Preoperative Chemotherapy in Patients With Operable Breast Cancer: Nine-Year Results From National Surgical Adjuvant Breast and Bowel Project B-18

Transcription

1 Preoperative Chemotherapy in Patients With Operable Breast Cancer: Nine-Year Results From National Surgical Adjuvant Breast and Bowel Project B-18 Norman Wolmark, Jiping Wang, Eleftherios Mamounas, John Bryant, Bernard Fisher National Surgical Adjuvant Breast and Bowel Project (NSABP) Protocol B-18 was initiated in 1988 to determine whether four cycles of doxorubicin/cyclophosphamide given preoperatively improve survival and disease-free survival (DFS) when compared with the same chemotherapy given postoperatively. Secondary aims included the evaluation of preoperative chemotherapy in downstaging the primary breast tumor and involved axillary lymph nodes, the comparison of lumpectomy rates and rates of ipsilateral breast tumor recurrence (IBTR) in the two treatment groups, and the assessment of the correlation between primary tumor response and outcome. Initially published findings were based on a follow-up of 5 years; this report updates results through 9 years of follow-up. There continue to be no statistically significant overall differences in survival or DFS between the two treatment groups. Survival at 9 years is 70% in the postoperative group and 69% in the preoperative group (P =.80). DFS is 53% in postoperative patients and 55% in preoperative patients (P =.50). A statistically significant correlation persists between primary tumor response and outcome, and this correlation has become statistically stronger with longer follow-up. Patients assigned to preoperative chemotherapy received notably more lumpectomies than postoperative patients, especially among patients with tumors greater than 5 cm at study entry. Although the rate of IBTR was slightly higher in the preoperative group (10.7% versus 7.6%), this difference was not statistically significant. Marginally statistically significant treatment-byage interactions appear to be emerging for survival and DFS, suggesting that younger patients may benefit from preoperative therapy, whereas the reverse may be true for older patients. [J Natl Cancer Inst Monogr 2001;30:96 102] The rationale for testing preoperative (neoadjuvant) chemotherapy in the treatment of patients with operable breast cancer has evolved from preclinical (1,2) and clinical (3 9) observations as well as from hypothetical considerations of tumor cell kinetics (10,11). Nonrandomized studies (12 15) have demonstrated that preoperative chemotherapy administration results in substantial rates of clinical response but in generally low rates of pathologic complete response. By reducing primary tumor size, preoperative chemotherapy allowed some patients who otherwise would have required a mastectomy to undergo breastconserving procedures. Since nonrandomized studies could not evaluate the relative efficacy of preoperative versus postoperative chemotherapy on overall survival (OS) and disease-free survival (DFS), several randomized trials (16 21) were implemented. Some of these trials (16,17), however, were not designed as direct comparisons of preoperative versus postoperative chemotherapy and, therefore, could not provide a definitive answer to the pivotal question of whether OS and DFS can be improved by administering chemotherapy before, rather than after, surgery. In 1988, the National Surgical Adjuvant Breast and Bowel Project (NSABP) initiated a randomized trial (B-18) to compare preoperative and postoperative chemotherapy in patients with operable breast cancer. The primary aim was to determine whether preoperative chemotherapy would result in improved OS and DFS relative to the same chemotherapy administered postoperatively. Secondary aims were to evaluate the response of the primary breast tumor and involved lymph nodes to preoperative chemotherapy, to correlate that response with outcome, and to determine whether preoperative chemotherapy would result in increased rates of breast-conserving surgery and decreased rates of ipsilateral breast tumor recurrence (IBTR). Findings with respect to local and regional response (20), 5-year outcome (21), compliance, and toxicity (21) have been published previously. This report updates the outcome results through 9 years of follow-up. PATIENTS AND METHODS Eligibility and Treatment Assignment Eligibility criteria and treatment have been described previously (20,21). In summary, eligible patients had operable, palpable