Sentinel Node Biopsy After Neoadjuvant Chemotherapy in Breast Cancer: Results From National Surgical Adjuvant Breast and Bowel Project Proto
http://www.100md.com
《临床肿瘤学》
the National Surgical Adjuvant Breast and Bowel Project
University of Pittsburgh Graduate School of Public Health
Allegheny General Hospital, Cancer Center, Pittsburgh, PA
Aultman Health Foundation, Canton, OH
Medical College of Virginia, Richmond, VA
Genesee Hospital, Rochester NY
Medical College of Wisconsin, Milwaukee, WI
The University of Texas, Health Sciences Center, Fredericksburg, TX
Centre Hspitalier de L'Universite de Montreal, PQ, Canada
ABSTRACT
PURPOSE: Experience with sentinel node biopsy (SNB) after neoadjuvant chemotherapy is limited. We examined the feasibility and accuracy of this procedure within a randomized trial in patients treated with neoadjuvant chemotherapy.
PATIENTS AND METHODS: During the conduct of National Surgical Adjuvant Breast and Bowel Project trial B-27, several participating surgeons attempted SNB before the required axillary dissection in 428 patients. All underwent lymphatic mapping and an attempt to identify and remove a sentinel node. Lymphatic mapping was performed with radioactive colloid (14.7%), with lymphazurin blue dye alone (29.9%), or with both (54.7%).
RESULTS: Success rate for the identification and removal of a sentinel node was 84.8%. Success rate increased significantly with the use of radioisotope (87.6% to 88.9%) versus with the use of lymphazurin alone (78.1%, P = .03). There were no significant differences in success rate according to clinical tumor size, clinical nodal status, age, or calendar year of random assignment. Of 343 patients who had SNB and axillary dissection, the sentinel nodes were positive in 125 patients and were the only positive nodes in 70 patients (56.0%). Of the 218 patients with negative sentinel nodes, nonsentinel nodes were positive in 15 (false-negative rate, 10.7%; 15 of 140 patients). There were no significant differences in false-negative rate according to clinical patient and tumor characteristics, method of lymphatic mapping, or breast tumor response to chemotherapy.
CONCLUSION: These results are comparable to those obtained from multicenter studies evaluating SNB before systemic therapy and suggest that the sentinel node concept is applicable following neoadjuvant chemotherapy.
INTRODUCTION
Based on a large number of single-institution and multicenter studies, sentinel node biopsy (SNB) is rapidly evolving as a replacement for level I and II axillary node dissection for staging the axilla of patients with operable breast cancer.1-11 Several randomized clinical trials have prospectively compared SNB with axillary node dissection in patients with either negative or positive sentinel nodes.9,12,13 Results from these trials will hopefully establish SNB as the standard of care for staging the axilla in patients with operable breast cancer.
During the past several years, the use of neoadjuvant chemotherapy has increased in patients with operable breast cancer. Although neoadjuvant chemotherapy results in an improvement in disease-free and overall survival similar to that obtained with adjuvant chemotherapy,14,15 there are some potential advantages to the neoadjuvant approach. The latter results in an increase in the rate of breast-conserving surgical procedures16,17 and offers the opportunity to assess in vivo chemosensitivity of breast tumors and eventually to correlate tumor response with long-term outcome.18
Despite the increasing use of both SNB and neoadjuvant chemotherapy in operable breast cancer, there is still limited information on the feasibility and accuracy of SNB following neoadjuvant chemotherapy. Several questions are unique to this setting. Does tumor response to neoadjuvant chemotherapy cause lymphatic scarring that could affect the drainage pattern, making sentinel node identification more difficult Does neoadjuvant chemotherapy have the same effect on involved sentinel nodes as it does on involved nonsentinel nodes So far, only small, single-institution studies have attempted to address these questions, and those have shown significant variability in the success rate of sentinel node identification and, more importantly, in the rate of false-negative sentinel nodes.19-25
The question of whether SNB is feasible and accurate following neoadjuvant chemotherapy is of significance, since neoadjuvant chemotherapy downstages axillary lymph nodes in a considerable proportion of patients (30% to 40%).16,26,27 If the value of SNB is demonstrated in this setting, a proportion of initially node-positive patients could eventually be spared from an axillary dissection if their involved sentinel nodes become uninvolved after the administration of neoadjuvant chemotherapy.
To address some of these questions, we evaluated our experience with SNB following neoadjuvant chemotherapy in National Surgical Adjuvant Breast and Bowel Project's (NSABP) trial B-27, which compared the effect of neoadjuvant doxorubicin/cyclophosphamide (AC) chemotherapy to that of neoadjuvant AC followed by neoadjuvant docetaxel and to that of neoadjuvant AC followed by adjuvant docetaxel. Before completing the study-mandated level I and II axillary dissection, 179 participating surgeons performed at least one lymphatic mapping and attempted SNB. Because SNB was not mandated in the study, there was no predefined protocol dictating the method of lymphatic mapping or the approach to SNB. Preliminary28 and updated results29 of this experience have been presented in abstract form, and in this report, we present in detail our final results.
PATIENTS AND METHODS
Eligibility
Patients eligible for the B-27 study had palpable, operable breast cancer diagnosed by fine-needle aspiration cytology or core needle biopsy. It was required that the tumor was movable in relationship to the skin, underlying muscle, and chest wall, and that it was free of any signs of locally advanced disease. Palpable axillary lymph nodes could not be fixed to each other or to underlying structures. Patients were stratified according to age ( 49 or 50 years), clinical tumor size ( 2.0, 2.1 to 4.0, or > 4.0 cm), and clinical nodal status (negative or positive).26
Treatment Regimens
Details regarding the treatment regimens have been reported previously in detail.26 Patients were randomly assigned to three groups: those in group 1 received four cycles of neoadjuvant AC followed by surgery (lumpectomy with axillary node dissection or modified radical mastectomy); those in group II received neoadjuvant AC as in group 1, followed by four cycles of docetaxel and then surgery; those in group 3 received neoadjuvant AC as in groups 1 and 2, followed by surgery, and then four cycles of adjuvant docetaxel. Tamoxifen was administered to all patients for five years and was initiated at the same time as chemotherapy. Patients undergoing lumpectomy also received breast radiotherapy. Regional radiotherapy or chest wall radiotherapy after mastectomy was prohibited.
Evaluation of Response to Neoadjuvant Chemotherapy
Breast tumor size and size of any palpable axillary nodes were obtained by physical examination before each cycle of chemotherapy and before surgery. Clinical tumor response was defined as complete if there was no evidence of palpable tumor in the breast and axillary lymph nodes. A reduction in total tumor size (breast primary + axillary nodes) by 50% at the time of surgery was considered to be a clinical partial response. An increase in total tumor size of > 50% (compared with pretreatment measurements) or the appearance of new suspicious ipsilateral axillary adenopathy was considered progressive disease. Tumors that did not meet the criteria for objective response or progression were considered to represent stable disease, and patients with either stable or progressive disease were considered clinical nonresponders. Surgical breast resection specimens were evaluated for pathologic tumor response. Patients who had no invasive cancer were considered to have had a pathologic complete response (pCR).
Review of the SNB Procedures
The records of patients participating in B-27 were reviewed to identify those who underwent SNB before the required axillary node dissection. Despite efforts to obtain complete information, 46 patients were excluded because of a missing operative or pathology report. The remaining 2,365 records were reviewed to identify cases in which lymphatic mapping was performed and an effort was made in surgery to identify at least one sentinel node. A total 428 such cases were identified, and information was abstracted about the method of lymphatic mapping used and the number and status of sentinel and nonsentinel nodes removed, as reported by the institutional pathologist. In the majority of the cases, nodal positivity was determined by hematoxylin and eosin staining. However, in a handful of cases, additional immunohistochemical staining was performed to further evaluate the status of sentinel nodes. Information was also obtained on the location of the sentinel nodes and the primary tumor within the breast.
