Oncologist. The. Breast Cancer. Trends for Inflammatory Breast Cancer: Is Survival Improving? The Oncologist 2007;12:

Transcription

1 The Oncologist Breast Cancer Trends for Inflammatory Breast Cancer: Is Survival Improving? ANA M. GONZALEZ-ANGULO, a BRYAN T. HENNESSY, b KRISTINE BROGLIO, c FUNDA MERIC-BERNSTAM, d MASSIMO CRISTOFANILLI, a SHARON H. GIORDANO, a THOMAS A. BUCHHOLZ, e AYSEGUL SAHIN, f S. EVA SINGLETARY, d AMAN U. BUZDAR, a GABRIEL N. HORTOBÁGYI a Departments of a Breast Medical Oncology, b Gynecology Oncology, c Biostatistics and Applied Mathematics, d Surgical Oncology, e Radiation Oncology, and f Pathology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, USA Key Words. Inflammatory breast cancer Trends Survival Prognosis ABSTRACT The purpose of this study was to evaluate whether the survival of women with inflammatory breast cancer (IBC) treated at our institution has improved over the past 30 years. Three-hundred ninety-eight patients with IBC were treated between 1974 and Patient characteristics and outcomes were tabulated and compared among decades of diagnosis. Survival outcomes were estimated with the Kaplan-Meier product limit method and compared among groups with the log-rank statistic. Cox proportional hazards models were fit to determine the association between year of diagnosis and survival outcomes after adjustment for patient and disease characteristics and treatments received. The median follow-up was 5.8 Disclosure of potential conflicts of interest is found at the end of this article. years (range, years). There were 238 recurrences and 236 deaths. The median recurrence-free survival (RFS) duration was 2.3 years and the median overall survival (OS) time was 4.2 years. In the models for RFS and OS, after adjustment for patient and disease characteristics, increasing year of diagnosis was not associated with a decrease in the risk for recurrence (hazard ratio, [HR], 1.00; 95% confidence interval [CI], ) or death (HR, 0.97; 95% CI, ). Our data show that there has not been an important change in the prognosis of patients with IBC in the last 30 years. Clinical trials focusing on the management of this aggressive disease are warranted. The Oncologist 2007;12: INTRODUCTION Inflammatory breast cancer (IBC) is the most aggressive form of primary breast cancer. It is rare, with an incidence rate of 1% 6% in the U.S. [1]. However, data from the Surveillance, Epidemiology, and End Results Program comparing trends and patterns for breast cancer revealed that between 1988 and 1990 and 1997 and 1999, IBC incidence rates (per 100,000 woman-years) increased from 2.0 to 2.5 (p.001) [2]. Compared with non-ibc stage III breast cancer, primary IBC is associated with lower overall survival (OS) rates [2, 3]. IBC is a clinical diagnosis characterized by the presence of symptoms and signs such as erythema, tenderness, edema, pain, and rarely, ulceration that rapidly extend to the entire breast [4, 5]. Pathologically, IBC is frequently diagnosed by the presence of cancer cells penetrating dermal Correspondence: Ana M. Gonzalez-Angulo, M.D., Department of Breast Medical Oncology, Unit 1354, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas , USA. Telephone: ; Fax: ; agonzalez@mdanderson.org Received February 15, 2007; accepted for publication May 22, AlphaMed Press /2007/$30.00/0 doi: /theoncologist The Oncologist 2007;12:

