence of infectious agent in cancer cases

Types of studies used for RR estimation

Laboratory method used for measurement of exposure

RR Geographical area Prevalence in cases (%)

Comments and strength of data

(Continued from previous page)

Epstein-Barr virus (EBV)

Hodgkin’s lymphoma (C81)

Cohort and case–control (>50 cases)

Detection of EBV gene products in tumour cells

>10 More developed regionsLess developed regions

40%Children:

90%Adults:

60%

Prevalence varies by age, area, and histological subtypeShape of the age distribution curve varies by area and study period. Limited data

Burkitt’s lymphoma† (C83.5)

Case–control and case series

EBV DNA in tumour cells

Sub-Saharan Africa: >20USA and Europe: NR‡Other regions: NR‡

Sub-Saharan AfricaUSA and EuropeOther regions

100%20%30%

Sub-Saharan Africa and New Guinea are considered endemic areas for Burkitt’s lymphoma in children (peak 4–7 years of age). Strong dataLimited data for USA and EuropeSparse data for other regions

Nasopharyngeal carcinoma (C11)


EBV DNA in tumour cells

>20 High-incidence and intermediate-incidence regionsLow-incidence regions (see appendix for details)

100%

80%

High-incidence and intermediate-incidence: more than 95% of carcinomas are undiff erentiated (type II), and nearly all are EBV-related. Strong dataLow-incidence: only one cohort study and no case–control studies are available. The distribution between diff erent types of carcinoma (types I, II, III) and the fraction attributable to EBV varies between countries and is largely unknown. Sparse data

Other non-Hodgkin lymphomas (C82–85, C96)

·· ·· ·· ·· .. Due to the heterogeneity of this group, estimation of RR and prevalence is not possible from published dataData suggest that most non-Hodgkin lymphomas arising in people with HIV infection are causally related to EBV

Human herpes virus type 8 (HHV-8)

Kaposi’s sarcoma† (C46) Cohort and case–control

HHV-8 DNA in tumour tissue

100 World 100% Mostly (but not exclusively) seen in HIV-infected populations, particularly in Africa. Strong data

Human T-cell lymphotropic virus type 1 (HTLV-1)

Adult T-cell leukaemia and lymphoma (ATL) (M9827, M9823)


·· NR‡ World 100% HTLV-1 seropositivity is required for diagnosis of ATL (HTLV-1 is a necessary cause). Strong data

Liver fl ukes (Opisthorchis viverrini, Clonorchis sinensis)

Cholangiocarcinoma (C22.1)

Case–control (in endemic areas)

Various 7·7 Endemic areas in southeast Asia NR‡ Estimates calculated prospectively using the same method as Parkin (2006).5 Limited data

Schistosoma haematobium

Bladder carcinoma (C67) Case–control (in endemic areas)

Various NR‡ Sub-Saharan AfricaEgypt, Sudan, and Yemen

41%42%

Assumption is that in endemic areas, all squamous-cell carcinomas could be attributed to S haematobium. Limited data

MALT=mucosa-associated lymphoid tissue. DLBC=diff use large B-cell. HBV=hepatitis B virus. HBsAG=hepatitis B surface antigen. HCV=hepatitis C virus. NR=not relevant. *International Classifi cation of Diseases (ICD)-10 code. †These subsites were not directly available in GLOBOCAN 2008; therefore, data from CI5-IX database were used to estimate corresponding incidence rates. ‡RRs were not available or not used in the calculation of PAF for these sites.

Table 1: Methods for calculation of the population attributable fraction (PAF), by infectious agent and cancer site or subsite*

For more on the CI5-IX database see http://ci5.iarc.fr

Articles


H pylori is, however, proportionally larger in more developed countries, and that of HCV and HBV is larger in less developed countries.

Three recent studies have reported country-specifi c estimates of PAFs for infection-related cancer in the UK,18 South Korea,19 and China.20 PAF estimates for China and South Korea were 25·9% and 21·2% respectively, in line with our estimates of 26·1% for China and 22·5% for east Asia. The estimate for the UK was 3·1%, which is lower than our regional estimate of 7·0% for Europe. We found similar estimates of PAF for North America (4·0%) and

Australia and New Zealand (3·3%), areas where infection prevalence is similar to the UK.

The main strengths of our approach are the use of the highest quality epidemiological evidence and incidence data available. The choice of infectious agents and cancer sites was taken from a review by an expert IARC working group.2 We chose only agents classifi ed as carcinogenic to humans by this group, and only cancer sites with suffi cient evidence of association with infection (see appendix). Including more cancer sites—eg EBV in gastric cancer, HBV in non-Hodgkin lymphoma, or HPV in oral cavity—is a possible approach, but it would be more subjective and potentially misleading, since the strength of the published evidence is controversial. The high threshold of evidence we used might have prevented us from addressing cancers for which evidence of an infectious link is rapidly emerging. Nevertheless, we aimed to not exaggerate the importance of infections in cancer. Estimates of relative risks and infection prevalence were always derived from the same review, or from systematic or comprehensive reviews that we updated when necessary. Cancer incidence data were derived from GLOBOCAN 2008, or calculated using a con sistent method when GLOBOCAN estimates were not available.

For most PAF calculations, we used estimates of infection prevalence based on case series rather than

Number of new cases in 2008

Number attributable to infection

PAF (%)

Africa

Sub-Saharan Africa 550 000 180 000 32·7%

North Africa and west Asia 390 000 49 000 12·7%

Asia

India 950 000 200 000 20·8%

Other central Asia 470 000 81 000 17·0%

China 2 800 000 740 000 26·1%

Japan 620 000 120 000 19·2%

Other east Asia 1 000 000 230 000 22·5%

America

South America† 910 000 150 000 17·0%

North America 1 600 000 63 000 4·0%

Europe 3 200 000 220 000 7·0%

Oceania

Australia and New Zealand 130 000 4200 3·3%

Other Oceania 8800 1600 18·2%

More developed regions‡ 5 600 000 410 000 7·4%

Less developed regions§ 7 100 000 1 600 000 22·9%

World 12 700 000 2 000 000 16·1%

PAF=population attributable fraction. *Numbers are rounded to two signifi cant digits. †Includes Mexico. ‡Total for Japan, North America, Europe, and Australia and New Zealand. §Total for all other regions.

Table 3: Number of new cancer cases* in 2008 attributable to infectious agents, by geographical region

Hepatitis B/C virusHuman papilloma virusHelicobacter pyloriOther infectious agents

2000

1500

1000

500

0

Num

ber o

f cas

es a

ttrib

utab

le to

infe

ctio

n (t

hous

ands

)

Less developed regions More developed regionsDevelopment status

Figure 2: Number of new cancer cases in 2008 attributable to infection, by infectious agent and development status

Less developed regions

More developed regions

World

Hepatitis B and C viruses 520 000 (32·0%) 80 000 (19·4%) 600 000 (29·5%)

Human papillomavirus 490 000 (30·2%) 120 000 (29·2%) 610 000 (30·0%)

Helicobacter pylori 470 000 (28·9%) 190 000 (46·2%) 660 000 (32·5%)

Epstein-Barr virus 96 000 (5·9%) 16 000 (3·9%) 110 000 (5·4%)

Human herpes virus type 8 39 000 (2·4%) 4100 (1·0%) 43 000 (2·1%)

Human T-cell lymphotropic virus type 1 660 (0·0%) 1500 (0·4%) 2100 (0·1%)

Opisthorchis viverrini and Clonorchis sinensis 2000 (0·1%) 0 (0·0%) 2000 (0·1%)

Schistosoma haematobium 6000 (0·4%) 0 (0·0%) 6000 (0·3%)

Total 1 600 000 (100·0%) 410 000 (100·0%) 2 000 000 (100·0%)

Data are number of new cancer cases attributed to a particular infectious agent (proportion of the total number of new cases attributed to infection that is attributable to a specifi c agent). *Numbers are rounded to two signifi cant digits.

Table 2: Number of new cancer cases* in 2008 attributable to infection, by infectious agent and development status

Articles


prevalence in the general population. This choice was made for two reasons. First, there are few large, high-quality, population-based surveys of infection prevalence that are representative of the general population. Such surveys tend to oversample young people or those at low risk for specifi c infections. Undertaking high-quality population-based surveys is a long and diffi cult process and has not been done at all in many countries, particularly in less developed regions. Cancer case series, which are available in less developed countries, are usually representative of the population served by the hospital. Patients with cancer need expert care, and the severity of the disease makes it very likely that these patients will seek appropriate treatment in specialised centres. Second, population-based surveys often use less sensitive or specifi c measurement methods than case series, because the best testing methods are often expensive, invasive, and less feasible on a large scale. For example, causation is diffi cult to determine from serology in cases of EBV-related or HPV-related cancers; more than 90% of the population is positive for EBV, and HPV serology is not site-specifi c. However, patients included in case series undergo many tests, often including direct detection of infectious agents and even gene expression in tumour tissue. Therefore, we are confi dent that calculating prevalence from case series increases the validity of PAF estimates.

The need to avoid the eff ect of time trends in infection prevalence when selecting the case series on which we based our calculations was also considered, but did not seem to be an issue, at least for the four main infections that drive most of the global PAF. Although improved living conditions have led to a steadily decreasing prevalence of H pylori infection in many populations, prevalence in gastric cancer cases from nested case–control studies or other epidemiological studies seems to be very stable, around 90%, with no detectable secular trends. The case series we selected for estimating the prevalence of hepatocellular carcinoma attributable to HBV and HCV used only second-generation or third-generation ELISA for detection of HCV, with a total of 37 000 cases from 132 studies published from 1992 to 2009. The low sensitivity of fi rst-generation anti-HCV ELISA has long been recognised and seems to diff er between cases and controls.13 Most studies of HPV-related cancers have been done in the past 15 years using DNA detection techniques, and relevant studies of oropharyngeal cancers are even more recent.

Our approach has several limitations. First, some uncertainty in cancer incidence and infection prevalence is inherent in our estimations of PAF. Our attempt to obtain global estimates of infection prevalence by pooling local data sources inevitably requires extrapolation to countries with sparse data on cancer incidence or

Number of new cases in 2008

Number attributable to infection

PAF (%) Number attributable to infection, by sex

Number attributable to infection, by age group

Number attributable to infection, by development status

Male Female <50 years 50–69 years ≥70 years Less developed regions

More developed regions

Carcinoma

Non-cardia gastric 870 000 650 000 74·7% 410 000 240 000 82 000 290 000 270 000 470 000 180 000

Liver† 750 000 580 000 76·9% 400 000 170 000 130 000 280 000 180 000 510 000 69 000

Cervix uteri 530 000 530 000 100·0% 0 530 000 250 000 220 000 59 000 450 000 77 000

Vulva 27 000 12 000 43·0% 0 12 000 1700 3900 6000 4100 7500

Anus 27 000 24 000 88·0% 11 000 13 000 5100 10 000 8300 12 000 12 000

Penis 22 000 11 000 50·0% 11 000 0 2500 4800 3500 7600 3200

Vagina 13 000 9000 70·0% 0 9000 2000 4000 3100 5700 3400

Oropharynx 85 000 22 000 25·6% 17 000 4400 4300 13 000 4600 6400 15 000

Nasopharynx 84 000 72 000 85·5% 49 000 23 000 31 000 32 000 9200 66 000 5900

Bladder 260 000 6000 2·3% 4600 1400 1200 3400 1400 6000 0

Lymphoma and leukaemia

Hodgkin’s 68 000 33 000 49·1% 20 000 13 000 23 000 6700 3500 23 000 10 000

Non-Hodgkin gastric 18 000 13 000 74·1% 7400 5800 3900 5000 4400 6500 6700

Burkitt 11 000 6800 62·5% 4000 2800 6300 290 210 6300 530

HCV-associated non-Hodgkin 360 000 29 000 8·2% 17 000 13 000 9500 11 000 8800 18 000 11 000

Adult T-cell 2100 2100 100·0% 1200 900 580 980 580 660 1500

Kaposi’s sarcoma 43 000 43 000 100·0% 29 000 14 000 30 000 7600 4700 39 000 4100

Total infectious-disease-related sites 3 200 000 2 000 000 64·4% 990 000 1 100 000 580 000 890 000 560 000 1 600 000 410 000

PAF=population attributable fraction. HCV=hepatitis C virus. *Numbers are rounded to two signifi cant digits. †Including cholangiocarcinoma.

Table 4: Number of new cancer cases* in 2008 attributable to infectious agents, by anatomic site

Articles


risk-factor prevalence. The main risk of this extrapolation is that estimates derived from a small amount of data are applied to a larger population that is substantially

diff erent, with possible amplifi cation of bias. The weighting scheme we used to estimate regional infection prevalence might give undue weight to larger surveys done in smaller countries.

A second limitation is that strong assumptions were required for the calculations. For example, we assumed that relative risks for infection were constant across populations and sexes—a common assumption in epidemiology. Generally, this assumption is not true when comparing populations or sexes with widely diff erent baseline cancer inci dence rates, since the multiplicative assumption behind relative risk estimation is only an approximation. Our PAF calculations were not strongly dependent on the constant relative risk assumption because we used the retrospective formula based on cancer cases, and the relative risks were uniformly large. A change in relative risk from 10 to 20, for example, makes only a 5% diff erence in the PAF estimate. For H pylori, HPV, HBV, and HCV, the order of magnitude of relative risk is generally constant worldwide when other known risk factors have been controlled for, so the assumption of a constant relative risk should not lead to substantial error.

A third limitation is the lack of high-quality epi-demiological data for some of the cancer sites in this study (eg, EBV-related cancers, such as nasopharyngeal carcinoma), in areas of low cancer incidence. We based our estimates on the most recent and least subjective evidence, but the lack of data inevitably leads to uncertainty in the estimates.

Some of the assumptions used in our calculations were conservative. We restricted the eff ect of HPV in head and neck cancer to the oropharynx and base of the tongue, where the epidemiological and mechanistic evidence for a causal eff ect is strongest. HPV might be associated with other head and neck cancers, but this is impossible to quantify with current data. For H pylori, we based PAF calculations on a relative risk of 5·9; although this is the best estimate available, evidence from prospective studies and studies using western blot rather than ELISA suggests that it might be higher. Such studies yielded relative risks greater than 10,21–25 which would increase the proportion of non-cardia gastric cancer attributable to H pylori from 75% to 90%. Likewise, we estimated that 75% of diff use large B-cell lymphoma of gastric location is due to H pylori, but the attributable fraction could be nearer to 100%. The discovery of new associations between infections, particularly viruses, and cancer has been anticipated; however, studies have either disclosed associations with very rare cancers (eg, Merkel-cell carcinoma) or are yet to provide conclusive results (eg, for cutaneous HPV types and non-melanomatous skin cancer). Nevertheless, undiscovered associations could exist, which is another reason to conclude that our results probably underestimate the true burden of infection-associated cancers.

Attributable-risk calculations can be done for any environmental exposure, but are most useful, from a

Panel 2: Research in context

Systematic reviewThe relation between infectious agents and cancer was the subject of a comprehensive literature review done as part of the IARC Monographs programme. Global cancer incidence and mortality data were synthesised by the GLOBOCAN project to provide estimates of the global burden of cancer. We used these data sources to estimate the global burden of cancer due to infection, relying on existing systematic reviews of the literature to provide the quantitative inputs (relative risk and infection prevalence) required for calculation of attributable fractions. Previous global estimates of the proportion of cancers attributable to infection were done for 19904 and for 2002.5 Country-specifi c estimates have been provided for the UK,18 South Korea,19 and China.20

InterpretationThe present review extends previous fi ndings by showing wide geographical variation in the fraction of cancers attributable to infection. It also underscores the importance of HPV, Helicobacter pylori, HBV, and HCV as cancer-related infectious agents. Since infections are an important and preventable cause of cancer worldwide, clinicians should support the implementation of available strategies for prevention—ie, vaccination against HBV and HPV, use of safe injection practices, and avoidance of parenteral treatment when oral treatment is available. Clinicians should also closely follow and, if possible, contribute to progress in areas where early detection of infection (eg, HPV) or treatment (eg, HCV and H pylori) could diminish cancer sequelae. Public health doctors and cancer-control specialists should appreciate the importance of infectious causes of cancer in diff erent regions and age groups, particularly in low-income and middle-income populations. The 2011 UN high-level meeting on non-communicable diseases highlighted the growing global agenda for prevention and control of non-communicable diseases. Although cancer is considered a major non-communicable disease, a sizable proportion of its causation is infectious and simple non-communicable disease paradigms will not be suffi cient.

<40 40−44 45−49 50−54 55−59 60−64 65−69 70+

Attr

ibut

able

frac

tion

(%)

30

20

15

10

5

0

35

25

Age group

Women in less developed regionsMen in less developed regionsWomen in more developed regionsMen in more developed regions

Figure 3: Relative percentage of new cancer cases in 2008 attributable to infection, by sex, age group, and development status

Articles


public health perspective, when relative risks are large and when interventions to reduce population exposure are feasible. Many infection-related cancers are preventable (panel 2), particularly those associated with HPV, H pylori, HBV, and HCV. Prophylactic vaccines have shown nearly 100% effi cacy in preventing precancerous lesions of the cervix due to HPV types 16 and 18, among previously uninfected individuals. In Taiwan, the incidence of hepatocellular carcinoma in children and adolescents has been sub stantially reduced by a combination of immunoglobulin given at birth to prevent vertical transmission from mother to child at birth and childhood HBV vaccination.1 The current WHO recommendation is to vaccinate all infants against HBV as soon as possible after birth.26 Although no vaccine is available for HCV, iatrogenic transmission can be avoided with safer practices for injection and blood transfusion, and preference for oral drug delivery over injections where available. H pylori is a treatable infection, although the feasibility, eff ectiveness, and safety of large-scale eradi cation of H pylori infection in diff erent age groups is not yet clear. Our fi nding that H pylori accounts for 46% of infection-associated cancers in more developed areas might refl ect lower investment in research on prevention of gastric cancer compared with cervical and liver cancers. Such considerations underscore the diff erence between what is theoretically preventable, according to the assumptions of the PAF calculations, and what is preventable in practice. The importance of time must also be acknowledged; preventing infection-associated cancers in 2008 would have required intervention programmes many decades earlier.

In view of the high mortality rate of infection-associated cancers, the fraction of cancer deaths attributable to infections is probably higher than the 16·1% that our study generated. Although a full investigation of cancer death due to infection is beyond the scope of this report, we can estimate the mortality burden by applying the PAFs to the 7·5 million cancer deaths that occurred in 2008. These calculations suggest that 1·5 million cancer deaths were attributable to infectious agents, or roughly one in fi ve deaths due to cancer worldwide.

ContributorsSF, DF, MP, and CDM conceived and designed the study. JF provided

cancer incidence estimates adapted from the GLOBOCAN 2008

database. MP, JV, and FB contributed to data collection and data

analysis. CDM and MP wrote the manuscript. All authors contributed to

the interpretation of data and approved the fi nal manuscript.


AcknowledgmentsCDM was funded by Fondation Innovations en Infectiologie (FINOVI)

and the Bill & Melinda Gates Foundation (BMGF), USA (grant number

35537). We thank Veronique Chabanis for technical assistance.

References1 Chang MH, You SL, Chen CJ, et al. Decreased incidence of

hepatocellular carcinoma in hepatitis B vaccines: a 20-year follow-up study. J Natl Cancer Inst 2009; 101: 1348–55.

2 IARC. Monographs on the evaluation of carcinogenic risks to humans, volume 100. A review of carcinogen—Part B: biological agents. Lyon: International Agency for Research on Cancer, 2011.


4 Pisani P, Parkin DM, Muñoz N, Ferlay J. Cancer and infection: estimates of the attributable fraction in 1990. Cancer Epidemiol Biomarkers Prev 1997; 6: 387–400.


6 Bouvard V, Baan R, Straif K, et al. A review of human carcinogens–Part B: biological agents. Lancet Oncol 2009; 10: 321–22.

7 UN. World population prospects: the 2008 revision highlights. New York; United Nations, 2009.

8 Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. GLOBOCAN 2008 v1.2, cancer incidence and mortality worldwide: IARC CancerBase 10. Lyon: International Agency for Research on Cancer, 2010.

9 Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB. Cancer incidence in fi ve continents, vol VIII. IARC scientifi c publications 155. Lyon: International Agency for Research on Cancer, 2002.

10 Curado MP, Edwards B, Shin HR, et al. Cancer incidence in fi ve continents, vol IX. IARC scientifi c publications 160. Lyon: International Agency for Research on Cancer, 2007.

11 Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. Am J Epidemiol 1974; 99: 325–32.

12 Helicobacter and Cancer Collaborative Group. Gastric cancer and Helicobacter pylori: a combined analysis of 12 case control studies nested within prospective cohorts. Gut 2001; 49: 347–53.

13 Donato F, Boff etta P, Puoti M. A meta-analysis of epidemiological studies on the combined eff ect of hepatitis B and C virus infections in causing hepatocellular carcinoma. Int J Cancer 1998; 75: 347–54.

14 Dondog B, Lise M, Dondov O, Baldandorj B, Franceschi S. Hepatitis B and C virus infections in hepatocellular carcinoma and cirrhosis in Mongolia. Eur J Cancer Prev 2011; 20: 33–39.

15 Dal Maso L, Franceschi S. Hepatitis C virus and risk of lymphoma and other lymphoid neoplasms: a meta-analysis of epidemiologic studies. Cancer Epidemiol Biomarkers Prev 2006; 15: 2078–85.

16 Backes DM, Kurman RJ, Pimenta JM, Smith JS. Systematic review of human papillomavirus prevalence in invasive penile cancer. Cancer Causes Control 2009; 20: 449–57.

17 De Vuyst H, Cliff ord GM, Nascimento MC, Madeleine MM, Franceschi S. Prevalence and type distribution of human papillomavirus in carcinoma and intraepithelial neoplasia of the vulva, vagina and anus: a meta-analysis. Int J Cancer 2009; 124: 1626–36.

18 Parkin DM. Cancers attributable to infections in the UK in 2010. Br J Cancer 2011; 105 (suppl 2): 49–56.

19 Shin A, Park S, Shin HR, et al. Population attributable fraction of infection-related cancers in Korea. Ann Oncol 2011; 22: 1435–42.

20 Xiang W, Shi JF, Li P, et al. Estimation of cancer cases and deaths attributable to infection in China. Cancer Causes Control 2011; 22: 1153–61.

21 Mitchell H, English DR, Elliott F, et al. Immunoblotting using multiple antigens is essential to demonstrate the true risk of Helicobacter pylori infection for gastric cancer. Aliment Pharmacol Ther 2008; 28: 903–10.

22 Ekstrom AM, Held M, Hansson LE, Engstrand L, Nyren O. Helicobacter pylori in gastric cancer established by CagA immunoblot as a marker of past infection. Gastroenterology 2001; 121: 784–91.

23 Brenner H, Arndt V, Stegmaier C, Ziegler H, Rothenbacher D. Is Helicobacter pylori infection a necessary condition for noncardia gastric cancer? Am J Epidemiol 2004; 159: 252–58.

24 Herrera V, Parsonnet J. Helicobacter pylori and gastric adenocarcinoma. Clin Microbiol Infect 2009; 15: 971–76.

25 Siman JH, Engstrand L, Berglund G, Forsgren A, Floren CH. Helicobacter pylori and CagA seropositivity and its association with gastric and oesophageal carcinoma. Scand J Gastroenterol 2007; 42: 933–40.

26 WHO. Weekly epidemiological report: hepatitis B vaccines. Geneva: World Health Organization, 2010.


Review


Department of Radiation Oncology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA (Z S Zumsteg MD, Prof M J Zelefsky MD)

Correspondence to:Prof Michael J Zelefsky, Department of Radiation Oncology, Memorial Sloan-Kettering Cancer Center, New York, NY 10065, [email protected]

Short-term androgen deprivation therapy for patients with intermediate-risk prostate cancer undergoing dose-escalated radiotherapy: the standard of care?Zachary S Zumsteg, Michael J Zelefsky

What is the best way to manage patients with intermediate-risk prostate cancer? One of the most controversial aspects of treatment is the role of short-term androgen deprivation therapy in combination with defi nitive radiotherapy. In two randomised trials of patients with mostly intermediate-risk prostate cancer, increased overall survival was reported when short-term androgen deprivation therapy was added to radiotherapy. However, radiation doses in these studies were far below the current standard of care. This limitation, in combination with the heterogeneous nature of the cancers classifi ed as intermediate risk, has complicated the application of these trial results to modern clinical practice. In this Review, we discuss clinical evidence for and against use of short-term androgen deprivation therapy with dose-escalated radiotherapy for patients with intermediate-risk prostate cancer.

IntroductionThe paradox of evidence-based medicine in a disease with a long natural history is that by the time a large phase 3 randomised controlled trial is designed, com-pletes accrual, and has adequate follow-up to provide meaningful results, the diagnostic and therapeutic interventions implemented have typically been replaced by newer techniques or novel modalities. Therefore, the control arm in the trial might not necessarily refl ect the current standard of care. Although trial fi ndings still provide very valuable insights, the latency between design and dissemination of results creates complexity when trying to apply outcomes to modern medical practice. This conundrum is especially evident for localised prostate cancer, an often indolent disease for which rapid advances in radiation technology and surgical technique have outpaced the ability of prospective trials to assess them.

An example of this diffi culty is evaluation of short-term androgen deprivation therapy in combination with radio therapy for patients with intermediate-risk prostate cancer. Findings of two randomised controlled trials showed an overall survival benefi t in this population with use of 4–6 months of androgen deprivation therapy in combination with defi nitive radiotherapy.1,2 However, the design of these trials incorporated radiation doses and techniques that would be judged inferior by modern standards, and adverse sequelae were associated with androgen deprivation therapy. Thus, some researchers suggest the need for androgen deprivation therapy could be obviated by implementation of dose-escalated radiation delivered via image-guided intensity-modulated radiotherapy or brachytherapy, either alone or in combin-ation with external beam radiotherapy. Notwithstand ing these considerations, use of androgen deprivation therapy with radiotherapy for intermediate-risk prostate cancer has increased greatly in the USA.3,4 In one analysis,4 neoadjuvant use of androgen deprivation therapy in patients with intermediate-risk prostate cancer also undergoing radiotherapy rose from 5% in

1989 to 85% in 2002. Within this context, we analyse evidence for and against use of short-term androgen deprivation therapy combined with radiation in the era of dose escalation for patients with intermediate-risk prostate cancer.