breast cancer (T1 3 N0 1 M0) diagnosed by fine-needle aspiration or core needle biopsy; open biopsy was not permitted. After stratification according to age ( 49 or 50 years of age), clinical tumor size ( 2.0, , or >5.0 cm), and clinical lymph node status (negative or positive), patients were randomly assigned to receive either surgery (lumpectomy and axillary lymph node dissection or modified radical mastectomy) followed by four cycles of doxorubicin (60 mg/m 2 )/cyclophosphamide (600 mg/m 2 ) (AC) chemotherapy every 21 days or the same chemotherapy followed by surgery. Before randomization, surgeons were required to disclose the intended surgical procedure (lumpectomy or mastectomy) without considering the pos- Affiliations of authors: N. Wolmark, National Surgical Adjuvant Breast and Bowel Project (NSABP), Pittsburgh, PA, and Department of Human Oncology, Allegheny General Hospital, Pittsburgh; J. Wang, NSABP Biostatistical Center and Department of Biostatistics, University of Pittsburgh; E. Mamounas, NSABP Breast Committee and Aultman Cancer Center, Canton, OH; J. Bryant, NSABP Biostatistical Center and Departments of Biostatistics and Statistics, University of Pittsburgh; B. Fisher, NSABP and Department of Surgery, University of Pittsburgh. Correspondence to: Norman Wolmark, M.D., 320 E. North Ave., Pittsburgh, PA ( nwolmark@wpahs.org). See Note following References. Oxford University Press 96 Journal of the National Cancer Institute Monographs No. 30, 2001

2 sible downstaging effect of preoperative chemotherapy. Patients 50 years old or older received 10 mg tamoxifen orally twice a day for 5 years, starting after chemotherapy, regardless of hormone receptor status. Patients undergoing lumpectomies received breast irradiation, either after recovering from surgery (preoperative group) or after recovering from postoperative chemotherapy (postoperative group). Accrual and Patient Characteristics The study opened in October 1988 and closed in April Patient characteristics are summarized in Table 1. Of the 1523 patients, 763 were randomly assigned to the preoperative chemotherapy group and 760 to the postoperative chemotherapy group. Twenty-one patients were declared ineligible (seven postoperative and 14 preoperative; these totals include one patient in each group determined to have been ineligible subsequent to the first report of outcome) (21). Three of these patients had not given informed consent, six had advanced disease at the time of randomization, and three others were found to have had an open biopsy. The remaining nine cases were attributed to a variety of eligibility infractions. Tumor Response The primary tumor and axillary lymph nodes were clinically assessed before randomization. For patients receiving preoperative chemotherapy, breast tumor and lymph node measurements were also obtained both before each cycle of chemotherapy and before surgery. Preoperative patients were considered evaluable Table 1. Patient eligibility, follow-up, and entry characteristics Eligibility, follow-up, and entry characteristics Postoperative AC Treatment group* Preoperative AC Total Eligibility Randomized Ineligible Eligible without follow-up Analyzed Follow-up of analyzed patients Mean time on study, y Characteristics of analyzed patients Age, % 49 y y y Menopausal status, % Premenopausal or perimenopausal Postmenopausal Unknown Race, % White Black Other Unknown Clinical tumor size, % 2.0 cm cm cm Mean tumor size ± standard deviation 3.5 ± ± ± 1.8 Clinical lymph node status, % Negative Positive *AC doxorubicin + cyclophosphamide. for response if they had received at least two cycles of preoperative chemotherapy, had bidimensional tumor measurements recorded at the beginning of cycle 1, and had at least one additional set of tumor measurements recorded after cycle 2. The absence of clinical evidence of tumor in the breast by physical examination was categorized as clinical complete response (ccr). A clinical partial response (cpr) was assigned if the product of the two largest perpendicular diameters of the breast tumor had decreased by 50% or more. Progressive disease (cp) was assigned if there was a 50% or greater increase in tumor size. Patients whose breast tumor did not meet the criteria for ccr, cpr, or cp were considered to have stable disease (cs). After surgery, patients achieving a ccr were assessed further for evidence of pathologic response. Patients with ccr were classified as pathologic complete responders (pcr) if there was no