Statistical Considerations
The primary end points used for statistical analysis were the success rate for identifying and removing a sentinel node in patients for whom an SNB was attempted (n = 428) and the false-negative rate of SNB in patients who were found to have at least one positive sentinel or nonsentinel node (n = 140). Comparisons of these rates were made by both analyzing simple proportions30 and performing logistic regression analyses31 adjusted for the stratification variables (age, clinical tumor size, and clinical nodal status). Comparisons of characteristics at the time of random assignment were made for patients who had an SNB attempted versus those who did not. Comparisons of the success and false-negative rates were made according to the calendar year of study entry, age at entry, clinical tumor size, clinical nodal status at random assignment, the number of sentinel node biopsies that the surgeon performed in the study, tumor location within the breast, and tumor response. When testing for heterogeneity of success or false-negative rates, 2 analyses were employed when at least 75% of the cells had expected counts of 5 or more. For cases in which this condition did not hold, exact tests were performed. When the interest was to investigate whether trends of success or false-negative rates were present, Fisher’s exact tests were performed according to the algorithm outlined in Agresti, Wackerly, and Boyett.32
RESULTS
Patient Population and Characteristics
Between December 1995 and December 2000, 2,411 patients were randomly assigned to NSABP Protocol B-27. In the 2,365 patients (98.1%) for whom operative and pathology reports were available, there were 428 (18.1%) who had lymphatic mapping and for whom an attempt was made to identify and remove a sentinel node. There were significant differences in the distribution of some of the patient and tumor characteristics between the group of patients who had an SNB attempted and the group of 1,937 patients (81.9%) who did not (Table 1). In general, at random assignment, patients in whom an SNB was attempted were less likely to have large tumors (P = .002) or clinically involved axillary nodes (P = .002) and more likely to be lumpectomy candidates as judged by their surgeon (P = .07).
Lymphatic Mapping Procedure
Lymphatic mapping was performed with radioactive colloid alone in 14.7% of the cases, with lymphazurin blue dye alone in 29.9%, or with both in 54.7%. In 0.7% of the cases, we were unable to determine the method of lymphatic mapping.
Success Rate of Sentinel Node Identification and Removal
Of the 428 patients in whom lymphatic mapping was attempted, at least one sentinel node was identified and removed in 363 (success rate, 84.8%; Table 2). The number of sentinel nodes removed ranged from one to 12, with only one removed in 40.5% of the patients, two in 27.0%, three in 14.6%, and four or more in 17.4%. In the 355 patients for whom sentinel node location was reported, the sentinel nodes were located exclusively in the axilla in 351 patients (98.9%). For the remaining four patients (1.1%), at least one sentinel node was found in both the axilla and the internal mammary chain.
The identification rate was significantly higher when radiocolloid was used for the lymphatic mapping than when lymphazurin blue dye alone was used (88.9% with radiocolloid alone, 87.6% with the combination of radiocolloid and lymphazurin blue dye, and 78.1% with lymphazurin blue dye alone; P = .03). There were no significant differences in the identification rate according to calendar year of entry to the study, age, clinical tumor size, clinical nodal status, tumor location within the breast, or number of sentinel node procedures that the surgeon performed in the study (Table 2). There was a nonsignificant trend toward increasing identification rate with calendar year of random assignment in the study (1996, 81.8%; 1997, 83.1%; 1998, 82.8%; 1999, 85.0%; and 2000, 89.9%; P for trend = .20). Similarly, there was a nonsignificant trend toward increased identification rate among surgeons who performed a higher number of sentinel node procedures in the study (81.7% for surgeons who performed one to four procedures v 88.4% for those who performed five or more procedures; P = .07).
Accuracy of Sentinel Node Status in Predicting the Status of the Axilla
Of the 363 patients in whom at least one sentinel node was identified and removed, 20 patients (5.5%) did not have the required axillary node dissection, leaving 343 patients in whom the accuracy of the sentinel node in correctly staging the axilla could be assessed (Table 3). At least one sentinel node was positive in 125 patients (36.4%), and these positive sentinel nodes were the only positive nodes in 70 (56.0%) of the 125. Of the 55 patients whose sentinel and nonsentinel nodes were positive, the median number of positive sentinel nodes was one and the median number of positive nonsentinel nodes was three (range, 1 to 21). Of the 218 patients with negative sentinel nodes, nonsentinel nodes were also negative in 203 patients (93.1%) and positive in 15 patients (6.9%). Thus, the sentinel node accurately predicted axillary nodal status in 328 of 343 patients with sentinel node biopsy (overall accuracy, 95.6%), in 125 of 140 node-positive patients (sensitivity, 89.3%), and in 203 of 218 patients with a negative SNB (negative predictive value, 93.1%). Of the 15 patients with a false-negative sentinel node, one nonsentinel node was positive in eight patients, two nodes in three patients, three nodes in two patients, and four or more nodes in two patients.
There were no significant differences in the rate of false-negative sentinel nodes according to calendar year of entry to the study, age, clinical tumor size, clinical nodal status, method of lymphatic mapping, tumor location within the breast, and number of surgeries that the surgeon performed in the study (Table 4). There was a nonsignificant trend toward higher false-negative rate with increasing clinical tumor size (5.0%, 9.7%, and 13.8% for tumors 2.0, 2.1-4.0, and > 4.0 cm, respectively; P = .33). Method of lymphatic mapping was also not significantly related to the false-negative rate (lymphazurin blue dye alone, 14.0%; radiocolloid alone, 5.0%; for radiocolloid in combination with lymphazurin blue dye, 9.3%; P = .50). There was a slight but nonsignificant heterogeneity in false-negative rate (P = .17) according to the location of the breast tumor. Finally, the number of sentinel node procedures that surgeons performed did not significantly affect the false-negative rate (12.3% for those who performed one to four surgeries and 9.0% for those who performed five or more surgeries; P = .71).
Accuracy of SNB According to Clinical and Pathologic Tumor Response
There were no significant differences in sentinel node false-negative rates according to clinical and pathologic breast tumor response to preoperative chemotherapy (P = .90, Table 5). Among patients who had pathologically positive nodes, the false-negative rate was 11.1% (one patients of nine) in those with clinical complete response and pCR in the breast; it was 8.9% (four patients of 45) in those with clinical complete response but residual invasive cancer on pathology; and it was 12.3% (nine patients of 73) among those with any other type of clinical response (partial response, stable disease, progressive disease).
Since previous results from B-27 have demonstrated that patients who achieve a pCR have the lowest rate of axillary nodal positivity (15.5%),26 we examined the overall sentinel node inaccuracy rate among all patients (pathologically node negative and node positive) according to their clinical and pathologic breast tumor response to preoperative chemotherapy. As expected, the inaccuracy of SNB was lowest among patients with pCR (one false negative in 58 patients, 1.7%) compared with those with clinical complete response but residual invasive cancer in the breast (four false negatives in 99 patients, 4.0%), and those with any other type of clinical response (nine false negatives in 165 patients, 5.5%). However, these differences were not statistically significant (P = .58). The one patient with a false-negative sentinel node after pCR in the breast had one sentinel node and seven nonsentinel nodes removed, of which only one was positive.
DISCUSSION
The use of SNB has increased dramatically during the past several years, and experienced surgeons are starting to perform this procedure in lieu of an axillary dissection. However, confirmation of equivalence between the two approaches is still pending until results from large randomized trials become available. It seems unlikely that the same rigorous process of randomized clinical trials will be followed with regard to demonstrating the value of SNB following neoadjuvant chemotherapy. Thus, it is probable that if the results from the currently ongoing randomized trials are favorable for SNB, surgeons will extrapolate the applicability of such results to patients who have received neoadjuvant chemotherapy. It is likely that the demonstration of feasibility and efficacy of sentinel node biopsy after neoadjuvant chemotherapy will rest primarily on the retrospective review of single-institution experiences and multicenter series in which SNB is followed by completion axillary dissection.
Our study is by far the largest to date of SNB following neoadjuvant chemotherapy. The number of patients included in our study is almost double the total number of patients included in seven single-institution series19-25 (Table 6) and is comparable in number of patients to some multicenter studies of sentinel node biopsy before systemic therapy6,33(Table 7). Thus, our results represent a significant proportion of the available evidence on the feasibility and accuracy of SNB following neoadjuvant chemotherapy.