Heterogeneity of intermediate-risk prostate cancerThe meaning of intermediate-risk prostate cancer varies somewhat depending on the defi nition used by a particular institution. The National Comprehensive Cancer Network (NCCN) risk classifi cation system defi nes intermediate-risk prostate cancer as having at least one of the following characteristics: clinical tumour stage T2b or T2c; a Gleason score of 7; or 10–20 μg/L of prostate-specifi c antigen (PSA); but without high-risk features (clinical stage T3 or higher, initial PSA >20 μg/L, or Gleason score 8–10).5 The NCCN guidelines also allow, but do not require, patients with multiple intermediate risk factors to be classifi ed as high, rather than intermediate, risk. However, other defi nitions of intermediate-risk disease have been proposed.6–8

Even within a given defi nition of intermediate-risk prostate cancer, substantial inherent heterogeneity and a wide range of clinical behaviour can be noted. For example, an 85-year-old man with clinical stage T1c prostate cancer, a Gleason score of 3+4=7 aff ecting one of 12 biopsy cores, and with PSA of 3·0 μg/L represents a diff erent clinical entity to a 45-year-old man with tumour at clinical stage T2c, a Gleason score of 4+3=7 in 12 of 12 cores, and a PSA of 19 μg/L, although both patients could be classifi ed as having intermediate-risk disease. With the Memorial Sloan-Kettering Cancer Center (MSKCC) nomogram,9 the 85-year-old man described above has an 82% probability of PSA recurrence-free survival at 10 years after dose-escalated external beam radiotherapy alone compared with 40% for the 45-year-old man. This clinical heterogeneity makes application of a uniform treatment standard diffi cult for what is a quite non-uniform category of disease.


Review

Moreover, researchers have suggested that more than 50% of variability in cure rates for prostate cancer (defi ned by absence of biochemical progression) is contributed to by clinical factors other than tumour stage, amount of PSA, and total Gleason score.10 Therefore, additional clinical factors have been proposed for incorporation into methods to predict risk, including proportion of positive biopsy cores,11 perineural invasion,12 pretreatment PSA velocity,13 and primary Gleason pattern.14 However, at this time, the best way to implement these clinical factors to guide therapeutic decision making for patients with intermediate-risk prostate cancer is unclear.

Randomised controlled trials of short-term androgen deprivation therapyFindings of major phase 3 randomised controlled trials published to date show consistently that addition of androgen deprivation therapy is benefi cial for patients with prostate cancer who are undergoing radiotherapy. Increases in biochemical progression-free survival and distant metastasis-free survival have been noted, as have reductions in prostate cancer-specifi c mortality.1,2,15–21 Moreover, fi ndings of most,1,2,15–19 but not all,20,21 studies show an overall survival benefi t with androgen deprivation therapy. However, these trials have consisted pre-dominantly of high-risk patients with a fairly high chance of occult lymph-node and distant metastases. The role of short-term (defi ned as 6 months or fewer) androgen deprivation therapy for men with lower risk prostate cancer is not as clear.

Two trials have assessed the benefi t of short-term androgen deprivation therapy in combination with radiotherapy for localised prostate cancer mostly of intermediate risk (table 1).1,2 In a trial from the Radiation Therapy Oncology Group (RTOG 94-08),1 1979 men were enrolled with clinical stage T1b–T2b prostate cancer and a PSA concentration of 20 μg/L or less. Patients were randomly allocated to receive radiotherapy alone or in combination with short-term androgen deprivation therapy, consisting of 4 months of neoadjuvant and concurrent combined androgen blockade. With a median follow-up of 9·1 years, short-term androgen deprivation therapy signifi cantly prolonged biochemical progression-free survival and distant metastasis-free survival and reduced prostate cancer-specifi c mortality, and was associated with a modest but signifi cant rise in overall survival at 10 years (62% vs 57%, p=0·03). In an unplanned subgroup analysis, improvements in overall survival and prostate cancer-specifi c mortality were only recorded in patients with intermediate-risk disease, comprising 54% of the population.1 No survival benefi t was noted in either low-risk or high-risk sub groups. These results were corroborated by data of a randomised trial from the Dana-Farber Cancer Institute that included patients mainly with intermediate-risk prostate cancer.2 In this trial, 206 men with T1b–T2b prostate cancer and either a Gleason score of at least 7, a concentration in serum of PSA greater than 10 μg/L (but not more than 40 μg/L), or MRI evidence of extra capsular extension or seminal vesicle invasion were randomly allocated to receive radiotherapy either alone or in combination with

Patients (n)

Patients at intermediate risk (n)

Median follow-up (years)

Androgen deprivation therapy comparison arms

Radiotherapy dose (Gy)*

Primary endpoint

Reported outcomes with short-term androgen deprivation therapy

Jones (2011)1 1979 1068† 9·1 0 vs 4 months 63·3 Overall survival Increased overall survival and biochemical progression-free survival, reduced prostate cancer-specifi c mortality and distant metastasis

D’Amico (2008)2 206 153† 7·6 0 vs 6 months 70·35 Biochemical progression-free survival

Prolonged overall survival and decreased prostate cancer-specifi c mortality

Denham (2011)17 818 130† 10·6 0 vs 3 vs 6 months 62·7 Prostate cancer-specifi c mortality and local control‡

Augmented overall survival and diminished prostate cancer-specifi c mortality and distant metastasis§

Roach (2008)20 456 Not reported¶ 11·9–13·2 0 vs 4 months 61·8–66·5 Local control Reductions in prostate cancer-specifi c mortality and distant metastasis, increases in biochemical progression-free survival and disease-free survival, but no improvements in overall survival or local control

Laverdière (2004)22 161 Not reported 5 0 vs 3 vs 10 months 64 Biochemical progression-free survival

Prolonged biochemical progression-free survival||

Dubray (2011)23 366 366 3·1 0 vs 4 months 80 Freedom from failure**

Increased biochemical progression-free survival, non-signifi cant rise in freedom from failure (p=0·09)

*All doses are normalised to the planning target volume to allow accurate comparisons. For doses prescribed to the centre of the prostate (ie, the isocentre), normalised radiation doses were calculated assuming a 5% dose reduction from the isocentre to the outside of the target. †D’Amico classifi cation.6 ‡Overall survival later replaced local control as a primary endpoint of this study. §Improvements recorded only with 6 months, not 3 months, of androgen deprivation therapy. ¶Probably only a few patients at intermediate risk were included; 70% had T3–4 tumours. ||No diff erence in outcome with 3 months versus 10 months of androgen deprivation therapy. **Combined clinical and biochemical failure.

Table 1: Randomised trials of radiotherapy and short-term versus no androgen deprivation therapy for localised prostate cancer


Review

short-term androgen deprivation therapy (consisting of 6 months of combined androgen blockade, divided into 2-month blocks of neoadjuvant, concurrent, and adjuvant treatment). Despite enrolment of far fewer patients in this study than in RTOG 94-08, a signifi cant increase in 8-year overall survival (74% vs 61%; p=0·01) and a reduction in prostate cancer-specifi c mortality (four vs 14 deaths from prostate cancer; p=0·007) were noted with short-term androgen deprivation therapy, at 7·6 years of follow-up.

Androgen deprivation therapy in the era of dose escalationThe results of RTOG 94-081 and the Dana-Farber trial2 indicate that short-term androgen deprivation therapy is benefi cial for patients with intermediate-risk prostate cancer also undergoing standard-dose radiotherapy. However, the radiation techniques and doses used in these trials—although consistent with the standard of care at the time the trials were designed—would be judged unacceptable by modern standards. In RTOG 94-08 a dose of 66·6 Gy was delivered to the centre of the prostate (also called the isocentre), meaning the periphery of the prostate was treated to even lower doses, and in the Dana-Farber trial, three-dimensional (3D) conformal radiotherapy was used to deliver 70·35 Gy to the planning target volume (the perimeter of the prostate with a margin).

Many technological improvements to delivery of radio-therapy have been adopted widely since these trials began accruing patients in the mid-1990s, including intensity-modulated radiotherapy, image-guided radio therapy, and brachytherapy either alone or in combination with external beam radiotherapy. Collectively, these advances allow delivery of much higher doses than were achievable

previously in a very conformal fashion, thereby more precisely treating the target while restricting toxic eff ects to healthy tissues. In fi ve clinical trials,24–28 all containing substantial numbers of patients with intermediate-risk disease, reductions in biochemical recurrence with increased radiation doses have been noted consistently, although no trial fi ndings showed improvement in prostate cancer-specifi c mortality or overall survival with dose escalation (table 2). Furthermore, fi ndings of a large, prospective, non-randomised, dose-escalation trial from MSKCC,31 in which doses as high as 86·4 Gy were used, showed that escalating radiation dose was associated with increased biochemical progression-free survival and lower rates of distant metastasis, whereas in a retrospective analysis of patients on RTOG dose-escalation protocols,32 an overall survival advantage was noted with dose escalation for patients with Gleason 8–10 prostate cancer. Therefore, some have questioned whether androgen deprivation therapy is necessary when using modern dose-escalated radiation techniques.33

Dose-escalated external beam radiotherapy and short-term androgen deprivation therapyUp to now, no mature results are available from randomised trials of short-term androgen deprivation therapy in patients with inter mediate-risk prostate cancer receiving dose-escalated external beam radiotherapy. The only randomised data we are aware of are from a preliminary analysis of the French trial GETUG 14,23 including 366 men with intermediate-risk prostate cancer treated with high-dose external beam radiotherapy (80 Gy), either alone or with 4 months of androgen deprivation therapy. Androgen deprivation therapy signifi cantly increased 3-year biochemical progression-free survival (97% vs 91%; p=0·04) but the

Patients (n)



Radiotherapy dose (Gy) comparison arms*

Androgen deprivation therapy

Primary endpoint Outcomes with dose escalation

Dearnaley (2007)24 843 264† 10 70·3 vs 60·8 3–6 months in 100%

Biochemical progression-free survival, local control, distant metastasis-free survival, overall survival, late toxic eff ects

Prolonged biochemical progression-free survival but not overall survival

Al-Mamgani (2008)25 669 182† 5·8 74·1 vs 64·6 6 months or 3 years in 21%

Freedom from failure (combined clinical and biochemical failure)

Rise in freedom from failure but not overall survival

Zietman (2010)26 393 144‡ 8·9 79·2 vs 70·2 None Biochemical progression-free survival Increased biochemical progression-free survival but not overall survival

Kuban (2008)27 301 139§ 8·7 74·1 vs 66·5 None Freedom from failure (combined clinical and biochemical failure)

Augmented freedom from failure but not overall survival; distant metastasis-free survival and prostate cancer-specifi c mortality saw non-signifi cant improvement

Beckendorf (2011)28 306 218‡ 5·1 70 vs 80 None Biochemical recurrence Decreased biochemical recurrence¶, overall survival not reported

*All doses are normalised to the planning target volume to allow more accurate comparisons. For doses prescribed to the centre of the prostate (ie, the isocentre), normalised radiation doses were calculated assuming a 5% dose reduction from the isocentre to the outside of the target. †Chism classifi cation.8 ‡D’Amico classifi cation.6 §NCCN classifi cation.5 ¶Decreased only according to ASTRO defi nition,29 not Phoenix.30

Table 2: Randomised trials of lower versus higher dose external beam radiotherapy


Review

primary endpoint of the trial (combined biochemical and local tumour control) did not diff er between groups (92% vs 86%, p=0·09). The trial was closed early because of poor accrual, and a fi nal analysis is planned in 2013. Researchers on a much larger trial, RTOG 08-15 (ClinicalTrials.gov identifi er, NCT00936390), are curr-ently enrolling patients with intermediate-risk prostate cancer who are undergoing either dose-escalated external beam radiation alone (79·2 Gy to 98% of the planning target volume) or a combination of brachytherapy and external beam radiotherapy and randomly allocating them to either 6 months of combined androgen blockade, initiated 2 months before radiotherapy, or no systemic therapy.

Conversely, the benefi t of dose escalation in a population uniformly receiving short-term androgen deprivation therapy was addressed by the Medical Research Council’s RT01 trial.24 843 men undergoing 3–6 months of androgen deprivation therapy were randomly assigned to receive either 64 Gy or 74 Gy radiotherapy to the isocentre. Dose escalation raised biochemical progression-free survival (71% vs 60% at 5 years; p=0·0007), but local progression, distant metastasis-free survival, and overall survival did not diff er between groups. The absence of overall survival benefi t with dose escalation was confi rmed in a 10-year analysis.34 Of note, the radiation doses used in RT01 are substantially lower than current standards for dose-escalated external beam radiotherapy. Assuming conservatively a 5% dose fall-off from the centre of the prostate to the perimeter, the dose-escalated group received 70·3 Gy to the target volume, which is identical to the conventional dose of external beam radiotherapy used in the Dana-Farber trial.2

Data from non-randomised studies of dose-escalated external beam radiotherapy in combination with androgen deprivation therapy are limited and confl icting (table 3).9,35–37 In an analysis of RTOG 94-06 (a phase 1–2 prospective dose-escalation trial using 3D conformal radiotherapy to deliver a mean dose of 78·5 Gy to the planning target volume),35 no benefi t was associated with androgen deprivation therapy, including in men

with intermediate-risk prostate cancer. Similar results have been published from retrospective series.36,37 However, a retrospective analysis from MSKCC,9 including 68% of patients receiving at least 81 Gy, showed signifi cantly increased 10-year PSA relapse-free survival in patients with intermediate-risk prostate cancer treated with androgen deprivation therapy. Compared with RTOG 94-06, the MSKCC study included more patients, had longer follow-up, and included men treated with doses as low as 64·8 Gy, which could account for the diff erences in outcome between these studies.

BrachytherapyBrachytherapy is a completely diff erent approach to dose-escalated radiotherapy. In this procedure, radioactive seeds with sharp fall-off s in dose are implanted directly into the prostate through the perineum. Unlike external beam radiotherapy, for which radiation must enter and exit the patient externally and traverse healthy tissue to reach the prostate, in brachytherapy, radiation is produced internally within the gland itself, allowing creation of very conformal dose distributions. Brachy therapy can be used either alone or in combination with external beam radiotherapy. By combining external and internal radiotherapy, biologically equivalent doses substantially exceeding those achievable by any other modern radiation technique can be delivered.38,39

As far as we know, the benefi t of neoadjuvant androgen deprivation therapy in combination with brachytherapy has not been studied in a randomised trial. Table 4 summarises data from retrospective series looking at the eff ects of androgen deprivation therapy and brachytherapy either alone or combined with external beam radiotherapy for men with intermediate-risk prostate cancer. Most series show no benefi t of tumour control with androgen deprivation therapy in patients undergoing brachy-therapy.36,37,40–49 Moreover, no benefi t was recorded with short-term androgen deprivation therapy and combined external beam radiotherapy and brachytherapy, possibly attributable to the high biologically equivalent doses delivered by this modality.36,42–49

Patients (n)



Number at intermediate risk receiving androgen deprivation therapy

Duration of androgen deprivation therapy (months)

Median radiotherapy dose (Gy)

Outcomes for patients at intermediate risk

Zelefsky (2011)9 2551 1074 8 456 (42%) 5 (median) 81 Prolonged biochemical progression-free survival with androgen deprivation therapy

Valicenti (2011)35 883 291 6·6–7·9 74 (25%) 2–6 ≥77·4 No benefi t in biochemical progression-free survival, disease-free survival, or overall survival with androgen deprivation therapy

Krauss (2011)36* 469 365 4·1 73 (28%) 6 (median) 75·6 No benefi t in biochemical progression-free survival, distant metastasis-free survival, or overall survival with androgen deprivation therapy

Ciezki (2004)37* 519 237 4 139 (59%) 6 (median) 78 No benefi t in biochemical progression-free survival with androgen deprivation therapy

*Only patients treated with external beam radiotherapy alone are included.

Table 3: Retrospective series combining short-term androgen deprivation therapy with dose-escalated external beam radiotherapy


Review

Thus, by contrast with external beam radiotherapy, data to support addition of short-term androgen deprivation therapy to brachytherapy are scarce at this time. Defi nitive conclusions about benefi t, or lack thereof, are diffi cult to draw from these data owing to their retrospective nature, substantial clinical imbalances in almost every series favouring patients not receiving hormonal therapy, scant standardisation of duration and timing of androgen deprivation therapy, and fairly short follow-up. Prospective randomised trials are needed to defi ne more precisely the therapeutic role of androgen deprivation therapy with brachytherapy (eg, NCT00936390 [RTOG 08-15]).

Interactions between short-term androgen deprivation therapy and radiotherapyPart of the uncertainty about the benefi t of short-term androgen deprivation therapy in combination with dose-escalated radiotherapy refl ects our incomplete understanding of the mechanisms by which these two modalities interact. A major mechanism postulated to underlie the benefi t of short-term androgen deprivation therapy during radiotherapy is increased local control via synergistic cytotoxic eff ects from radiosensitisation of prostate cancer cells.50 The importance of this local interaction between treatments is highlighted by the contrast in fi ndings between randomised trials of

hormonal therapy preceding radical prostatectomy, all failing to show even a biochemical recurrence benefi t,51,52 and trials of androgen deprivation therapy and radiation, in which improved outcomes have been noted.1,2,17,20 Furthermore, fi ndings of randomised trials indicate that improved local control leads to prolonged overall survival in prostate cancer.53–55

Understanding the local eff ects of short-term androgen deprivation therapy in combination with dose-escalated radiation, thus, is crucial. Unfortunately, ascertainment of the true incidence of local recurrence after radiotherapy in prostate cancer is diffi cult, because clinical examination and post-radiotherapy imaging both have low sensitivity and specifi city.56 Therefore, to ascertain local failure rates more accurately, some researchers have assessed biopsy samples taken about 2 years after completion of radiotherapy.1,38,57–61 Post-treatment biopsy data from prospective studies show that standard-dose radiotherapy (64·8–70·2 Gy) provides fairly poor local control, with persistent prostate cancer seen in 28–65% of biopsy samples 2 years after treatment.1,56–58,61 On the other hand, strikingly fewer positive post-treatment biopsy samples are noted with neoadjuvant and concurrent androgen deprivation therapy.1,57,59 For example, fi ndings of a small randomised trial from Canada, in which patients received radiation consisting of 64·8 Gy to the isocentre, showed

Patients (n)



Proportion receiving androgen deprivation therapy

Radiotherapy technique Outcomes with androgen deprivation therapy

Ciezki (2004)37* 386 91 4 64% Low-dose rate No benefi t in biochemical progression-free survival

Lee (2002)40 201 66 3·5 66% Low-dose rate Prolonged biochemical progression-free survival†

Ash (2005)41 667 238 2·6 52% Low-dose rate No benefi t in biochemical progression-free survival in patients with intermediate-risk disease

Stock (2010)42 432 432 4·7 81% Low-dose rate and external beam radiotherapy No benefi t in biochemical progression-free survival

Dattoli (2010)43 321 157 10·6 45% Low-dose rate and external beam radiotherapy No benefi t in biochemical progression-free survival

Demanes (2009)44 411 188 6·4 49% High-dose rate and external beam radiotherapy No benefi t in biochemical progression-free survival, local control, distant metastasis-free survival, and prostate cancer-specifi c mortality

Martinez (2005)45 934 Not reported 4·4 44% High-dose rate and external beam radiotherapy No benefi t in biochemical progression-free survival, prostate cancer-specifi c mortality, and overall survival

Potters (2000)46 263 Not reported 3·8 50% Low-dose rate or low-dose rate and external beam radiotherapy

No benefi t in biochemical progression-free survival (matched-pair analysis)

Merrick (2006)47 938 425 5·4 41% Low-dose rate or low-dose rate and external beam radiotherapy

No diff erence in biochemical progression-free survival, prostate cancer-specifi c mortality, and overall survival for patients with intermediate-risk disease

Beyer (2005)48 2378 787 4·1 20% Low-dose rate or low-dose rate and external beam radiotherapy

No benefi t in prostate cancer-specifi c mortality, reduced overall survival

Zelefsky (2011)49 1466 563 4·1 31% Low-dose rate or low-dose rate and external beam radiotherapy or high-dose rate and external beam radiotherapy

No benefi t in biochemical progression-free survival

Krauss (2011)36* 575 417 5 47% Low-dose rate or high-dose rate or high-dose rate and external beam radiotherapy

Increased biochemical progression-free survival for brachytherapy alone, no benefi t for external beam radiotherapy and brachytherapy

*Only patients receiving brachytherapy, or brachytherapy with external beam radiotherapy, are included. †Benefi t restricted to patients with low-dose implants (dose to 90% of the prostate <140 Gy for iodine-125, <100 Gy for palladium-103).

Table 4: Retrospective series with short-term androgen deprivation therapy in patients treated with brachytherapy


Review

persistent prostate cancer on biopsy 2 years after treat-ment in 65%, 28%, and 5% of patients receiving 0, 3, or 9 months of androgen deprivation therapy, respectively.59

These fi ndings are important for putting available randomised data into context. For example, in RTOG 94-08,1 androgen deprivation therapy lowered the proportion of patients with persistent prostate cancer on biopsy 2 years after completion of external beam radiotherapy, from 39% to 20%. Thus, assuming local synergy between radiation and androgen deprivation therapy is the central factor accounting for this prolonged survival, would a signifi cant absolute benefi t in survival still be seen with a combination of short-term androgen deprivation therapy and a radiation regimen that provides a substantially better rate of local control than that used in RTOG 94-08?33 Even if short-term androgen deprivation therapy produces a fi xed reduction in the relative risk of local failure independent of radiation dose, any resulting benefi t in overall survival would be diminished, if increased rates of absolute local control are achieved with radiation alone. This idea has relevant clinical precedent in radiation oncology. For example, although post-mastectomy radiotherapy decreases the relative risk of locoregional recurrence by about two-thirds in all breast cancer patients, it does not increase survival for patients with negative lymph nodes because of their low absolute risk of locoregional recurrence.62

Although the diminishing benefi t of androgen deprivation therapy with escalating radiation dose is a logical thought, the argument is based on the assumption that modern dose-escalated radiation regimens provide such a high degree of local control that further incre-mental improvements from treatments such as androgen deprivation therapy would be clinically insignifi cant. This assumption might be overly optimistic. Even with doses of external beam radiotherapy of 81 Gy or higher in a series from MSKCC, 33% of patients had positive 2-year post-treatment biopsy samples when men receiving androgen deprivation therapy were excluded.57 Combined external beam radiotherapy and brachytherapy, on the other hand, achieves negative post-treatment biopsy specimens in up to 94% of patients.38,63 Thus, patients treated with this combined approach could derive a smaller absolute benefi t from androgen deprivation therapy than individuals receiving dose-escalated external beam radiotherapy.

Alternative explanations for the benefi cial eff ects of short-term androgen deprivation therapy combined with radiotherapy have been suggested. Androgen deprivation therapy improves distant control, especially with long-term treatment; for example, in a randomised trial of patients with lymph-node metastases, immediate indefi nite androgen deprivation therapy after prosta-tectomy increased overall survival compared with delayed treatment.64 Thus, sterilisation of subclinical distant micrometastases could account for the survival benefi t seen in trials combining androgen deprivation

therapy with external beam radiotherapy, but not with dose escalation.

However, short-term androgen deprivation therapy seems less eff ective for control of distant disease than does long-term treatment. Findings of two large randomised trials in patients with locally advanced or high-risk disease showed that survival with short-term androgen deprivation therapy was inferior to that noted with long-term treatment,19,21 and no survival advantage was recorded in the high-risk subgroup treated with short-term androgen deprivation therapy in RTOG 94-08.1 Moreover, as mentioned, no benefi t has been noted in trials of short-term androgen deprivation therapy and prostatectomy, although participants in these trials usually have lower risk disease than those in radiotherapy trials and, thus, might be less likely to benefi t from androgen deprivation therapy-mediated sterilisation of distant micrometastases.

Collectively, these data suggest that androgen depriv-ation therapy has both local and distant antitumour eff ects when combined with radiotherapy. However, the relative importance of these mechanisms in a given patient probably depends on both duration of androgen deprivation therapy and patterns of failure. For men with intermediate-risk prostate cancer treated with short-term androgen deprivation therapy, does increased survival stem predominantly from augmented eradication of local disease, which could be far greater in size than that achieved by the 10–15% rise in radiation dose used in dose-escalation trials, or from combined local and systemic eff ects? In view of the important therapeutic implications, further investigation is needed.

Duration and timing of androgen deprivation therapyIn view of these considerations, optimisation of timing and duration of androgen deprivation therapy is important. Typically, short-term androgen deprivation therapy is given before and during radiation, on the basis of preclinical data.65 However, fi ndings of a randomised trial comparing this approach with adjuvant androgen deprivation therapy showed no substantial diff erence in outcome, although an interaction between timing of androgen deprivation therapy and size of the radiation fi eld was postulated.66

The optimum duration for androgen deprivation therapy has been studied in men with high-risk and locally advanced prostate cancer. Findings of RTOG 92-0221 and a trial from the European Organisation for Research and Treatment of Cancer (EORTC 22961)19 showed that 4–6 months of androgen deprivation therapy were inferior to 28–36 months of treatment in this population. Although these trials did include some patients with intermediate-risk cancer by the NCCN defi nition, especially in RTOG 92-02, they formed only a small proportion of the study population. In a Canadian trial,67 361 patients—including 155 (43%) with intermediate-risk disease—were randomly


Review

allocated to either 3 or 8 months of neoadjuvant androgen deprivation therapy before conventional dose radiation (66 Gy).67 No benefi t was seen with the longer course of hormonal therapy either overall or in the intermediate-risk subgroup. However, high-risk patients had signifi cantly longer 7-year disease-free survival with the longer course of hormonal therapy (33% vs 59%, p=0·01). In another trial by the Trans-Tasman Radiation Oncology Group (TROG 96.01),17 6 months (but not 3 months) of androgen deprivation therapy and radiation decreased distant progression and prostate cancer-specifi c mortality and boosted overall survival at 10 years, compared with radiation alone. Only 130 (16%) of 802 patients in this study had intermediate-risk disease. Taken together, fi ndings of TROG 96.01 and the Canadian trial suggest that 6–8 months of neoadjuvant androgen deprivation therapy are superior to 3 months of neoadjuvant treatment for men with high-risk prostate cancer, but they provide scant evidence that longer term androgen deprivation ther apy further improves outcomes in intermediate-risk prostate cancer.