histologic evidence of invasive carcinoma on pathologic examination of the surgical specimen and as pathologic nonresponders (pinv) otherwise. These findings were those reported by the institutional pathologists. Outcome Measures OS was defined as the time from study entry to death from any cause. DFS was defined as the time from randomization to local, regional, or distant treatment failure; occurrence of contralateral breast cancer; other second primary cancer; or death without evidence of breast or second primary cancer. Patients who became inoperable before surgery or in whom the tumor could not be completely resected were counted as local treatment failures. Recurrence-free survival (RFS) was defined as the time from randomization to local, regional, or distant treatment failure. In the calculation of RFS, occurrences of contralateral breast cancer, other second primary cancers, and deaths without evidence of recurrence were treated as censoring events. Statistical Methodology Treatment comparisons included in this report were based on the cohort of eligible patients with follow-up. Substantively identical findings were obtained when ineligible patients also were included in the analyses. Patients were analyzed according to their assigned treatment regardless of compliance or crossover. Survival curves were estimated using the Kaplan Meier method, and treatment comparisons were made using the logrank test stratified according to age, clinical lymph node status, and clinical tumor size as reported at randomization. The Cox proportional hazards model was used to compute relative risks (RRs) and 95% confidence intervals (CIs), to examine the effect of prognostic variables, and to test for interactions between treatment and covariates. Treatment comparison of rates of IBTR was based on the occurrence of IBTRs as first events. The Mantel Haenszel approach was used to control for patient age and clinical tumor size and was based on the Poisson occurrences model. In preoperative patients, correlation between primary tumor response and subsequent outcome is clinically relevant primarily because it might enable one to distinguish patients who, after surgery, had an excellent prognosis from those whose prognosis was poor and who, therefore, might be candidates for additional therapy. For this reason, in correlation analyses, the outcome variables OS, DFS, and RFS were measured from the date of the surgery to the time of the event, and the analyses were restricted to eligible preoperative patients who were evaluable for response, had undergone surgery, and were clinically free of dis- Journal of the National Cancer Institute Monographs No. 30,

3 ease as of the date of surgery. Of 682 such patients, 247 (36%) had primary tumor responses that were classified as ccr, 295 (43%) were cpr, 118 (17%) were cs, and 22 (3%) were cp. Because few patients experienced cp, the cs and cp categories were combined in these analyses. Patients in the combined category are referred to as clinical nonresponders (cnr). Statistical tests of association between clinical tumor response and outcome variables assumed an ordinal relationship between response categories. The tests were obtained by computing a response score for each patient (1 ccr, 2 cpr, and 3 cnr) and introducing this score as a covariate in Cox proportional hazards models. Of the 247 patients with complete clinical responses, 88 (13% of 682) were further classified in terms of pathologic response as pcr, and 159 (23% of 682) were pinv. Tests for association between overall primary tumor response and outcome variables were obtained by assigning an ordinal response score to each patient (1 pcr, 2 pinv, 3 cpr, and 4 cnr) and introducing this score into a proportional hazards model. Tests for association were carried out both ignoring and controlling for other prognostic variables. Results presented here are based on data received at the NSABP Biostatistical Center as of June 30, The mean time on study is 9.5 years. All P values are two-sided. RESULTS Survival There have been 218 deaths in the postoperative group and 221 in the preoperative group. There continues to be no statistically significant difference in survival between the two groups (P.80; RR 1.02; 95% CI 0.84 to 1.21). The 5-year survival was 81% in the postoperative group and 80% in the preoperative group. The 9-year survival was 70% in postoperative patients and 69% in preoperative patients (Fig. 1). Disease-Free Survival There have been 338 events