NCNJPM

Lednsain

C

R2Fc(MCCgM

©P

eoadjuvant Chemotherapy in Invasive Breastancer Results in a Lower Axillary Lymphode Count

ulie Bélanger, MD, Geneviève Soucy, MD, Lucas Sidéris, MD, FRCSC, Guy Leblanc, MD, FRCSC,ierre Drolet, MD, FRCP, Andrew Mitchell, MD, Yves-Eugène Leclerc, MD, FRCSC, Julie Beaudet, MD,ichel-Pierre Dufresne, MD, Pierre Dubé, MD, MSC, FACS, FRCSC

BACKGROUND: It is essential to have the highest level of confidence in axillary staging assessment. Manysurgeons and pathologists believe that fewer lymph nodes are present in axillary dissectionspecimens of women treated by neoadjuvant chemotherapy. Consequently, the purpose of thisstudy was to compare the lymph node counts of axillary dissection specimens from patientshaving received neoadjuvant chemotherapy with those of patients treated with primaryoperation.

STUDY DESIGN: A retrospective analysis of a prospective database from our institution identified 283 womenwith invasive breast cancer who underwent level I and II axillary lymph node dissections.Women from the neoadjuvant chemotherapy group (n � 107) were compared with those fromthe primary surgery group (n � 176). The total number of lymph nodes harvested was consid-ered as a continuous variable, but also dichotomized into two categories (� 10 and � 10). Itscorrelation with the different variables was analyzed.

RESULTS: The median number of lymph nodes retrieved in the neoadjuvant chemotherapy group was10.0 (range 0 to 38) compared with 12.5 (range 0 to 30) in the control group (p�0.002). Therewere also significantly more patients with fewer than 10 lymph nodes recovered in the neoad-juvant group (45 versus 28%, p�0.007). Logistic regression showed that neoadjuvant chemo-therapy was the only factor associated with retrieval of fewer than 10 lymph nodes.

CONCLUSIONS: This study suggests that administration of neoadjuvant chemotherapy to breast cancer patientsresults in a reduced number of lymph nodes retrieved in the axillary dissection specimens. (J Am

Coll Surg 2008;206:704–708. © 2008 by the American College of Surgeons)

avmaNcfisitwltlwkq

ymph node staging by axillary dissection is the most pow-rful prognosis factor in invasive breast cancer. It also re-uces locoregional recurrence.1-3 A minimum of 10 lymphodes is required in the axillary lymph node dissectionpecimen to provide accurate pathologic staging.3-5 Indeed,mathematical model developed by Kiricuta and Tausch5

n 1992 determined that the retrieval of 10 axillary lymphodes was the cutoff value to allow 90% specificity for

ompeting Interests Declared: None.

eceived August 22, 2007; Revised October 28, 2007; Accepted October 31,007.rom the Departments of General Surgery (Bélanger, Soucy), Surgical On-ology (Sidéris, Leblanc, Leclerc, Dubé), Anesthesiology (Drolet), PathologyMitchell), Medical Oncology (Beaudet), and Radiology (Dufresne),

aisonneuve-Rosemont Hospital, University of Montréal, Montréal, PQ,anada.orrespondence address: Lucas Sidéris, MD, FRCSC, Department of Sur-ery, Maisonneuve-Rosemont Hospital, 5415 boulevard de l’Assomption,

aontréal, Quebec, Canada, H1T 2M4.

7042008 by the American College of Surgeons

ublished by Elsevier Inc.

xillary lymph node dissection. In recent years, neoadju-ant chemotherapy in breast cancer has become the treat-ent of choice for patients with locally advanced disease

nd is increasingly being used for early stage disease.6-8

eoadjuvant chemotherapy, in the context of breast can-er, results in an increased rate of conservative surgery, of-ers the opportunity to assess chemosensitivity of the tumorn vivo, and correlates with outcomes for women demon-trating complete pathologic response.1,7,9,10 In addition, its well known that neoadjuvant treatment can downstagehe lymph node status for breast cancer, but it is not clearhether it also reduces the total number of harvested

ymph nodes available for pathologic assessment. In prac-ice, many surgeons and pathologists believe that fewerymph nodes are found in axillary dissection specimens ofomen treated with neoadjuvant chemotherapy. To ournowledge, only one retrospective study has addressed thisuestion; it revealed that 13% of patients treated by neo-

djuvant chemotherapy had fewer than 10 lymph nodes at

ISSN 1072-7515/08/$34.00doi:10.1016/j.jamcollsurg.2007.10.016

aganbb

MAptoi2agcujcdwl

fiptnnn

m

emnbTsfacsaEiswotew

mhmdcommtegdsap

twwiaitalds

RU

FwftFpdc

705Vol. 206, No. 4, April 2008 Bélanger et al Neoadjuvant Chemotherapy in Invasive Breast Cancer

xillary dissection compared with 3% in the operation-firstroup (p � 0.003).11 The primary goal of our study was tossess the impact of neoadjuvant therapy on the lymphode counts of women undergoing axillary dissection forreast cancer. We also looked for other variables that coulde associated with lower lymph node counts.