The overall success rate of 84.8% that we observed for sentinel node identification and removal is similar to the identification rate reported in single-institution and multicenter studies evaluating SNB before systemic therapy.2,6,10 Our results are also concordant with those from a recent meta-analysis of 69 SNB studies, which showed that a 90% or greater success rate of identifying and removing at least one sentinel node was achieved in only 45% of the studies.11 Our finding that there was a higher identification rate when the lymphatic mapping included radioactive colloid is in agreement with results of several other studies showing that sentinel node identification rates are generally higher when colloid is included either alone5,6,34 or with blue dye,3,10 as opposed to when blue dye is used alone.4,35
We found no significant differences in the sentinel node identification rate according to age, clinical tumor size, clinical nodal status, or tumor location within the breast. Some multicenter studies of SNB before systemic therapy have demonstrated that older age (> 50 years) is a significant predictor of a decreased rate of sentinel node identification,6,33 but others have not.10 Similar to what we found, tumor size has not been found to be a significant predictor of the sentinel node identification rate in studies in which SNB was performed at the time of diagnosis.6,10,33 In our study, however, we did observe a non–statistically significant trend towards higher false-negative rates with increasing clinical tumor size (5.0%, 9.7%, and 13.8% for tumors 2.0, 2.1-4.0, and > 4.0 cm, respectively; P = .33). It is possible that such differences could have reached statistical significance if more patients were included in our study. In NSABP B-27, 45% of the patients had tumors measuring 4 cm before neoadjuvant chemotherapy. Such patients would not be considered good candidates for SNB at the time of diagnosis. However, it is reassuring that the inaccuracy rate in our study was lowest for patients achieving a pCR. Thus, until more data become available, for patients who present with large tumors and are about to undergo neoadjuvant chemotherapy, it is prudent to consider offering SNB alone only to those who demonstrate a pCR in the breast. Interestingly, in one of the latter studies,6 a prior excisional biopsy significantly decreased the sentinel node identification rate, indicating that there may be a potential advantage to performing SNB following neoadjuvant chemotherapy when the primary tumor is intact. With regard to tumor location within the breast, a lower identification rate was seen with medial lesions in two studies of SNB before systemic therapy,6,33 but it reached statistical significance in only one.6 Finally, although we found no significant differences in the identification rate by calendar year of random assignment or by the number of SNBs performed in the study by each surgeon, it was reassuring to observe that there were nonsignificant trends of progressively increasing success rates with calendar year and by the number of SNBs that a surgeon performed in the study. A surgeon's experience with the procedure has been shown to be a significant predictor of identification rate in some multicenter studies of SNB before systemic therapy.6,33
The false-negative rate we observed compares favorably with that of other multicenter trials of SNB before systemic therapy that have shown false-negative rates between 7% and 13% (Table 7). 6,10,33 An overview analysis of 69 SNB studies including more than 10,000 patients reported an overall false-negative rate of 8%.11 Recently, Veronesi et al12 reported a false-negative rate of 9% as part of a smaller randomized trial in which SNB alone was compared with the same procedure followed by an axillary node dissection. Although we found no significant differences in the false-negative rate according to any of the factors examined, there were nonsignificant trends toward higher false-negative rates with increasing tumor size and with the use of blue dye alone as the sole method of lymphatic mapping. In fact, the false-negative rate in our study when isotope was included for lymphatic mapping was 8.4% as opposed to 14.0% when isotope was not used. In the previously mentioned trials, several factors have been examined as potential predictors of false-negative rate (age, tumor location, tumor size, type of prior biopsy, surgeons' experience, type of colloid, injection time, number of positive nodes, and number of sentinel nodes). Given the small number of false-negative sentinel nodes in each of these trials, conclusions about factors affecting sentinel node false-negative rates are tentative at best. Only tumor location has been shown to correlate significantly with false-negative rates, but the results are not consistent among studies (outer and upper-outer location was associated with higher false-negative rate in two studies,6,10 but medial location resulted in higher false-negative rate in the third study).33
Until our study, only small, single-institution studies have examined the efficacy of lymphatic mapping and the accuracy of SNB after neoadjuvant chemotherapy (Table 6).19-25 These studies generally included patients with operable as well as locally advanced breast cancer and reported sentinel node identification rates between 84% and 94%. Similar to our experience, sentinel node identification rates were generally higher when radiocolloid was included for the lymphatic mapping. The rates of a false-negative sentinel node were quite variable (0% to 33%), leading to different conclusions about the accuracy of the procedure in this setting. However, the small size of each of these studies can easily account for the wide variability of the estimates. When one examines all of these studies combined, the overall identification rate is 86.6%, which is comparable to our experience and to that reported by other single-institution7 or multicenter series6,10,33 of SNB before systemic therapy. The overall false-negative rate when all the studies are combined is 12.5%, which is only slightly higher than the 10.7% false-negative rate disclosed in our study and the rates seen in studies of SNB before systemic therapy.1,2,7,10 Our data, together with those from all the single-institution series combined, indicate that when large data sets of SNB after neoadjuvant chemotherapy are compared with multicenter series of SNB before systemic therapy, the identification rates and false-negative rates are comparable between the two approaches.
There are several limitations as well as strengths in our study. Limitations include the lack of a predetermined protocol dictating the method of lymphatic mapping, the specific surgical approach to the SNB procedure, and the standardized pathologic assessment of the sentinel nodes. As a result, several different methods were used for lymphatic mapping, SNB, and pathologic evaluation of the specimen with various degrees of success. Furthermore, since performing an SNB was entirely at the discretion of the individual surgeon and an axillary dissection was to follow in all cases, there was no requirement for the surgeons to look diligently for all sentinel nodes as they would if an SNB protocol were in place or if SNB were not going to be followed by completion axillary dissection.
Our strengths include the fact that these data were collected uniformly as part of a prospective randomized trial that included a large number of patients undergoing SNB, thus increasing the reliability of the estimates of sentinel node identification and false-negative rates. In addition, since no particular training or certification was required for the surgeon in order to have his/her cases included in the study, our results should be more representative of the "real world" experience with SNB rather than of the experience of surgeons highly trained in the procedure.
Establishing the feasibility and accuracy of sentinel node biopsy after neoadjuvant chemotherapy is of considerable importance as we attempt to eliminate axillary dissections in the majority of patients with operable breast cancer. Neoadjuvant chemotherapy with anthracycline, cyclophosphamide, and taxane-containing regimens has been shown to sterilize involved axillary nodes in approximately 35% to 40% of patients.26,27 More active regimens that may hopefully increase the conversion rate even further are currently being developed. Thus, for patients with documented positive axillary nodes (positive fine-needle aspiration) or for those with a high likelihood of axillary nodal involvement (clinically palpable nodes, large breast primaries), neoadjuvant chemotherapy and sentinel node biopsy could potentially spare the patient axillary dissection.
Since in our study patients who achieved pCR had the lowest chance of having involved axillary nodes (15.5%) and, as a result, the lowest risk for a negative sentinel node to be inaccurate (1.7%), it makes intuitive sense that if one were to start using SNB alone in patients who have received neoadjuvant chemotherapy, it would be safer to do so first in patients who have achieved pCR after neoadjuvant chemotherapy.
Some have proposed that patients who are considered candidates for neoadjuvant chemotherapy should have an SNB performed before rather than after neoadjuvant chemotherapy.36,37 The proponents of this approach argue that information on the status of the axillary nodes can be obtained without the potential confounding effects of prior neoadjuvant chemotherapy, and sentinel node-negative patients may not require completion axillary dissection after neoadjuvant chemotherapy. Such an approach, however, does not take advantage of the downstaging effect of neoadjuvant chemotherapy in the axillary nodes. Patients with a positive sentinel node will generally require an axillary node dissection either before or after neoadjuvant chemotherapy. Furthermore, this approach commits the patient to two surgical procedures irrespective of the status of the sentinel node and contributes minimally to the decision to use neoadjuvant chemotherapy, since this decision is usually made based on clinical patient and tumor characteristics and not on the pathologic status of the axillary nodes. Finally, knowing what the status of the sentinel node is before neoadjuvant chemotherapy is of little or no value relative to the subsequent decision to administer additional adjuvant chemotherapy or adjuvant loco-regional radiotherapy.
In summary, the results of our study support the notion that the SNB concept is applicable in breast cancer patients with operable disease who have received neoadjuvant chemotherapy. The rates for sentinel node identification and false-negative sentinel nodes that we observed are similar to those that have been seen when SNB is carried out before systemic therapy. Thus, should SNB become established as the standard method for staging the axilla, it will be reasonable to utilize this technique in patients who have received prior neoadjuvant chemotherapy, expanding the utility of both interventions.