in the postoperative group and 323 in the preoperative group. There was no difference in DFS between the two groups (P.50; RR 0.95; 95% CI 0.88 to 1.10). The 5-year DFS was 67% for both treatment groups. The 9-year DFS was 53% in the postoperative group and 55% in the preoperative group (Fig. 1). First Reported Sites of Treatment Failure As has been reported through 5 years of follow-up, there continue to be no statistically significant differences in the rates of treatment failure at any specific site (Table 2). Although there was a trend toward a higher rate of IBTR with preoperative chemotherapy, this difference was not statistically significant (P.12): There were 34 (7.6%) IBTRs among 448 patients who underwent lumpectomy in the postoperative group and 54 (10.7%) among 503 such patients in the preoperative group. There was a strong correlation between age and rate of IBTR (P.00003), with higher IBTR rates in women less than 50 years of age (13.1%) when compared with the rates of those 50 years of age or over (5.2%) (Table 3). Of note is the fact that women 50 years of age or older at randomization received tamoxifen, whereas those under 50 years of age did not. Clinical tumor size did not appear to correlate with the rate of IBTR (P.59; Table 3). Although patients with a complete pathologic response (pcr) appeared to have a somewhat lower rate of IBTR than the remaining patients, the association between primary tumor response and IBTR rate was not statistically significant (P.12; Table 3). A marginally statistically significant increase (P.04) was reported initially in the rate of IBTR found in patients who were converted from proposed mastectomy to lumpectomy after preoperative chemotherapy when compared with those patients who had a lumpectomy as initially planned before randomization (21). This trend persists through 9 years of follow-up. The rate of IBTR is 11/69 (15.9%) in preoperative patients downstaged to lumpectomies, as compared with 43/434 (9.9%) in preoperative patients who received lumpectomies as originally planned. The difference, however, is explained partially by corresponding differences between the age distribution of downstaged patients to that of patients having lumpectomies as planned and is no longer statistically significant after controlling for patient age and initial clinical tumor sizes (P.14). Subset Analyses There was no evidence for treatment-by-covariate interaction for either clinical lymph node status or clinical tumor size. Treatment-by-age interaction, however, was marginally statistically significant for both OS and DSF (P.04 for OS; P.053 for DFS). In women 49 years old or younger, there appeared to be an advantage for preoperative chemotherapy; at 9 years of follow-up, OS was 71% versus 65% and DFS was 55% versus 46% in favor of patients treated with preoperative chemotherapy. Conversely, in women 50 years old or older, there seemed to be an advantage in favor of postoperative chemotherapy; at 9 years of follow-up, OS was 75% versus 67% and DFS was 60% versus Fig. 1. Overall survival and diseasefree survival according to treatment through 9 years of follow-up (Postop postoperative chemotherapy; Preop preoperative chemotherapy). 98 Journal of the National Cancer Institute Monographs No. 30, 2001

4 Type and site of failure Table 2. First reported sites of treatment failure Postoperative AC Treatment group Preoperative AC No. % No. % Clinically inoperable Gross residual disease IBTR only* Other local recurrence Regional recurrence Distant metastasis (except opposite breast) Second primary cancer (except opposite breast) Opposite breast cancer Dead, no evidence of disease Total first events Alive, event free Total No. of patients *Percentages for ipsilateral breast tumor recurrence (IBTR) are based on the numbers of patients who received lumpectomies. Clinical factor Table 3. Clinical factors associated with ipsilateral breast tumor recurrence (IBTR) Postoperative AC, % of patients with IBTR Treatment group Preoperative AC, % of patients with IBTR 56% in favor of postoperative chemotherapy. Within either age group, however, the preoperative versus postoperative treatment comparison did not achieve statistical significance for either OS or DSF (in younger women, RR 0.85 and P.22 for OS and RR 0.85 and P.11 for DFS; in older women, RR 1.28 and P.08 for OS and RR 1.09 and P.44 for DFS). Association Between Clinical Response and Outcome Total, % of patients with IBTR Age, y Clinical tumor size <3 cm cm