Toxic eff ects of androgen deprivation therapyIf androgen deprivation therapy were a completely benign well-tolerated treatment with few side-eff ects, its use in patients undergoing dose-escalated radiotherapy would be much less controversial. Unfortunately, toxic eff ects associated with androgen deprivation therapy are a major factor contributing to hesitation by some clinicians to recommend this treatment routinely for all men with intermediate-risk disease treated with modern radiation techniques. Androgen deprivation therapy can be associated with deleterious medical and quality-of-life sequelae, including hot fl ushes, sexual dysfunction, decreased libido, mood changes, muscle loss, anaemia, osteoporosis and fracture,68 diabetes mellitus,69 and the metabolic syndrome.69

In a prospective study comparing quality-of-life out-comes in patients treated with surgery, external beam radiotherapy, or brachytherapy,70 androgen deprivation therapy caused signifi cant reductions in vitality and hormonal quality-of-life outcomes (eg, fatigue, hot fl ushes, depression, gynaecomastia, and weight change). Furthermore, recovery of sexual function was signifi -cantly worse in patients receiving androgen deprivation therapy and external beam radiotherapy compared with those receiving external beam radiotherapy alone. Moreover, these quality-of-life diff erences persisted for more than 2 years, the maximum time reported in the study, even though 94% of patients receiving androgen deprivation therapy were treated for less than 1 year. Thus, although, typically, most side-eff ects related to short-term androgen depriv ation therapy are temporary, longer-lasting adverse events are possible, at least partly due to the fact that testos terone concentrations can take months or years after androgen deprivation therapy to return to normal amounts, especially in older men.71

In addition to the morbidity inherent to androgen deprivation therapy itself, researchers have investigated whether this treatment also amplifi es toxic eff ects from radiotherapy in healthy tissue, in view of its radiation-modulating properties. Indeed, in RTOG 92-02,21 long-term androgen deprivation therapy signifi cantly increased the risk of grade 3 or higher late radiation toxic eff ects compared with short-term treatment (10% vs 7%, p=0·0269). Furthermore, fi ndings of a multivariate analysis from the prospective quality-of-life study mentioned previously70 showed that acute urinary irritation was exacerbated in men receiving androgen deprivation therapy in conjunction with either external beam radiotherapy or brachytherapy, although these diff erences resolved within 6 months of treatment. Furthermore, androgen deprivation therapy was an independent predictor of urinary incontinence after brachytherapy in this study.70 On the other hand, at least with short-term androgen deprivation therapy, data of several other randomised trials indicate no increase in acute or late radiation-induced urinary or gastrointestinal toxic eff ects with hormonal treatment.1,17,20,72

Perhaps the most controversial and potentially danger-ous adverse eff ect of androgen deprivation therapy is a purported increase in cardiovascular morbidity and mortality. Data are confl icting but a full analysis is outside of the scope of our Review. Findings of major randomised controlled trials of androgen deprivation therapy and radiation do not show an increase in incidence of cardiovascular mortality from androgen deprivation therapy with 8–10 years of follow-up, nor do data from a meta-analysis.19,20,73–75 By contrast, fi ndings of several large retrospective population-based series indicate that cardiovascular morbidity is amplifi ed with androgen deprivation therapy.69,76,77 However, in view of biases inherent to these types of retrospective studies, prospective validation is necessary to outweigh the evidence from randomised trials and change practice patterns.

In a post-hoc subgroup analysis of the trial from the Dana-Farber Cancer Institute,2 a survival benefi t with short-term androgen deprivation therapy was only noted in men without relevant cardiovascular morbidities, suggesting that even with short-term treatment, in creased cardiovascular toxic eff ects could counteract the benefi cial eff ects of androgen deprivation therapy with respect to prostate cancer in a population with high baseline cardiovascular risk. Although these results should be regarded as hypothesis-generating in view of their retrospective nature and the few patients included, they emphasise that the optimum treatment for a given patient should ideally be personalised on the basis of charac teristics not only of the prostate cancer but also of the patient.

Conclusions and future directionsShould short-term androgen deprivation therapy be the standard of care combined with dose-escalated


Review

radiotherapy for all patients with intermediate-risk prostate cancer? Clinicians supporting this view argue that every randomised clinical trial undertaken to date has shown a benefi t with addition of short-term androgen deprivation therapy to radiotherapy, including two trials containing substantial numbers of men with intermediate-risk prostate cancer. Furthermore, although fi ndings of trials of androgen deprivation therapy have shown reductions in disease-specifi c mortality, and prolonged overall survival in some cases, data from dose-escalation trials have not, leading some to conclude that the evidence of benefi t for androgen deprivation therapy is much stronger than that for dose escalation. However, the adverse sequelae of androgen deprivation therapy are incontrovertible, and few would argue that sparing patients who are unlikely to derive relevant benefi t from this treatment should be a priority.

Therefore, in view of the low radiation doses used in RTOG 94-081 and the Dana-Farber trial,2 the substantial clinical heterogeneity of patients with intermediate-risk prostate cancer, and the adverse eff ects associated with androgen deprivation therapy, a one-size-fi ts-all treatment algorithm on the basis of risk classifi cation alone might not be the most appropriate therapeutic approach. We propose a risk-adaptive strategy based on a nuanced interpretation of available data (table 5). For patients with favourable intermediate-risk disease, dose-escalated radiotherapy alone might be suffi cient treatment, especially in view of the eff ects of stage migration and Gleason score infl ation over the past two decades.78,79 However, for men with unfavourable intermediate-risk disease, dose-escalated external beam radiotherapy with 4–6 months of androgen deprivation therapy should be judged a standard of care.

The role of brachytherapy in patients with intermediate-risk prostate cancer is less defi ned and prospective evidence is scarce. In general, we deem brachytherapy alone an acceptable option for selected patients with favourable features of intermediate-risk prostate cancer and low-volume disease (table 5), but this strategy is controversial and needs prospective validation. Combined brachytherapy and external beam radiotherapy could be considered for patients with unfavourable intermediate-risk disease. Although fi ndings of retrospective series have not shown a benefi t from addition of androgen

deprivation therapy to combined brachytherapy and external beam radiotherapy in men with intermediate-risk disease,36,42–49 in view of the paucity of randomised evidence and fairly poor-quality retrospective data, addition of short-term androgen deprivation therapy to this modality for patients with unfavourable features could be considered on the basis of extrapolation of data from trials of external beam radiotherapy.

Ascertainment of the best use of short-term androgen deprivation therapy combined with dose-escalated radiation for patients with intermediate-risk prostate cancer will need randomised evidence, but are better answers forthcoming? RTOG 08-15 is years away from meeting its accrual goal, and GETUG 14, which was not powered to detect diff erences in distant metastasis or overall survival in the fi rst place, was closed early because of poor accrual.23 Therefore, defi nitive answers are years away from arriving.

The most imperative question might not be whether patients with intermediate-risk disease benefi t from short-term androgen deprivation therapy with dose-escalated radiotherapy but, rather, can we more accurately ascertain which patients from this heterogeneous group derive enough benefi t to justify the toxic eff ects? Findings of genomic studies suggest that prostate cancers vary widely in transcriptional output of androgen receptors and, thus, substantial variation in tumour response to androgen deprivation therapy might also take place.80 Ongoing and future clinical trials should incorporate basic and translational research to uncover mechanisms of interaction between radiation and androgen deprivation therapy and identify molecular predictors of

Favourable intermediate-risk prostate cancer* Unfavourable intermediate-risk prostate cancer†

Clinical characteristics One intermediate risk factorGleason score of 3+4=7 or less<50% positive biopsy cores

Several intermediate risk factors5,7

Gleason score of 4+3=714

≥50% positive biopsy cores11

Recommended radiation options Dose-escalated external beam radiotherapy aloneBrachytherapy alone in select cases (eg, ≤3 positive cores, none with >50% involvement)

Dose-escalated external beam radiotherapy and short-term androgen deprivation therapyCombined brachytherapy and external beam radiotherapy with or without short-term androgen deprivation therapy

*All these criteria are required. †Any of these criteria can be met.

Table 5: Memorial Sloan-Kettering Cancer Center treatment algorithm for defi nitive radiotherapy in patients with intermediate-risk prostate cancer


We identifi ed references for this Review by searching PubMed with the terms: “intermediate risk prostate cancer”; “dose escalated” or “dose escalation”; and “androgen deprivation therapy”. We also searched the reference lists of articles identifi ed by this strategy and included any we deemed relevant. For retrospective single-institution series, we only included the publication with the longest follow-up. Only papers in English were included. We undertook the last literature search in December, 2011.


Review

recurrence patterns. Furthermore, novel androgen-suppressing agents—eg, abiraterone and MDV3100 (Medivation, San Francisco, CA, USA)—show augmented ability to target the androgen axis even in castration-resistant prostate cancer,81,82 thus they could also have potent synergy with radiotherapy. Through increased understanding of the synergistic interaction between radiation and androgen deprivation therapy and expan-sion of available androgen-suppressing agents, treatment for intermediate-risk prostate cancer will hopefully move away from a homogeneous therapeutic strategy to an individualised approach that boosts oncol og ical outcome while diminishing toxic eff ects.

ContributorsZSZ and MJZ had the original idea for this Review and wrote it

together. ZSZ did the literature search, selected relevant articles, and

produced the tables.


References1 Jones CU, Hunt D, McGowan DG, et al. Radiotherapy and

short-term androgen deprivation for localized prostate cancer. N Engl J Med 2011; 365: 107–18.

2 D’Amico AV, Chen MH, Renshaw AA, et al. Androgen suppression and radiation vs radiation alone for prostate cancer: a randomized trial. JAMA 2008; 299: 289–95.

3 Cooperberg MR, Grossfeld GD, Lubeck DP, Carroll PR. National practice patterns and time trends in androgen ablation for localized prostate cancer. J Natl Cancer Inst 2003; 95: 981–89.

4 Park S, Meng MV, Elkin EP, et al. Androgen deprivation use with external beam radiation for prostate cancer: results from CaPSURE. J Urol 2005; 174: 1802–07.

5 Mohler J, Bahnson RR, Boston B, et al. NCCN clinical practice guidelines in oncology: prostate cancer. J Natl Compr Canc Netw 2010; 8: 162–200.

6 D’Amico AV, Whittington R, Malkowicz SB, et al. Biochemical outcome after radical prostatectomy, external beam radiation therapy, or interstitial radiation therapy for clinically localized prostate cancer. JAMA 1998; 280: 969–74.

7 Zelefsky MJ, Leibel SA, Gaudin PB, et al. Dose escalation with three-dimensional conformal radiation therapy aff ects the outcome in prostate cancer. Int J Radiat Oncol Biol Phys 1998; 41: 491–500.

8 Chism DB, Hanlon AL, Horwitz EM, et al. A comparison of the single and double factor high-risk models for risk assignment of prostate cancer treated with 3D conformal radiotherapy. Int J Radiat Oncol Biol Phys 2004; 59: 380–85.

9 Zelefsky MJ, Pei X, Chou JF, et al. Dose escalation for prostate cancer radiotherapy: predictors of long-term biochemical tumor control and distant metastases-free survival outcomes. Eur Urol 2011; 60: 1133–39.

10 D’Amico AV, Whittington R, Malkowicz SB, et al. Pretreatment nomogram for prostate-specifi c antigen recurrence after radical prostatectomy or external-beam radiation therapy for clinically localized prostate cancer. J Clin Oncol 1999; 17: 168–72.

11 D’Amico AV, Renshaw AA, Cote K, et al. Impact of the percentage of positive prostate cores on prostate cancer-specifi c mortality for patients with low or favorable intermediate-risk disease. J Clin Oncol 2004; 22: 3726–32.

12 Feng FY, Qian Y, Stenmark MH, et al. Perineural invasion predicts increased recurrence, metastasis, and death from prostate cancer following treatment with dose-escalated radiation therapy. Int J Radiat Oncol Biol Phys 2011; 81: e361–67.

13 D’Amico AV, Renshaw AA, Sussman B, Chen MH. Pretreatment PSA velocity and risk of death from prostate cancer following external beam radiation therapy. JAMA 2005; 294: 440–47.

14 Stark JR, Perner S, Stampfer MJ, et al. Gleason score and lethal prostate cancer: does 3 + 4 = 4 + 3? J Clin Oncol 2009; 27: 3459–64.

15 Bolla M, Van Tienhoven G, Warde P, et al. External irradiation with or without long-term androgen suppression for prostate cancer with high metastatic risk: 10-year results of an EORTC randomised study. Lancet Oncol 2010; 11: 1066–73.

16 Pilepich MV, Winter K, Lawton CA, et al. Androgen suppression adjuvant to defi nitive radiotherapy in prostate carcinoma: long-term results of phase III RTOG 85-31. Int J Radiat Oncol Biol Phys 2005; 61: 1285–90.

17 Denham JW, Steigler A, Lamb DS, et al. Short-term neoadjuvant androgen deprivation and radiotherapy for locally advanced prostate cancer: 10-year data from the TROG 96.01 randomised trial. Lancet Oncol 2011; 12: 451–59.

18 Granfors T, Modig H, Damber JE, Tomic R. Long-term followup of a randomized study of locally advanced prostate cancer treated with combined orchiectomy and external radiotherapy versus radiotherapy alone. J Urol 2006; 176: 544–47.

19 Bolla M, de Reijke TM, Van Tienhoven G, et al. Duration of androgen suppression in the treatment of prostate cancer. N Engl J Med 2009; 360: 2516–27.

20 Roach M III, Bae K, Speight J, et al. Short-term neoadjuvant androgen deprivation therapy and external-beam radiotherapy for locally advanced prostate cancer: long-term results of RTOG 8610. J Clin Oncol 2008; 26: 585–91.

21 Horwitz EM, Bae K, Hanks GE, et al. Ten-year follow-up of Radiation Therapy Oncology Group protocol 92-02: a phase III trial of the duration of elective androgen deprivation in locally advanced prostate cancer. J Clin Oncol 2008; 26: 2497–504.

22 Laverdière J, Nabid A, De Bedoya LD, et al. The effi cacy and sequencing of a short course of androgen suppression on freedom from biochemical failure when administered with radiation therapy for T2-T3 prostate cancer. J Urol 2004; 171: 1137–40.

23 Dubray BM, Beckendorf V, Guerif S, et al. Does short-term androgen depletion add to high-dose radiotherapy (80 Gy) in localized intermediate-risk prostate cancer? Intermediate analysis of GETUG 14 randomized trial (EU-20503/NCT00104741). Proc Am Soc Clin Oncol 2011; 29 (suppl): 4521 (abstr).

24 Dearnaley DP, Sydes MR, Graham JD, et al, on behalf of the RT01 collaborators. Escalated-dose versus standard-dose conformal radiotherapy in prostate cancer: fi rst results from the MRC RT01 randomised controlled trial. Lancet Oncol 2007; 8: 475–87.

25 Al-Mamgani A, van Putten WL, Heemsbergen WD, et al. Update of Dutch multicenter dose-escalation trial of radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 2008; 72: 980–88.

26 Zietman AL, Bae K, Slater JD, et al. Randomized trial comparing conventional-dose with high-dose conformal radiation therapy in early-stage adenocarcinoma of the prostate: long-term results from Proton Radiation Oncology Group/American College of Radiology 95-09. J Clin Oncol 2010; 28: 1106–11.

27 Kuban DA, Tucker SL, Dong L, et al. Long-term results of the M D Anderson randomized dose-escalation trial for prostate cancer. Int J Radiat Oncol Biol Phys 2008; 70: 67–74.

28 Beckendorf V, Guerif S, Le Prisé E, et al. 70 Gy versus 80 Gy in localized prostate cancer: 5-year results of GETUG 06 randomized trial. Int J Radiat Oncol Biol Phys 2011; 80: 1056–63.

29 American Society for Therapeutic Radiology and Oncology Consensus Panel. Consensus statement: guidelines for PSA following radiation therapy. Int J Radiat Oncol Biol Phys 1997; 37: 1035–41.

30 Roach M III, Hanks GE, Thames H Jr, et al. Defi ning biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: recommendations of the RTOG-ASTRO Phoenix Consensus conference. Int J Radiat Oncol Biol Phys 2006; 65: 965–74.

31 Zelefsky MJ, Yamada Y, Fuks Z, et al. Long-term results of conformal radiotherapy for prostate cancer: impact of dose escalation on biochemical tumor control and distant metastases-free survival outcomes. Int J Radiat Oncol Biol Phys 2008; 71: 1028–33.

32 Valicenti R, Lu J, Pilepich M, Asbell S, Grignon D. Survival advantage from higher-dose radiation therapy for clinically localized prostate cancer treated on the Radiation Therapy Oncology Group trials. J Clin Oncol 2000; 18: 2740–46.

33 Créhange G, Bolla M, Maingon P. Radiotherapy and androgen deprivation for prostate cancer. N Engl J Med 2011; 365: 1354.


Review

34 Dearnaley DP, Jovic G, Syndikus I, et al. Escalated-dose conformal radiotherapy for localised prostate cancer: long-term overall survival results from the MRC RT01 randomised controlled trial, on behalf of the RT01 investigators. Presented at the 2011 European Multidisciplinary Cancer Conference; Stockholm, Sweden; Sept 23–27, 2011.

35 Valicenti RK, Bae K, Michalski J, et al. Does hormone therapy reduce disease recurrence in prostate cancer patients receiving dose-escalated radiation therapy? An analysis of Radiation Therapy Oncology Group 94-06. Int J Radiat Oncol Biol Phys 2011; 79: 1323–29.

36 Krauss D, Kestin L, Ye H, et al. Lack of benefi t for the addition of androgen deprivation therapy to dose-escalated radiotherapy in the treatment of intermediate- and high-risk prostate cancer. Int J Radiat Oncol Biol Phys 2011; 80: 1064–71.

37 Ciezki JP, Klein EA, Angermeier K, et al. A retrospective comparison of androgen deprivation (AD) vs no AD among low-risk and intermediate-risk prostate cancer patients treated with brachytherapy, external beam radiotherapy, or radical prostatectomy. Int J Radiat Oncol Biol Phys 2004; 60: 1347–50.

38 Stone NN, Stock RG, Cesaretti JA, Unger P. Local control following permanent prostate brachytherapy: eff ect of high biologically eff ective dose on biopsy results and oncologic outcomes. Int J Radiat Oncol Biol Phys 2010; 76: 355–60.

39 Pieters BR, van de Kamer JB, van Herten YR, et al. Comparison of biologically equivalent dose-volume parameters for the treatment of prostate cancer with concomitant boost IMRT versus IMRT combined with brachytherapy. Radiother Oncol 2008; 88: 46–52.

40 Lee LN, Stock RG, Stone NN. Role of hormonal therapy in the management of intermediate- to high-risk prostate cancer treated with permanent radioactive seed implantation. Int J Radiat Oncol Biol Phys 2002; 52: 444–52.

41 Ash D, Al-Qaisieh B, Bottomley D, et al. The impact of hormone therapy on post-implant dosimetry and outcome following iodine-125 implant monotherapy for localised prostate cancer. Radiother Oncol 2005; 75: 303–06.

42 Stock RG, Yalamanchi S, Hall SJ, Stone NN. Impact of hormonal therapy on intermediate risk prostate cancer treated with combination brachytherapy and external beam irradiation. J Urol 2010; 183: 546–50.

43 Dattoli M, Wallner K, True L, Bostwick D, Cash J, Sorace R. Long-term outcomes for patients with prostate cancer having intermediate and high-risk disease, treated with combination external beam irradiation and brachytherapy. J Oncol 2010; published online July 1. DOI:10.1155/2010/471375.

44 Demanes DJ, Brandt D, Schour L, Hill DR. Excellent results from high dose rate brachytherapy and external beam for prostate cancer are not improved by androgen deprivation. Am J Clin Oncol 2009; 32: 342–47.

45 Martinez AA, Demanes DJ, Galalae R, et al. Lack of benefi t from a short course of androgen deprivation for unfavorable prostate cancer patients treated with an accelerated hypofractionated regime. Int J Radiat Oncol Biol Phys 2005; 62: 1322–31.

46 Potters L, Torre T, Ashley R, Leibel S. Examining the role of neoadjuvant androgen deprivation in patients undergoing prostate brachytherapy. J Clin Oncol 2000; 18: 1187–92.

47 Merrick GS, Butler WM, Wallner KE, et al. Androgen-deprivation therapy does not impact cause-specifi c or overall survival after permanent prostate brachytherapy. Int J Radiat Oncol Biol Phys 2006; 65: 669–77.

48 Beyer DC, McKeough T, Thomas T. Impact of short course hormonal therapy on overall and cancer specifi c survival after permanent prostate brachytherapy. Int J Radiat Oncol Biol Phys 2005; 61: 1299–305.

49 Zelefsky MJ, Chou JF, Pei X, et al. Predicting biochemical tumor control after brachytherapy for clinically localized prostate cancer: the Memorial Sloan-Kettering Cancer Center experience. Brachytherapy 2011; published online Sept 17. DOI:10.1016/j.brachy.2011.08.003.

50 Wo JY, Zietman AL. Why does androgen deprivation enhance the results of radiation therapy? Urol Oncol 2008; 26: 522–29.

51 Klotz LH, Goldenberg SL, Jewett MA, et al. Long-term followup of a randomized trial of 0 versus 3 months of neoadjuvant androgen ablation before radical prostatectomy. J Urol 2003; 170: 791–94.

52 Soloway MS, Pareek K, Sharifi R, et al. Neoadjuvant androgen ablation before radical prostatectomy in cT2bNxMo prostate cancer: 5-year results. J Urol 2002; 167: 112–16.

53 Warde P, Mason M, Ding K, et al, for the NCIC CTG PR.3/MRC UK PR07 investigators. Combined androgen deprivation therapy and radiation therapy for locally advanced prostate cancer: a randomised, phase 3 trial. Lancet 2011; 378: 2104–11.

54 Widmark A, Klepp O, Solberg A, et al. Endocrine treatment, with or without radiotherapy, in locally advanced prostate cancer (SPCG-7/SFUO-3): an open randomised phase III trial. Lancet 2009; 373: 301–08.

55 Bill-Axelson A, Holmberg L, Ruutu M, et al. Radical prostatectomy versus watchful waiting in early prostate cancer. N Engl J Med 2011; 364: 1708–17.

56 Crook J, Robertson S, Collin G, et al. Clinical relevance of trans-rectal ultrasound, biopsy, and serum prostate-specifi c antigen following external beam radiotherapy for carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1993; 27: 31–37.

57 Zelefsky MJ, Reuter VE, Fuks Z, et al. Infl uence of local tumor control on distant metastases and cancer related mortality after external beam radiotherapy for prostate cancer. J Urol 2008; 179: 1368–73.

58 Crook J, Malone S, Perry G, et al. Postradiotherapy prostate biopsies: what do they really mean? Results for 498 patients. Int J Radiat Oncol Biol Phys 2000; 48: 355–67.

59 Laverdière J, Gomez JL, Cusan L, et al. Benefi cial eff ect of combination hormonal therapy administered prior and following external beam radiation therapy in localized prostate cancer. Int J Radiat Oncol Biol Phys 1997; 37: 247–52.

60 Zapatero A, Mínguez R, Nieto S, et al. Post-treatment prostate biopsies in the era of three-dimensional conformal radiotherapy: what can they teach us? Eur Urol 2009; 55: 902–09.

61 Pollack A, Zagars GK, Smith LG, et al. Preliminary results of a randomized radiotherapy dose-escalation study comparing 70 Gy with 78 Gy for prostate cancer. J Clin Oncol 2000; 18: 3904–11.

62 Early Breast Cancer Trialists’ Collaborative Group (EBCTCG). Eff ects of radiotherapy and of diff erences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005; 366: 2087–106.

63 Bachand F, Martin AG, Beaulieu L, et al. An eight-year experience of HDR brachytherapy boost for localized prostate cancer: biopsy and PSA outcome. Int J Radiat Oncol Biol Phys 2009; 73: 679–84.

64 Messing EM, Manola J, Yao J, et al, on behalf of the Eastern Cooperative Oncology Group study EST 3886. Immediate versus deferred androgen deprivation treatment in patients with node-positive prostate cancer after radical prostatectomy and pelvic lymphadenectomy. Lancet Oncol 2006; 7: 472–79.

65 Zietman AL, Nakfoor BN, Prince EA, Gerweck LE. The eff ect of androgen deprivation on an androgen sensitive tumor: an in vivo and in vitro study. Cancer J Sci Am 1997; 3: 31–36.

66 Lawton CA, DeSilvio M, Roach M III, et al. An update of the phase III trial comparing whole pelvic to prostate only radiotherapy and neoadjuvant to adjuvant total androgen suppression: updated analysis of RTOG 94-13, with emphasis on unexpected hormone/radiation interactions. Int J Radiat Oncol Biol Phys 2007; 69: 646–55.

67 Crook J, Ludgate C, Malone S, Perry G, et al. Final report of multicenter Canadian Phase III randomized trial of 3 versus 8 months of neoadjuvant androgen deprivation therapy before conventional-dose radiotherapy for clinically localized prostate cancer. Int J Radiat Oncol Biol Phys 2009; 73: 327–33.

68 Shahinian VB, Kuo YF, Freeman JL, Goodwin JS. Risk of fracture after androgen deprivation for prostate cancer. N Engl J Med 2005; 352: 154–64.

69 Keating NL, O’Malley AJ, Smith MR. Diabetes and cardiovascular disease during androgen deprivation therapy for prostate cancer. J Clin Oncol 2006; 24: 4448–56.

70 Sanda MG, Dunn RL, Michalski J, et al. Quality of life and satisfaction with outcome among prostate-cancer survivors. N Engl J Med 2008; 358: 1250–61.