ETHODSprospective database from Maisonneuve-Rosemont Hos-

ital, University of Montréal, collected from January 2003o June 2006, provided the patient population. An analysisf the medical records identified 1,021 patients treated fornvasive breast cancer during this 3-year interval. Of these,83 women with nonmetastatic disease underwent level Ind II axillary lymph node dissections as part of their sur-ical treatment.3 Among them, 107 received neoadjuvanthemotherapy. Different chemotherapy protocols weresed, as illustrated in Figure 1. After completion of neoad-

uvant chemotherapy, patients underwent definitive breastancer surgery and a level I and II axillary lymph nodeissection by one of the four surgical oncologists. Noomen in the neoadjuvant group underwent sentinel

ymph node biopsy.12

A control group of patients was derived retrospectivelyrom the same database and included 176 women withnvasive breast cancer treated initially by operation. Theseatients also underwent level I and II lymph node dissec-ion by the same group of surgeons. Patients treated byeoadjuvant chemotherapy were excluded. The lymphode counts in the control group included both the senti-el and nonsentinel nodes.A total of 107 neoadjuvant chemotherapy and 176 pri-

igure 1. Neoadjuvant chemotherapy regimens used. Trastuzumabas used in combination with one of the chemotherapy regimens

or women who are HER2/neu positive. Taxanes include pacli-axel or docetaxel, which were used after a regimen of AC, CMF,AC or FEC. AC, doxorubicin and cyclophosphamide; CMF, cyclo-hosphamide, methotrexate, and fluorouacil; FAC, 5-fluorouacil,oxorubicin, and cyclophosphamide; FEC, 5-fluorouacil, epirubi-in, and cyclophosphamide.

ary surgery patients were included. Male patients were m

xcluded from the study, as were women known to haveetastatic disease before operation. Patients treated with

eoadjuvant hormone therapy were also excluded fromoth groups. Patient and tumor characteristics are listed inable 1. Pathologic evaluation of histologic subtype, tumorize, lymphovascular invasion, and nodal staging were per-ormed on the surgical resection specimens. When avail-ble, tumor type and grade, estrogen and progesterone re-eptors status, and HER2/neu status of the biopsypecimens were recorded before chemotherapy; if not avail-ble, results from the surgical specimens were recorded.strogen and progesterone receptors status was determined

mmunohistochemically, with a threshold of 10% oftained tumor cells considered positive. HER2/neu statusas also evaluated by immunohistochemistry, with the flu-rescence in situ hybridation (FISH) technique used whenhe immunohistochemical analysis was interpreted asquivocal. Results that remained doubtful were pooledith negative results for analysis.All axillary lymph nodes retrieved from surgical speci-ens were analyzed by one of the four pathologists at our

ospital. The contents of the axillary lymph node speci-ens were examined by blunt manual dissection; chemical

issolving of fatty tissue was not used in our center. Afterareful gross dissection of the specimen, each lymph node,r a representative sample of larger lymph nodes, was sub-itted for hematoxylin and phloxin staining followed byicroscopic evaluation. Cytokeratin immunohistochemis-

ry was used to confirm the presence of neoplastic cells inquivocal cases. Some women, only in the surgery-firstroup, had sentinel lymph node operations before axillaryissection. Each sentinel lymph node harvested was seriallyectioned, and multiple levels were stained with phloxinnd stained with an immunohistochemical stain againstancytokeratin.Univariate comparisons were performed between pa-

ients who received neoadjuvant chemotherapy and thoseho underwent primary operations. Continuous variablesere analyzed with Student t-test, and ordinal and categor-

cal variables were compared with Mann-Whitney U testnd Fisher or chi-square tests, respectively. To evaluate theirndependency, all variables that showed a difference in ei-her one of the two sets of comparisons were introduced innominal logistic regression model, with the number of

ymph nodes (fewer than 10 versus 10 or more) as theependent variable. A p value less than 0.05 was consideredignificant.

