Pancreatic Cancer

a report by

Rob e r t C Mo e s i n g e r , MD

Assistant Professor in the Section of Surgical Oncology in the Department of Surgery at the University of

Maryland School of Medicine

Pancreatic adenocarcinoma is one of the great challengesto oncologic physicians and surgeons. Statistics for the USare typical. In 2006, it is expected that there will be33,730 new cases of pancreatic adenocarcinoma with32,300 deaths. This corresponds to a survival rate ofapproximately 4%. Resectional surgery, followed byadjuvant therapy, has been the best hope for long-termsurvival, but the results are still disappointing. Amongpatients with resectable disease, median, three-year andfive-year survival is approximately 25 months, 35% and20%, respectively, with multimodality treatment. Thesesurvival data are from series with large numbers ofpatients and are probably appropriate benchmarks againstwhich to measure further trials.

Neoadjuvant chemoradiation therapy has positivelyaltered the natural history of several cancers, especiallyesophageal and rectal cancer, and many investigators arelooking at neoadjuvant therapy in the treatment ofpancreatic cancer. I hope to give a brief overview ofwhere we stand in 2006 in terms of neoadjuvantchemoradiation therapy for pancreatic adenocarcinoma.

Chemoradiation therapy prior to pancreatic cancersurgery has a number of potential advantages overpostoperative therapy. First, pancreatic cancer can beconsidered a systemic illness in most cases andneoadjuvant therapy provides systemic treatment withindays of diagnosis. Second, patients undergo systemictherapy prior to the huge physiologic insult of apancreatic resection. They are therefore more likely tocomplete their therapy and less likely to havecomplications or delays associated with being in therelatively debilitated postoperative state. Followingpancreatic cancer surgery, 25% of patients’ adjuvanttherapy is not received, delayed or truncated due tocomplications, morbidity or patient refusal.Third, patientswhose disease progresses during neoadjuvant therapy self-select for aggressive tumor biology,never come to surgeryand avoid the morbidity of a pancreatic resection duringtheir short time of survival. Fourth, the ‘sterilization’ ofthe operative field by neoadjuvant therapy might decreasethe risk of disseminating microscopic tumor duringresection and also result in fewer positive margins. Fifth,radiation therapy generally is more effective in well-

vascularized preoperative tissue compared to relativelyhypoxic postoperative tissues. Sixth, bowel in theirradiated field is resected at the time of surgery,decreasing long-term effects of irradiated bowel.

A disadvantage is that these patients usually requiregreater pretreatment preparation than patients who areexpeditiously explored. Prior to neoadjuvant therapy atissue diagnosis (usually fine-needle aspiration (FNA) viaendoscopic ultrasound) is needed and biliary obstructionmust be controlled (usually via an endoscopic biliarystent).Many patients also undergo diagnostic laparoscopyto rule out occult metastatic disease. There is also thepossibility of making the surgery more difficult becauseof an irradiated field, however there is evidence thatneoadjuvant therapy may actually decrease pancreaticleak complications following surgery. The biggestpotential concern is the possibility that a resectablepatient may progress to an unresectable patient duringthe course of therapy, thus missing the ‘window ofopportunity’ for curative resection. In the end, theclinical relevance of these possible advantages anddisadvantages can only be determined through well-controlled, randomized, prospective trials whichdetermine overall morbidity and survival.

Challenges to determining if neoadjuvant therapy istruly advantageous to patients include:

1. lack of agreement among oncologists as to the mostappropriate neoadjuvant regimen, particularly giventhat current adjuvant therapies have modest responserates at best;

2. bias and fear among surgeons, patients and physiciansthat “expeditious exploration and resection wheneverpossible is the only hope for cure;” and

3. difficulty in accurately staging patients prior totherapy, particularly in determining involvement ofthe SMV/Portal Vein complex.

In evaluating neoadjuvant therapy trials, it isimportant to distinguish those reports in whichlocally unresectable cancers were treated, and after a

Neoad juvant Chemorad ia t ion for Pancreat i c Cancer

Robert C Moesinger, MD, is anAssistant Professor in the Section ofSurgical Oncology in theDepartment of Surgery at theUniversity of Maryland School ofMedicine. His interest is alimentarytract surgery and particularlybiliary and pancreatic pathology. Hegraduated from the University ofUtah Medical School in 1992 andcompleted the Halsted SurgicalResidency at Johns Hopkins Hospitalin 2000.

U S O N C O L O G I C A L D I S E A S E 2 0 0 6 51

moesinger_BOOK.qxp 11/7/06 2:12 pm Page 51

DOI: 10.17925/OHR.2006.00.01.51

clinical response (downstaging), patients wereexplored vs patients with radiographically resectabletumors who were treated with chemoradiation priorto planned surgery.