71 D’Amico AV, Chen MH, Renshaw AA, et al. Interval to testosterone recovery after hormonal therapy for prostate cancer and risk of death. Int J Radiat Oncol Biol Phys 2009; 75: 10–15.

72 D’Amico AV, Manola J, Loff redo M, et al. 6-month androgen suppression plus radiation therapy vs radiation therapy alone for patients with clinically localized prostate cancer: a randomized controlled trial. JAMA 2004; 292: 821–27.

73 Efstathiou JA, Bae K, Shipley WU, et al. Cardiovascular mortality after androgen deprivation therapy for locally advanced prostate cancer: RTOG 85-31. J Clin Oncol 2009; 27: 92–99.


Review

74 Efstathiou JA, Bae K, Shipley WU, et al. Cardiovascular mortality and duration of androgen deprivation for locally advanced prostate cancer: analysis of RTOG 92-02. Eur Urol 2008; 54: 816–23.

75 Nguyen PL, Je Y, Schutz FA, et al. Association of androgen deprivation therapy with cardiovascular death in patients with prostate cancer: a meta-analysis of randomized trials. JAMA 2011; 306: 2359–66.

76 D’Amico AV, Denham JW, Crook J, et al. Infl uence of androgen suppression therapy for prostate cancer on the frequency and timing of fatal myocardial infarctions. J Clin Oncol 2007; 25: 2420–25.

77 Tsai HK, D’Amico AV, Sadetsky N, et al. Androgen deprivation therapy for localized prostate cancer and the risk of cardiovascular mortality. J Natl Cancer Inst 2007; 99: 1516–24.

78 Albertsen PC, Hanley JA, Barrows GH, et al. Prostate cancer and the Will Rogers phenomenon. J Natl Cancer Inst 2005; 97: 1248–53.

79 Gallina A, Chun FK, Suardi N, et al. Comparison of stage migration patterns between Europe and the USA: an analysis of 11 350 men treated with radical prostatectomy for prostate cancer. BJU Int 2008; 101: 1513–18.

80 Taylor BS, Schultz N, Hieronymus H, et al. Integrative genomic profi ling of human prostate cancer. Cancer Cell 2010; 18: 11–22.

81 Scher HI, Beer TM, Higano CS, et al, and the Prostate Cancer Foundation/Department of Defense Prostate Cancer Clinical Trials Consortium. Antitumour activity of MDV3100 in castration-resistant prostate cancer: a phase 1–2 study. Lancet 2010; 375: 1437–46.

82 De Bono JS, Logothetis CJ, Molina A, et al. Abiraterone and increased survival in metastatic prostate cancer. N Engl J Med 2011; 364: 1995–2005.


Review


*These authors contributed equally

Toxicology Unit, Medical Research Council, Lancaster Road, Leicester, UK (Y W Kong PhD, D Ferland-McCollough PhD, T J Jackson, M Bushell DPhil); and University of Nottingham Medical School, Queen’s Medical Centre, Nottingham, UK (T J Jackson)

Correspondence to:Dr Martin Bushell, MRC Toxicology Unit, Lancaster Road, Leicester LE1 9HN, UK [email protected]

microRNAs in cancer managementYi W Kong*, David Ferland-McCollough*, Thomas J Jackson, Martin Bushell

Since the identifi cation of microRNAs (miRNAs) in 1993, and the subsequent discovery of their highly conserved nature in 2000, the amount of research into their function—particularly how they contribute to malignancy—has greatly increased. This class of small RNA molecules control gene expression and provide a previously unknown control mechanism for protein synthesis. As such, it is unsurprising that miRNAs are now known to play an essential part in malignancy, functioning as tumour suppressors and oncogenes. This Review summarises the present understanding of how miRNAs operate at the molecular level; how their dysregulation is a crucial part of tumour formation, maintenance, and metastasis; how they can be used as biomarkers for disease type and grade; and how miRNA-based treatments could be used for diverse types of malignancies.

IntroductionMicroRNAs (miRNAs) are small non-coding RNAs (21–23 nucleotides) encoded in the genome of plants, invertebrates, and vertebrates. These small molecules mainly bind imperfectly to the 3 untranslated region of target messenger RNAs (mRNAs).1 They negatively regulate gene expression post-transcriptionally by inhibiting translation and causing degradation of target mRNA. More than a thousand miRNAs exist in the human genome and each one can potentially regulate hundreds of mRNAs. miRNAs play an important part in many cellular processes, such as diff erentiation, proliferation, apoptosis, and stress response. Addition-ally, they are key regulators in many diseases—eg, neurological disorders, heart disease, vascular diseases, viral infection, and cancer.1

The discovery of miRNAs led to a worldwide research eff ort to establish their roles in cancer. The results have been impressive, collectively showing that miRNAs form central nodal points in cancer development pathways. miRNAs regulate molecular pathways in cancer by targeting various oncogenes and tumour suppressors,1 and have a role in cancer-stem-cell biology, angiogenesis, the epithelial–mesenchymal transition, metastasis, and drug resistance. Loss of one miRNA can have substantial eff ects; dysregulation can cause tumorigenesis (fi gure 1). Because miRNA-based regulation is dependent on expression of its mRNA targets, which are not always ubiquitously expressed, an miRNA can have eff ects specifi c to cells types and conditions.

Researchers are beginning to uncover the complex role that miRNAs have in malignant disease. However, little detailed, in-vivo work has been reported. An improved understanding of miRNA mechanisms in tumorigenesis and cancer maintenance would thus provide invaluable information about key cancer pathways, cancer diagnostics, and disease prognosis. Importantly, this knowledge could be used in the development of anticancer therapies.

Regulation of cancer pathwaysGene regulation and mutationsIn 2002, the fi rst report about the role of miRNAs in cancer established that the gene cluster containing the

miRNAs miR-15 and miR-16 is deleted in most people with chronic lymphocytic leukaemia (CLL).2 Further studies have shown that miR-15 and miR-16 act as tumour suppressors by targeting the oncogene BCL2, which encodes a protein involved in cell survival.3

Conversely, miR-21 is an excellent example of an onco-genic miRNA (a so-called oncomir). It is overexpressed in most cancers—eg, breast cancer, colorectal cancer, lung cancer, pancreatic cancer, glioblastoma, neuro-blastoma, leukaemia, and lymphoma.4–7 Overexpression of this miRNA can cause tumour growth, maintenance, and survival in vivo.8 Importantly, these tumours are completely dependent on expression of miR-21.8 miR-21 targets several tumour-suppressor genes, including PTEN, to increase proliferation and decrease apoptosis.9,10 miR-21 could also promote tumour migration by targeting pro-invasion genes.11

Modulation of miRNA expression is increasingly thought to be an important mechanism by which tumour-suppressor proteins and oncoproteins exert some of their eff ects. The proto-oncogene MYC transcriptionally activates the miR-17-92 cluster—a polycistronic transcript of miRNAs 17, 18a, 19a, 20a, 19b-1, and 92a-1.12 The oncogenicity of miR-17-92 is attributed mainly to miR-19, which inhibits PTEN to activate AKT signalling and promote cancer-cell survival (fi gure 2).13 MYC also represses transcription of many tumour-suppressor miRNAs, including the let-7 family.14 Reduced expression of let-7 miRNAs is recorded in many cancers,1 and correlates with poor survival. The let-7 family targets the enzyme RAS, which is involved in cell growth, diff erentiation, and survival. let-7a, let-7c, and let-7g are deleted in lung cancer tumours, and many miRNAs of the let-7 family are located in the genomic region frequently deleted in patients with lung cancer.15

The tumour suppressor P53 transcriptionally induces the miR-34 family of miRNAs in response to DNA damage (fi gure 2). miR-34a is expressed from one genomic site, but miR-34b and miR-34c are produced together from one primary transcript at another site.16,17 Loss of miR-34a expression is associated with metastasis and recurrence of prostate cancer.18 Restoration of miR-34 expression in pancreatic cancer cells substantially inhibited clonogenic cell growth and invasion, induced


Review

apoptosis, and sensitised cells to chemotherapy and radiation. Similarly, studies of prostate cancer stem cells show that reintroduction of miR-34a into tumours and cancer cell lines leads to corresponding reductions of CD44 and tumour size.19 Expression of miR-34b and miR-34c is lost through deletion or hypermethylation, or is downregulated, in 90% of colorectal cancers.20,21 In their absence, both P53-dependent and P38-MAPK-dependent responses to DNA damage are attenuated, leading to oncogenesis.21,22 Importantly, MYC is a target of miR-34b and miR-34c, substantiating the inter connected nature of miRNA expression in malignancies (fi gure 2).

Gene regulation is not the only way in which miRNAs are implicated in malignancies (fi gure 1). Mutations of miRNA binding sites in the 3 untranslated region (UTR) of oncogenes are correlated with an increased risk of cancer. For example, a single nucleotide poly-morphism in the 3 UTR of the KRAS oncogene signifi -cantly increases risk of non-small-cell lung cancer.23

Additionally, genetic variation in the miR-221 binding site in the 3 UTR of KIT is associated with a heightened risk of acral melanoma.24

Cancer stem cellsCancer stem cells—also called tumour-initiating cells—are a small population of cells within a tumour that are thought to arise from somatic stem cells. They have enhanced self-renewal capacities, tumorigenic potential, and expression of unique surface markers;25 and are often essential for tumour maintenance, treatment resistance, tumour progression, and distant metastasis. Recent studies suggest that miRNAs are involved in regulating properties of these cells (fi gure 3). Some miRNAs involved in stem-cell regulation are miR-296, miR-134, miR-470, and the miR-34 family, which targets genes essential for pluripotency and stem-cell function (eg, Oct4, NANOG, SOX2, NOTCH, and BCL2).26

In breast cancer, let-7 and miR-30 are important for the regulation of cancer stem cells. let-7 is downregulated in breast-cancer stem cells.27 When overexpressed in mice, it can reduce numbers of undiff erentiated cells in vitro and inhibit cell proliferation, tumour formation, and metastasis.28 miR-30 expression is also low and can inhibit the self-renewal ability of breast-cancer stem cells. Antagonism of miR-30 by antisense oligonucleotides enhances self-renewal, tumour regeneration, and metastasis in diff erentiated breast cancer cells. The combination of both let-7 and miR-30 inhibits self-renewal and mammosphere formation in breast-cancer stem cells.28

AngiogenesisPromotion of angiogenesis by tumours necessitates activation of proliferation and migration pathways in vascular smooth muscle cells. In nasopharyngeal carcinoma cells, the miR-15a/16-1 cluster regulates angiogenesis (fi gure 3) by targeting the angiogenic factors VEGFA and MET.29 miR-145 blocks migration of vascular smooth muscle cells by inhibiting FLI1,30 and miR-143 by targeting the versican protein involved in migration induced by platelet-derived growth factor.31 The miR-143 cluster is downregulated in cancers of the prostate, colon, gastric, and bladder, and in CLL and B-cell lymphomas.32

Epithelial–mesenchymal transition and metastasisEpithelial cells undergo several molecular changes during the epithelial–mesenchymal transition to assume a mesenchymal cell phenotype necessary for tumour metastasis and progression (fi gure 3).33 Expression of E-cadherin (CDH1) is essential for retaining an epithelial cell type. The miR-200 family, miR-27, and miR-205 inhibit ZEB1 and ZEB2, which are repressors of E-cadherin expression. Similarly, expression of miR-200 in breast cancers is positively correlated with concen trations of E-cadherin. The

Figure 1: Regulation of tumorigenesis by miRNAsTumorigenesis can be regulated by miRNAs at diff erent levels. Upregulation of oncogenic miRNAs reduces expression of tumour-suppressor protein, but downregulation of tumour-suppressing miRNAs results in an increased production of oncogenic protein. Loss-of-function mutations in tumour-suppressing miRNAs and mutation of the target section of oncogene mRNA can cause tumorigenesis, because expression of oncogenic proteins is no longer regulated. Loss-of-function mutations in oncogenic miRNAs and mutations in tumour-suppressor mRNA would increase expression of tumour-suppressor proteins and hence reduce tumorigenesis. miRNA=microRNA.

3´ 5´

3´ 5´

3´ 5´ 3´ 5´

3´ 5´

3´ 5´

Oncogenic miRNAs

Tumour-suppressing miRNAs

Upregulation• More transcription• Gene amplification• Hypomethylation

Downregulation• Less transcription• Genomic deletion• Hypermethylation

MutationsLoss of function

Tumour-suppressing miRNA

miRNA resistance

Oncogene mRNATumour-suppressor mRNA

Tumour-supressor mRNA AAA

Oncogene mRNAAAA

Oncogenic miRNA

3´ 5´

3´ 5´

3´ 5´ 3´ 5´

3´ 5´

3´ 5´

AAA

AAA

AAA

AAA

miRNA resistance

MutationsLoss of function

Oncogene mRNATumour-suppressor mRNA

Tumorigenesis


Review

importance of this association is supported by evidence that restoration of miR-200 expression is suffi cient to reverse the transition (ie, mesenchymal to epithelial) in a kidney-derived cell line.34 However, other miRNAs are needed to prevent the epithelial–mesenchymal transition. In pancreatic epithe lial cells, the expression of miRNAs from the miR-30 family inversely correlates with the mesenchymal pheno type.35 In mesenchymal-like ovarian cancer cell lines, overexpression of miR-429 reverses the transition.36

Clinical potentialDiagnosis and classifi cationThe development of cytogenetic, immunological, and molecular methods to supplement the traditional morphological classifi cation systems has refi ned the identifi cation of cancer subtypes.37 For instance, tran-scriptome profi ling studies of diff erent cancer types with large-scale genomic approaches showed that a 97-gene expression profi le is better for classifi cation of breast cancer histological grade than are lymph-node status and tumour size.37 Gene-expression profi ling has also led to development of breast-cancer prognostic profi ling tests, which have been approved for clinical use.38

As with mRNA, whole-genome miRNA profi ling has shown that miRNA expression changes substantially in most human cancers.39 miRNA expression signatures provide a more accurate method of cancer subtype classifi cation than does expression profi ling of an entire group of known protein-coding RNAs.40 miRNA profi les can contribute to the diagnostic and prognostic classi-fi cation of human malignancies (table).2,41–48

The use of the traditional, gold standard, histological examination in the diagnosis and classifi cation of cancers can be limited by availability of adequately preserved tissue and the possibility of subjective interpretation by pathologists. Global cytogenetic techniques are typically time-consuming and protein screening depends on prevention of protein degradation. By contrast, miRNAs show resistance to degradation and their expression levels can be established in a few hours with as little as 10 ng of total RNA; therefore, miRNAs are ideal as both diagnostic and prognostic indicators in clinical settings.

Screenings of resected tumours and biopsy samples have identifi ed miRNA signatures that can be used to diff erentiate between malignant and benign condi tions in several organs. Seven miRNAs are diff erentially expressed in biopsied pancreatic ductal adenocarcinomas compared with benign and healthy tissues.49 A multi centre trial50 showed that use of signatures from these miRNAs gives greater accuracy than does conventional cytology. Screening of 27 thyroid ectomies was shown to have sensitivity, and negative and positive predictive values all of at least 98%. One miRNA was overexpressed in 19 of 20 malignant samples.51

miRNA expression can also be used to identify well characterised genotypes—eg, to distinguish between

sporadic and germ-line tumours,47 or to identify genomic instability within a tumour. Genomic instability is poten tially informative because many miRNAs are located in fragile sites.52 The 13q14.3 region is a fragile site and is the most common chromosomal abnormality in CLL, (deleted in about 55% of cases). The miR-15/16 cluster resides in a 30 kb region of loss within this area.

Figure 2: Oncogenic and tumour-suppressing miRNAs in the DNA damage responseDNA damage can lead to the upregulation of the miR-34 family through the activation of the P38 MAP-kinase and P53 pathways. The tumour suppressor miR-34 blocks MYC translation, thus inhibiting MYC-dependent transcription of oncogenes such as the oncogenic miR-17-92 miRNA cluster. P53 also enables miRNA processing through interactions with the primary miRNA processing enzyme Drosha. miRNA=microRNA.

miR-34

P38

DNA damage

miR-17-92 cluster

Carcinogenesis

P53

MYC

Drosha

miRNA processing

PTEN

AKT

Figure 3: Regulation of cancer progression at diff erent stages by microRNAs

Cancer stem cell Epithelial–mesenchymal transition Invasion

miR-27, miR-30, miR-200, miR-205, miR-429

miR-15a/16-1, miR-143, miR-145, miR-340

let-7, miR-30, miR-200

miR-9 miR-9, miR-181


Review

These two miRNAs are deleted or down regulated in 68% of patients with CLL.2

Furthermore, CLL can be classifi ed into fi ve diff erent categories according to genetic instabilities (deletion of chromosomal regions 11q, 13q, and 17p; trisomy 12; and normal karyotype). Studies show that the specifi c expression pattern of 32 individual miRNAs can diff er-entiate between these diff erent chromosomal abnor-malities (table).42 Other forms of genomic instability can be inferred with miRNA profi ling (table).

Prognostic indicatorsMany identifi ed genotypes are associated with distinct prognoses (table). However, some miRNA expression profi les show new and predictive prognostic power. Calin and colleagues41 reported that miRNA expression profi les in CLL can establish prognosis. By comparing two main groups of patients with good prognosis (low ZAP70 expression, mutated immunoglobulin variable region genes) and poor prognosis (high ZAP70, immunoglobulin variable region genes not mutated), they identifi ed 13 miRNAs with variable expression according to prognosis (table).41

In lung cancer, researchers focusing on the miR-183 family of miRNAs (miR-183, miR-182 and miR-96) have recorded a link between the expression of these miRNAs and progression of non-small-cell lung cancer. Expression of the miR-183 family can also help to identify tumours with heightened lymph-node metastasis; an increased likelihood of progression to stage II, III, or IV; and poor survival (table).44

Many biomarkers are prognostic indicators for breast cancer. The two most common are the oestrogen receptor, expression of which predicts a good outcome; and HER2, which when over expressed is associated with poor outcome, anti-oestrogen treatment resistance, and metastasis.53 Although identifi cation of these two markers has been invaluable for breast cancer management, other molecular methods are needed for classifi cation of subtypes. Analysis of 93 breast-cancer samples showed that specifi c patterns of miRNA expression were associated with breast-tumour subtypes of diff erent clinical outcomes.45 One investigation45 showed that expression of 31 miRNAs was signifi cantly associated with clinical factors. Over expression of 17 miRNAs was connected with oestrogen-receptor-positive stage I or II breast cancers with good clinical outcome. 14 were overexpressed in grade III, oestrogen-receptor-negative malignancies of basal-like subtype, seven of which were also associated with HER2-overexpressing cancers.45 Import antly, six were related to the same clinical outcomes in another independent dataset (table).46

Although the expression pattern of several miRNAs can give useful information about stage and prognosis, one miRNA alone can have accurate predictive power. A study46 of 328 human miRNAs screened for in patients with breast cancer identifi ed miR-210 as a strong prognostic biomarker. The predictive eff ect of miR-210 was so strong that the researchers noted that the overexpression of miR-210 alone allowed prediction of prognosis to the same level as a 76-gene mRNA signature test (GENE76). Overexpression of miR-210 is associated with an increased risk of recur-rence and a reduced chance of relapse-free survival.46 In CLL with trisomy 12, when expression of immuno globulin variable region genes and ZAP70 does not predict clinical outcome, the overexpression of one miRNA, miR-181, can predict disease progression (table).42

miRNA profi ling can identify treatment-resistant cancers. As with overexpression of HER2 in breast cancer, overexpression of four miRNAs have been linked to colorectal cancers resistant to EGFR antagonists.48 Screening for miRNAs could enable treatment to be tailored to individual patients and thus increase the chances of survival. Evidently, further investigation is necessary to confi rm the validity of miRNA patterns reported to be prognostic markers. However, the reproducibility of some previous studies gives hope that miRNA profi ling tests will be used in conjunction with gene-expression tests to provide more appropriate treatments.52

Prognosis or outcome

microRNA Expression Reference

Chronic lymphocytic leukaemia

High ZAP70; immunoglobulin variable region not mutated

Poor miR-15amir-16-1miR-16-2miR-195miR-221miR-23bmiR-155miR-24-1miR-146miR-223miR-29a-2miR-29b-2miR-29c

UpUpUpUpUpUpUpUpUpDownDownDownDown

Calin41

13q14.3 deletion Indolent miR-15amiR-16-1

DownDown

Calin2

17p deletion Aggressive miR-130bmiR-129-3pmiR-632miR-768-5pmiR-638miR-453miR-29b/cmiR-181b/c/dmiR-342-3pmiR-223miR-181amiR-367

UpUpUpUpUpUpDownDownDownDownDownDown

Visone42

Trisomy 12 Disease progression miR-181 Up Visone42

Lung cancer

Non-squamous non-small-cell lung cancer (stages II, III, and IV) and lymph-node metastasis

Poor miR-205miR-96miR-182miR-183

UpUpUpUp

Lebanony43

Zhu44

Zhu44

Zhu44

(Continues on next page)


Review

Biomarkers in body fl uidsAlthough miRNA profi ling of tumours could be an excellent prognostic test, invasive procedures such as biopsy, aspiration biopsy, or excision surgery of already-existing tumours are necessary to obtain samples. An ideal biomarker should be accessible with non-invasive methods, sensitive enough to detect early presence of tumours before clinical symptoms present, and absent or low in healthy, tumour-free individuals.54 Stable, degradation-resistant, extracellular miRNAs could be detected in body fl uids such as serum, urine, saliva, tears, breast milk, seminal fl uid, and faecal matter. These miRNAs are associated with various pathophysiological conditions.54

Although studies have suggested that circulating miRNAs could originate from non-specifi c release by lysed or necrotised cells,55 others have shown that miRNAs are actively transported outside of cells packaged as exosomes56 or so-called free miRNAs bound to Argonaut proteins.57 Intracellular and secreted miRNA populations are diff erent, suggesting a selective process is involved in extracellular miRNA secretion.58 Several investigations have shown that extracellular miRNAs play a major part in intercellular communication,56,59 and that they can contribute to carcinogenesis.60 Their stability, low complexity, and the availability of inexpensive methods of detection and profi ling could make extracellular miRNAs ideal bio markers for various diseases.54

Human plasma contains roughly 308 μg/L of RNA, in which about 350 diff erent miRNA sequences can be detected.54 The fi rst instance of detection of tumour-associated miRNAs in the serum was reported in patients with diff use large B-cell lymphoma.61 The serum level of three miRNAs was already known to be increased in these tumours (miR-155, miR-210, and miR-21), and the investigators linked high expression of miR-21 in the serum of patients with improved relapse-free survival.61

This study focused on the prognostic value of circulating miRNAs, but others are identifying extra cellular miRNA patterns that enable the detection of early-stage cancers. One breast cancer study comparing the circulating miRNA profi le in healthy women with that of those with early-stage breast cancer62 showed that downregulation of miR-181a and miR-1304 is associated with the disease. miR-195 and let-7a concentrations are higher in the plasma of patients with breast cancer than in healthy individuals.62 After tumour-excision surgery, postoperative serum concentrations of both these miRNAs return to values reported in the healthy control group.59 Researchers investigating the plasma levels of miRNAs in 74 patients diagnosed with diff erent types of lung cancer63 identifi ed three miRNAs (miR-155, miR-197, and miR-182) that were at a higher concentration in patients with cancer than in control individuals. The concentrations were not associated with any specifi c type of lung cancer or any particular stage, so the miRNAs are purely diagnostic biomarkers. The investigators also reported that the serum

concentrations of miR-155, miR-197 and miR-182 decreased after chemotherapy in most clinically responsive patients.63

Stable miRNAs are detectable in urine. Cystoscopy is the most sensitive method of screening for bladder cancer, but is quite invasive and can cause discomfort to the patients. Urine cytology of specifi c markers can also be used but this method does not have the sensitivity

Prognosis or outcome

microRNA Expression Reference

(Continued from previous page)

Breast cancer

Stages I and II; oestrogen-receptor-positive disease

Good miR-126*miR-136miR-100miR-99amiR-145miR-10amiR-199amiR-130amiR-30a-3pmiR-30a-5pmiR-224miR-214miR-342let-7a/b/c/f

UpUpUpUpUpUpUpUpUpUpUpUpUpUp

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45; Rothe46


Stage III; oestrogen-receptor-negative disease; basal-like

Poor miR-150miR-142-3pmiR-142-5pmiR-148amiR-106a/bmiR-18amiR-93miR-155miR-25miR-187miR-135b

UpUpUpUpUpUpUpUpUpUpUp

Blenkiron45

Blenkiron45

Blenkiron45




Blenkiron45


Blenkiron45

Blenkiron45

Blenkiron45

HER2+ Poor

High recurrence risk

miR-150miR-142-3pmiR-142-5pmiR-148amiR-106amiR-25miR-187miR-155miR-210

UpUpUpUpUpUpUpUpUp

Blenkiron45

Blenkiron45

Blenkiron45


Blenkiron45

Blenkiron45

Blenkiron45

Blenkiron45

Rothe46

Colorectal cancer

Lynch syndrome vs sporadic disease

·· miR-30a*miR-16-2*miR-362-5pmiR-1238miR-622

DownDownDownDownDown

Lanza47

Microsatellite instability vs microsatellite stable

Good miR-938miR-615-5pmiR-1184miR-551amiR-622miR-17-5pmiR-192*miR-337-3p

UpUpUpUpUpDownDownDown

Lanza47

KRAS mutated Poor miR-127-3pmiR-92amiR-486-3pmiR-378

UpUpUpDown

Mosakhani48

Table: microRNA prognostic indicators for diff erent cancers


Review

for detection of low-grade bladder cancer,64 driving the search for new specifi c and sensitive biomarkers. A proof-of-concept study65 showed that miR-126, miR-182 and miR-199a concentrations are signifi cantly increased in the urine of patients with stage 1 bladder cancer, as well as in those with stage 2 and stage 3 disease. Additionally, urine miRNA profi ling can identify miR-96 and miR-183 as biomarkers for urothelial carcinoma. The concentration of these markers in urine continues to increase as malignancy progresses and decreases after tumour-excision operations.65

For colorectal cancer, patients older than 60 years are off ered screening every 2 years with non-invasive faecal occult blood testing,66 with a subsequent colonoscopic investigation for those with a positive result. Unfortunately, the test has only 33–50% sensitivity; use of novel biomarkers would prevent many healthy individuals undergoing an expensive, painful procedure with the risk of serious medical complications.67 miR-21 and miR-106 are upregulated in the stool of patients with colorectal neoplasia (colorectal cancer or adenoma). miR-144* is increased in the faeces of patients with colorectal cancer with a sensitivity of 74% and a specifi city of 87%, and is thus another potential biomarker to help diagnosis of colorectal cancer.68 Samples extracted with the faecal blood-test kits could also be used to extract miRNAs; clinicians could undertake both tests simultaneously to increase the sensitivity of diagnosis.67

miRNAs as cancer treatmentThe development of new treatments has contributed substantially to increased 5-year survival and the reduction in overall mortality rates.62,69 However, although the classifi cation of cancers has become increasingly diver sifi ed, the variety and specifi city of treatment options has lagged behind. The same treatment is indicated for cancers with similar clinical phenotypes, even though substantial variation exists in the molecular phenotype that will aff ect treatment success. With the progress in profi ling, cancer treatments can now be customised for each individual. More than 2000 gene-therapy-based clinical trials are in progress for various illnesses, but only one is of miRNA treatment.32

Treatment can be targeted to miRNAs in two ways: miRNA reduction and miRNA replacement (fi gure 4). miRNA reduction treatment uses inactivating miRNAs that are upregulated or overexpressed in tumours. miRNA replacement treatment involves the reintro-duction of the miRNAs that are downregulated or deleted in the tumours. New miRNA-based treatments would need to have selective and accurate delivery of the agents to the target tumours to increase the therapeutic potential and reduce possible side-eff ects.