Authors' Disclosures of Potential Conflicts of Interest
The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Consultant/Advisory Role: Eleftherios P. Mamounas, AstraZeneca, Aventis. Honoraria: Eleftherios P. Mamounas, Aventis; Harry D. Bear, Aventis; D. Lawrence Wickerham, AstraZeneca. Research Funding: Christopher B. Caldwell, NSABP. For a detailed description of these categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section of Information for Contributors found in the front of each issue.
Acknowledgment
We thank Barbara C. Good, PhD, for editorial assistance with this manuscript.
NOTES
Supported by Public Health Service Grants U10CA-12027, U10CA-69974, U10CA-37377 and U10CA-69651 from the National Cancer Institute, National Institutes of Health Department of Health and Human Services.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Krag DN, Weaver DL, Alex JC, et al: Surgical resection and radiolocalization of the sentinel lymph node in breast cancer using a gamma probe. Surg Oncol 2:335-340, 1993
Giuliano AE, Kirgan DM, Guenther JM, et al: Lymphatic mapping and sentinel lymphadenectomy for breast cancer. Ann Surg 220:391-401, 1994
Albertini JJ, Lyman GH, Cox C, et al: Lymphatic mapping and sentinel node biopsy in the patient with breast cancer. JAMA 276:1818-1822, 1996
Giuliano AE, Jones RC, Brennan M, et al: Sentinel lymphadenectomy in breast cancer. J Clin Oncol 15:2345-2350, 1997
Veronesi U, Paganelli G, Galimberti V, et al: Sentinel-node biopsy to avoid axillary dissection in breast cancer with clinically negative lymph-nodes. Lancet 349:1864-1867, 1997
Krag D, Weaver D, Ashikaga T, et al: The sentinel node in breast cancer—A multicenter validation study. N Engl J Med 339:941-946, 1998
O'Hea BJ, Hill AD, El-Shirbiny AM, et al: Sentinel lymph node biopsy in breast cancer: Initial experience at Memorial Sloan-Kettering Cancer Center. J Am Coll Surg 186:423-427, 1998
Bedrosian I, Reynolds C, Mick R, et al: Accuracy of sentinel lymph node biopsy in patients with large primary breast tumors. Cancer 88:2540-2545, 2000
Giuliano AE, Haigh PI, Brennan MB, et al: Prospective observational study of sentinel lymphadenectomy without further axillary dissection in patients with sentinel node-negative breast cancer. J Clin Oncol 18:2553-2559 [Erratum: J Clin Oncol 18:3877, 2000]
McMasters KM, Tuttle TM, Carlson DJ, et al: Sentinel lymph node biopsy for breast cancer: A suitable alternative to routine axillary dissection in multi-institutional practice when optimal technique is used. J Clin Oncol 18:2560-2566, 2000
Kim T, Agboola O, Lyman GH, et al: Lymphatic mapping and sentinel lymph node sampling in breast cancer: A meta-analysis. Proc Am Soc Clin Oncol 21:36a, 2002 (abstr 139)
Veronesi U, Paganelli G, Viale G, et al: A randomized comparison of sentinel-node biopsy with routine axillary dissection in breast cancer. N Engl J Med 349:546-553, 2003
Clarke D, Mansel RE, group OboAt: The learning curve in sentinel node biopsy (SNB) in breast cancer: Results from the ALMANAC trial. Br Ca Res Treat 69:212, 2001 (abstr 15)
Fisher B, Bryant J, Wolmark N, et al: Effect of preoperative chemotherapy on the outcome of women with operable breast cancer. J Clin Oncol 16:2672-2685, 1998
van der Hage JA, van de Velde CJ, Julien JP, et al: Preoperative chemotherapy in primary operable breast cancer: Results from the European Organization for Research and Treatment of Cancer trial 10902. J Clin Oncol 19:4224-4237, 2001
Fisher B, Brown A, Mamounas E, et al: Effect of preoperative chemotherapy on local-regional disease in women with operable breast cancer: Findings from National Surgical Adjuvant Breast and Bowel Project B-18. J Clin Oncol 15:2483-2493, 1997
Bonadonna G, Veronesi U, Brambilla C, et al: Primary chemotherapy to avoid mastectomy in tumors with diameters of three centimeters or more. J Natl Cancer Inst 82:1539-1545, 1990
Fisher B, Mamounas EP: Preoperative chemotherapy: A model for studying the biology and therapy of primary breast cancer. J Clin Oncol 13:537-540, 1995
Breslin TM, Cohen L, Sahin A, et al: Sentinel lymph node biopsy is accurate after neoadjuvant chemotherapy for breast cancer. J Clin Oncol 18:3480-3486, 2000
Nason KS, Anderson BO, Byrd DR, et al: Increased false negative sentinel node biopsy rates after preoperative chemotherapy for invasive breast carcinoma. Cancer 89:2187-2194, 2000
Julian TB, Patel N, Dusi D, et al: Sentinel lymph node biopsy after neoadjuvant chemotherapy for breast cancer. Am J Surg 182:407-410, 2001
Fernandez A, Cortes M, Benito E, et al: Gamma probe sentinel node localization and biopsy in breast cancer patients treated with a neoadjuvant chemotherapy scheme. Nucl Med Commun 22:361-366, 2001
Stearns V, Ewing CA, Slack R, et al: Sentinel lymphadenectomy after neoadjuvant chemotherapy for breast cancer may reliably represent the axilla except for inflammatory breast cancer. Ann Surg Oncol 9:235-242, 2002
Haid A, Tausch C, Lang A, et al: Is sentinel lymph node biopsy reliable and indicated after preoperative chemotherapy in patients with breast carcinoma Cancer 92:1080-1084, 2001
Miller AR, Thomason VE, Yeh IT, et al: Analysis of sentinel lymph node mapping with immediate pathologic review in patients receiving preoperative chemotherapy for breast carcinoma. Ann Surg Oncol 9:243-247, 2002
Bear HD, Anderson S, Brown A, et al: The effect 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 21:4165-4174, 2003
Gianni L, Baselga J, Eiermann W, et al: First report of the European Cooperative Trial in operable breast cancer (ECTO): Effects of primary systemic therapy (PST) on local-regional disease. Proc Am Soc Clin Oncol 21:34a, 2002 (abstr 132)
Mamounas E, Brown A, Smith R, et al.: Sentinel node biopsy (SNB) following neoadjuvant chemotherapy in breast cancer (BC): Results from NSABP B-27. Proceedings Society of Surgical Oncology:21, 2001 (abstr 51)
Mamounas E, Brown A, Smith R, et al.: Accuracy of sentinel node biopsy after neoadjuvant chemotherapy in breast cancer: Updated results from NSABP B-27. Proc Am Soc Clin Oncol 21:36a, 2002 (abstr 140)
Snedecor G, Cochran W: Statistical Methods (ed 7). Ames, Iowa, The Iowa State University Press, 1980, pp 124-128
Hosmer DW, Lemeshow S: Applied Logistic Regression. New York, NY, John Wiley & Sons, Inc, 1989
Agresti A, Wackerly D, Boyett JM: Exact conditional tests for cross-classifications: Approximation of attained significance levels. Psychometrika 44:75-84, 1979
Tafra L, Lannin DR, Swanson MS, et al: Multicenter trial of sentinel node biopsy for breast cancer using both technetium sulfur colloid and isosulfan blue dye. Ann Surg 233:51-59, 2001
Pijpers R, Meijer S, Hoekstra OS, et al: Impact of lymphoscintigraphy on sentinel node identification with technetium-99m-colloidal albumin in breast cancer. J Nucl Med 38:366-368, 1997
Guenther JM, Krishnamoorthy M, Tan LR: Sentinel lymphadenectomy for breast cancer in a community managed care setting. Cancer J Sci Am 3:336-340, 1997
Sabel MS, Schott AF, Kleer CG, et al: Sentinel node biopsy prior to neoadjuvant chemotherapy. Am J Surg 186:102-105, 2003
Zirngibl C, Steinfeld-Birg D, Vogt H, et al: Sentinel lymph node biopsy before neoadjuvant chemotherapy—Conservation of breast and axilla. Br Ca Res Treat 76:S129, 2002 (suppl 1; abstr 516)(Eleftherios P. Mamounas, )
University of Pittsburgh Graduate School of Public Health
Allegheny General Hospital, Cancer Center, Pittsburgh, PA
Aultman Health Foundation, Canton, OH
Medical College of Virginia, Richmond, VA
Genesee Hospital, Rochester NY
Medical College of Wisconsin, Milwaukee, WI
The University of Texas, Health Sciences Center, Fredericksburg, TX
Centre Hspitalier de L'Universite de Montreal, PQ, Canada
ABSTRACT
PURPOSE: Experience with sentinel node biopsy (SNB) after neoadjuvant chemotherapy is limited. We examined the feasibility and accuracy of this procedure within a randomized trial in patients treated with neoadjuvant chemotherapy.