Clinical and pathologic tumor response ccr* N/A 9.8 pcr N/A 6.7 pinv N/A 11.5 cpr N/A 11.8 cnr (csd and cpd) N/A 13.0 Procedure after preoperative chemotherapy Lumpectomy vs. planned mastectomy N/A 15.9 Lumpectomy as planned N/A 9.9 *ccr clinical complete response; pcr pathologic complete response; pinv pathologic nonresponders; cpr clinical partial response; cnr clinical nonresponders; csd clinical stable disease; cpd clinical progressive disease; N/A not applicable. Patients in the preoperative chemotherapy group were categorized according to clinical response (ccr, cpr, or cnr). Through 9 years of follow-up, there continues to be an apparent association between clinical response and outcome. This association now has become statistically significant not only for DFS and RFS (as was the case through 5 years of follow-up) but also for OS (OS: P.005; DFS: P.0008; RFS: P.0002). OS at 9 years was 78% in patients with ccr, 67% in patients with cpr, and 65% in those with cnr. The rates of DFS were 64%, 54%, and 46%, respectively. The statistically significant association between clinical response and outcome persisted after adjustment for clinical tumor size at randomization, clinical lymph node status, and age at randomization (OS: P.04; DFS: P.004; RFS: P.0008). Association Between Pathologic Response and Outcome Similar to the results through 5 years of follow-up, the outcome for patients who achieved a pcr continues to be superior to that of those with a ccr with residual invasive cancer on pathologic examination (pinv) or to those patients failing to achieve a ccr (Fig. 2). At 9 years, the OS rate for patients achieving a pcr was 85% as compared with 73% for patients with pinv. For DFS, the respective rates were 75% and 58%. Overall primary tumor response graded as pcr, pinv, cpr, or cnr was strongly associated with all outcome measures (OS: P.0008; DFS: P.00005; RFS: P.0002). These associations persisted after adjustment for clinical tumor size at randomization, clinical lymph node status, and age at randomization (OS: P.006; DFS: P.0004; RFS: P.00006). After adjustment for the other prognostic variables, patients with pcr had a 50% reduction in the risk of death when compared with the group as a whole (RR 0.50; 95% CI 0.32 to 0.78), those with pinv had an 8% increase (RR 1.08; 95% CI 0.81 to 1.42), those with cpr had a 28% increase (RR 1.28; 95% CI 1.01 to 1.62), and those with cnr had a 45% increase (RR 1.45; 95% CI 1.11 to 1.90). Prognostic Importance of Primary Tumor Response After Controlling for Pathologic Lymph Node Status As expected, pathologic lymph node status was a strong predictor of outcome in both preoperative and postoperative patients (P<.0001 for OS, DFS, and RFS in either cohort). In patients treated with preoperative chemotherapy, the resulting pathologic lymph node status was also, not surprisingly, related to primary tumor response. The Spearman correlation between the number of involved lymph nodes and primary tumor response (pcr, pinv, cpr, or cnr) was 0.22 (P<.0001). To provide a formal test of the hypothesis that primary tumor response contributes prognostic information beyond that provided by pathologic lymph node status, proportional hazards models were fitted to the preoperative cohort, including a score variable representing primary tumor response after stratifying for pathologic lymph node status (0, 1 3, or 4). Results demonstrated that primary tumor response does contribute additional prognostic information over and above pathologic lymph node status (OS: P.06; DFS: P.006; RFS: P.004). Conversely, pathologic lymph node status was also strongly prognostic even after controlling for primary tumor response (P<.0001 for OS, DFS, and RFS). DISCUSSION The mature results of the B-18 trial presented here continue to support the conclusions of previous reports (21). They demonstrate that, through 9 years of follow-up, the outcome for patients Journal of the National Cancer Institute Monographs No. 30,