2 Gonzalez-Angulo, Hennessy, Broglio et al. 905 lymphatic channels causing the inflammatory signs [6]. The management of IBC has evolved over the past 30 years. Prior to 1974, IBC was essentially a death sentence, with a 5% 5-year survival rate [7 9]. At our institution, we introduced multidisciplinary management for IBC in 1974, when the first protocol dedicated exclusively to IBC patients began [10]. Some studies have shown that a multidisciplinary approach can improve both local control and survival, with about 30% of patients living 5 years [11, 12]. Neoadjuvant chemotherapy is considered to be the main component of treatment. The presence of pathological residual disease in the breast and lymph nodes following neoadjuvant chemotherapy is considered an important adverse prognostic factor [11 13]. Therapeutic advances clearly have been making an impact on the survival of women with early-stage breast cancer. The Early Breast Cancer Trialists Collaborative Group has performed meta-analyses on all known randomized trials [14, 15]. These aggregate data have established the benefit of adjuvant chemotherapy and endocrine therapy in women with lymph node positive and lymph node negative disease, the superiority of polychemotherapy compared with single-agent regimens, and the superiority of anthracycline-containing regimens and has established the optimal duration of both adjuvant chemotherapy and tamoxifen therapy. More recently, large randomized clinical trials have demonstrated that the addition of a taxane to an anthracycline-based adjuvant regimen results in longer disease-free survival (DFS) and OS times [16 18]. Despite the clear stepwise advances that are being made in the adjuvant treatment of breast cancer, a favorable effect of new therapies on the survival of patients with IBC has not been established. Improvement in survival over time would suggest that, in aggregate, new treatments are helping women with IBC live longer [19]. Therefore, the current analysis was designed to evaluate whether the survival of women with IBC treated at The University of Texas M. D. Anderson Cancer Center (Houston, TX) has improved over the past 30 years and to identify the clinical and pathological characteristics of IBC that predict for survival. PATIENTS AND METHODS Patients The institutional tumor registry and the Breast Medical Oncology database were reviewed and 398 patients with nonmetastatic IBC were identified as treated at M.D. Anderson Cancer Center between January 1974 and April Patients were treated either under specific IBC-designed protocols or with the best known multidisciplinary management off protocol [12, 19 22]. The institutional review board (IRB) of M.D. Anderson approved the individual clinical trials and all participants had provided written informed consent prior to enrollment. In addition, the IRB also approved our retrospective review of the medical records for the purposes of this report. Diagnosis of invasive breast cancer was made by coreneedle biopsy or fine-needle aspiration of the breast tumor. All pathologic specimens were prospectively reviewed by dedicated breast pathologists at M. D. Anderson. Clinical stage was defined by the sixth edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual. The histologic type of all tumors was defined according to the World Health Organization s classification system [23]. Tumor grade was defined according to the modified Black s nuclear grading system [24]. Immunohistochemical analysis to determine estrogen receptor (ER) and progesterone receptor (PgR) status was performed using standard procedures on 4- m sections of paraffin-embedded tissues stained with monoclonal antibodies: 6F11 (Novacastra Laboratories Ltd., Burlingame, CA) for ER and 1A6 (Novacastra Laboratories Ltd.) for PgR. Nuclear staining 10% was considered a positive result. Prior to 1993, the dextran-coated charcoal ligand-binding method was used to determine ER and/or PgR. Pathological complete response (pcr) was defined as no evidence of invasive carcinoma in the breast and the axillary lymph nodes at the time of surgery. Statistical Methods Patients were divided into four groups by their decade of diagnosis. Patient characteristics were tabulated and compared among decade of diagnosis groups with the Kruskal- Wallis test or the 2 test as appropriate. The OS duration was calculated from the date of diagnosis to the date of death from any cause. The recurrence-free survival (RFS) time was calculated from the date of diagnosis to the date of first distant or local disease recurrence or last follow-up. Patients who died before experiencing disease recurrence were considered censored. Survival outcomes were estimated with the Kaplan- Meier product limit method and compared among groups with the log-rank statistic. A series of multivariable Cox proportional hazards models were fit to determine the association between year of diagnosis and survival outcomes after adjustment for patient and disease characteristics and treatments received. Statistical analyses were carried out using S-Plus 7.0 (Insightful Corporation, Seattle, WA) software.