PATIENTS AND METHODS: During the conduct of National Surgical Adjuvant Breast and Bowel Project trial B-27, several participating surgeons attempted SNB before the required axillary dissection in 428 patients. All underwent lymphatic mapping and an attempt to identify and remove a sentinel node. Lymphatic mapping was performed with radioactive colloid (14.7%), with lymphazurin blue dye alone (29.9%), or with both (54.7%).
RESULTS: Success rate for the identification and removal of a sentinel node was 84.8%. Success rate increased significantly with the use of radioisotope (87.6% to 88.9%) versus with the use of lymphazurin alone (78.1%, P = .03). There were no significant differences in success rate according to clinical tumor size, clinical nodal status, age, or calendar year of random assignment. Of 343 patients who had SNB and axillary dissection, the sentinel nodes were positive in 125 patients and were the only positive nodes in 70 patients (56.0%). Of the 218 patients with negative sentinel nodes, nonsentinel nodes were positive in 15 (false-negative rate, 10.7%; 15 of 140 patients). There were no significant differences in false-negative rate according to clinical patient and tumor characteristics, method of lymphatic mapping, or breast tumor response to chemotherapy.
CONCLUSION: These results are comparable to those obtained from multicenter studies evaluating SNB before systemic therapy and suggest that the sentinel node concept is applicable following neoadjuvant chemotherapy.
INTRODUCTION
Based on a large number of single-institution and multicenter studies, sentinel node biopsy (SNB) is rapidly evolving as a replacement for level I and II axillary node dissection for staging the axilla of patients with operable breast cancer.1-11 Several randomized clinical trials have prospectively compared SNB with axillary node dissection in patients with either negative or positive sentinel nodes.9,12,13 Results from these trials will hopefully establish SNB as the standard of care for staging the axilla in patients with operable breast cancer.
During the past several years, the use of neoadjuvant chemotherapy has increased in patients with operable breast cancer. Although neoadjuvant chemotherapy results in an improvement in disease-free and overall survival similar to that obtained with adjuvant chemotherapy,14,15 there are some potential advantages to the neoadjuvant approach. The latter results in an increase in the rate of breast-conserving surgical procedures16,17 and offers the opportunity to assess in vivo chemosensitivity of breast tumors and eventually to correlate tumor response with long-term outcome.18
Despite the increasing use of both SNB and neoadjuvant chemotherapy in operable breast cancer, there is still limited information on the feasibility and accuracy of SNB following neoadjuvant chemotherapy. Several questions are unique to this setting. Does tumor response to neoadjuvant chemotherapy cause lymphatic scarring that could affect the drainage pattern, making sentinel node identification more difficult Does neoadjuvant chemotherapy have the same effect on involved sentinel nodes as it does on involved nonsentinel nodes So far, only small, single-institution studies have attempted to address these questions, and those have shown significant variability in the success rate of sentinel node identification and, more importantly, in the rate of false-negative sentinel nodes.19-25
The question of whether SNB is feasible and accurate following neoadjuvant chemotherapy is of significance, since neoadjuvant chemotherapy downstages axillary lymph nodes in a considerable proportion of patients (30% to 40%).16,26,27 If the value of SNB is demonstrated in this setting, a proportion of initially node-positive patients could eventually be spared from an axillary dissection if their involved sentinel nodes become uninvolved after the administration of neoadjuvant chemotherapy.
To address some of these questions, we evaluated our experience with SNB following neoadjuvant chemotherapy in National Surgical Adjuvant Breast and Bowel Project's (NSABP) trial B-27, which compared the effect of neoadjuvant doxorubicin/cyclophosphamide (AC) chemotherapy to that of neoadjuvant AC followed by neoadjuvant docetaxel and to that of neoadjuvant AC followed by adjuvant docetaxel. Before completing the study-mandated level I and II axillary dissection, 179 participating surgeons performed at least one lymphatic mapping and attempted SNB. Because SNB was not mandated in the study, there was no predefined protocol dictating the method of lymphatic mapping or the approach to SNB. Preliminary28 and updated results29 of this experience have been presented in abstract form, and in this report, we present in detail our final results.
PATIENTS AND METHODS
Eligibility
Patients eligible for the B-27 study had palpable, operable breast cancer diagnosed by fine-needle aspiration cytology or core needle biopsy. It was required that the tumor was movable in relationship to the skin, underlying muscle, and chest wall, and that it was free of any signs of locally advanced disease. Palpable axillary lymph nodes could not be fixed to each other or to underlying structures. Patients were stratified according to age ( 49 or 50 years), clinical tumor size ( 2.0, 2.1 to 4.0, or > 4.0 cm), and clinical nodal status (negative or positive).26
Treatment Regimens
Details regarding the treatment regimens have been reported previously in detail.26 Patients were randomly assigned to three groups: those in group 1 received four cycles of neoadjuvant AC followed by surgery (lumpectomy with axillary node dissection or modified radical mastectomy); those in group II received neoadjuvant AC as in group 1, followed by four cycles of docetaxel and then surgery; those in group 3 received neoadjuvant AC as in groups 1 and 2, followed by surgery, and then four cycles of adjuvant docetaxel. Tamoxifen was administered to all patients for five years and was initiated at the same time as chemotherapy. Patients undergoing lumpectomy also received breast radiotherapy. Regional radiotherapy or chest wall radiotherapy after mastectomy was prohibited.
Evaluation of Response to Neoadjuvant Chemotherapy
Breast tumor size and size of any palpable axillary nodes were obtained by physical examination before each cycle of chemotherapy and before surgery. Clinical tumor response was defined as complete if there was no evidence of palpable tumor in the breast and axillary lymph nodes. A reduction in total tumor size (breast primary + axillary nodes) by 50% at the time of surgery was considered to be a clinical partial response. An increase in total tumor size of > 50% (compared with pretreatment measurements) or the appearance of new suspicious ipsilateral axillary adenopathy was considered progressive disease. Tumors that did not meet the criteria for objective response or progression were considered to represent stable disease, and patients with either stable or progressive disease were considered clinical nonresponders. Surgical breast resection specimens were evaluated for pathologic tumor response. Patients who had no invasive cancer were considered to have had a pathologic complete response (pCR).
Review of the SNB Procedures
The records of patients participating in B-27 were reviewed to identify those who underwent SNB before the required axillary node dissection. Despite efforts to obtain complete information, 46 patients were excluded because of a missing operative or pathology report. The remaining 2,365 records were reviewed to identify cases in which lymphatic mapping was performed and an effort was made in surgery to identify at least one sentinel node. A total 428 such cases were identified, and information was abstracted about the method of lymphatic mapping used and the number and status of sentinel and nonsentinel nodes removed, as reported by the institutional pathologist. In the majority of the cases, nodal positivity was determined by hematoxylin and eosin staining. However, in a handful of cases, additional immunohistochemical staining was performed to further evaluate the status of sentinel nodes. Information was also obtained on the location of the sentinel nodes and the primary tumor within the breast.