ESULTSnivariate comparisons showed that patients from the pri-

ary surgery group were significantly older than the neo-

aAawgHwa

la(Tltpt

pmnchdtmtt(

DAm

T

C

nMTT

N

H

G

H

LR

*

706 Bélanger et al Neoadjuvant Chemotherapy in Invasive Breast Cancer J Am Coll Surg

djuvant patients (mean 57 versus 52 years, p�0.0004).lso, tumors from patients in the neoadjuvant chemother-py group had a higher histologic grade. In contrast, thereas no significant difference between the two groups re-arding size of tumor, TNM classification,13 invasiveness,ER2/neu status, and histologic subtype. Finally, thereere fewer positive hormone receptors tumors in the neo-

djuvant chemotherapy group.The median number of lymph nodes retrieved by axil-

ary lymph node dissection in the neoadjuvant chemother-py group was 10.0 (range 0 to 38) compared with 12.5range 0 to 30) in the primary surgery group (p�0.0023).here were significantly more patients with fewer than 10

ymph nodes recovered in the neoadjuvant group (45%)han in the primary surgery group (28%), with a�0.0067 (Fig. 2). Among women who had axillary me-

able 1. Patients Characteristics

haracteristicNeoadjuvant groupn

107ean age, y (SD) 52 (10)

umor size, cm (SD) 2.2 (1.9)umor stageT0 14T1 34T2 48T3 4T4 7ode stageNp0 47Np1 37Np2 18Np3 5ER2/neuPositive 27Negative 59Doubtful 1

rade1 22 313 71istologyCanalar 96Lobular 8Mixed 2

ymphovascular invasion 29eceptor status*Positive 67Negative 39

Receptor status includes estrogen and progesterone receptors.

astasis at preoperative ultrasonographic evaluation, a m

athologic complete remission of the axilla (pCR) to che-otherapy (39.6%) was not associated with a lower lymph

ode count (median 11 versus 10 for pCR; p�0.22). Be-ause age, histologic grade, hormonal receptor status, andaving received neoadjuvant chemotherapy or not showedifferences in at least one of the two sets of comparisons,hey were introduced in a nominal logistic regressionodel. The only independent factor associated with re-

rieval of fewer than 10 lymph nodes during axillary dissec-ion was having received neoadjuvant chemotherapyp�0.005; Table 2).

ISCUSSIONs has been previously shown, neoadjuvant systemic che-otherapy may sterilize axillary metastases in approxi-

Primary surgery groupp Valuen %

17657 (11) 0.00042.1 (1.5) 0.764

0.7271 1

93 5467 394 27 4

0.66671 4082 4714 89 5

0.68824 1781 5738 27

0.000130 1764 3779 46

0.938156 8915 94 2

55 32 0.5020.001

133 8230 18

%

13324547

4435175

2868

1

23068

9082

28

6337

ately 25% of patients,7,14 but it is not clear whether it can

aaomccntt

tcsntpiwlvcccanf

diusgnsavnnWntdog

bollicv3lloIbngmd1c

mlstasierhsbas

Fbm

TEV

AGRN

707Vol. 206, No. 4, April 2008 Bélanger et al Neoadjuvant Chemotherapy in Invasive Breast Cancer

lso affect the total number of lymph nodes harvested atxillary dissection. Similar to Newman and colleagues,11 webtained fewer lymph nodes in axillary dissection speci-ens of women treated by neoadjuvant chemotherapy

ompared with those treated by primary operation. Indeed,omparisons between both the total number of lymphodes harvested and the ratio of patients exhibiting fewerhan 10 axillary lymph nodes were significantly different inhe two groups.

After neoadjuvant chemotherapy, microscopic examina-ion of lymph nodes may reveal accumulations of histio-ytes and display considerable lymphoid depletion, fibro-is, and hyalinization.15,16 So we can suppose that lymphodes may shrink or disappear secondary to lymphocyticoxicity and fibrotic involution. In addition, fibrotic re-lacement in the axilla changes its architecture and resultsn a more difficult area to dissect correctly during surgery,hich can also contribute to lowering the total number of

ymph nodes harvested. We can also hypothesize that pre-ious dissection from sentinel lymph node biopsy couldhallenge the surgical technique and eventually lower theount of lymph nodes collected because of the fibrosisaused by the cicatrization. Because no woman in the neo-djuvant chemotherapy group underwent sentinel lymphode biopsy, this strengthens the lower lymph node countound in the neoadjuvant chemotherapy group.