5 Fig. 2. Comparison of outcome of patients treated with preoperative chemotherapy according to primary breast tumor response. treated with preoperative chemotherapy is similar to the outcome for those treated with standard adjuvant chemotherapy. These results do not support the Goldie Coldman hypothesis, which proposes that, as a tumor cell population increases, an ever-expanding number of drug-resistant phenotypic variants arises that are more difficult to eradicate with chemotherapy. Two smaller European trials (16,17) that compared preoperative with postoperative chemotherapy had outcome results discordant with those of B-18. These trials demonstrated a survival advantage for preoperative chemotherapy with no differences in DFS. In both trial designs, however, there were imbalances in the systemic and local therapy administered to the two groups. Although all patients in the preoperative chemotherapy group received chemotherapy, only lymph node-positive patients did so in the postoperative chemotherapy group. Similarly, more patients received surgery in the postoperative chemotherapy group than in the preoperative chemotherapy group, with a resulting increase in the rate of local recurrence in the latter. The outcome results of another trial (19), conducted at the Royal Marsden Hospital in England, were similar to our results. In that trial, a total of 309 patients were randomly assigned either to receive four preoperative cycles of chemoendocrine therapy followed by four postoperative cycles of the same therapy or to receive all eight cycles of therapy postoperatively. At a median follow-up of 48 months, there were no statistically significant differences between the two groups in terms of local relapse, metastatic relapse, or OS. Study B-18 continues to demonstrate a statistically significant association between clinical/pathologic tumor response to preoperative chemotherapy and long-term outcome. This association does not support the Skipper concept, in which the response of a primary tumor to chemotherapy may not necessarily reflect the response of micrometastatic disease. Furthermore, it suggests that the underlying biologic factors required for pathologic complete response may also confer true chemosensitivity to micrometastatic disease, allowing long-term improvement in outcome as opposed to a temporary delay in recurrence. This is in contrast to the metastatic disease setting, where ccr generally results in only temporary prolongation in time to progression. As reported previously (20), administration of preoperative chemotherapy resulted in statistically significantly more lumpectomies, particularly among patients with tumors greater than 5 cm in diameter at randomization. This was accompanied by a statistically nonsignificant increase in the rate of IBTR (10.7% in the preoperative chemotherapy group versus 7.6% in the postoperative chemotherapy group). This can be attributed partially to the fact that the former group contained some downstaged patients who may have been at higher risk for local recurrence irrespective of the assigned treatment arm. Although, in the present report, IBTR rates were not statistically significantly associated with the initial clinical tumor size, the rates were somewhat higher in patients for whom a mastectomy was planned at the time of randomization but for whom a lumpectomy was performed after preoperative chemotherapy (15.9%), as opposed to those for whom a lumpectomy was planned from the beginning (9.9%). The noted increase in IBTR rates in patients under 50 years of age when compared with those in patients 50 years of age or older is not surprising. Younger patients generally have more aggressive disease than older patients; this results in a higher rate of local and systemic recurrence. But even so, the largest part of the difference probably is caused by the inherent design of the study, whereby tamoxifen was administered only to patients 50 years old or older at randomization, irrespective of estrogen receptor status. Randomized trials have shown convincingly that tamoxifen markedly reduces the rate of IBTR after lumpectomy and breast radiotherapy in both older and younger women (22). The observed marginally statistically significant interaction between treatment effect and age at randomization is enigmatic. In the B-18 data, patients under 50 years of age appeared to show a greater benefit from preoperative chemotherapy than from postoperative chemotherapy. In contrast, patients 50 years old or older appeared to benefit more from postoperative chemotherapy. The most likely explanation for these results is that 100 Journal of the National Cancer Institute Monographs No. 30, 2001