Statistical Considerations
The primary end points used for statistical analysis were the success rate for identifying and removing a sentinel node in patients for whom an SNB was attempted (n = 428) and the false-negative rate of SNB in patients who were found to have at least one positive sentinel or nonsentinel node (n = 140). Comparisons of these rates were made by both analyzing simple proportions30 and performing logistic regression analyses31 adjusted for the stratification variables (age, clinical tumor size, and clinical nodal status). Comparisons of characteristics at the time of random assignment were made for patients who had an SNB attempted versus those who did not. Comparisons of the success and false-negative rates were made according to the calendar year of study entry, age at entry, clinical tumor size, clinical nodal status at random assignment, the number of sentinel node biopsies that the surgeon performed in the study, tumor location within the breast, and tumor response. When testing for heterogeneity of success or false-negative rates, 2 analyses were employed when at least 75% of the cells had expected counts of 5 or more. For cases in which this condition did not hold, exact tests were performed. When the interest was to investigate whether trends of success or false-negative rates were present, Fisher’s exact tests were performed according to the algorithm outlined in Agresti, Wackerly, and Boyett.32
RESULTS
Patient Population and Characteristics
Between December 1995 and December 2000, 2,411 patients were randomly assigned to NSABP Protocol B-27. In the 2,365 patients (98.1%) for whom operative and pathology reports were available, there were 428 (18.1%) who had lymphatic mapping and for whom an attempt was made to identify and remove a sentinel node. There were significant differences in the distribution of some of the patient and tumor characteristics between the group of patients who had an SNB attempted and the group of 1,937 patients (81.9%) who did not (Table 1). In general, at random assignment, patients in whom an SNB was attempted were less likely to have large tumors (P = .002) or clinically involved axillary nodes (P = .002) and more likely to be lumpectomy candidates as judged by their surgeon (P = .07).
Lymphatic Mapping Procedure
Lymphatic mapping was performed with radioactive colloid alone in 14.7% of the cases, with lymphazurin blue dye alone in 29.9%, or with both in 54.7%. In 0.7% of the cases, we were unable to determine the method of lymphatic mapping.
Success Rate of Sentinel Node Identification and Removal
Of the 428 patients in whom lymphatic mapping was attempted, at least one sentinel node was identified and removed in 363 (success rate, 84.8%; Table 2). The number of sentinel nodes removed ranged from one to 12, with only one removed in 40.5% of the patients, two in 27.0%, three in 14.6%, and four or more in 17.4%. In the 355 patients for whom sentinel node location was reported, the sentinel nodes were located exclusively in the axilla in 351 patients (98.9%). For the remaining four patients (1.1%), at least one sentinel node was found in both the axilla and the internal mammary chain.
The identification rate was significantly higher when radiocolloid was used for the lymphatic mapping than when lymphazurin blue dye alone was used (88.9% with radiocolloid alone, 87.6% with the combination of radiocolloid and lymphazurin blue dye, and 78.1% with lymphazurin blue dye alone; P = .03). There were no significant differences in the identification rate according to calendar year of entry to the study, age, clinical tumor size, clinical nodal status, tumor location within the breast, or number of sentinel node procedures that the surgeon performed in the study (Table 2). There was a nonsignificant trend toward increasing identification rate with calendar year of random assignment in the study (1996, 81.8%; 1997, 83.1%; 1998, 82.8%; 1999, 85.0%; and 2000, 89.9%; P for trend = .20). Similarly, there was a nonsignificant trend toward increased identification rate among surgeons who performed a higher number of sentinel node procedures in the study (81.7% for surgeons who performed one to four procedures v 88.4% for those who performed five or more procedures; P = .07).
Accuracy of Sentinel Node Status in Predicting the Status of the Axilla
Of the 363 patients in whom at least one sentinel node was identified and removed, 20 patients (5.5%) did not have the required axillary node dissection, leaving 343 patients in whom the accuracy of the sentinel node in correctly staging the axilla could be assessed (Table 3). At least one sentinel node was positive in 125 patients (36.4%), and these positive sentinel nodes were the only positive nodes in 70 (56.0%) of the 125. Of the 55 patients whose sentinel and nonsentinel nodes were positive, the median number of positive sentinel nodes was one and the median number of positive nonsentinel nodes was three (range, 1 to 21). Of the 218 patients with negative sentinel nodes, nonsentinel nodes were also negative in 203 patients (93.1%) and positive in 15 patients (6.9%). Thus, the sentinel node accurately predicted axillary nodal status in 328 of 343 patients with sentinel node biopsy (overall accuracy, 95.6%), in 125 of 140 node-positive patients (sensitivity, 89.3%), and in 203 of 218 patients with a negative SNB (negative predictive value, 93.1%). Of the 15 patients with a false-negative sentinel node, one nonsentinel node was positive in eight patients, two nodes in three patients, three nodes in two patients, and four or more nodes in two patients.
There were no significant differences in the rate of false-negative sentinel nodes according to calendar year of entry to the study, age, clinical tumor size, clinical nodal status, method of lymphatic mapping, tumor location within the breast, and number of surgeries that the surgeon performed in the study (Table 4). There was a nonsignificant trend toward higher false-negative rate with increasing clinical tumor size (5.0%, 9.7%, and 13.8% for tumors 2.0, 2.1-4.0, and > 4.0 cm, respectively; P = .33). Method of lymphatic mapping was also not significantly related to the false-negative rate (lymphazurin blue dye alone, 14.0%; radiocolloid alone, 5.0%; for radiocolloid in combination with lymphazurin blue dye, 9.3%; P = .50). There was a slight but nonsignificant heterogeneity in false-negative rate (P = .17) according to the location of the breast tumor. Finally, the number of sentinel node procedures that surgeons performed did not significantly affect the false-negative rate (12.3% for those who performed one to four surgeries and 9.0% for those who performed five or more surgeries; P = .71).
Accuracy of SNB According to Clinical and Pathologic Tumor Response
There were no significant differences in sentinel node false-negative rates according to clinical and pathologic breast tumor response to preoperative chemotherapy (P = .90, Table 5). Among patients who had pathologically positive nodes, the false-negative rate was 11.1% (one patients of nine) in those with clinical complete response and pCR in the breast; it was 8.9% (four patients of 45) in those with clinical complete response but residual invasive cancer on pathology; and it was 12.3% (nine patients of 73) among those with any other type of clinical response (partial response, stable disease, progressive disease).
Since previous results from B-27 have demonstrated that patients who achieve a pCR have the lowest rate of axillary nodal positivity (15.5%),26 we examined the overall sentinel node inaccuracy rate among all patients (pathologically node negative and node positive) according to their clinical and pathologic breast tumor response to preoperative chemotherapy. As expected, the inaccuracy of SNB was lowest among patients with pCR (one false negative in 58 patients, 1.7%) compared with those with clinical complete response but residual invasive cancer in the breast (four false negatives in 99 patients, 4.0%), and those with any other type of clinical response (nine false negatives in 165 patients, 5.5%). However, these differences were not statistically significant (P = .58). The one patient with a false-negative sentinel node after pCR in the breast had one sentinel node and seven nonsentinel nodes removed, of which only one was positive.
DISCUSSION
The use of SNB has increased dramatically during the past several years, and experienced surgeons are starting to perform this procedure in lieu of an axillary dissection. However, confirmation of equivalence between the two approaches is still pending until results from large randomized trials become available. It seems unlikely that the same rigorous process of randomized clinical trials will be followed with regard to demonstrating the value of SNB following neoadjuvant chemotherapy. Thus, it is probable that if the results from the currently ongoing randomized trials are favorable for SNB, surgeons will extrapolate the applicability of such results to patients who have received neoadjuvant chemotherapy. It is likely that the demonstration of feasibility and efficacy of sentinel node biopsy after neoadjuvant chemotherapy will rest primarily on the retrospective review of single-institution experiences and multicenter series in which SNB is followed by completion axillary dissection.
Our study is by far the largest to date of SNB following neoadjuvant chemotherapy. The number of patients included in our study is almost double the total number of patients included in seven single-institution series19-25 (Table 6) and is comparable in number of patients to some multicenter studies of sentinel node biopsy before systemic therapy6,33(Table 7). Thus, our results represent a significant proportion of the available evidence on the feasibility and accuracy of SNB following neoadjuvant chemotherapy.