The total number of lymph nodes retrieved from axillaryissection may be influenced by age, tumor and nodal stag-

ng, extent of axillary dissection, and histopathologic eval-ation.17 It is noteworthy that the mean age of the primaryurgery group was higher than that of the neoadjuvantroup. As described by Schaapveld and associates,17 theumber of total lymph nodes retrieved from axillary dis-ection decreases with advancing age. So, this difference inge between the two groups does not explain why we har-ested fewer axillary lymph nodes in women treated byeoadjuvant chemotherapy. When looking at tumor andodal staging, differences were not significant (Table 1).e did not look for differences in the number of lymph

odes harvested between the four different surgeons. Al-hough they were not blinded, they all performed a stan-ard technique of axillary lymph node dissection includingnly levels I and II for all women in both groups. All sur-

able 2. Nominal Logistic Regression Model; Results offfect Likelihood Ratiosariable p Value

ge 0.9054rade 0.8142eceptor status 0.8035eoadjuvant chemotherapy 0.0051

ical axillary specimens were analyzed in a similar manner b

y pathologists by blunt dissection of the specimen with-ut a fat dissolving technique, serial sectioning, or radio-ogic evaluation. This technique of gross and visual patho-ogic examination of specimens can explain the discrepancyn percentages of fewer than 10 lymph nodes in our studyompared with that of Neuman and coworkers11 (45%ersus 13% in the neoadjuvant group and 28% versus.5% in the operation-first group). In Neuman and col-

eagues’ study, gross dissection of the specimens was fol-owed by serial sectioning of the adipose tissue in intervalsf 2 to 3 mm. In addition, some patients underwent a levelII axillary dissection, with a possible upgrade in the num-er of lymph nodes collected. Our data concerning theumber of lymph nodes obtained in the operation-firstroup is within the range reported in literature. Indeed, theedian number of lymph nodes harvested from axillary

issection in operation-first patients varied from 10 to7,4,17-19 with the percentage of fewer than 10 lymph nodesollected varying between 8%4 and 44%.19

Axillary lymph node involvement is an important deter-inant of tumor prognosis. A greater number of positive

ymph nodes is associated with a higher risk of developingystemic breast recurrence.2 Also, in neoadjuvant chemo-herapy, a complete pathologic axillary response is associ-ted with a favorable prognosis.20 As a matter of fact, intudies of neoadjuvant chemotherapy for breast cancer us-ng the multivariate method, the extent of metastatic dis-ase in the axilla is consistently the most significant variableelated to outcomes.2,10,21,22 So it is important to have aigh level of confidence with regard to axillary staging as-essment. But studies that determined the minimum num-er of lymph nodes necessary for reliable assessment ofxillary status were not done in the context of primaryystemic therapy. So retrieving fewer than 10 lymph nodes

igure 2. Lymph node count in axillary dissection of women treatedy neoadjuvant chemotherapy compared with those treated by pri-ary surgery, p � 0.0067. Black bar, neoadjuvant therapy; white

ar, primary surgery.

ic

trsruwaasb

A

S

A

A

D

C

R

1

1

1

1

1

1

1

1

1

1

2

2

2

708 Bélanger et al Neoadjuvant Chemotherapy in Invasive Breast Cancer J Am Coll Surg

n axillary dissections of women treated by neoadjuvanthemotherapy may not represent inadequate nodal staging.

In conclusion, this study suggests that the administra-ion of neoadjuvant chemotherapy in breast cancer patientsesults in a reduced number of lymph nodes in the surgicalpecimen when compared with the number in patients noteceiving neoadjuvant chemotherapy. Nevertheless, it isncertain whether this affects the staging or prognosis ofomen suffering from invasive breast cancer. More studies

ssessing the number of lymph nodes in surgical specimensfter neoadjuvant chemotherapy are needed to evaluate theignificance and safety concerning adequate staging ofreast cancer patients.

uthor Contributions

tudy conception and design: Bélanger, Soucy, Sidéris, Le-blanc, Dubé, Leclerc

cquisition of data: Bélanger, Soucy, Sidéris, Leblanc, Leclerc,Dubé

nalysis and interpretation of data: Bélanger, Soucy, Sidéris,Leblanc, Drolet, Mitchell, Beaudet, Dufresne, Dubé,Leclerc

rafting of manuscript: Bélanger, Soucy, Sidéris, Leblanc,Dubé

ritical revision: Bélanger, Soucy, Sidéris, Leblanc, Drolet,Mitchell, Beaudet, Dufresne, Dubé, Leclerc

EFERENCES

1. Kaufmann M, Hortobagyi GN, Goldhirsch A, et al. Recom-mendations from an international expert panel on the use ofneoadjuvant (primary) systemic treatment of operable breastcancer: an update. J Clin Oncol 2006;24:1940–1949.