Two major diffi culties have impeded development of miRNA-based treatments. First, RNA has low stability in vivo; miRNA introduced into mice via the tail vein is cleared from the circulatory system in 30 min.70 Unmodifi ed,

saline-formulated, double-stranded RNA injected intra-venously undergoes RNase-mediated degradation and rapid renal excretion. Thus, the half-life of RNA needs to be increased for miRNA treatment to be eff ective. An increase could be achieved by improved miRNA stability or by protection from RNA-hostile environments.

Substitution of phosphodiester by phosphorothioate in the RNA backbone, and of the ribose moieties to 2 -O-methyl or other 2 substitutions confer substantial nuclease resistance.71 Locked nucleic acid (LNA) is a modifi ed RNA nucleotide that has an extra bridge con-necting the 2 oxygen and 4 carbon. These modifi cations increase the thermostability of LNA-RNA duplexes, increase target specifi city, and are resistant to exonucle-ases and endonucleases, thereby improving stability of miRNAs in vitro and in vivo.72 Finally, introduced RNA can be conjugated to a cholesterol moiety, increasing stability in the circulation. For example, an antagonist of miR-16 with 2 O-methyl modifi ed nucleotides and cholesterol linked to its end via a hydroxyprolinol linkage is stable and effi ciently silences miRNA expression.73 Protection from hostile environ ments is typically achieved by encasing of LNA or miRNA mimics in nanoparticles to form micelle-like structures (fi gure 4).

The second diffi culty is how to ensure tumour-specifi c delivery and retention of miRNAs. Because many mRNAs are targeted by one miRNA, off -target eff ects are likely to be substantial. Eff orts to achieve targeted delivery could be further hindered by fi rst-pass metabolism and rapid localisation of small molecules delivered systemically to the kidney and liver.74 Targeted delivery to specifi c tissues can be done when tumour-specifi c ligands are linked to nanoparticles, which can be directed to tumour cells via active or passive targeting. Passive targeting uses the size of nanoparticles and tumour vasculature properties to selectively deliver the load to specifi c cells. Tumour blood vessels have large pores (200 nm to 1·2 μm)75 compared with normal vessels, allowing nanoparticles to accumulate inside tumours. Nanoparticles between 15 nm and 100 nm are the best for systemic delivery.76

Active targeting of nanoparticles necessitates their conjugation with diff erent compounds that have a specifi c affi nity to tumours. Various cancer-associated cell-surface proteins (eg, HER2,77 EGFR,78 and CA-12579) and hyaluronic acid80 could potentially be used for this conjugation. Hyaluronic acid is a polysaccharide that binds to the cancer-stem-cell marker CD44, which is overexpressed in various tumour cells. Investigators have developed a poly(ethylene glycol) hyaluronic acid nanoparticle (P-HA-NP) that can be used to deliver doxorubicin and camptothecin to mouse cells.81 The nanoparticles were internalised successfully into cancer cells (SCC7 and MDA-3T3), but rarely taken up by normal fi broblasts (NIH-3T3). Tumour-bearing mice were systemically treated with camptothecin de livered with P-HA-NP, and tumour growth was success fully stopped for at least 35 days. These nanoparticles could


Review

possibly be modifi ed to carry miRNAs to specifi cally targeted cancer stem cells.81 Furthermore, integrin α-ν-β-3-targeted nanoparticles used to deliver anti-miR-132 had promising results in inhibition of angiogenesis and breast tumour metastasis.82

The goal of miRNA replacement is the reintroduction of miRNAs depleted in cancer cells and the reactivation of cellular pathways that drive a therapeutic response. Trang and co-workers70 concluded that a chemically synthesised miR-34a packaged into a lipid-based delivery vehicle and given locally or systemically could block tumour growth in mouse models of non-small-cell lung cancer. miR-34a accumulated in the tumour tissue, resulting in down regulation of its direct targets. Systemic delivery of formulated miR-34a did not induce a rise in cytokines or liver and kidney enzymes in serum, suggesting that the formulation is well tolerated and that side-eff ects could be negligible. Similar results were noted in another study of a mouse model of non-small-cell lung cancer with activated KRAS;70 an miR-34a or let-7 mimic miRNA formed a complex with a neutral lipid emulsion, which led to a 60% reduction in tumour area.70 Additionally, systemic delivery of atelocollagen-conjugated miR-16 in a mouse xenograft model of prostate cancer inhibited metastasis into bone.83

miRNA reduction with LNA has been done with anti-miR-21 to treat autoimmune splenomegaly in mice with systemic lupus erythematosus;84 miR-21 is upregulated in all subsets of lupus mouse lymphocytes. Targeting of miR-21 with an intraperitoneally injected LNA-based antagonist decreases manifestation of cardinal systemic lupus erythematosus and leads to the reversal of spleno-megaly.84 The only miRNA-based treatment tested in people is anti-miR-122 for the treatment of infection with hepatitis C virus (HCV). miR-122 is an essential miRNA for HCV replication and is predominantly in the liver.85 In 2010, data from a drug trial of an intravenously delivered LNA to remove miR-122 and stop infection in chimpanzees was reported.86 LNA given to chronically infected chim panzees once a week for 12 weeks led to a reduction in viral load in the serum and the liver. The decline in the liver was sustained for 13 weeks after withdrawal of treatment. The liver histological status improved greatly, with decreased disruption of hepatocellular sinuses and cords, and interferon-regulated genes were down regulated. No resistance or side-eff ects were noted in the treated animals.86

These results led to the fi rst clinical trials of miRNA-based treatments in people. A phase 1 trial87 in 77 healthy volunteers established that anti-miR-122 is safe and

Figure 4: miRNA-based treatmentIn miRNA reduction treatment (A), single-stranded LNA molecules (anti-miRNAs) bind to miRNAs complementarily, preventing the miRNAs from binding to target mRNAs. In miRNA replacement treatment (B), tumour suppressive miRNAs that are lost during carcinogenesis are reintroduced by use of a miRNA mimic. These double-stranded miRNA mimics can either be modifi ed on the complementary strand or encapsulated in nanoparticles to increase their stability. The delivery of LNA and miRNA mimics can be improved with nanoparticles conjugated to antibodies or cancer-specifi c ligands. LNA=locked nucleic acid.

3´ 5´

3´ 5´3´ 5´

AAAAAA3´ 5´

AAAAAA3´ 5´

3´ 5´

3´ 5´

3´ 5´

3´ 5´3´ 5´

3´ 5´3´ 5´

3´ 5´3´ 5´

B

A

Cell membrane

LNAmiRNAModified RNATarget mRNANanoparticleAntibodyCancer-specific ligands


Review

identifi ed no dose-limiting toxicities. The subsequent phase 2a trial87 assessed the safety and antiviral activity of subcutaneous anti-miR-122 given at doses of 3 mg/kg, 5 mg/kg, and 7 mg/kg every week for 29 days; patients were followed up until week 18. A dose-dependent, long reduction in HCV RNA that continued to fall after completion of treatment without any serious adverse eff ects was recorded. Notably, one participant had undetectable concentrations of HCV RNA 10 weeks after the last dose. This trial87 drew attention to the potential of LNAs for cancer treatment.

ConclusionThe discovery of miRNAs has changed the way control of gene expression is thought about and, perhaps more importantly, has set a precedent for the development of new diagnostic methods and treatments of diseases, including malignancies. In the short term, the validation of prog nostic and diagnostic miRNA panels in large cohort studies would enable their introduction into the clinic setting. Tailoring of treatment regimens to specifi c cancers would maximise the likelihood of success. In the long term, work to identify major miRNA targets and to develop safe and specifi c methods of delivery of miRNA-based treatments could allow modulation of miRNAs to become a central feature of cancer treatment and management.

ContributorsYWK and DF-M did the literature search, wrote and edited the report,

and designed the fi gures. TJJ edited the report. MB wrote and edited

the report.

Confl icts of interestYWK and MB have fi led a patent relating to microRNA delivery with

nanoparticles conjugated to cancer-targeting anti-CD4 antibodies. The

other authors declare that they have no confl icts of interest.

AcknowledgmentsMB is funded by the Medical Research Council. We thank

Ania Wilczynska and Samantha Johnston for reading and commenting

on this Review.

References1 Zhang B, Pan X, Cobb GP, Anderson TA. microRNAs as oncogenes

and tumor suppressors. Dev Biol 2007; 302: 1–12.

2 Calin GA, Dumitru CD, Shimizu M, et al. Frequent deletions and down-regulation of micro-RNA genes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia. Proc Natl Acad Sci USA 2002; 99: 15524–29.

3 Cimmino A, Calin GA, Fabbri M, et al. miR-15 and miR-16 induce apoptosis by targeting BCL2. Proc Natl Acad Sci USA 2005; 102: 13944–49.

4 Si ML, Zhu S, Wu H, Lu Z, Wu F, Mo YY. miR-21-mediated tumor growth. Oncogene 2007; 26: 2799–803.


Data for this Review were identifi ed by searches of PubMed with the term “microRNA” in combination with “cancer”, “in vivo”, “circulating”, “systemic delivery”, “therapeutics”, or “LNA”. Reference lists of reports identifi ed were also searched. Only articles published in English and relevant to the Review were included. We used no date restrictions.

5 Iorio MV, Ferracin M, Liu CG, et al. MicroRNA gene expression deregulation in human breast cancer. Cancer Res 2005; 65: 7065–70.

6 Chan JA, Krichevsky AM, Kosik KS. MicroRNA-21 is an antiapoptotic factor in human glioblastoma cells. Cancer Res 2005; 65: 6029–33.

7 Schetter AJ, Leung SY, Sohn JJ, et al. MicroRNA expression profi les associated with prognosis and therapeutic outcome in colon adenocarcinoma. JAMA 2008; 299: 425–36.

8 Medina PP, Nolde M, Slack FJ. OncomiR addiction in an in vivo model of microRNA-21-induced pre-B-cell lymphoma. Nature 2010; 467: 86–90.

9 Gabriely G, Wurdinger T, Kesari S, et al. MicroRNA 21 promotes glioma invasion by targeting matrix metalloproteinase regulators. Mol Cell Biol 2008; 28: 5369–80.

10 Darido C, Georgy SR, Wilanowski T, et al. Targeting of the tumor suppressor GRHL3 by a miR-21-dependent proto-oncogenic network results in PTEN loss and tumorigenesis. Cancer Cell 2011; 20: 635–48.

11 Sarkar J, Gou D, Turaka P, Viktorova E, Ramchandran R, Raj JU. MicroRNA-21 plays a role in hypoxia-mediated pulmonary artery smooth muscle cell proliferation and migration. Am J Physiol Lung Cell Mol Physiol 2010; 299: L861–71.

12 O’Donnell KA, Wentzel EA, Zeller KI, Dang CV, Mendell JT. c-Myc-regulated microRNAs modulate E2F1 expression. Nature 2005; 435: 839–43.

13 Olive V, Bennett MJ, Walker JC, et al. miR-19 is a key oncogenic component of mir-17-92. Genes Dev 2009; 23: 2839–49.

14 Chang TC, Zeitels LR, Hwang HW, et al. Lin-28B transactivation is necessary for Myc-mediated let-7 repression and proliferation. Proc Natl Acad Sci USA 2009; 106: 3384–49.

15 Boyerinas B, Park SM, Hau A, Murmann AE, Peter ME. The role of let-7 in cell diff erentiation and cancer. Endocr Relat Cancer 2010; 17: F19–36.

16 Chang TC, Wentzel EA, Kent OA, et al. Transactivation of miR-34a by p53 broadly infl uences gene expression and promotes apoptosis. Mol Cell 2007; 26: 745–52.

17 Corney DC, Flesken-Nikitin A, Godwin AK, Wang W, Nikitin AY. MicroRNA-34b and microRNA-34c are targets of p53 and cooperate in control of cell proliferation and adhesion-independent growth. Cancer Res 2007; 67: 8433–38.

18 Watahiki A, Wang Y, Morris J, et al. MicroRNAs associated with metastatic prostate cancer. PLoS One 2011; 6: e24950.

19 Liu C, Kelnar K, Liu B, et al. The microRNA miR-34a inhibits prostate cancer stem cells and metastasis by directly repressing CD44. Nat Med 2011; 17: 211–15.

20 Toyota M, Suzuki H, Sasaki Y, et al. Epigenetic silencing of microRNA-34b/c and B-cell translocation gene 4 is associated with CpG island methylation in colorectal cancer. Cancer Res 2008; 68: 4123–32.

21 Kress TR, Cannell IG, Brenkman AB, et al. The MK5/PRAK kinase and Myc form a negative feedback loop that is disrupted during colorectal tumorigenesis. Mol Cell 2011; 41: 445–57.

22 Cannell IG, Kong YW, Johnston SJ, et al. p38 MAPK/MK2-mediated induction of miR-34c following DNA damage prevents Myc-dependent DNA replication. Proc Natl Acad Sci USA 2010; 107: 5375–80.

23 Chin LJ, Ratner E, Leng S, et al. A SNP in a let-7 microRNA complementary site in the KRAS 3 untranslated region increases non-small cell lung cancer risk. Cancer Res 2008; 68: 8535–40.

24 Godshalk SE, Paranjape T, Nallur S, et al. A variant in a microRNA complementary site in the 3 UTR of the KIT oncogene increases risk of acral melanoma. Oncogene 2011; 30: 1542–50.

25 Soltanian S, Matin MM. Cancer stem cells and cancer therapy. Tumour Biol 2011; 32: 425–40.

26 Takebe N, Harris PJ, Warren RQ, Ivy SP. Targeting cancer stem cells by inhibiting Wnt, Notch, and Hedgehog pathways. Nat Rev Clin Oncol 2011; 8: 97–106.

27 Shimono Y, Zabala M, Cho RW, et al. Downregulation of miRNA-200c links breast cancer stem cells with normal stem cells. Cell 2009; 138: 592–603.

28 Yu F, Deng H, Yao H, Liu Q, Su F, Song E. Mir-30 reduction maintains self-renewal and inhibits apoptosis in breast tumor-initiating cells. Oncogene 2010; 29: 4194–204.


Review

29 Hua Z, Lv Q, Ye W, et al. MiRNA-directed regulation of VEGF and other angiogenic factors under hypoxia. PLoS One 2006; 1: e116.

30 Larsson E, Fredlund Fuchs P, Heldin J, et al. Discovery of microvascular miRNAs using public gene expression data: miR-145 is expressed in pericytes and is a regulator of Fli1. Genome Med 2009; 1: 108.

31 Wang X, Hu G, Zhou J. Repression of versican expression by microRNA-143. J Biol Chem 2010; 285: 23241–50.

32 Yue J. miRNA and vascular cell movement. Adv Drug Deliv Rev 2011; 63: 616–22.

33 Gavert N, Ben-Ze’ev A. Epithelial-mesenchymal transition and the invasive potential of tumors. Trends Mol Med 2008; 14: 199–209.

34 Gregory PA, Bert AG, Paterson EL, et al. The miR-200 family and miR-205 regulate epithelial to mesenchymal transition by targeting ZEB1 and SIP1. Nat Cell Biol 2008; 10: 593–601.

35 Ozcan S. MiR-30 family and EMT in human fetal pancreatic islets. Islets 2009; 1: 283–85.

36 Chen J, Wang L, Matyunina LV, Hill CG, McDonald JF. Overexpression of miR-429 induces mesenchymal-to-epithelial transition (MET) in metastatic ovarian cancer cells. Gynecol Oncol 2011; 121: 200–05.

37 Sotiriou C, Wirapati P, Loi S, et al. Gene expression profi ling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis. J Natl Cancer Inst 2006; 98: 262–72.

38 Kim C, Paik S. Gene-expression-based prognostic assays for breast cancer. Nat Rev Clin Oncol 2010; 7: 340–47.

39 Calin GA, Croce CM. MicroRNA-cancer connection: the beginning of a new tale. Cancer Res 2006; 66: 7390–94.

40 Gusev Y, Brackett DJ. MicroRNA expression profi ling in cancer from a bioinformatics prospective. Expert Rev Mol Diagn 2007; 7: 787–92.

41 Calin GA, Ferracin M, Cimmino A, et al. A microRNA signature associated with prognosis and progression in chronic lymphocytic leukemia. N Engl J Med 2005; 353: 1793–801.

42 Visone R, Rassenti LZ, Veronese A, et al. Karyotype-specifi c microRNA signature in chronic lymphocytic leukemia. Blood 2009; 114: 3872–79.

43 Lebanony D, Benjamin H, Gilad S, et al. Diagnostic assay based on hsa-miR-205 expression distinguishes squamous from nonsquamous non-small-cell lung carcinoma. J Clin Oncol 2009; 27: 2030–37.

44 Zhu W, Liu X, He J, Chen D, Hunag Y, Zhang YK. Overexpression of members of the microRNA-183 family is a risk factor for lung cancer: a case control study. BMC Cancer 2011; 11: 393.

45 Blenkiron C, Goldstein LD, Thorne NP, et al. MicroRNA expression profi ling of human breast cancer identifi es new markers of tumor subtype. Genome Biol 2007; 8: R214.

46 Rothe F, Ignatiadis M, Chaboteaux C, et al. Global microRNA expression profi ling identifi es MiR-210 associated with tumor proliferation, invasion and poor clinical outcome in breast cancer. PLoS One 2011; 6: e20980.

47 Lanza G, Ferracin M, Gafa R, et al. mRNA/microRNA gene expression profi le in microsatellite unstable colorectal cancer. Mol Cancer 2007; 6: 54.

48 Mosakhani N, Sarhadi VK, Borze I, et al. MicroRNA profi ling diff erentiates colorectal cancer according to KRAS status. Genes Chromosomes Cancer 2012; 51: 1–9.

49 Szafranska AE, Doleshal M, Edmunds HS, et al. Analysis of microRNAs in pancreatic fi ne-needle aspirates can classify benign and malignant tissues. Clin Chem 2008; 54: 1716–24.

50 Karim BO, Ali SZ, Landolfi JA, et al. Cytomorphologic diff erentiation of benign and malignant mammary tumors in fi ne needle aspirate specimens from irradiated female Sprague-Dawley rats. Vet Clin Pathol 2008; 37: 229–36.

51 Mazeh H, Mizrahi I, Halle D, et al. Development of a microRNA-based molecular assay for the detection of papillary thyroid carcinoma in aspiration biopsy samples. Thyroid 2011; 21: 111–18.

52 Calin GA, Sevignani C, Dumitru CD, et al. Human microRNA genes are frequently located at fragile sites and genomic regions involved in cancers. Proc Natl Acad Sci USA 2004; 101: 2999–3004.

53 Freudenberg JA, Wang Q, Katsumata M, Drebin J, Nagatomo I, Greene MI. The role of HER2 in early breast cancer metastasis and the origins of resistance to HER2-targeted therapies. Exp Mol Pathol 2009; 87: 1–11.

54 Weber JA, Baxter DH, Zhang S, et al. The microRNA spectrum in 12 body fl uids. Clin Chem 2010; 56: 1733–41.

55 Wang K, Zhang S, Marzolf B, et al. Circulating microRNAs, potential biomarkers for drug-induced liver injury. Proc Natl Acad Sci USA 2009; 106: 4402–07.

56 Valadi H, Ekstrom K, Bossios A, Sjostrand M, Lee JJ, Lotvall JO. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol 2007; 9: 654–59.

57 Arroyo JD, Chevillet JR, Kroh EM, et al. Argonaute2 complexes carry a population of circulating microRNAs independent of vesicles in human plasma. Proc Natl Acad Sci USA 2011; 108: 5003–08.

58 Pigati L, Yaddanapudi SC, Iyengar R, et al. Selective release of microRNA species from normal and malignant mammary epithelial cells. PLoS One 2010; 5: e13515.

59 Heneghan HM, Miller N, Lowery AJ, Sweeney KJ, Newell J, Kerin MJ. Circulating microRNAs as novel minimally invasive biomarkers for breast cancer. Ann Surg 2010; 251: 499–505.

60 Kogure T, Lin WL, Yan IK, Braconi C, Patel T. Intercellular nanovesicle-mediated microRNA transfer: a mechanism of environmental modulation of hepatocellular cancer cell growth. Hepatology 2011; 54: 1237–48.

61 Lawrie CH, Gal S, Dunlop HM, et al. Detection of elevated levels of tumour-associated microRNAs in serum of patients with diff use large B-cell lymphoma. Br J Haematol 2008; 141: 672–75.

62 Zhao H, Shen J, Medico L, Wang D, Ambrosone CB, Liu S. A pilot study of circulating miRNAs as potential biomarkers of early stage breast cancer. PLoS One 2010; 5: e13735.

63 Zheng D, Haddadin S, Wang Y, et al. Plasma microRNAs as novel biomarkers for early detection of lung cancer. Int J Clin Exp Pathol 2011; 4: 575–86.

64 Hanke M, Hoefi g K, Merz H, et al. A robust methodology to study urine microRNA as tumor marker: microRNA-126 and microRNA-182 are related to urinary bladder cancer. Urol Oncol 2010; 28: 655–61.

65 Yamada Y, Enokida H, Kojima S, et al. MiR-96 and miR-183 detection in urine serve as potential tumor markers of urothelial carcinoma: correlation with stage and grade, and comparison with urinary cytology. Cancer Sci 2011; 102: 522–29.

66 National Health Service. NHS bowel cancer screening programme. http://www.cancerscreening.nhs.uk/bowel/index.html (accessed March 5, 2012).

67 Link A, Balaguer F, Shen Y, et al. Fecal microRNAs as novel biomarkers for colon cancer screening. Cancer Epidemiol Biomarkers Prev 2010; 19: 1766–74.

68 Kalimutho M, Del Vecchio Blanco G, Di Cecilia S, et al. Diff erential expression of miR-144* as a novel fecal-based diagnostic marker for colorectal cancer. J Gastroenterol 2011; 46: 1391–402.

69 Berrino F, De Angelis R, Sant M, et al. Survival for eight major cancers and all cancers combined for European adults diagnosed in 1995–99: results of the EUROCARE-4 study. Lancet Oncol 2007; 8: 773–83.

70 Trang P, Wiggins JF, Daige CL, et al. Systemic delivery of tumor suppressor microRNA mimics using a neutral lipid emulsion inhibits lung tumors in mice. Mol Ther 2011; 19: 1116–22.

71 Jensen TB, Langkjaer N, Wengel J. Unlocked nucleic acid (UNA) and UNA derivatives: thermal denaturation studies. Nucleic Acids Symp Ser (Oxf) 2008: 52: 133–34.

72 Vester B, Wengel J. LNA (locked nucleic acid): high-affi nity targeting of complementary RNA and DNA. Biochemistry 2004; 43: 13233–41.

73 van Solingen C, Seghers L, Bijkerk R, et al. Antagomir-mediated silencing of endothelial cell specifi c microRNA-126 impairs ischemia-induced angiogenesis. J Cell Mol Med 2009; 13: 1577–85.

74 Nishina K, Unno T, Uno Y, et al. Effi cient in vivo delivery of siRNA to the liver by conjugation of alpha-tocopherol. Mol Ther 2008; 16: 734–40.

75 Maeda H, Fang J, Inutsuka T, Kitamoto Y. Vascular permeability enhancement in solid tumor: various factors, mechanisms involved and its implications. Int Immunopharmacol 2003; 3: 319–28.

76 Brannon-Peppas L, Blanchette JO. Nanoparticle and targeted systems for cancer therapy. Adv Drug Deliv Rev 2004; 56: 1649–59.


Review

77 Wu X, Liu H, Liu J, et al. Immunofl uorescent labeling of cancer marker Her2 and other cellular targets with semiconductor quantum dots. Nat Biotechnol 2003; 21: 41–46.

78 Nida DL, Rahman MS, Carlson KD, Richards-Kortum R, Follen M. Fluorescent nanocrystals for use in early cervical cancer detection. Gynecol Oncol 2005; 99 (suppl 1): S89–94.

79 Wang HZ, Wang HY, Liang RQ, Ruan KC. Detection of tumor marker CA125 in ovarian carcinoma using quantum dots. Acta Biochim Biophys Sin (Shanghai) 2004; 36: 681–86.

80 Choi KY, Yoon HY, Kim JH, et al. Smart nanocarrier based on PEGylated hyaluronic acid for cancer therapy. ACS Nano 2011; 5: 8591–99.

81 Eliaz RE, Szoka FC Jr. Liposome-encapsulated doxorubicin targeted to CD44: a strategy to kill CD44-overexpressing tumor cells. Cancer Res 2001; 61: 2592–601.