6 they occurred by chance alone and that a true interaction between treatment and age does not exist. Alternatively, the overview analyses of the Early Breast Cancer Trialists Collaborative Group (3) indicate that the effects of chemotherapy are most apparent in younger women, so it is not inconceivable that the benefit of preoperative chemotherapy relative to postoperative treatment could be age dependent as well. To the extent that younger patients present more often than older patients with estrogen receptor-negative tumors, this conjecture is consistent with a recent International Breast Cancer Study Group retrospective analysis (23) suggesting that there may be a preferential benefit to early initiation of adjuvant chemotherapy in premenopausal patients whose tumors do not express the estrogen receptor. Because of limitations on the assay of hormone receptors in the early years of the B-18 study, data are not available to address this issue. In any case, although intriguing explanations and hypotheses may be invoked, until additional data are forthcoming, the interpretation of the findings remains speculative. On the basis of both the results presented here and those reported previously (20,21), preoperative chemotherapy can be used instead of postoperative adjuvant chemotherapy, and its use would be most appropriate for patients who wish to preserve their breasts but who have tumors too large for breastconserving surgery. Another potential advantage of preoperative chemotherapy is the resulting classification of patients in different categories of clinical and pathologic tumor response, which can be used as a prognostic factor for outcome and as a guideline for further locoregional and systemic therapy (24). The development of taxanes and the demonstration of their marked antitumor activity in patients with advanced breast cancer provided the opportunity to examine further some of the concepts that have emerged from the B-18 trial. The NSABP recently completed accrual to Protocol B-27, a randomized trial designed to determine whether the preoperative or postoperative administration of docetaxel after preoperative AC therapy will prolong OS and DFS rates when compared with four courses of preoperative AC therapy alone (25,26). Equally important are the secondary aims of this trial, which are to determine whether the administration of preoperative docetaxel after preoperative AC therapy will further increase the clinical and pathologic response rates of primary breast tumors, whether it will result in further axillary lymph node downstaging, and whether it will increase the use of lumpectomy. A comparison of the group receiving postoperative docetaxel after preoperative AC therapy with the group receiving preoperative AC alone will identify whether any improvement in outcome will be evident in subgroups of patients, i.e., in patients with residual positive lymph nodes after preoperative AC. Two ancillary studies to the B-27 trial evaluate serum and tumor biomarkers as they relate to outcome and response to preoperative AC and/or docetaxel chemotherapy. Thus, it will be possible, using the collected materials, to evaluate the prognostic and predictive value of a panel of biomarkers, including HER2, p53, p-glycoprotein, bcl-2 Ki67, and array-based comparative genomic hybridization. Perhaps the greatest potential of preoperative adjuvant therapy is yet to be realized. This is a unique setting in which the tumor is readily accessible while the patient is undergoing treatment. Thus, a potentially powerful tool could become available whereby molecularly characterized tumor discriminants could be correlated with the efficacy of preoperative adjuvant treatment and, more importantly, with subsequent survival. Although it is premature to suggest that objective tumor regression during the course of adjuvant therapy is a definitive surrogate marker for eventual patient outcome, the data for NSABP Protocol B-18 suggest that this is a distinct possibility. REFERENCES (1) Gunduz N, Fisher B, Saffer EA. Effect of surgical removal on the growth and kinetics of residual tumor. Cancer Res 1979;39: (2) Fisher B, Gunduz N, Saffer EA. Influence of the interval between primary tumor removal and chemotherapy on kinetics and growth of metastases. Cancer Res 1983;43: (3) Early Breast Cancer Trialists Collaborative Group. Polychemotherapy for early breast cancer: an overview of the randomised trials. Lancet 1998; 352: (4) Early Breast Cancer Trialists Collaborative Group. Favourable and unfavourable effects on long-term survival of radiotherapy for early breast cancer: an overview of the randomised trials. Lancet 2000;355: (5) Fisher B, Redmond C, Poisson R, Margolese R, Wolmark N, Wickerham L, et al. Eight-year results of a randomized clinical trial comparing total mastectomy and lumpectomy