3 906 Inflammatory Breast Cancer Trends RESULTS Forty-six patients were diagnosed from , 92 patients were diagnosed from , 156 patients were diagnosed from , and 104 patients were diagnosed from Table 1 shows the tabulation of patient characteristics by decade of diagnosis group. There were large amounts of missing data among patients diagnosed in the 1970s and the amount of missing data decreased as the decade of diagnosis became more recent. Patient characteristics were not equally distributed by decade of diagnosis. Note that the presented p-values were calculated excluding patients with unknown characteristics. The proportions of patients with hormonal receptor (HR)-positive disease, ductal histology, and lymphovascular invasion (LVI) tended to be lower in the more recent decades of diagnosis. The proportions of patients with high nuclear grade and node-positive disease tended to be higher in the more recent decades of diagnosis. Because of the large amount of missing data among patients diagnosed in the 1970s and the potential bias associated with this, p- values were also calculated excluding this group of patients. For human epidermal growth factor receptor (HER)-2 status, the p-value was recalculated excluding patients from both the 1970s and the 1980s. These results are not presented because they were no different. Treatments received were also not consistent across the decade groups. The use of taxanes in the neoadjuvant or adjuvant setting has increased as well as the proportion of patients receiving surgery. However, the proportion of patients achieving a pcr was not significantly different among the decade of diagnosis groups. Almost all patients had received anthracycline-based neoadjuvant chemotherapy: 395 (98%) as either four cycles of AC: doxorubicin (60 mg/m 2 ) and cyclophosphamide (600 mg/m 2 ) i.v. on day 1; four to six cycles of VACP: vincristine (1 mg/m 2 ), doxorubicin (50 mg/m 2 ), and cyclophosphamide (500 mg/m 2 ) i.v. on day 1 and prednisone (100 mg/m 2 ) orally on days 1 5 of treatment alone or followed by four cycles of MV: methotrexate (75 mg/m 2 ) i.v. on day 1 and vinblastine (1.7 mg/m 2 per day) i.v. continuous infusion on days 1 5; or four to 12 cycles of FAC: 5-fluorouracil (500 mg/m 2 ) i.v. on days 1 and 4, doxorubicin (50 mg/m 2 ) i.v. by continuous infusion over 72 hours on days 1 3 or bolus on day 1, and cyclophosphamide (500 mg/m 2 ) i.v. on day 1 with or without Bacille Calmette-Guerin (BCG) injections. One hundred seventy patients (43%) received additional taxane therapy for 12 weeks, either as weekly paclitaxel ( mg/m 2 i.v. on day 1, dose reduced to 80 mg/m 2 i.v. on day 1), every 3-week paclitaxel ( mg/m 2 i.v. on day 1), or every 3-week docetaxel ( mg/m 2 i.v. on day 1) as part of primary chemotherapy in clinical trials. Six patients (2%) received consolidation with high-dose chemotherapy and bone marrow transplant. Over the study period the approach to locoregional therapy evolved. In the 1970s, patients were offered definitive radiation therapy. Starting in the 1980s locoregional therapy was determined by a multidisciplinary group of physicians, and was based on clinical response to neoadjuvant chemotherapy. Patients who had a complete or partial clinical response underwent a modified radical mastectomy and postoperative radiation therapy. Patients with progressive disease on neoadjuvant chemotherapy with inoperable disease received radiation therapy and were re-evaluated for the feasibility of surgery. A modified radical mastectomy was performed on 354 patients (89%). Thirty-eight patients (9%) received definitive radiotherapy as part of their treatment protocol and six patients (2%) had disease progression that required palliative radiation. One hundred forty patients from the 294 patients with known HR status were HR positive; of these 43 (31%) received adjuvant tamoxifen for 5 years. Tamoxifen was not offered to premenopausal patients with ER/PgR-positive disease until Radiotherapy was delivered at the completion of primary chemotherapy and after surgery in all patients. The median follow-up among patients who were still alive at their last follow-up was 5.8 years for the entire group (range, years). The median follow-up by decade was: 22.2 years (range, years) for the 1970s, 17.8 years (range years) for the 1980s, 7.8 years (range, years) for the 1990s, and 2.5 years (range, years) for the 2000s. The median follow-up for all patients was 3.3 years (range, years); 3.1 years (range, years) for the 1970s, 3.6 years (range, years) for the 1980s, 4.4 years (range, years) for the 1990s, and 2.0 years (range, years) for the 2000s. Table 2 shows the OS estimates. Two hundred thirty-six patients had died and the median OS duration was 4.2 years. Unadjusted for differences in patient characteristics and treatments, there is not a clear trend in OS by decade of diagnosis. The Kaplan-Meier OS curve for all patients is shown in Figure 1A, and OS by decade of diagnosis is shown in Figure 1B and C. HR status and nuclear grade were significantly associated with OS in these patients. Patients who received taxanes, surgery, or achieved a pcr had a significantly longer OS time than patients who did not. RFS estimates are shown in Table 2. Two hundred thirty-eight patients had experienced a recurrence and the median time to disease recurrence was 2.3 years. In the univariate setting, there was no statistically significant dif-