82 Anand S, Majeti BK, Acevedo LM, et al. MicroRNA-132-mediated loss of p120RasGAP activates the endothelium to facilitate pathological angiogenesis. Nat Med 2010; 16: 909–14.

83 Takeshita F, Patrawala L, Osaki M, et al. Systemic delivery of synthetic microRNA-16 inhibits the growth of metastatic prostate tumors via downregulation of multiple cell-cycle genes. Mol Ther 2010; 18: 181–87.

84 Garchow BG, Encinas OB, Leung YT, et al. Silencing of microR6-21 in vivo ameliorates autoimmune splenomegaly in lupus mice. EMBO Mol Med 2011; 3: 605–15.

85 Jopling CL, Yi M, Lancaster AM, Lemon SM, Sarnow P. Modulation of hepatitis C virus RNA abundance by a liver-specifi c microRNA. Science 2005; 309: 1577–81.

86 Lanford RE, Hildebrandt-Eriksen ES, Petri A, et al. Therapeutic silencing of microRNA-122 in primates with chronic hepatitis C virus infection. Science 2010; 327: 198–201.

87 Janssen HL, Reesink HW, Zeuzem S, et al. A randomized, double-blind, placebo (Plb) controlled safety and anti-viral proof of concept study of miravirsen (Mir), an oligonucleotide targeting Mir-122, in treatment naive patients with genotype 1 (Gt1) chronic Hcv infection. Hepatology 2011; 54 (suppl S1): 1430A (abstr).


Personal View

Optimising radiation treatment decisions for patients who receive neoadjuvant chemotherapy and mastectomyKaren E Hoff man, Elizabeth A Mittendorf, Thomas A Buchholz

Whereas randomised clinical trials have established which patients might benefi t from postmastectomy radiation therapy after upfront surgery, no such data exist for guiding decisions on who might benefi t from postmastectomy radiation therapy after upfront chemotherapy. Insight must be drawn from non-randomised data to provide such guidance. Early data suggest that both extent of disease at presentation and response to neoadjuvant chemotherapy predict the risk of locoregional recurrence, and can be used to tailor recommendations for postmastectomy radiation therapy. Randomised clinical trial data are needed to assess whether postmastectomy radiation therapy can be safely omitted in selected women with good response to neoadjuvant chemotherapy.

IntroductionNeoadjuvant chemotherapy is widely used in the treatment of locally advanced, operable breast cancer and is increasingly used for treatment of women with early stage breast cancer. Randomised clinical trials have established that neoadjuvant chemotherapy given before surgical removal of breast cancer provides equivalent survival outcomes to chemotherapy after upfront surgical resection.1 Neoadjuvant chemotherapy allows physicians to observe tumour response and modify chemotherapy plans if necessary. Additionally, administration of chemo-therapy before surgery decreases tumour size, thereby increasing the proportion of women who can undergo breast-conserving surgery rather than mast ectomy. However, many women who receive neo adjuvant chemo-therapy undergo mastectomy, either by patient choice or because breast-conserving surgery is not feasible. In this Personal View, we review the literature assessing radiation therapy after neoadjuvant chemo therapy and mast ectomy, and identify subgroups of women who might benefi t from such therapy.

Postmastectomy radiation therapy after upfront surgery improves overall survival Administration of postmastectomy radiation therapy to appropriately selected women reduces locoregional recurrence and improves breast-cancer survival.2 The landmark trials establishing the benefi t of postmast-ectomy radiation therapy for women with high-risk pathological features were designed and completed before the widespread adoption of neoadjuvant chemotherapy into clinical practice. Women enrolled in Danish3,4 and Canadian5 randomised trials established the benefi t of post mastectomy radiation therapy; they underwent upfront, modifi ed radical mastectomy and received adjuvant chemotherapy (cyclophosphamide, metho trexate, and fl uorouracil) or tamoxifen. The benefi t of post mastectomy radiation for patients with positive lymph nodes who received upfront surgery and adjuvant chemotherapy was confi rmed in a meta-analysis by the Early Breast Cancer Trialists’ Collaborative Group,6 which included data on 9933 women from 25 trials comparing adjuvant chemotherapy and mastectomy with and

without postmastectomy radiation therapy. Admin is-tration of postmastectomy radiation therapy for patients with positive lymph nodes led to a 17% reduction in isolated locoregional recurrence at 5 years (22·8% vs 5·8%), which resulted in a 5·4% reduction in breast-cancer mortality at 15 years (60·1% vs 54·7%, p=0·0002).6 This benefi t was not seen among patients with negative lymph nodes, in whom postmastectomy radiation therapy resulted in only a 4·0% reduction in 5-year locoregional recurrence (6·3% vs 2·3%), and did not improve 15-year breast-cancer mortality (31·1% vs 27·7%, p=0·18). Therefore, after upfront mastectomy, post-mastectomy radiation therapy improves locoregional control and breast-cancer survival for women with a clinically signifi cant risk of locoregional recurrence, but is not necessary for women with low risk of recurrence.

Analysis of data from prospective randomised trials has defi ned those women most likely to benefi t from radiation therapy after upfront mastectomy and adjuvant chemotherapy. As outlined in consensus treatment guidelines, it is generally accepted that women with four or more positive lymph nodes at resection or pathological stage III disease benefi t from postmastectomy radiation therapy.7–10 Opinions vary as to the benefi t of post-mastectomy radiation therapy in women who undergo upfront mastectomy for tumour (T) stage 1 or 2 tumours with one to three involved lymph nodes.7–9 However, there is mounting evidence supporting postmastectomy radiation therapy for women with one to three involved lymph nodes,11–13 especially those with adverse patho-logical features that increase the risk of locoregional recurrence, such as lymphovascular space invasion and high-grade disease.14–16

The magnitude of locoregional recurrence risk that is necessary to derive benefi t from postmastectomy radiation therapy is a matter of debate; however, a reanalysis of women enrolled in Danish randomised trials suggests that women with a 5-year risk of locoregional recurrence as low as 11% might obtain a survival benefi t from postmastectomy radiation therapy.17 In this subgroup of women with good-prognosis disease, postmastectomy radiation provided an 11% improvement in 5-year locoregional failure (11% without vs 0% with;


Department of Radiation Oncology (K E Hoff man MD, Prof T A Buchholz MD) and Department of Surgical Oncology (E A Mittendorf MD), University of Texas MD Anderson Cancer Center, Houston, TX, USA

Correspondence to: Dr Karen Hoff man, Department of Radiation Oncology, University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 1202, Houston, TX 77030, USAkhoff [email protected]


Personal View

p=0·0004), which resulted in an 11% improvement in 15-year breast-cancer survival (67% without vs 78% with; p=0·06). Detailed discussion of selection of patients for postmastectomy radiation therapy after upfront mastectomy is beyond the scope of this Personal View; however, awareness of the clinical context is important, in view of the limited data available to guide use of postmastectomy radiation therapy after neoadjuvant chemotherapy and surgical resection.

Neoadjuvant chemotherapy modifi es pathological extent of diseaseThe key information used to discern the need for radiation therapy after upfront surgery—pathological extent of disease at resection—is altered by administration of neoadjuvant chemotherapy. Pathological extent of disease is modifi ed in 80–90% of patients who receive neoadjuvant chemotherapy.18 Generally, 20–40% of women with lymph-node positive disease are converted to lymph-node negative disease after neoadjuvant chemo-therapy,19 and this percentage might be sub stan tially higher in women with HER2-overexpressing tumours who receive trastuzumab-based neoadjuvant chemo-therapy. One series reported that 74% of patients with positive lymph nodes at presentation converted to lymph-node negative disease after trastuzumab-based neo-adjuvant chemotherapy.20

The risk of locoregional recurrence, based on pathological extent of disease at resection, is diff erent after neoadjuvant chemotherapy than after upfront surgery. A retrospective analysis of patients who received mastectomy and no radiation therapy at MD Anderson Cancer Center (Houston, TX, USA) from 1974 to 1988 compared locoregional recurrence in 150 patients who received neoadjuvant chemotherapy with 1031 women who received upfront surgery. Across all pathological tumour sizes at resection, locoregional recurrence rates were higher in patients who received neoadjuvant chemo therapy than in those who had upfront surgery.21 For example, in patients with tumour size 2·1–5 cm at resection, 5-year locoregional recurrence was 36% for patients in the neoadjuvant group (in whom this tumour size represented residual disease after chemotherapy) compared with only 15% for those who had upfront surgery (p<0·001). Similarly, patients with four or more involved lymph nodes at the time of surgery were more likely to experience 5-year locoregional recurrence if they had neoadjuvant chemotherapy before surgery than if they underwent upfront surgery (53% vs 23%, p<0·001).

Since neoadjuvant chemotherapy modifi es the pathological extent of disease, it is essential to accurately document the clinical extent of disease in the breast and lymph-node basins before chemotherapy. At MD Anderson Cancer Center, all of our patients are assessed with clinical examination, breast mammography, and ultra sonography of the breast and regional nodal basins

(axillary, infraclavicular, supraclavicular, and internal mammary chain). Fine-needle aspiration biopsy is performed on suspicious lymph nodes to confi rm nodal involvement. Taken together, these studies allow determination of the extent of local and regional disease. Selected patients with extensive nodal involvement also undergo cross-sectional imaging to assess for distant metastasis. For patients with clinically involved axillary lymph nodes, the current standard of practice is to perform axillary dissection after neoadjuvant chemo-therapy. Patients with clinically negative axillary lymph nodes are candidates for sentinel lymph-node biopsy (SLNB). Opinions vary regarding the timing of SLNB relative to administration of neoadjuvant chemotherapy. We perform SLNB after completion of neoadjuvant chemotherapy, with excellent sentinel lymph-node (SLN) identifi cation rate (97·4%) and low false-negative rate (5·7%).22 Additionally, in patients presenting with clinical T2N0 or T3N0 tumours, if SLNB is done after neoadjuvant chemotherapy, fewer patients have a positive SLN and fewer undergo completion axillary lymph-node dissection. Nodal response to neoadjuvant chemo therapy also has important prognostic information, including the risk of locoregional recurrence, which can inform post-mast ectomy radiation treatment decisions.

Available data to guide postmastectomy treatment decisionsAlthough randomised clinical trials have established which patients might benefi t from postmastectomy radiation therapy after upfront surgery, there are no randomised trial data comparing outcomes with and without post mast ectomy radiation therapy after neoadjuvant chemo therapy and mastectomy. Instead, insights from non-randomised data provide guidance on who might benefi t from post mast ectomy radiation therapy in this setting, and who should be spared the morbidity of radiation treatment. Most of the retrospective data assessing outcomes with and without postmastectomy radiation therapy are from patients treated at MD Anderson.21,23–29 Most patients included in these analyses were treated on prospective trials investigating chemotherapy regimens and received doxorubicin-based neoadjuvant chemo therapy, but sub-sequent administration of adjuvant cytotoxic chemo-therapy and tamoxifen was not consistent across the cohort. When interpreting these studies, it should be remembered that some patients are included in several publications, and that there have been advances over time in clinical staging, radiographic imaging, chemo-therapeutic agents, and surgical techniques that could aff ect locoregional control rates. Information is also available on outcomes of patients with operable breast cancer treated on the NSABP B-18 and B-27 studies. NSABP B-18 compared preoperative versus post-operative chemotherapy with doxorubicin and cyclo-phosphamide, and concluded that preoperative


Personal View

chemotherapy provided equivalent disease-free and overall survival.18,30 NSABP B-27 reported that the addition of docetaxel to preoperative doxorubicin and cyclo phosphamide improved disease-free and overall survival.18,31 Patients enrolled in the NSABP studies who underwent mastectomy were prohibited from receiving radiation therapy, and therefore provide insight on locoregional recurrence risk for patients who do not receive postmastectomy radiation therapy.19,32 Patients in the NSABP trials generally had earlier stage tumours than patients in the MD Anderson retrospective studies. Only 43% of the 1070 patients enrolled in the preoperative chemotherapy group of the NSABP trials who underwent mastectomy had clinical tumour size greater than 5 cm, and only 35% had clinical lymph-node involvement at diagnosis.32 Potential diff erences in distribution of tumour stage and how the patients were followed (retrospective analysis vs prospectively followed on a clinical trial) should be considered when interpreting the study results and comparing outcomes between the MD Anderson and NSABP reports.

Subgroups of patients with high locoregional recurrence risk after upfront mastectomyTo benefi t from postmastectomy radiation therapy, there must be a clinically signifi cant risk of locoregional recurrence without radiation therapy. An analysis of 150 women from prospective trials at MD Anderson from 1974 to 1998 who received neoadjuvant chemotherapy and mastectomy without postmastectomy radiation found that both clinical extent of disease before and pathological extent of disease after neoadjuvant chemo therapy aff ect the risk of locoregional recurrence.23 Women with more advanced T stage before chemotherapy, more advanced combined clinical stage before chemo therapy, larger tumour size after chemotherapy, and increasing number of involved lymph nodes after chemo therapy had higher rates of locoregional recurrence. Consideration of individual subgroups of women showed that women with clinical stage IIB or higher disease at presentation (American Joint Committee on Cancer [AJCC] 1988 staging), women with pathologically involved lymph nodes after neoadjuvant chemotherapy, and women with pathologically negative lymph nodes but clinical T3 or T4 disease at presentation generally had at least a 15% cumulative risk of 5-year locoregional recurrence (table)—a risk threshold that would be considered high enough by many clinicians to off er postmastectomy radiation therapy. Similar to the MD Anderson results, patients on the NSABP B-18 or B-27 trials who received neoadjuvant chemotherapy, mastectomy, and no radiation had higher rates of locoregional recurrence if they presented with more advanced T stage (>5 cm vs <5 cm) or nodal involvement before chemotherapy.32 NSABP patients with lymph-node involvement after neoadjuvant chemo therapy had a 14·9% cumulative risk of 8-year locoregional recurrence (fi gure 1).19

Subgroups of patients who benefi t from postmastectomy radiation therapyTo confer benefi t, not only must there be a clinically signifi cant risk of locoregional recurrence, but also post-mastectomy radiation therapy must decrease this risk. Retro spective studies have shown that postmastectomy radiation therapy—when compared with no post mast-ectomy radiation—was associated with improved out-comes for selected patients with non-metastatic, non-infl ammatory breast cancer who received neo-adjuvant doxorubicin-based chemotherapy and mast-ectomy at MD Anderson from 1974 through 2000.25 542 (80%) of 676 patients in the study received post mast-ectomy radiation therapy, and median follow-up was 69 months. Postmastectomy radiation was associated with reduced 10-year locoregional recurrence for patients with clinical T3 disease at presentation (22% without vs 8% with, p=0·002), clinical T4 disease at presentation (46% vs 15%, p<0·0001), clinical N2–3 disease at presentation (40% vs 12%, p<0·0001; fi gure 2A), pathological tumour size 2·1–5·0 cm at resection (31% vs 14%, p=0·002), pathological tumour size greater than 5·0 cm at resection (52% vs 13%, p=0·001), or involvement of four or more nodes at resection (59% vs 16%, p<0·0001). A non-signifi cant improvement in locoregional recurrence was noted for patients with clinical N1 disease (14% vs 9%, p=0·062), and improvement was seen for patients with clinical N0 disease (23% vs 10%, p=0·014). Furthermore,

Number of patients 5-year locoregional recurrence

AJCC 1988 stage at diagnosis

IIA 21 5%

IIB 44 16%

IIIA 35 17%

IIIB 38 50%

IV* 11 79%

Clinical tumour stage and pathological nodal status

T1–2, negative lymph nodes 19 5%

T3–4, negative lymph nodes 23 34%

T1–2, positive lymph nodes 42 13%

T3–4, positive lymph nodes 64 36%

Pathological tumour size and lymph-node status

<2 cm, negative lymph nodes 21 10%

2·1–5·0 cm, negative lymph nodes 14 49%

>5·0 cm, negative lymph nodes 5 20%

<2 cm, positive lymph nodes 52 20%

2·1–5·0 cm, positive lymph nodes 41 30%

>5·0 cm, positive lymph nodes 9 63%

Adapted from reference 23. AJCC=American Joint Committee on Cancer. *Ipsilateral supraclavicular disease without systemic metastasis.

Table: Locoregional recurrence among patients from MD Anderson who received neoadjuvant chemotherapy, mastectomy, and no postmastectomy radiation therapy, stratifi ed by clinical and pathological disease characteristics


Personal View

postmastectomy radiation was associated with improved 10-year cancer-specifi c survival for patients with stage IIIB or higher disease at presentation (AJCC 1988 staging; 22% without vs 44% with, p=0·002), clinical T4 tumours (24% vs 45%, p=0·007), clinical N2–3 disease

(27% vs 49%, p=0·024), or four or more positive lymph nodes at resection (18% vs 44%, p=0·005).

Data regarding the benefi t of postmastectomy radiation for patients with clinical T3 node-negative disease off er confl icting conclusions. A retrospective analysis of 162 patients with T3N0 disease treated from 1985 to 2004 at MD Anderson suggests a benefi t from postmastectomy radiation therapy.29 Median tumour size for patients in this cohort was 6 cm, and 73% (n=119) received radiation. Overall, radiation was associated with improved 5-year locoregional recurrence, from 24% to 4% (p<0·001; fi gure 2D). 45% (n=73) of the patients were found to have pathological nodal involvement after neoadjuvant chemotherapy, and therefore had radiographically occult lymph nodes at presentation. Among 73 patients with clinical T3N0 disease and pathological nodal involvement after chemotherapy, radiation was associated with signifi cant improvement in 5-year loco regional recurrence (53% vs 5%, p<0·001). Among 89 patients with T3N0 disease and no pathological nodal involvement after chemotherapy, there was borderline signifi cant improvement in 5-year locoregional recurrence associated with radiation (14% vs 2%, p=0·06). However, locoregional recurrence rates among patients with T3N0 disease who did not receive radiation are higher in the MD Anderson study than in the NSABP B-18 and B-27 trials. In the NSABP trials, 10-year cumulative locoregional failure was 14·0% among 179 women with T3N0 disease and nodal disease after chemotherapy, 11·8% among 95 women with no nodal involvement and residual invasive disease in the breast after chemotherapy, and 6·2% among 16 women with no nodal involvement and no residual disease in the breast after chemotherapy.32 Based on information from both sources, patients with clinical T3N0 disease and nodal involvement after neoadjuvant chemotherapy have suffi cient risk of locoregional recurrence to warrant consideration of post-mast ectomy radiation, and MD Anderson data suggest that this therapy improves locoregional control. Additional information is needed to assess whether post-mast ectomy radiation therapy can be omitted for patients with clinical T3N0 at presentation and no nodal involve-ment after neoadjuvant chemotherapy. MD Anderson data suggest that postmastectomy radiation therapy lowers locoregional recurrence risk for these patients; however, NSABP data suggest that recurrence risk might be suffi ciently low to omit postmastectomy radiation therapy for patients who show pathological complete response to neoadjuvant chemotherapy.

There is also limited information on the benefi t of postmastectomy radiation therapy for very young patients. A retrospective analysis of MD Anderson patients by Garg and colleagues24 focused on 107 patients younger than 35 years with clinical stage IIA–IIIC disease (AJCC 2003 staging) who received doxorubicin-based neoadjuvant chemotherapy and mastectomy from 1975 to 2005. 75% (n=80) of the patients received

Lymph-node negativeand pCR

Lymph-node negativeand no pCR

Lymph-node positive

5·9%8·0%

14·9%

0

5

10

15

208-

year

loco

regi

onal

recu

rrenc

e (%

)

Figure 1: Locoregional recurrence among patients in NSABP B-18 and NSABP B-27 trials who received neoadjuvant chemotherapy, mastectomy, and no postmastectomy radiation therapy, stratifi ed by pathological extent of breast and lymph-node disease at resection Pathological complete response (pCR) was defi ned as no invasive tumour in the breast. Reproduced with permission from T Mamounas.19

No PMRT

PMRT

A

No PMRT

PMRT

B

No PMRT

PMRT

C

0 15 30 45

No PMRT

PMRT

40·0%

12·0%

33·0%

7·3%

37·0%

12·0%

24·0%

4·0%

D

Locoregional recurrence (%)

Figure 2: Locoregional recurrence for subgroups of patients from MD Anderson who received neoadjuvant chemotherapy and mastectomy with or without postmastectomy radiation therapy (PMRT) 10-year locoregional recurrence for patients with clinical N2–N3 disease (A); 10-year locoregional recurrence for patients with clinical stage III disease and pathological complete response to chemotherapy (B); 5-year locoregional recurrence for patients younger than 35 years and with clinical stage II or III disease (C); and 5-year locoregional recurrence for patients with clinical T3N0 disease (D).24,25,27,29


Personal View

postmastectomy radiation therapy. Despite more adverse features among these patients, postmastectomy radiation was associated with improved locoregional control (5-year recurrence 37% without vs 12% with, p=0·001; fi gure 2C) and overall survival (5-year survival 48% vs 67%, p=0·031). Some improvement in 5-year survival was seen across most clinical stages, although not all signifi cantly, including women with IIB disease (n=24, 56% without vs 92% with, p=0·033), IIIA disease (n=27, 50% vs 66%, p=0·154), IIIB disease (n=43, 17% vs 60%, p=0·064), and IIIC disease (n=7, 0% vs 33%, p=0·56). Thus, MD Anderson data suggest that women younger than 35 years with clinical stage IIB disease or higher might benefi t from post mastectomy radiation therapy.

Response to neoadjuvant chemotherapy aff ects recurrence riskClinical response to neoadjuvant chemotherapy modulates locoregional recurrence risk. In the NSABP B-18 and B-27 trials, 8-year cumulative incidence of locoregional failure after mastectomy without radiation was 14·9% in 447 patients with lymph-node involvement after neo adjuvant chemotherapy, but only 5·9% in 68 patients with pathological complete response and no lymph-node involvement after chemotherapy (fi gure 1).19 Since clinical response to neoadjuvant chemotherapy predicts loco regional recurrence risk, this response can help tailor postmastectomy radiation treatment decisions.

McGuire and colleagues27 did a retrospective analysis of 106 patients treated at MD Anderson from 1982 to 2002 for non-infl ammatory, non-metastatic breast cancer who had a pathological complete response to neoadjuvant chemotherapy. 28% had clinical stage II disease (AJCC 2003), 32% had stage IIIA, 25% had stage IIIB, and 12% had stage IIC; 68% received postmastectomy radiation. With a median follow-up of 62 months, postmastectomy radiation was associated with improvement in locoregional recurrence for patients with stage III disease and pathological complete response. 10-year locoregional recurrence was 7·3% in the 62 stage III patients who received postmastectomy radiation therapy compared with 33% in the 12 patients who did not (p=0·04; fi gure 2B), despite a greater proportion of patients with more adverse features in the group that received radiation. Admin istration of postmastectomy radiation did not aff ect locoregional control for the 30 patients with stage II disease and pathological complete response. No locoregional recurrence was observed in these patients regardless of radiation therapy; however, patient numbers were limited. Le Scodan and colleagues33 also assessed the eff ect of postmastectomy radiation therapy in patients with clinical stage II and stage III disease (AJCC 1988) and no nodal involvement after anthracycline-based neoadjuvant chemotherapy. Results were analysed for 134 patients treated at Rene Huguenin Cancer Center (Saint-Cloud, France) from 1990 to 2004 with median follow-up of 91·4 months (range 11·6–218). 62% of the

patients had clinical stage II disease, 37% had clinical stage III disease, and 58% received postmastectomy radiation therapy. After adjustment for diff erences in patient and tumour characteristics between the two groups, admin istration of radiation was not signifi cantly associated with improved locoregional recurrence-free survival ([RFS]; adjusted hazard ratio 0·37, 95% CI 0·09–1·61; p=0·18). Similar to the MD Anderson fi ndings, postmastectomy radiation therapy was not associated with improved locoregional control for patients with clinical stage II disease and good response to chemo therapy. Estimated 5-year locoregional RFS for the 83 patients with stage II disease was 97·4% with radiation and 92·9% without radiation (p>0·4). By contrast with the MD Anderson fi ndings, admin istration of post mast ectomy radiation was not associated with improved locoregional control for patients with clinical stage III disease and good response to chemotherapy. Estimated 5-year locoregional RFS for the 50 patients with stage III disease was 94·7% without radiation and 90·9% with radiation (p>0·2). These data suggest that patients with clinical stage II disease and pathological complete response after neoadjuvant chemo therapy have a low risk of locoregional failure and might not need postmast ectomy radiation therapy, whereas data for patients with clinical stage III disease and pathological complete response after neoadjuvant chemo therapy are mixed.

Additional prospective data are needed to guide postmastectomy treatment decisionsAlthough retrospective series and non-randomised clinical trial data suggest postmastectomy radiation therapy might not be necessary for selected women who achieve a good response to neoadjuvant chemotherapy, high-quality pro spective data are needed to determine if postmastectomy radiation therapy can be safely omitted without com promising local control or breast-cancer survival.

The Radiotherapy After Primary CHEMotherapy for breast cancer (RAPCHEM) trial (NCT01279304) at the Netherlands Cancer Institute (Amsterdam, Netherlands) is a non-randomised prospective trial, enrolling patients with clinical T1–T2 invasive breast cancer with one or more pathologically proven axillary lymph nodes who convert to node-negative disease after neoadjuvant chemotherapy (ypT0–2 ypN0 disease). Patients undergoing mastectomy do not receive postmastectomy radiation and are followed to determine 5-year locoregional recurrence. The Alliance for Clinical Trials in Oncology has proposed a randomised trial to examine the role of postmastectomy radiation for patients with lymph-node-positive breast cancer who convert to node-negative disease at resection after neo adjuvant chemotherapy. Women with clinical T0–T3, lymph-node-positive disease who undergo mastectomy and convert to lymph-node-negative disease after neo adjuvant


Personal View

chemotherapy will be randomised to post mast ectomy radiation therapy or to observation.

Once completed, these trials will help to determine if postmastectomy radiation therapy can be safely omitted for women with clinical T1 or T2 disease at diagnosis and no nodal involvement after neoadjuvant chemotherapy, and will provide the fi rst prospective data on whether radiation treatment can be individualised based on disease response to neoadjuvant chemotherapy.