In terms of converting or downstaging locallyunresectable tumors into resectable disease, there aremany anecdotal reports and small series with a varietyof agents. Based on a recent review of availableliterature, possibly 20% of locally unresectable patientsmay become resectable, and of these only about 50%will actually have negative margins (see Table 1). Aninteresting exception is the small series of Wanebo et al.in which nine of 11 patients underwent resectionfollowing a regimen of cisplatin, 5-FU and radiation.However, two-thirds of these patients required a majorvascular resection and reconstruction (portal vein,superior mesenteric artery (SMA) or both) in order toobtain negative margins, and so these patients weretechnically not downstaged. Similarly, in a report byFogelman et al. (Columbia) using a novel neoadjuvantstrategy of gemcitabine, capecitabine, taxotere andradiation therapy, 12 of 17 (71%) patients underwentsuccessful resection but 50% of those required vascularreconstruction. Until there is a breakthrough inadjuvant therapy, it is unlikely that more than a handfulof patients will have successful downstaging of a locallyaggressive cancer following neoadjuvant therapy.

The most important question to be answered is the rolefor neoadjuvant therapy in patients who appear to beradiographically resectable at the time of presentation.Asummary of large series is provided in Table 2. (Thereare many smaller published series of neoadjuvanttherapy for pancreatic cancer from North America andEurope, but I have chosen to focus on the largest of theavailable series.) Perhaps the earliest center to offerneoadjuvant chemoradiation therapy for localized(radiographically resectable) pancreatic cancer was theFox Chase Cancer Center. Their 20-year experiencedemonstrates the safety and practicality of aneoadjuvant approach. Although median survival of

those patients treated preoperatively was statistically nobetter than those patients who had postoperativetherapy only, they note that as many as 86% of theneoadjuvant patients were considered “marginallyresectable” (as opposed to clearly resectable) at the timeof preoperative therapy whereas only 21% of patientswithout preoperative therapy were “marginallyresectable”. This selection bias makes analysis verydifficult and may obscure a small survival benefit inpreoperatively treated patients.This observation is likelytrue for the ECOG study published in 1988.

The MD Anderson experience uses strict computedtomography (CT) resectability criteria for studyinclusion, categorizing patients as “radiographicallyresectable” if the SMA and celiac axis are completelytumor free and the SMV/PV complex is patent).Encroachment on SMV/PV, however, was not anexclusion criterion and though their overall resectionrate was up to 74% in the gemcitabine group, 43% ofpatients required vascular resection and reconstruction,suggesting they were actually “marginally resectable”.The Duke experience uses the same CT resectabilitycriteria in categorizing patients as “potentially resectable”versus “locally advanced” and both groups were offeredneoadjuvant therapy. As expected, resectability rate wasmuch better in the potentially resectable group, butsurvival curves were the same through 24 months.

Conclusions that can be drawn from these series arethat neoadjuvant therapy is safe. Patients who beginwith radiologically resectable disease do better thanpatients who begin with locally advanced disease.Thereis a slight trend that gemcitabine may be more effectivethan 5-FU based therapy in terms of achieving asuccessful resection which is consistent with widespreadbelief that gemcitabine is advantageous over 5-FU inpancreatic cancer patients. ■

This article is continued, with references, in the ReferenceSection on the website supporting this briefing(www.touchbriefings.com).

Pancreatic Cancer

52 U S O N C O L O G I C A L D I S E A S E 2 0 0 6

Table 1:

Institution or Trial Time Frame Number of Patients Agents Used Resection Rate Median Survival

Fox Chase 1986-1996 89 XRT, 5-FU +/- Mito C 45% 20 monthsFox Chase 1996-2003 63 XRT & Gemcitabine 65% 20 monthsECOG (PD-289) 1990-1998 53 XRT, 5-FU & Mito C 45% 16 monthsMD Anderson 1990-1998 205 XRT, 5-FU or paclitaxel 60% 21 monthsMD Anderson 1998-2006 86 XRT, Gemcitabine 74% 21 monthsDuke 1994-2004 96^ XRT, 5-FU* Cisplatin#, 18% 23 months

88` +/- Mito C 55% (both groups combined)*In some patients capecitabine (Xeloda) was substituted for 5-FU.

#Cisplatin was eventually excluded secondary to toxicity.

^ “potentially resectable”

` “locally advanced”

moesinger_BOOK.qxp 11/7/06 2:12 pm Page 52

Do your patients know

that there is a

CROSSWALK to get

individual information about

their cancer ?

UNVEILING TUMOR BEHAVIOR

Mam

maP

rintTM

, C

upPr

intTM

and

Age

ndia

TMar

e re

gist

ered

tra

dem

arks

of

Age

ndia

BV,

Am

ster

dam

,Th

e N

ethe

rland

s

For more detailed information please visit our website: www.agendia.com

CUPPRINTTM

“A gene expression test to identify an unknown primary cancer”MG Erlander et al, ASCO 2004 [poster]

MAMMAPRINTTM

“A gene expression signature as a predictor of survival in

breast cancer”New England Journal of Medicine, Vol. 347, No. 25, 19 December 2002

Agendia_ad.qxp 12/7/06 3:34 pm Page 53