with or without irradiation in the treatment of breast cancer. N Engl J Med 1989;320: (6) Fisher B, Anderson S, Redmond CK, Wolmark N, Wickerham DL, Cronin WM. Reanalysis and results after 12 years of follow-up in a randomized clinical trial comparing total mastectomy with lumpectomy with or without irradiation in the treatment of breast cancer. N Engl J Med 1995;333: (7) Perloff M, Lesnick GJ. Chemotherapy before and after mastectomy in stage III breast cancer. Arch Surg 1982;117: (8) Schick P, Goodstein J, Moor J, Butler J, Senter KL. Preoperative chemotherapy followed by mastectomy for locally advanced breast cancer. J Surg Oncol 1983;22: (9) Sorace RA, Bagley CS, Lichter AS, Danforth DN Jr, Wesley MW, Young RC, et al. The management of nonmetastatic locally advanced breast cancer using primary induction chemotherapy with hormonal synchronization followed by radiation therapy with or without debulking surgery. World J Surg 1985;9: (10) Skipper HE. Kinetics of mammary tumor cell growth and implications for therapy. Cancer 1971;28: (11) Goldie JH, Coldman AJ. A mathematic model for relating the drug sensitivity of tumors to their spontaneous mutation rate. Cancer Treat Rep 1979; 63: (12) Bonadonna G, Veronesi U, Brambilla C, Ferrari L, Luini A, Greco M, et al. Primary chemotherapy to avoid mastectomy in tumors with diameters of three centimeters or more. J Natl Cancer Inst 1990;82: (13) Tubiana-Hulin M, Malek M, Briffod M, et al. Preoperative chemotherapy of operable breast cancer (stage IIIA). Prognostic factors of distant recurrence. Eur J Cancer 1993;29A(Suppl 6):S76. (14) Belembaogo E, Feillel V, Chollet P, Cure H, Verrelle P, Kwiatkowski F, et al. Neoadjuvant chemotherapy in 126 operable breast cancers. Eur J Cancer 1992;28A: (15) Smith IE, Jones AL, O Brien ME, McKinna JA, Sacks N, Baum M. Primary medical (neo-adjuvant) chemotherapy for operable breast cancer. Eur J Cancer 1993;29A: (16) Mauriac L, Durand M, Avril A, Dilhuydy JM. Effects of primary chemotherapy in conservative treatment of breast cancer patients with operable tumors larger than 3 cm. Results of a randomized trial in a single centre. Ann Oncol 1991;2: (17) Scholl SM, Fourquet A, Asselain B, Pierga JY, Vilcoq JR, Durand JC, et al. Neoadjuvant versus adjuvant chemotherapy in premenopausal patients with tumours considered too large for breast conserving surgery: preliminary results of a randomised trial: S6. Eur J Cancer 1994;30A: (18) Powles TJ, Hickish TF, Makris A, Ashley SE, O Brien ME, Tidy VA, et al. Randomized trial of chemoendocrine therapy started before or after surgery for treatment of primary breast cancer. J Clin Oncol 1995;13: (19) Makris A, Powles TJ, Ashley SE, Chang J, Hickish T, Tidy VA, et al. A reduction in the requirements for mastectomy in a randomized trial of neoadjuvant chemoendocrine therapy in primary breast cancer. Ann Oncol 1998;9: Journal of the National Cancer Institute Monographs No. 30,

7 (20) Fisher B, Brown A, Mamounas E, Wieand S, Robidoux A, Margolese RG, 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 1997;15: (21) Fisher B, Bryant J, Wolmark N, Mamounas E, Brown A, Fisher ER, et al. Effect of preoperative chemotherapy on the outcome of women with operable breast cancer. J Clin Oncol 1998;16: (22) Fisher B, Costantino J, Redmond C, Poisson R, Bowman D, Couture J, et al. A randomized clinical trial evaluating tamoxifen in the treatment of patients with node-negative breast cancer who have estrogen-receptorpositive tumors. N Engl J Med 1989;320: (23) Colleoni M, Bonetti M, Coates AS, Castiglione-Gertsch M, Gelber RD, Price K, et al. Early start of adjuvant chemotherapy may improve treatment outcome for premenopausal breast cancer patients with tumors not expressing estrogen receptors. The International Breast Cancer Study Group. J Clin Oncol 2000;18: (24) Fisher B, Mamounas EP. Preoperative chemotherapy: a model for studying the biology and therapy of primary breast cancer. J Clin Oncol 1995;13: (25) Mamounas EP. NSABP Protocol B-27. Preoperative doxorubicin plus cyclophosphamide followed by preoperative or postoperative docetaxel. Oncology (Huntingt) 1997;(6 Suppl 6): (26) Mamounas EP. Overview of National Surgical Adjuvant Breast Project neoadjuvant chemotherapy studies. Semin Oncol 1998;25(2 Suppl 3): NOTE Supported by Public Health Service grants U10CA12027, U10CA69651, U10CA37377, and U10CA69974 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services. 102 Journal of the National Cancer Institute Monographs No. 30, 2001