2. Meric F MN, Buzdar AU. Prognostic implications of patholog-ical lymph node status after preoperative chemotherapy for op-erable T3N0Mo breast cancer. Ann Surg Oncol 2000;7:435–440.

3. Thain SK, Olivotto IA, Bouchard F, et al. Axillary dissection.The Steering Committee on Clinical Practice Guidelines for theCare and Treatment of Breast Cancer. Can Med Assoc J 1998;10:158.

4. Somner JE, Dixon JM, Thomas JS. Node retrieval in axillarylymph node dissections: recommendations for minimum num-bers to be confident about node negative status. J Clin Pathol2004;57:845–848.

5. Kiricuta CI, Tausch TJ. A mathematical model of axillary lymphnode involvement based on 1446 complete axillary dissectionsin patients with breast carcinoma. Cancer 1992;69:2496–2501.

6. von Minckwitz G, Blohmer JU, Raab G, et al. In vivo

chemosensitivity-adapted preoperative chemotherapy in pa-

tients with early-stage breast cancer: the GEPARTRIO pilotstudy. Ann Oncol 2005;16:56–63.

7. Mamounas EP. NSABP protocol B-27. Preoperative doxorubi-cin plus cyclophosphamide followed by preoperative or postop-erative docetaxel. Oncology (Williston Park) 1997;11:37–40.

8. Chen AM, Meric-Bernstam F, Hunt KK, et al. Breast conserva-tion after neoadjuvant chemotherapy: the M.D. Anderson Can-cer Center experience. J Clin Oncol 2004;22:2303–2312.

9. von Minckwitz G, Raab G, Caputo A, et al. Doxorubicin withcyclophosphamide followed by docetaxel every 21 days com-pared with doxorubicin and docetaxel every 14 days as preoper-ative treatment in operable breast cancer: the GEPARDUOstudy of the German Breast Group. J Clin Oncol 2005;3:2676–2685.

0. Fisher ER, Wang J, Bryant J, et al. Pathobiology of preoperativechemotherapy: findings from the National Surgical AdjuvantBreast and Bowel (NSABP) protocol B-18. Cancer 2002;95:681–695.

1. Neuman H Carey LA, Ollila DW, et al. Axillary lymph nodecount is lower after neoadjuvant chemotherapy. Am J Surg2006;191:827–829.

2. Lyman GH, Giuliano AE, Somerfield MR, et al. American So-ciety of Clinical Oncology guideline recommendations for sen-tinel lymph node biopsy in early-stage breast cancer. J ClinOncol 2005;23:7703–7720.

3. AJCC cancer staging manual, 6th ed. New York: Springer-Ver-lag; 2002.

4. Sachelarie I, Grossbard ML, Chadha M, et al. Primary systemictherapy of breast cancer. Oncologist 2006;11:574–589.

5. Kuroi K, Toi M, Tsuda H, et al. Issues in the assessment of thepathologic effect of primary systemic therapy for breast cancer.Breast Cancer 2006;13:38–48.

6. Newman L, Pernick NL, Adsay V, et al. Histopathologic evi-dence of tumor regression in the axillary lymph nodes of patientstreated with preoperative chemotherapy correlates with breastcancer outcome. Ann Surg Oncol 2003;10:734–739.

7. Schaapveld M, Otter R, De Vries EG, et al. Variability in axillarylymph node dissectionfor breast cancer. J Surg Oncol 2004;87:4–12.

8. Salama JK, Heimann R, Lin F, et al. Does the number of lymphnodes examined in patients with node-negative breast carci-noma have prognostic significance? Cancer 2004;103:664–671.

9. Wier L, Speers C, D’yachkova Y, Olivotto IA. Prognostic signif-icance of the number of axillary lymph nodes removed in pa-tients with node-negative breast cancer. J Clin Oncol 2002;20:1793–1799.

0. Hennessy BT, Hortobagyi GN, Rouzier R, et al. Outcome afterpathologic complete eradication of cytologically proven breastcancer axillary node metastases following primary chemother-apy. J Clin Oncol 2005;23:9304–9311.

1. Rouzier R, Extra JM, Klijanienko J, et al. Incidence and prog-nostic significance of complete axillary downstaging after pri-mary chemotherapy in breast cancer patients with T1 to T3tumors and cytologically proven axillary metastatic lymphnodes. J Clin Oncol 2002;20:1304–1310.

2. Fisher B, Bryant J, Wolmark N, et al. Effect of preoperativechemotherapy on the outcome of women with operable breast

cancer. J Clin Oncol 1998;16:2672–2685.