The overall success rate of 84.8% that we observed for sentinel node identification and removal is similar to the identification rate reported in single-institution and multicenter studies evaluating SNB before systemic therapy.2,6,10 Our results are also concordant with those from a recent meta-analysis of 69 SNB studies, which showed that a 90% or greater success rate of identifying and removing at least one sentinel node was achieved in only 45% of the studies.11 Our finding that there was a higher identification rate when the lymphatic mapping included radioactive colloid is in agreement with results of several other studies showing that sentinel node identification rates are generally higher when colloid is included either alone5,6,34 or with blue dye,3,10 as opposed to when blue dye is used alone.4,35
We found no significant differences in the sentinel node identification rate according to age, clinical tumor size, clinical nodal status, or tumor location within the breast. Some multicenter studies of SNB before systemic therapy have demonstrated that older age (> 50 years) is a significant predictor of a decreased rate of sentinel node identification,6,33 but others have not.10 Similar to what we found, tumor size has not been found to be a significant predictor of the sentinel node identification rate in studies in which SNB was performed at the time of diagnosis.6,10,33 In our study, however, we did observe a non–statistically significant trend towards higher false-negative rates with increasing clinical tumor size (5.0%, 9.7%, and 13.8% for tumors 2.0, 2.1-4.0, and > 4.0 cm, respectively; P = .33). It is possible that such differences could have reached statistical significance if more patients were included in our study. In NSABP B-27, 45% of the patients had tumors measuring 4 cm before neoadjuvant chemotherapy. Such patients would not be considered good candidates for SNB at the time of diagnosis. However, it is reassuring that the inaccuracy rate in our study was lowest for patients achieving a pCR. Thus, until more data become available, for patients who present with large tumors and are about to undergo neoadjuvant chemotherapy, it is prudent to consider offering SNB alone only to those who demonstrate a pCR in the breast. Interestingly, in one of the latter studies,6 a prior excisional biopsy significantly decreased the sentinel node identification rate, indicating that there may be a potential advantage to performing SNB following neoadjuvant chemotherapy when the primary tumor is intact. With regard to tumor location within the breast, a lower identification rate was seen with medial lesions in two studies of SNB before systemic therapy,6,33 but it reached statistical significance in only one.6 Finally, although we found no significant differences in the identification rate by calendar year of random assignment or by the number of SNBs performed in the study by each surgeon, it was reassuring to observe that there were nonsignificant trends of progressively increasing success rates with calendar year and by the number of SNBs that a surgeon performed in the study. A surgeon's experience with the procedure has been shown to be a significant predictor of identification rate in some multicenter studies of SNB before systemic therapy.6,33
The false-negative rate we observed compares favorably with that of other multicenter trials of SNB before systemic therapy that have shown false-negative rates between 7% and 13% (Table 7). 6,10,33 An overview analysis of 69 SNB studies including more than 10,000 patients reported an overall false-negative rate of 8%.11 Recently, Veronesi et al12 reported a false-negative rate of 9% as part of a smaller randomized trial in which SNB alone was compared with the same procedure followed by an axillary node dissection. Although we found no significant differences in the false-negative rate according to any of the factors examined, there were nonsignificant trends toward higher false-negative rates with increasing tumor size and with the use of blue dye alone as the sole method of lymphatic mapping. In fact, the false-negative rate in our study when isotope was included for lymphatic mapping was 8.4% as opposed to 14.0% when isotope was not used. In the previously mentioned trials, several factors have been examined as potential predictors of false-negative rate (age, tumor location, tumor size, type of prior biopsy, surgeons' experience, type of colloid, injection time, number of positive nodes, and number of sentinel nodes). Given the small number of false-negative sentinel nodes in each of these trials, conclusions about factors affecting sentinel node false-negative rates are tentative at best. Only tumor location has been shown to correlate significantly with false-negative rates, but the results are not consistent among studies (outer and upper-outer location was associated with higher false-negative rate in two studies,6,10 but medial location resulted in higher false-negative rate in the third study).33
Until our study, only small, single-institution studies have examined the efficacy of lymphatic mapping and the accuracy of SNB after neoadjuvant chemotherapy (Table 6).19-25 These studies generally included patients with operable as well as locally advanced breast cancer and reported sentinel node identification rates between 84% and 94%. Similar to our experience, sentinel node identification rates were generally higher when radiocolloid was included for the lymphatic mapping. The rates of a false-negative sentinel node were quite variable (0% to 33%), leading to different conclusions about the accuracy of the procedure in this setting. However, the small size of each of these studies can easily account for the wide variability of the estimates. When one examines all of these studies combined, the overall identification rate is 86.6%, which is comparable to our experience and to that reported by other single-institution7 or multicenter series6,10,33 of SNB before systemic therapy. The overall false-negative rate when all the studies are combined is 12.5%, which is only slightly higher than the 10.7% false-negative rate disclosed in our study and the rates seen in studies of SNB before systemic therapy.1,2,7,10 Our data, together with those from all the single-institution series combined, indicate that when large data sets of SNB after neoadjuvant chemotherapy are compared with multicenter series of SNB before systemic therapy, the identification rates and false-negative rates are comparable between the two approaches.
There are several limitations as well as strengths in our study. Limitations include the lack of a predetermined protocol dictating the method of lymphatic mapping, the specific surgical approach to the SNB procedure, and the standardized pathologic assessment of the sentinel nodes. As a result, several different methods were used for lymphatic mapping, SNB, and pathologic evaluation of the specimen with various degrees of success. Furthermore, since performing an SNB was entirely at the discretion of the individual surgeon and an axillary dissection was to follow in all cases, there was no requirement for the surgeons to look diligently for all sentinel nodes as they would if an SNB protocol were in place or if SNB were not going to be followed by completion axillary dissection.
Our strengths include the fact that these data were collected uniformly as part of a prospective randomized trial that included a large number of patients undergoing SNB, thus increasing the reliability of the estimates of sentinel node identification and false-negative rates. In addition, since no particular training or certification was required for the surgeon in order to have his/her cases included in the study, our results should be more representative of the "real world" experience with SNB rather than of the experience of surgeons highly trained in the procedure.
Establishing the feasibility and accuracy of sentinel node biopsy after neoadjuvant chemotherapy is of considerable importance as we attempt to eliminate axillary dissections in the majority of patients with operable breast cancer. Neoadjuvant chemotherapy with anthracycline, cyclophosphamide, and taxane-containing regimens has been shown to sterilize involved axillary nodes in approximately 35% to 40% of patients.26,27 More active regimens that may hopefully increase the conversion rate even further are currently being developed. Thus, for patients with documented positive axillary nodes (positive fine-needle aspiration) or for those with a high likelihood of axillary nodal involvement (clinically palpable nodes, large breast primaries), neoadjuvant chemotherapy and sentinel node biopsy could potentially spare the patient axillary dissection.
Since in our study patients who achieved pCR had the lowest chance of having involved axillary nodes (15.5%) and, as a result, the lowest risk for a negative sentinel node to be inaccurate (1.7%), it makes intuitive sense that if one were to start using SNB alone in patients who have received neoadjuvant chemotherapy, it would be safer to do so first in patients who have achieved pCR after neoadjuvant chemotherapy.
Some have proposed that patients who are considered candidates for neoadjuvant chemotherapy should have an SNB performed before rather than after neoadjuvant chemotherapy.36,37 The proponents of this approach argue that information on the status of the axillary nodes can be obtained without the potential confounding effects of prior neoadjuvant chemotherapy, and sentinel node-negative patients may not require completion axillary dissection after neoadjuvant chemotherapy. Such an approach, however, does not take advantage of the downstaging effect of neoadjuvant chemotherapy in the axillary nodes. Patients with a positive sentinel node will generally require an axillary node dissection either before or after neoadjuvant chemotherapy. Furthermore, this approach commits the patient to two surgical procedures irrespective of the status of the sentinel node and contributes minimally to the decision to use neoadjuvant chemotherapy, since this decision is usually made based on clinical patient and tumor characteristics and not on the pathologic status of the axillary nodes. Finally, knowing what the status of the sentinel node is before neoadjuvant chemotherapy is of little or no value relative to the subsequent decision to administer additional adjuvant chemotherapy or adjuvant loco-regional radiotherapy.
In summary, the results of our study support the notion that the SNB concept is applicable in breast cancer patients with operable disease who have received neoadjuvant chemotherapy. The rates for sentinel node identification and false-negative sentinel nodes that we observed are similar to those that have been seen when SNB is carried out before systemic therapy. Thus, should SNB become established as the standard method for staging the axilla, it will be reasonable to utilize this technique in patients who have received prior neoadjuvant chemotherapy, expanding the utility of both interventions.