ConclusionThousands of women worldwide now receive neoadjuvant chemotherapy as standard treatment for operable breast cancer. Although randomised clinical trials have established which patients might benefi t from post mast ectomy radiation therapy after upfront surgery, no randomised trial data exist to defi ne which women benefi t from postmastectomy radiation after neoadjuvant chemotherapy. Instead, insights from non-randomised data provide guidance on who might benefi t and who can be spared the morbidity of radiation treatment. The data suggest that disease at presentation and response to neoadjuvant chemotherapy can be used to tailor post mastectomy radiation treatment recommendations; however, the available single-arm clinical trial data and retrospective institutional data have inherent limitations.

Consistent with conclusions from randomised trials of postmastectomy radiation therapy after upfront surgery, retrospective data suggest that administration of post-mast ectomy radiation after neoadjuvant chemo therapy improves locoregional control for women with clinical T3–T4 disease at presentation, clinical N2–N3 disease at presentation, and node-positive disease at resection. Response to neoadjuvant chemotherapy can also predict locoregional recurrence risk and might help select women for postmastectomy radiation therapy. Outcomes of women who receive neoadjuvant chemotherapy, mast-ectomy, and no postmastectomy radiation show locoregional recurrence rates that are suffi ciently high (at least 15%) among women with lymph-node-positive disease at resection to support use of postmastectomy radiation therapy, and locoregional recurrence rates that

are suffi ciently low (<10%) among women with clinical T1–T2 disease and no lymph-node involvement at resection to support omission of postmastectomy radiation. Although there is an absence of randomised data to guide postmastectomy radiation therapy treatment decisions, the available retrospective data provide insights that help guide appropriate design of future prospective studies. Randomised clinical trials are needed to assess whether postmastectomy radiation therapy can be safely omitted in women with good response to chemotherapy, without compromising local control or breast-cancer survival.

At this time, at MD Anderson, we routinely administer postmastectomy radiation therapy after neoadjuvant chemotherapy to women with clinical stage III disease and clinical T3 disease at presentation, and most patients with lymph-node involvement at the time of surgical resection. We consider postmastectomy radiation after neoadjuvant chemotherapy for selected patients with clinical stage II disease, including those with lymph-node involvement at resection and patients with features suggesting high-risk of local recurrence, such as young age, oestrogen-receptor-negative disease, and poor response to chemotherapy.

Contributors All authors contributed equally to the manuscript.


AcknowledgmentsWe thank Greg Pratt for assisting with the literature search.

References1 Mieog JS, van der Hage JA, van de Velde CJ. Neoadjuvant

chemotherapy for operable breast cancer. Br J Surg 2007; 94: 1189–200.

2 Hoff man K, Buchholz T. How improved local-regional therapy impacts survival. Curr Breast Cancer Rep 2010; 2: 83–89.

3 Overgaard M, Hansen PS, Overgaard J, et al. Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. Danish Breast Cancer Cooperative Group 82b Trial. N Engl J Med 1997; 337: 949–55.

4 Overgaard M, Jensen MB, Overgaard J, et al. Postoperative radiotherapy in high-risk postmenopausal breast-cancer patients given adjuvant tamoxifen: Danish Breast Cancer Cooperative Group DBCG 82c randomised trial. Lancet 1999; 353: 1641–48.

5 Ragaz J, Olivotto IA, Spinelli JJ, et al. Locoregional radiation therapy in patients with high-risk breast cancer receiving adjuvant chemotherapy: 20-year results of the British Columbia randomized trial. J Natl Cancer Inst 2005; 97: 116–26.

6 Clarke M, Collins R, Darby S, et al. Eff ects of radiotherapy and of diff erences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005; 366: 2087–106.

7 Harris JR, Halpin-Murphy P, McNeese M, Mendenhall NP, Morrow M, Robert NJ. Consensus Statement on postmastectomy radiation therapy. Int J Radiat Oncol Biol Phys 1999; 44: 989–90.

8 Recht A, Edge SB, Solin LJ, et al. Postmastectomy radiotherapy: clinical practice guidelines of the American Society of Clinical Oncology. J Clin Oncol 2001; 19: 1539–69.

9 Taylor ME, Haff ty BG, Rabinovitch R, et al. ACR appropriateness criteria on postmastectomy radiotherapy expert panel on radiation oncology-breast. Int J Radiat Oncol Biol Phys 2009; 73: 997–1002.

10 National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology. Breast Cancer v.1.2011. http://www.nccn.org/professionals/physician_gls/pdf/breast.pdf (accessed Sept 9, 2011).


References for this Personal View were obtained from searches of Medline and Embase, using MeSH and keyword search terms for “breast cancer”, “neoadjuvant chemotherapy”, and “radiotherapy”. Articles published up to October, 2011, with a focus on outcomes after neoadjuvant chemotherapy and mastectomy with or without radiotherapy were included. There was no lower date limit and no language restrictions. Reports on outcomes of women with infl ammatory breast cancer were excluded. Relevant published abstracts from scientifi c meetings were also considered.


Personal View

11 Whelan TJ, Ackerman I, Chapman JW, et al. NCIC-CTG MA.20: an intergroup trial of regional nodal irradiation in early breast cancer. Proc Am Soc Clin Oncol 2011; 29 (suppl): abstr LBA1003.

12 Marks LB, Zeng J, Prosnitz LR. One to three versus four or more positive nodes and postmastectomy radiotherapy: time to end the debate. J Clin Oncol 2008; 26: 2075–77.

13 Overgaard M, Nielsen HM, Overgaard J. Is the benefi t of postmastectomy irradiation limited to patients with four or more positive nodes, as recommended in international consensus reports? A subgroup analysis of the DBCG 82 b&c randomized trials. Radiother Oncol 2007; 82: 247–53.

14 Katz A, Strom EA, Buchholz TA, et al. Locoregional recurrence patterns after mastectomy and doxorubicin-based chemotherapy: implications for postoperative irradiation. J Clin Oncol 2000; 18: 2817–27.

15 Recht A, Gray R, Davidson NE, et al. Locoregional failure 10 years after mastectomy and adjuvant chemotherapy with or without tamoxifen without irradiation: experience of the Eastern Cooperative Oncology Group. J Clin Oncol 1999; 17: 1689–700.

16 Wallgren A, Bonetti M, Gelber RD, et al. Risk factors for locoregional recurrence among breast cancer patients: results from International Breast Cancer Study Group Trials I through VII. J Clin Oncol 2003; 21: 1205–13.

17 Kyndi M, Overgaard M, Nielsen HM, Sorensen FB, Knudsen H, Overgaard J. High local recurrence risk is not associated with large survival reduction after postmastectomy radiotherapy in high-risk breast cancer: a subgroup analysis of DBCG 82 b&c. Radiother Oncol 2009; 90: 74–79.

18 Rastogi P, Anderson SJ, Bear HD, et al. Preoperative chemotherapy: updates of National Surgical Adjuvant Breast and Bowel Project Protocols B-18 and B-27. J Clin Oncol 2008; 26: 778–85.

19 Mamounas T. Sentinel node biopsy after neoadjuvant chemotherapy: the pros. http://ctep.cancer.gov/highlights/docs/mamounas.pdf (accessed Sept 9, 2011).

20 Dominici LS, Negron Gonzalez VM, Buzdar AU, et al. Cytologically proven axillary lymph node metastases are eradicated in patients receiving preoperative chemotherapy with concurrent trastuzumab for HER2-positive breast cancer. Cancer 2010; 116: 2884–89.

21 Buchholz TA, Katz A, Strom EA, et al. Pathologic tumor size and lymph node status predict for diff erent rates of locoregional recurrence after mastectomy for breast cancer patients treated with neoadjuvant versus adjuvant chemotherapy. Int J Radiat Oncol Biol Phys 2002; 53: 880–88.

22 Hunt KK, Yi M, Mittendorf EA, et al. Sentinel lymph node surgery after neoadjuvant chemotherapy is accurate and reduces the need for axillary dissection in breast cancer patients. Ann Surg 2009; 250: 558–66.

23 Buchholz TA, Tucker SL, Masullo L, et al. Predictors of local-regional recurrence after neoadjuvant chemotherapy and mastectomy without radiation. J Clin Oncol 2002; 20: 17–23.

24 Garg AK, Oh JL, Oswald MJ, et al. Eff ect of postmastectomy radiotherapy in patients <35 years old with stage II-III breast cancer treated with doxorubicin-based neoadjuvant chemotherapy and mastectomy. Int J Radiat Oncol Biol Phys 2007; 69: 1478–83.

25 Huang EH, Tucker SL, Strom EA, et al. Postmastectomy radiation improves local-regional control and survival for selected patients with locally advanced breast cancer treated with neoadjuvant chemotherapy and mastectomy. J Clin Oncol 2004; 22: 4691–99.

26 Huang EH, Tucker SL, Strom EA, et al. Predictors of locoregional recurrence in patients with locally advanced breast cancer treated with neoadjuvant chemotherapy, mastectomy, and radiotherapy. Int J Radiat Oncol Biol Phys 2005; 62: 351–57.

27 McGuire SE, Gonzalez-Angulo AM, Huang EH, et al. Postmastectomy radiation improves the outcome of patients with locally advanced breast cancer who achieve a pathologic complete response to neoadjuvant chemotherapy. Int J Radiat Oncol Biol Phys 2007; 68: 1004–09.

28 Oh JL, Dryden MJ, Woodward WA, et al. Locoregional control of clinically diagnosed multifocal or multicentric breast cancer after neoadjuvant chemotherapy and locoregional therapy. J Clin Oncol 2006; 24: 4971–75.

29 Nagar H, Mittendorf EA, Strom EA, et al. Local-regional recurrence with and without radiation therapy after neoadjuvant chemotherapy and mastectomy for clinically staged T3N0 breast cancer. Int J Radiat Oncol Biol Phys 2011; 81: 782–87.

30 Fisher B, Brown A, Mamounas E, et al. Eff ect of preoperative chemotherapy on local-regional disease in women with operable breast cancer: fi ndings from National Surgical Adjuvant Breast and Bowel Project B-18. J Clin Oncol 1997; 15: 2483–93.

31 Bear HD, Anderson S, Brown A, et al. The eff ect on tumor response of adding sequential preoperative docetaxel to preoperative doxorubicin and cyclophosphamide: preliminary results from National Surgical Adjuvant Breast and Bowel Project Protocol B-27. J Clin Oncol 2003; 21: 4165–74.

32 Mamounas E, Anderson S, Bear H, et al. Predictors of locoregional failure (LRF) in patients receiving neoadjuvant chemotherapy (NC): results from combined analysis of NSABP B-18 and NSABP B-27. 2010 Breast Cancer Symposium; Oct 1–3, 2010; Washington, DC, USA. Abstr 90.

33 Le Scodan R, Selz J, Stevens D, et al. Radiotherapy for stage II and stage III breast cancer patients with negative lymph nodes after preoperative chemotherapy and mastectomy. Int J Radiat Oncol Biol Phys 2012; 82: 1–7.


Articles

Prevalence of mismatch repair-defi cient crypt foci in Lynch syndrome: a pathological studyMatthias Kloor*, Cathrin Huth*, Anita Y Voigt, Axel Benner, Peter Schirmacher, Magnus von Knebel Doeberitz, Hendrik Bläker

SummaryBackground Lynch syndrome is an inherited tumour predisposition syndrome caused by germline mutations of DNA mismatch repair (MMR) genes. Mutation carriers have a high risk of developing colorectal cancer, but do not present with polyposis, a typical feature of other colorectal cancer syndromes such as familial adenomatous polyposis, in which polyposis refl ects the high frequency of biallelic APC gene inactivation. We asked whether in Lynch syndrome biallelic inactivation of MMR genes occurred at a similar frequency to that of APC gene, and whether MMR inactivation resulted in detectable lesions within the intestinal mucosa.

Methods Resections done for small and large bowel cancer between January, 2002, and January, 2011, were retrieved. We systematically analysed non-tumorous mucosa from carriers of a Lynch syndrome mutation (set 1: ten patients) and control patients without Lynch syndrome (set 1: nine patients) for MMR protein expression (MLH1, MSH2, and EPCAM) with immunohistochemistry. We validated the fi ndings in an independent sample set (set 2: 30 Lynch syndrome patients, 79 controls). We did an analysis of microsatellite instability by PCR analysis to test lesions for mismatch repair defi ciency. We applied a Poisson regression model to analyse the distribution of MMR-defi cient crypt foci counts and a Fisher’s exact test to compare the prevalence of these foci between mutation carriers and control patients.

Findings 20 crypt foci with no MMR protein expression were detected in 20·1 cm² of non-tumorous mucosa from Lynch syndrome patients (set 1), an additional fi ve were detected upon resectioning of two samples. In an independent validation set (set 2), two MMR-defi cient crypt foci were noted in 2·2 cm² of mucosa. No MMR-defi cient crypt foci were noted in non-tumorous mucosa from control patients without evidence for Lynch syndrome (set 1: 3·7 cm², set 2: 4·8 cm²). Microsatellite instability was detected in all seven MMR-defi cient crypt foci analysed. A subset of these foci displayed unusual architectural and cytological abnormalities, although they had no polypous or adenomatous appearance.

Interpretation We identifi ed a novel type of lesion, the MMR-defi cient crypt focus, as the manifestation of biallelic MMR gene inactivation in Lynch syndrome. The abundance of MMR-defi cient crypt foci indicates a high frequency of biallelic MMR gene inactivation, which is in sharp contrast with the low number of clinically manifest cancers in Lynch syndrome. This discrepancy suggests that most MMR-defi cient crypt foci do not progress to cancer. We propose Lynch syndrome as a unique model syndrome for studying initial steps of MMR defi ciency, tumour initiation and, possibly, elimination.

Funding German Cancer Aid and German Research Foundation.

IntroductionLynch syndrome, clinically referred to as hereditary non-polyposis colorectal cancer (HNPCC), is one of the most frequent inherited cancer syndromes. Lynch syndrome is caused by germline mutations in one of the DNA mismatch repair (MMR) genes—mainly MLH1 and MSH2 1—occurring with an estimated allele frequency of one in 350. Recently, deletions aff ecting EPCAM, a gene located immediately upstream of MSH2, have been identifi ed as an alternative mechanism of MSH2 silen-cing in Lynch syndrome.2 Tumours developing in the context of Lynch syndrome are characterised by high microsatellite instability, which is the result of functional MMR defi ciency caused by biallelic MMR gene inacti-vation through a somatic alteration as a second mutation event. Inactivation of MMR genes commonly leads to loss of MMR protein expression in tumour cells.

Im munohistochemistry reliably detects this phenom-enon3,4 and, by showing the absence of a specifi c MMR component, can guide the geneticist in the selection of MMR genes for mutation analysis.

Carriers of a Lynch syndrome mutation have a lifetime risk of developing colorectal cancer of about 50%.5,6 However, these carriers do not show a clearly increased frequency of adenomatous polyps,7–9 which are common precursor lesions of colorectal cancer. This is in sharp contrast with other predisposition syndromes for colorectal cancer, which are characterised by a signifi cantly increased number of polyps, such as the familial adenomatous polyposis coli syndrome. In this syndrome, biallelic inactivation of the APC tumour suppressor gene triggers the formation of an aberrant crypt focus, which can develop into an adenomatous polyp and eventually cancer.10,11

Lancet Oncol 2012; 13: 598–606


DOI:10.1016/S1470-2045(12)70109-2


*These authors contributed equally to the study

Department of Applied Tumour Biology (M Kloor MD, C Huth,

A Y Voigt MSc, Prof M von Knebel Doeberitz

MD), Department of General Pathology

(Prof P Schirmacher MD, Prof H Bläker MD), Institute of

Pathology, University Hospital Heidelberg, Heidelberg,

Germany; and Cooperation Unit Applied Tumour Biology

(M Kloor, C Huth, A Y Voigt, Prof M von Knebel Doeberitz), Department of Biostatistics

(A Benner MSc), DKFZ (German Cancer Research Center),

Heidelberg, Germany

Correspondence to:Dr Matthias Kloor, Department

of Applied Tumour Biology, Institute of Pathology, University

Hospital Heidelberg, Im Neuenheimer Feld,

69120 Heidelberg, Germanymatthias.kloor@med.

uni-heidelberg.de

Articles


By contrast, the consequences of biallelic inactivation of the MMR system in non-tumoral mucosa from carriers of a Lynch syndrome mutation are unknown. Currently, the inactivation of the MMR system is commonly regarded as an early event, but not as the fi rst step on the route from non-tumoral colonic mucosa towards adenoma and eventually colorectal cancer.12 This idea is supported by the fi nding that a substantial part of adenomas from carriers of a Lynch syndrome mutation are MMR-profi cient.13–15 Instead, there is evidence that MMR defi ciency accelerates progression from adenoma to carcinoma in carriers of a Lynch syndrome mutation.8,16,17

The present systematic study was initiated to investigate biallelic inactivation of DNA mismatch repair genes and to determine the prevalence of MMR-defi cient crypt foci in the intestinal mucosa from carriers of a Lynch syndrome mutation.

MethodsSamplesParaffi n-embedded archival tissues (194 paraffi n blocks, referred to as set 1) from 24 resections done for small bowel or large bowel cancers between January, 2002, and January, 2011, were retrieved from the Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany (table 1), including a local registry of small bowel cancers established by one of the authors (HB). 15 resection specimens were from ten patients with Lynch syndrome (four carriers of an MLH1 mutation, three carriers of an MSH2 mutation, and three carriers of an EPCAM deletion), nine resection specimens were from nine patients without evidence for Lynch syndrome (seven with colorectal cancer, two with small bowel cancer). We analysed non-tumorous mucosa from the resection margins, other sites routinely investigated in pathological work-up (eg, Bauhin’s valve, sites of anastomoses), and from tumour-adjacent non-neoplastic tissue for changes in MMR protein expression. As a validation set (set 2), we investigated 109 colorectal cancer paraffi n blocks—30 colorectal cancers with high microsatellite instability from carriers of a Lynch syndrome mutation and 79 microsatellite-stable colorectal cancers from patients without Lynch syndrome—that had been sent to the Department of Applied Tumour Biology for Lynch syndrome diagnostics between 1998, and 2011, for the presence of altered MMR protein expression in non-tumorous mucosa.

All patients provided informed and written consent for participation in the study of the German HNPCC Consortium. The study was approved by the Institutional Ethics Committee (vote number 220/2002).

ProceduresWe calculated the amount of mucosal surface analysed by multiplying the measured length of the mucosa

stretch by mean crypt diameter and mean number of crypts evaluable for MMR protein expression per section.

Age at diagnosis (years)

Large bowel Small bowel

Surface (mm²)

MMR-defi cient crypt foci

Surface (mm²)

MMR-defi cient crypt foci

LS1, man, MSH2 mutation

a: subtotal colon 55 49 0 13 0

b: jejunal segment/sigmoid 67 64 1 126 2

LS2, man, MSH2 (EPCAM) mutation

a: right-sided colon 53 32 0 21 1

b: anastomosis 54 95 0 90 1


Whipple operation (duodenal cancer) 37 ·· ·· 13 0

LS4, woman, MLH1 mutation

a: anastomosis / jejunal segment 70 237 2 308 0

b: anastomosis / rectum 70 77 1 21 0


a: total colon 42 62 1 24 0

b: jejunal segment 48 ·· ·· 220 2


Right-sided colon 54 71 0 8 0

LS7, man, MLH1 mutation



a: right-sided colon, ileum 53 25 0 67 1

b: Whipple operation (duodenal cancer) 53 ·· ·· 102 2


Right sided colon 64 43 1 25 0


Subtotal colon 52 102 1 11 0

C1, man


C2, woman

Sigmoid 68 65 0 ·· ··

C3, woman

Small bowel segment 73 ·· ·· 11 0

C4, man

Right-sided colon 71 45 0 ·· ··

C5, woman

Duodenum 47 ·· ·· 42 0

C6, man

Rectum 55 41 0 ·· ··

C7, man

Small bowel segment 59 ·· ·· 11 0

C8, woman

Sigmoid 50 86 0 ·· ··

C9, man

Whipple operation 81 ·· ·· 25 0

Table 1: Distribution of MMR-defi cient crypt foci in intestinal mucosa for patients with Lynch syndrome (LS1–10) and control patients (C1–9; set 1) and by resected segment

Surface = length × crypt diameter × 4

Articles


A detailed description of the calculation of mucosal surface and the estimation of crypt numbers is provided in the appendix. We measured the length of the mucosa stretch at the level of the lamina muscularis mucosae with Aperio software (Spectrum Version 11.0.0.725, Aperio Technologies, Vista, CA, USA) after scanning of the slides.

Immunohistochemistry was done on 5 μm paraffi n sections as described previously.18 Briefl y, the slides were

pretreated by boiling for 10 min with a microwave in Dako target retrieval buff er (pH 9, Dako, Hamburg, Germany) before application of monoclonal antibodies specifi c for MSH2 (clone FE11, dilution 1:100, Calbiochem, Darmstadt, Germany), MLH1 (clone G168-15, dilution 1:100, BD Pharmingen, Heidelberg, Germany), EPCAM (clone BerEP4, dilution 1:100, Dako), or Ki-67 (clone MIB1, dilution 1:100, Dako). An immunoperoxidase method was used to visualise the antibodies by labelling them with a chromogen (3-amino-9-ethylcarbazole, Dako).

DNA was isolated from paraffi n-embedded tissue (immunohistochemically stained sections or haema-toxylin/eosin-stained serial sections) after manual or laser-assisted microdissection with the DNeasy tissue kit (Qiagen, Hilden, Germany) according to the manu-facturer’s instructions.

Microsatellite instability was analysed in lesions with suffi cient DNA isolated by PCR fragment length analysis in the non-coding mononucleotide markers BAT25, BAT26, and CAT25 as described previously.19 We analysed the coding microsatellite located within TGFBR2 for frameshift mutations, and scored as instability profi les showing substantial deviation (peak ratio lower than 0·5 or higher than 2) as described previously.20 PCR products were separated on an ABI3100 sequencer (Applied Biosystems, Darmstadt, Germany), and fragment lengths were identifi ed with the GeneScan software (Version 3.7, Applied Biosystems).

Statistical analysisSince we can assume that the number of MMR-defi cient crypt foci follows a Poisson distribution, we applied a Poisson regression model to analyse the distribution of MMR-defi cient crypt foci counts with respect to mucosal area and number of crypts, respectively. To account for patient-related clusters we used mixed-eff ects models.

We used a Fisher’s exact test to compare the incidence of MMR-defi cient crypt foci between carriers of Lynch syndrome mutation and control patients. We did all statistical analyses with R, version 2.14.1.21

Role of the funding sourceThe sponsors of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all of the data and the fi nal responsibility to submit for publication.

ResultsIn total, we studied 705·1 cm of mucosa (319·4 cm large bowel, 385·7 cm small bowel) from carriers of a Lynch syndrome mutation and 244·9 cm (205·6 cm large bowel, 39·3 cm small bowel) from patients with cancer and without evidence for Lynch syndrome. There was a median of four crypts per section with a mean crypt diameter of 90 μm in the large bowel and 70 μm in the


Control 1–9 Control 10–88

LS 11–40LS 1–10

Set 1 Set 2MMR-deficient foci

Imm

unoh

istoc

hem

istry

2·52 cm²

1·14 cm²

9·26 cm²

10·79 cm²

4·78 cm²

2·20 cm²

Imm

unoh

istoc

hem

istry

Validation by PCR

0 0

2

0

11*

9*

Detection of microsatellite instability in 7 of 7 MMR-deficient crypt foci

Figure 1: Schematic overview of study designMMR=DNA mismatch repair. *Five additional MMR-defi cient crypt foci were detected upon resampling. Non-tumoral colon and small bowel mucosa from patients with Lynch syndrome (set 1: LS1–10, set 2: LS11–40) and from control colorectal cancer patients without evidence for Lynch syndrome was analysed for the presence of MMR-defi cient crypt foci using immunohistochemical MMR protein expression analysis. From seven out of 27 MMR-defi cient crypt foci, suffi cient amounts of DNA could be isolated to perform MSI analysis. MSI in at least one out of four markers was detected in all seven MMR-defi cient crypt foci.

C

A

D

B

MLH1

MSH2

MSH2

MLH1

Figure 2: Monocryptic MMR-defi cient crypt foci Objective magnifi cation is 10× for all panels. Except for absence of MMR expression, monocryptic MMR-defi cient crypt foci are commonly indistinguishable from the surrounding MMR-profi cient crypts (A–D). In an MMR-defi cient crypt focus from a patient with MSH2-associated Lynch syndrome (LS2), MLH1 expression is retained (A, arrows) whereas the expression of MSH2 is lost (B). Vice versa, an MMR-defi cient crypt focus from a patient with MLH1-associated Lynch syndrome (LS4) expresses MSH2 (C, arrows), but not MLH1 (D) proteins.

Articles


small bowel. We calculated the following surface areas: 22·3 cm² mucosa from Lynch syndrome mutation carriers (set 1: 20·1 cm²; set 2: 2·2 cm²; fi gure 1) and 8·4 cm² from cancer patients without evidence for Lynch syndrome (set 1: 3·7 cm²; set 2: 4·8 cm²). Figure 1 shows mucosal surface areas examined for the large and small bowel.

In set 1, immunohistochemistry for MLH1 and MSH2 in mucosa from patients with Lynch syndrome mutations showed 20 crypt foci without either MLH1 or MSH2 expression (table 1). 11 of these foci lacking MLH1 and MSH2 expression were detected in 9·26 cm² of large bowel mucosa (corresponding to an estimated number of 114 000 crypts), and nine were detected in 10·8 cm² of small bowel mucosa (220 000 crypts). These lesions will be referred to as MMR-defi cient crypt foci hereafter. MMR-defi cient crypt foci were identifi ed in the mucosa from eight of the ten patients with a Lynch syndrome mutation. The MMR-defi cient crypt foci consisted of a varying number of crypts, ranging from one to about 20 (fi gure 2, appendix). The incidence of MMR-defi cient crypt foci was higher, although not signifi cantly, in the large bowel than in the small bowel (per area: incidence rate ratio [IRR] 1·60, 95% CI 0·66–3·90, p=0·30, Wald test; per crypt: IRR 1·98, 0·75–5·21, p=0·17).