Authors' Disclosures of Potential Conflicts of Interest
The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Consultant/Advisory Role: Eleftherios P. Mamounas, AstraZeneca, Aventis. Honoraria: Eleftherios P. Mamounas, Aventis; Harry D. Bear, Aventis; D. Lawrence Wickerham, AstraZeneca. Research Funding: Christopher B. Caldwell, NSABP. For a detailed description of these categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section of Information for Contributors found in the front of each issue.
Acknowledgment
We thank Barbara C. Good, PhD, for editorial assistance with this manuscript.
NOTES
Supported by Public Health Service Grants U10CA-12027, U10CA-69974, U10CA-37377 and U10CA-69651 from the National Cancer Institute, National Institutes of Health Department of Health and Human Services.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Krag DN, Weaver DL, Alex JC, et al: Surgical resection and radiolocalization of the sentinel lymph node in breast cancer using a gamma probe. Surg Oncol 2:335-340, 1993
Giuliano AE, Kirgan DM, Guenther JM, et al: Lymphatic mapping and sentinel lymphadenectomy for breast cancer. Ann Surg 220:391-401, 1994
Albertini JJ, Lyman GH, Cox C, et al: Lymphatic mapping and sentinel node biopsy in the patient with breast cancer. JAMA 276:1818-1822, 1996
Giuliano AE, Jones RC, Brennan M, et al: Sentinel lymphadenectomy in breast cancer. J Clin Oncol 15:2345-2350, 1997
Veronesi U, Paganelli G, Galimberti V, et al: Sentinel-node biopsy to avoid axillary dissection in breast cancer with clinically negative lymph-nodes. Lancet 349:1864-1867, 1997
Krag D, Weaver D, Ashikaga T, et al: The sentinel node in breast cancer—A multicenter validation study. N Engl J Med 339:941-946, 1998
O'Hea BJ, Hill AD, El-Shirbiny AM, et al: Sentinel lymph node biopsy in breast cancer: Initial experience at Memorial Sloan-Kettering Cancer Center. J Am Coll Surg 186:423-427, 1998
Bedrosian I, Reynolds C, Mick R, et al: Accuracy of sentinel lymph node biopsy in patients with large primary breast tumors. Cancer 88:2540-2545, 2000
Giuliano AE, Haigh PI, Brennan MB, et al: Prospective observational study of sentinel lymphadenectomy without further axillary dissection in patients with sentinel node-negative breast cancer. J Clin Oncol 18:2553-2559 [Erratum: J Clin Oncol 18:3877, 2000]
McMasters KM, Tuttle TM, Carlson DJ, et al: Sentinel lymph node biopsy for breast cancer: A suitable alternative to routine axillary dissection in multi-institutional practice when optimal technique is used. J Clin Oncol 18:2560-2566, 2000
Kim T, Agboola O, Lyman GH, et al: Lymphatic mapping and sentinel lymph node sampling in breast cancer: A meta-analysis. Proc Am Soc Clin Oncol 21:36a, 2002 (abstr 139)
Veronesi U, Paganelli G, Viale G, et al: A randomized comparison of sentinel-node biopsy with routine axillary dissection in breast cancer. N Engl J Med 349:546-553, 2003
Clarke D, Mansel RE, group OboAt: The learning curve in sentinel node biopsy (SNB) in breast cancer: Results from the ALMANAC trial. Br Ca Res Treat 69:212, 2001 (abstr 15)
Fisher B, Bryant J, Wolmark N, et al: Effect of preoperative chemotherapy on the outcome of women with operable breast cancer. J Clin Oncol 16:2672-2685, 1998
van der Hage JA, van de Velde CJ, Julien JP, et al: Preoperative chemotherapy in primary operable breast cancer: Results from the European Organization for Research and Treatment of Cancer trial 10902. J Clin Oncol 19:4224-4237, 2001
Fisher B, Brown A, Mamounas E, et al: Effect of preoperative chemotherapy on local-regional disease in women with operable breast cancer: Findings from National Surgical Adjuvant Breast and Bowel Project B-18. J Clin Oncol 15:2483-2493, 1997
Bonadonna G, Veronesi U, Brambilla C, et al: Primary chemotherapy to avoid mastectomy in tumors with diameters of three centimeters or more. J Natl Cancer Inst 82:1539-1545, 1990
Fisher B, Mamounas EP: Preoperative chemotherapy: A model for studying the biology and therapy of primary breast cancer. J Clin Oncol 13:537-540, 1995
Breslin TM, Cohen L, Sahin A, et al: Sentinel lymph node biopsy is accurate after neoadjuvant chemotherapy for breast cancer. J Clin Oncol 18:3480-3486, 2000
Nason KS, Anderson BO, Byrd DR, et al: Increased false negative sentinel node biopsy rates after preoperative chemotherapy for invasive breast carcinoma. Cancer 89:2187-2194, 2000
Julian TB, Patel N, Dusi D, et al: Sentinel lymph node biopsy after neoadjuvant chemotherapy for breast cancer. Am J Surg 182:407-410, 2001
Fernandez A, Cortes M, Benito E, et al: Gamma probe sentinel node localization and biopsy in breast cancer patients treated with a neoadjuvant chemotherapy scheme. Nucl Med Commun 22:361-366, 2001
Stearns V, Ewing CA, Slack R, et al: Sentinel lymphadenectomy after neoadjuvant chemotherapy for breast cancer may reliably represent the axilla except for inflammatory breast cancer. Ann Surg Oncol 9:235-242, 2002
Haid A, Tausch C, Lang A, et al: Is sentinel lymph node biopsy reliable and indicated after preoperative chemotherapy in patients with breast carcinoma Cancer 92:1080-1084, 2001
Miller AR, Thomason VE, Yeh IT, et al: Analysis of sentinel lymph node mapping with immediate pathologic review in patients receiving preoperative chemotherapy for breast carcinoma. Ann Surg Oncol 9:243-247, 2002
Bear HD, Anderson S, Brown A, et al: The effect 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 21:4165-4174, 2003
Gianni L, Baselga J, Eiermann W, et al: First report of the European Cooperative Trial in operable breast cancer (ECTO): Effects of primary systemic therapy (PST) on local-regional disease. Proc Am Soc Clin Oncol 21:34a, 2002 (abstr 132)
Mamounas E, Brown A, Smith R, et al.: Sentinel node biopsy (SNB) following neoadjuvant chemotherapy in breast cancer (BC): Results from NSABP B-27. Proceedings Society of Surgical Oncology:21, 2001 (abstr 51)
Mamounas E, Brown A, Smith R, et al.: Accuracy of sentinel node biopsy after neoadjuvant chemotherapy in breast cancer: Updated results from NSABP B-27. Proc Am Soc Clin Oncol 21:36a, 2002 (abstr 140)
Snedecor G, Cochran W: Statistical Methods (ed 7). Ames, Iowa, The Iowa State University Press, 1980, pp 124-128
Hosmer DW, Lemeshow S: Applied Logistic Regression. New York, NY, John Wiley & Sons, Inc, 1989
Agresti A, Wackerly D, Boyett JM: Exact conditional tests for cross-classifications: Approximation of attained significance levels. Psychometrika 44:75-84, 1979
Tafra L, Lannin DR, Swanson MS, et al: Multicenter trial of sentinel node biopsy for breast cancer using both technetium sulfur colloid and isosulfan blue dye. Ann Surg 233:51-59, 2001
Pijpers R, Meijer S, Hoekstra OS, et al: Impact of lymphoscintigraphy on sentinel node identification with technetium-99m-colloidal albumin in breast cancer. J Nucl Med 38:366-368, 1997
Guenther JM, Krishnamoorthy M, Tan LR: Sentinel lymphadenectomy for breast cancer in a community managed care setting. Cancer J Sci Am 3:336-340, 1997
Sabel MS, Schott AF, Kleer CG, et al: Sentinel node biopsy prior to neoadjuvant chemotherapy. Am J Surg 186:102-105, 2003
Zirngibl C, Steinfeld-Birg D, Vogt H, et al: Sentinel lymph node biopsy before neoadjuvant chemotherapy—Conservation of breast and axilla. Br Ca Res Treat 76:S129, 2002 (suppl 1; abstr 516)(Eleftherios P. Mamounas, )