MLH1-negative crypt foci were exclusively noted in carriers of an MLH1 mutation, whereas MSH2-negative crypt foci only occurred in carriers of an MSH2 mutation. All MLH1-negative crypt foci analysed showed regular MSH2 expression, and all MSH2-negative crypt foci analysed showed regular MLH1 expression (fi gure 2). No MMR-defi cient crypt foci were detected in resection specimens from patients with cancer without Lynch syndrome. The diff erence between the density of MMR-defi cient crypt foci in patients with Lynch syn drome compared with control patients with cancer was signifi cant (Fisher’s exact test, p=0·0044).

To confi rm the frequency of MMR-defi cient crypt foci in the non-tumorous mucosa from carriers of Lynch syndrome mutation, resection specimens from two samples in set 1—LS1 b and LS4 a—were re-examined by immunohistochemistry after removal of a layer of 500 μm from the respective paraffi n blocks, allowing for the assessment of an additional area of 3·01 cm² mucosa from the large bowel and 4·34 cm² from the small bowel. With this approach, we identifi ed fi ve additional MMR-defi cient crypt foci, three located in the large bowel and two in the small bowel. These fi ve newly identifi ed MMR-defi cient crypt foci were novel and did not overlap with the lesions identifi ed in the initial analysis.

The detection of multiple MMR-defi cient crypt foci in the mucosa from carriers of Lynch syndrome mutation prompted us to reassess tissue sections that had previously been immunohistochemically analysed for the expression of MLH1 and MSH2 in the context of Lynch syndrome diagnostics (set 2). Two MMR-defi cient crypt foci, which

had not been noticed during previous routine Lynch syndrome diagnostics, were noted in 2·2 cm² mucosa from the 30 carriers of Lynch syndrome mutation. By contrast, no MLH1-negative or MSH2-negative crypt foci were observed in the 79 sections (corresponding to an area of 4·8 cm²) obtained from patients without any evidence of Lynch syndrome (fi gure 1).

In summary, we report the occurrence of about one MMR-defi cient crypt focus per 1 cm² mucosa from carriers of Lynch syndrome mutation. We computed a percent change in the incidence rate of MMR-defi cient crypt foci by about 67% per 1 cm² increase in mucosa from carriers of a Lynch syndrome mutation. We noted a tendency towards a higher density of MMR-defi cient

Section Type MSI testing

LS1

Lesion 1 b Oligocryptic ··


Lesion 3 b Monocryptic ··

Lesion 4 b (re-sectioning) Oligocryptic ··

LS2

Lesion 1 a Polycryptic Yes

Lesion 1 b Monocryptic ··

LS4

Lesion 1 a Polycryptic Yes

Lesion 2 a Monocryptic ··

Lesion 3 a (re-sectioning) Polycryptic ··

Lesion 4 a (re-sectioning) Polycryptic ··

Lesion 5 a (re-sectioning) Polycryptic Yes

Lesion 6 a (re-sectioning) Oligocryptic ··


LS5

Lesion 1 a Oligocryptic Yes


Lesion 2 ·· Monocryptic ··

LS7

Lesion 1 ·· Oligocryptic ··


Lesion 3 ·· Polycryptic Yes


LS8

Lesion 1 a Monocryptic ··



LS9

Lesion 1 ·· Monocryptic ··

LS10

Lesion 1 ·· Polycryptic Yes

LS11

Lesion 1 ·· Oligocryptic Yes

LS12


Table 2: MMR-defi cient crypt foci

Articles


crypt foci in large bowel with 1·6 times more MMR-defi cient crypt foci observed in the same surface area than in small bowel. No obvious diff erence was detected when comparing mucosa in the vicinity of the tumour from patients LS11–LS40 and remote from the tumour as in patients LS1–LS10.

We classifi ed the 27 detected MMR-defi cient crypt foci into three phenotypic groups based on the number of adjacent crypts sharing the same immunohistochemistry properties. Monocryptic foci containing a single MMR-defi cient crypt were observed in six cases; oligocryptic foci consisting of two to four adjacent MMR-defi cient crypts were observed in 14 cases and polycryptic foci consisting of fi ve to 19 MMR-defi cient crypts were found in the remaining seven cases (table 2, appendix).

13 of the 27 MMR-defi cient crypt foci (one of six monocryptic, fi ve of 14 oligocryptic, and all of the polycryptic foci) displayed unusual morphological changes. The remaining foci, except for their absence of MMR protein expression, were morphologically indistinguishable from MMR-profi cient crypts (fi gure 2). The most consistent abnormal histological fi nding was nuclear enlargement of cells at the crypt bottom. Maturation towards the crypt top, however, was retained. Architectural changes were noticed in oligocryptic and polycryptic lesions. Non-dilated crypts showed basal

branching and duplications (fi gure 3). Lesion 5 from patient LS4 displayed the complete morphological spectrum of MMR-defi cient crypt foci (fi gure 4). Crypts histologically indistinguishable from non-tumoral mucosa were found at one border of the lesion whereas the opposite border was composed of irregularly branched and duplicated crypts with nuclear enlargement at the crypt bottom. Congruent with retained cell maturation towards the crypt top, Ki-67 analysis showed a physio-logical pattern of proliferation not increased compared with non-tumoral mucosa (fi gure 4C).

In carcinomas, loss of MMR expression is associated with a defi ciency of the DNA mismatch repair system and consequently, with microsatellite instability. To investigate whether consequences of DNA MMR dys function were detectable in MMR expression-negative crypt foci, we isolated DNA from eight large oligocryptic or polycryptic lesions and from corresponding sur-rounding crypts with retained MMR protein expression. Seven lesions were successfully analysed and found to have microsatellite instability, with three of the seven showing instability at the coding A10 microstatellite located within TGFBR2 gene (fi gure 3, table 3). None of the MMR-profi cient surrounding mucosa crypts displayed microsatellite instability at non-coding micro-satellites (more than 50 DNA specimens were analysed).

Figure 3: Morphologic and molecular characteristics of polycryptic MMR-defi cient crypt fociPolycryptic MMR-defi cient crypt foci display architectural abnormalities. The lesions show crypt fi ssions (A, B) and branching with extension along a horizontal axis (C, D). MMR-defi cient crypt foci in EPCAM deletion carriers can display loss of either MSH2 (E) or EPCAM (F). Microsatellite instability analysis after tissue microdissection showed microsatellite instability in the non-coding markers BAT25, BAT26, and CAT25 in the lesions compared with surrounding non-tumoral tissue (G). Coding microsatellite instability was detected in the A10 stretch located in TGFBR2 in lesions from patients LS4 and LS2 (G). Arrows show borders of MMR-defi cient crypt foci, red arrows show additional peaks indicative of microsatellite instability.

BAT25 BAT26 CAT25 TGFBR2



Normaltissue

Lesions

Normaltissue

Lesions

Normaltissue

Lesions

100 105 110 105 110 115 140 145 150 145 150

100 105 110 105 110 115 140 145 150 145 150

100 105 110 105 110 115 140 145 150 145 150

GBA

DC

FE

LS7

LS4

LS2

Articles


In the MMR-defi cient crypt foci from patients LS2 and LS11 (EPCAM deletion carriers), the absence of MSH2 protein expression was accompanied by the absence of EPCAM protein expression.

DiscussionWe describe a novel lesion in Lynch syndrome, which we termed the MMR-defi cient crypt focus, based on the loss of MMR protein expression and the observation of microsatellite instability in the subset of lesions analysed for MMR defi ciency by PCR-based DNA fragment length analysis. Based on the data presented in this study, about one MMR-defi cient crypt focus is estimated to be detected per 1 cm² of non-tumorous mucosa from carriers of Lynch syndrome mutations. This fi nding highlights the abundance of secondary somatic muta-tions in activating the MMR system in Lynch syndrome patients, suggesting that these lesions occur in thousands throughout the intestinal tract of these carriers.

MMR-defi cient crypt foci were present in all carriers of a Lynch syndrome mutation from whom more than 100 mm² surface of intestinal mucosa could be analysed; hence they are a common phenomenon in this disease, and not restricted to a subset of mutation carriers and conditions. However, it has to be conceded that only patients with manifest tumours could be included in the study, and that the power of the study with regard to analysis of the eff ect of additional variables potentially related to the occurrence of MMR-defi cient crypt foci is limited. Further studies are therefore needed to establish the frequency of MMR-defi cient crypt foci in carriers of a Lynch syndrome mutation without a history of tumours, and to elucidate a potential interdependence between the frequency of MMR-defi cient crypt foci and variables such as sex and age, before the diagnostic relevance of MMR-defi cient crypt foci will be made clear. Based on the data of our study, however, the probability of detecting at least one MMR-defi cient crypt focus in a colon biopsy sample seems to be very low, estimated to be about 1% for colon biopsies of 1 mm² size. Our fi ndings also warrant further investigations as to whether MMR-defi cient cells can be observed in extra colonic tissues like endometrial mucosa or, independent from Lynch syndrome, in non-tumoral colonic epithelium from patients with sporadic colorectal cancer with high microsatellite instability.

MMR-defi cient crypt foci diff er from aberrant crypt foci, which are morphologically characterised early precursor lesions of adenomas22,23 and have been shown to present with various genetic abnormalities including microsatellite instability in a small fraction of cases.24–28 MMR-defi cient crypt foci are non-elevated and do not have widened luminal openings, which are characteristics of aberrant crypt foci.10 Furthermore, epithelial cells of MMR-defi cient crypt foci are neither dysplastic nor hyperplastic. Therefore, it is not surprising that, unlike aberrant crypt foci, MMR-defi cient crypt foci, despite their high frequency in

Lynch syndrome, have escaped detection by high magnifi cation chromoscopic endos copy or methylene blue staining of resected colon specimens.27

Histological assessment with routine stains failed to identify most MMR-defi cient foci. We became aware of potentially altered MMR protein expression in non-tumoral mucosa from Lynch syndrome mutation carriers in the context of our previous study29 on EPCAM protein expression in tumours from patients with Lynch syndrome and with EPCAM germline deletions.

B

A

C

Figure 4: Staining of a complex MMR-defi cient crypt focusObjective magnifi cation is 10× for all panels. (A) Haematoxylin and eosin stain. (B) Corresponding MLH1 immunohistochemical stain. (C) Ki-67 immuno histochemistry. In the haematoxylin and eosin stain, the MMR-defi cient crypts at the left border (arrows on the left) of the lesion are indistinguishable from the adjacent MMR profi cient crypts on the basis of morphological criteria. The right border (arrows on the right) is characterised by distorted crypts colonised by cells with nuclear enlargement. A crypt fi ssion is noted at the right border and causes an expansive and “pushing” appearance of the lesion. By contrast with conventional adenomas, the enterocytes of the MMR-defi cient crypt focus fully mature towards the luminal surface and, like in the normal crypt, lose the capacity to proliferate (C).

Articles


Using EPCAM immunohistochemistry, we detected an EPCAM-negative crypt focus that also proved to be MSH2-negative and MMR-defi cient as shown by analysis of microsatellite instability.

Functional MMR inactivation in crypt foci with no MMR protein expression was confi rmed by PCR-based analysis of microsatellite DNA sequences. We thus propose the MMR-defi cient crypt focus as a morpho-logical explanation for the occurrence of low-degree microsatellite instability in non-tumorous mucosa re-ported in previous studies.30,31 By diluting DNA isolated from non-tumoral mucosa to single-cell genome equivalents, Parsons and colleagues30 showed a signifi cant percentage of positivity for microsatellite instability in the diluted DNA samples obtained from an MLH1 germline mutation carrier, whereas microsatellite instability was not detected in undiluted DNA.30 The authors concluded that a few MMR-defi cient cells were present in the non-neoplastic mucosa in Lynch syndrome, but not detectable amid a large background of MMR-profi cient cells in undiluted DNA extracts. Our data suggest that MMR-defi cient crypt foci, rather than individual MMR-defi cient cells scattered throughout the mucosa, might contribute to the presence of micro-satellite instability in non-tumoral colonic mucosa.

The potential neoplastic nature of MMR-defi cient crypt foci will need careful assessment in further studies. MMR-defi cient crypt foci seem to be clonal lesions colonising complete crypts, which is suggestive of a clonal origin from a single ancestor cell. This fi nding is consistent with the concept that most non-tumoral adult intestinal crypts originate from four-to-six stem cells,32–34 but are themselves mostly clonal, because they are colonised by descendants of a single stem cell that is stochastically lost and replaced.35 The hypothesis that MMR-defi cient crypt foci might be neoplastic is supported by morphological characteristics suggestive of neoplastic growth: polycryptic MMR-defi cient crypt foci presented with aberrant branching and fi ssion of crypts, as well nuclear enlargement at the crypt bottom. The neoplastic concept is also compatible with molecular changes noted in MMR-defi cient crypt foci, namely MMR defi ciency-induced mutations of crucial tumour suppressor genes such as TGFBR2. TGFBR2 mutations had previously been

described as rather late events in microsatellite instability tumorigenesis by some studies,36 whereas others detected a high frequency of TGFBR2 mutations already in early colorectal adenomas with high microsatellite instability.20 The detection of TGFBR2 mutations in MMR-defi cient crypt foci shows that TGFBR2 mutations can occur very early after biallelic inactivation of MMR genes; however, the signifi cance of these mutations for the progression or regression of these lesions remains to be elucidated.

The abundance of MMR-defi cient crypt foci is in sharp contrast with the low number of adenomas or carcinomas becoming clinically manifest in Lynch syndrome.6 This disparity clearly shows that most, if not all, of these lesions will not progress to malignancy, potentially indicating elimination of a subset of MMR-defi cient crypt cells after stochastic loss of MMR-defi cient stem cells and replacement by MMR-profi cient stem cells within aff ected crypts.35 Second somatic events inacti-vating MMR in Lynch syndrome induce the formation of MMR-defi cient crypt foci at a very high frequency, but MMR inactivation alone is insuffi cient to induce the formation of adenomatous polyps. Accordingly, cancer development in Lynch syndrome apparently depends on other oncogenic events such as the inactivation of the APC gene. These events have been postulated to either precede MMR inactivation—as assumed in current models12,37,38—or follow MMR inactivation, if occurring within an MMR-defi cient crypt focus. MMR-defi cient crypt foci might thus provide the explanation to why molecular clock studies suggested the occurrence of MMR defi ciency in healthy appearing colon before transformation.39 Models of colorectal carcinogenesis in Lynch syndrome should be adapted accordingly.

This novel discovery of MMR-defi cient crypt foci provides a plausible explanation for the presence of T-cell immune responses against MMR defi ciency-specifi c frameshift peptide antigens in carriers of a Lynch syndrome mutation. MMR defi ciency might lead to mutations of microsatellites located in gene-encoding regions, causing shifts of the translational reading frame and giving rise to the generation of specifi c frameshift neoantigens.40 Frameshift neoantigen-specifi c immune responses are readily detectable in many patients with Lynch syndrome-associated cancers, but have surprisingly also been reported in carriers of a Lynch syndrome mutation without a previous history of tumours.41,42 The abundant occurrence of MMR-defi cient crypt foci in Lynch syndrome might explain an early exposure of the immune system towards MMR defi ciency-induced frameshift neoantigens. The auto-vaccination hypothesis indicates that early sensitisation of the immune system towards MMR defi ciency-induced frameshift neoanti gens mediated by the recurrent development of MMR-defi cient crypt foci might be one reason for the comparatively low penetrance of Lynch syndrome with respect to colorectal cancer. Conversely, MMR-defi cient crypt foci might be eliminated directly by the immune system, potentially


LS2 a lesion 1 wt –5 wt –1 wt

LS4 a lesion 5 –1 –1 –1 –1 wt

LS4 a lesion 1 –3 wt –6 wt –7 wt –1 wt

LS5 a lesion 1 wt –3 wt wt wt

LS7 lesion 3 –5 wt –4 wt

LS10 lesion 1 wt –5 wt –3 wt

LS11 lesion 1 wt –1 –2 wt

wt=wild type. Negative numbers indicate numbers of deleted nucleotides.

Table 3: Allele patterns of microsatellite markers in MMR-defi cient crypt foci

Articles


explaining why only very few of these lesions might progress to clinically manifest tumours. Lynch syndrome thus probably represents a unique model for the study of the immunoediting concept43 in humans, also in comparison with fi ndings in mouse models of immuno-genic colorectal cancer,44 and studies aiming at the characterisation of a potential host’s immune response against MMR-defi cient crypt foci are underway.

We report the identifi cation of a novel, previously unrecognised lesion, the MMR-defi cient crypt focus, which results from a second somatic mutation event which inactivates the MMR system in Lynch syndrome. On a molecular level, the MMR-defi cient crypt focus may play a similar role as the aberrant crypt focus in familial adenomatous polyposis; however, it does not share the properties as a potent cancer precursor inherent in the aberrant crypt foci. MMR-defi cient crypt foci are characterised by microsatellite instability and occur at an exceedingly high frequency in mucosa from carriers of a Lynch syndrome mutation, indicating that, by contrast with the low number of colorectal cancers, focal inactivation of the MMR system is a highly frequent event in Lynch syndrome.

ContributorsMK, CH, and HB contributed to the conception and design of the study.

PS and MvKD gave administrative support. MK, PS, MvKD, and HB

provided study materials or patients. MK, CH, AYV, and HB participated

in the collection and assembly of data. MK, CH, AYV, AB, and HB

participated in data analysis and interpretation. MK, CH, AYV, AB, PS,

MvKD, and HB participated in the writing of the report. MK, CH, AYV,

AB, PS, MvKD, and HB approved the fi nal report.


AcknowledgmentsWe thank Beate Kuchenbuch, Stefanie Kellner, Petra Höfl er, and

Daniel Baumgärtner for technical assistance. We thank Johannes Gebert

for careful reading of the manuscript. This work was funded partly by

grants of the Deutsche Krebshilfe (grant number 70-2369) and the

Deutsche Forschungsgemeinschaft (DFG, grant numbers BL-554/3 and

DFG-KFO 227).

References1 de la Chapelle A. Microsatellite instability. N Engl J Med 2003;

349: 209–10.

2 Ligtenberg MJ, Kuiper RP, Chan TL, et al. Heritable somatic methylation and inactivation of MSH2 in families with Lynch syndrome due to deletion of the 3’ exons of TACSTD1. Nat Genet 2009; 41: 112–17.

3 Thibodeau SN, French AJ, Roche PC, et al. Altered expression of hMSH2 and hMLH1 in tumors with microsatellite instability and genetic alterations in mismatch repair genes. Cancer Res 1996; 56: 4836–40.

4 Lindor NM, Burgart LJ, Leontovich O, et al. Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J Clin Oncol 2002; 20: 1043–48.

5 Quehenberger F, Vasen HF, van Houwelingen HC. Risk of colorectal and endometrial cancer for carriers of mutations of the hMLH1 and hMSH2 gene: correction for ascertainment. J Med Genet 2005; 42: 491–96.

6 Engel C, Rahner N, Schulmann K, et al. Effi cacy of annual colonoscopic surveillance in individuals with hereditary nonpolyposis colorectal cancer. Clin Gastroenterol Hepatol 2010; 8: 174–82.

7 Liljegren A, Barker G, Elliott F, et al. Prevalence of adenomas and hyperplastic polyps in mismatch repair mutation carriers among CAPP2 participants: report by the colorectal adenoma/carcinoma prevention programme 2. J Clin Oncol 2008; 26: 3434–39.

8 Lindgren G, Liljegren A, Jaramillo E, Rubio C, Lindblom A. Adenoma prevalence and cancer risk in familial non-polyposis colorectal cancer. Gut 2002; 50: 228–34.

9 de Jong AE, Morreau H, van PM, et al. The role of mismatch repair gene defects in the development of adenomas in patients with HNPCC. Gastroenterology 2004; 126: 42–48.

10 Roncucci L, Stamp D, Medline A, Cullen JB, Bruce WR. Identifi cation and quantifi cation of aberrant crypt foci and microadenomas in the human colon. Hum Pathol 1991; 22: 287–94.

11 Takayama T, Ohi M, Hayashi T, et al. Analysis of K-ras, APC, and beta-catenin in aberrant crypt foci in sporadic adenoma, cancer, and familial adenomatous polyposis. Gastroenterology 2001; 121: 599–611.

12 Fearon ER. Molecular genetics of colorectal cancer. Annu Rev Pathol 2011; 6: 479–507.

13 Giuff re G, Muller A, Brodegger T, et al. Microsatellite analysis of hereditary nonpolyposis colorectal cancer-associated colorectal adenomas by laser-assisted microdissection: correlation with mismatch repair protein expression provides new insights in early steps of tumorigenesis. J Mol Diagn 2005; 7: 160–70.

14 Pino MS, Mino-Kenudson M, Wildemore BM, et al. Defi cient DNA mismatch repair is common in Lynch syndrome-associated colorectal adenomas. J Mol Diagn 2009; 11: 238–47.

15 de la Chapelle A, Hampel H. Clinical relevance of microsatellite instability in colorectal cancer. J Clin Oncol 2010; 28: 3380–87.

16 Lynch HT, Smyrk T, Jass JR. Hereditary nonpolyposis colorectal cancer and colonic adenomas: aggressive adenomas? Semin Surg Oncol 1995; 11: 406–10.

17 Rijcken FE, Hollema H, Kleibeuker JH. Proximal adenomas in hereditary non-polyposis colorectal cancer are prone to rapid malignant transformation. Gut 2002; 50: 382–86.

18 Kloor M, Sutter C, Wentzensen N, et al. A large MSH2 Alu insertion mutation causes HNPCC in a German kindred. Hum Genet 2004; 115: 432–38.

19 Findeisen P, Kloor M, Merx S, et al. T25 repeat in the 3’ untranslated region of the CASP2 gene: a sensitive and specifi c marker for microsatellite instability in colorectal cancer. Cancer Res 2005; 65: 8072–78.

20 Woerner SM, Kloor M, Mueller A, et al. Microsatellite instability of selective target genes in HNPCC-associated colon adenomas. Oncogene 2005; 24: 2525–35.

21 R Development Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2011.

22 Gupta AK, Schoen RE. Aberrant crypt foci: are they intermediate endpoints of colon carcinogenesis in humans? Curr Opin Gastroenterol 2009; 25: 59–65.

23 Alrawi SJ, Schiff M, Carroll RE, et al. Aberrant crypt foci. Anticancer Res 2006; 26: 107–19.

24 Smith AJ, Stern HS, Penner M, et al. Somatic APC and K-ras codon 12 mutations in aberrant crypt foci from human colons. Cancer Res 1994; 54: 5527–30.

25 Heinen CD, Shivapurkar N, Tang Z, Groden J, Alabaster O. Microsatellite instability in aberrant crypt foci from human colons. Cancer Res 1996; 56: 5339–41.

26 Augenlicht LH, Richards C, Corner G, Pretlow TP. Evidence for genomic instability in human colonic aberrant crypt foci. Oncogene 1996; 12: 1767–72.

27 Pedroni M, Sala E, Scarselli A, et al. Microsatellite instability and mismatch-repair protein expression in hereditary and sporadic colorectal carcinogenesis. Cancer Res 2001; 61: 896–99.

28 Greenspan EJ, Cyr JL, Pleau DC, et al. Microsatellite instability in aberrant crypt foci from patients without concurrent colon cancer. Carcinogenesis 2007; 28: 769–76.

29 Kloor M, Voigt AY, Schackert HK, Schirmacher P, von Knebel Doeberitz M, Blaker H. Analysis of EPCAM protein expression in diagnostics of Lynch syndrome. J Clin Oncol 2011; 29: 223–27.

30 Parsons R, Li GM, Longley M, et al. Mismatch repair defi ciency in phenotypically normal human cells. Science 1995; 268: 738–40.

31 Dietmaier W, Gansbauer S, Beyser K, et al. Microsatellite instability in tumor and nonneoplastic colorectal cells from hereditary non-polyposis colorectal cancer and sporadic high microsatellite-instable tumor patients. Pathobiology 2000; 68: 227–31.

Articles


32 Marshman E, Booth C, Potten CS. The intestinal epithelial stem cell. Bioessays 2002; 24: 91–98.

33 Bjerknes M, Cheng H. Colossal crypts bordering colon adenomas in Apc(Min) mice express full-length Apc. Am J Pathol 1999; 154: 1831–34.

34 Humphries A, Wright NA. Colonic crypt organization and tumorigenesis. Nat Rev Cancer 2008; 8: 415–24.

35 Simons BD, Clevers H. Stem cell self-renewal in intestinal crypt. Exp Cell Res 2011; 317: 2719–24.

36 Grady WM, Rajput A, Myeroff L, et al. Mutation of the type II transforming growth factor-beta receptor is coincident with the transformation of human colon adenomas to malignant carcinomas. Cancer Res 1998; 58: 3101–04.

37 de la Chapelle A, Hampel H. Clinical relevance of microsatellite instability in colorectal cancer. J Clin Oncol 2010; 28: 3380–87.

38 Boland CR, Goel A. Microsatellite instability in colorectal cancer. Gastroenterology 2010; 138: 2073–87.

39 Shibata D. When does MMR loss occur during HNPCC progression? Cancer Biomark 2006; 2: 29–35.

40 Saeterdal I, Bjorheim J, Lislerud K, et al. Frameshift-mutation-derived peptides as tumor-specifi c antigens in inherited and spontaneous colorectal cancer. Proc Natl Acad Sci USA 2001; 98: 13255–60.

41 Schwitalle Y, Kloor M, Eiermann S, et al. Immune response against frameshift-induced neopeptides in HNPCC patients and healthy HNPCC mutation carriers. Gastroenterology 2008; 134: 988–97.

42 Reuschenbach M, Kloor M, Morak M, et al. Serum antibodies against frameshift peptides in microsatellite unstable colorectal cancer patients with Lynch syndrome. Fam Cancer 2010; 9: 173–79.

43 Dunn GP, Bruce AT, Ikeda H, Old LJ, Schreiber RD. Cancer immunoediting: from immunosurveillance to tumor escape. Nat Immunol 2002; 3: 991–98.

44 Czéh M, Loddenkemper C, Shalapour S, et al. The immune response to sporadic colorectal cancer in a novel mouse model. Oncogene 2010; 29: 6591–602.

The Lancet Oncology

Documents

Transcript of The Lancet Oncology