4 Gonzalez-Angulo, Hennessy, Broglio et al. 907 Table 1. Patient characteristics by decade of diagnosis 1970s 1980s 1990s 2000s n % n % n % n % p-value n Age Minimum Median Maximum Estrogen receptor Negative 0 0.0% % % %.02 Positive 4 8.7% % % % Unknown % % % 2 1.9% Progesterone receptor Negative 1 2.2% % % %.68 Positive 1 2.2% % % % Unknown % % % 3 2.9% HER-2/neu Negative 0 0.0% % %.45 Positive 1 1.1% % % Unknown % % % 2 1.9% Lymphovascular invasion No 4 8.7% 5 5.4% % %.0004 Yes % % % % Unknown % % % 4 3.8% Modified Black s nuclear grade % 0 0.0% 1 0.6% 2 1.9% % % % % % % % % Unknown % % 5 3.2% 4 3.8% Node N % % % 6 5.8%.0001 N % % % % N % % % % N % 8 8.7% % % Unknown 0 0.0% 0 0.0% 1 0.6% 2 1.9% Neo/adjuvant taxane No % % % %.0001 Yes 0 0.0% 0 0.0% % % Unknown 0 0.0% 1 1.1% 3 1.9% 1 1.0% Received surgery No % 1 1.1% 2 1.3% 0 0.0%.0001 Yes % % % % Unknown 0 0.0% 0 0.0% 2 1.3% 0 0.0% Pathologic complete response No % % % %.40 Yes 2 4.3% 9 9.8% % % Unknown % 1 1.1% 6 3.8% 0 0.0% Abbreviation: HER-2, human epidermal growth factor receptor.

5 908 Inflammatory Breast Cancer Trends Table 2. Survival estimates n of patients n of events Median (years) 5-Year % estimate 95% CI 10-Year % estimate 95% CI p-value Overall survival estimates All By decade of diagnosis 1970s s s s Age ER Negative Positive PgR Negative Positive HER-2/neu Negative Positive LVI No Yes MNG N stage N N1 N Taxanes No Yes pcr No Yes Surgery No Yes Recurrence-free survival estimates All By decade of diagnosis 1970s s s s (continued)

6 Gonzalez-Angulo, Hennessy, Broglio et al. 909 Table 2. Continued n of patients n of events Median (years) 5-Year % estimate 95% CI 10-Year % estimate 95% CI p-value Age ER Negative Positive PgR Negative Positive HER-2/neu Negative Positive LVI No Yes MNG N stage N N1 N Taxanes No Yes pcr No Yes Surgery No Yes Abbreviations: CI, confidence interval; ER, estrogen receptor; HER-2, human epidermal growth factor receptor 2; LVI, lymphovascular invasion; MNG, modified nuclear Black s grade; pcr, pathologic complete response; PgR, progesterone receptor. ference in RFS by decade of diagnosis. The Kaplan-Meier RFS curves for all patients are shown in Figure 2A and shown by decade of diagnosis in Figure 2B and C. Patients with HR-positive disease, no LVI, and nuclear grade 1 or 2 had a significantly longer RFS duration than patients with HR-negative disease, LVI, or nuclear grade 3 disease. Patients who achieved a pcr also had a longer RFS time than those who did not. The multivariate analysis is presented as a single model in which we included the known prognostic factors and excluded therapy-related variables that could be confounded with diagnosis year. Table 3 shows the OS and RFS model results. In the OS and RFS models, after adjustment for patient and disease characteristics and excluding treatment-related characteristics, increasing year of diagnosis tended to be associated with an increase in the hazard of death. However, this association did not achieve statistical significance. DISCUSSION Prior to 1974, IBC was essentially a death sentence, with 5-year survival rates of 5% and a median survival time of 1.2 years [7 9]. Our retrospective data suggest that patients with IBC continue to have a poor prognosis, as demonstrated by the lower OS and RFS rates through the decades. Interestingly, when RFS and OS were analyzed, there were

7 910 Inflammatory Breast Cancer Trends Figure 1. Overall survival for all patients (A), by decade of diagnosis (B), and by decade of diagnosis truncated at 5 years (C). Figure 2. Recurrence-free survival for all patients (A),by decade of diagnosis (B), and by decade of diagnosis truncated at 5 years (C). no significant differences in survival by decade. This important observation suggests that advances in the therapy of non-ibc have not affected the prognosis of IBC since multidisciplinary management was introduced, despite advances in systemic and locoregional treatments. This review has the longest follow-up of any paper on IBC describing the clinical course of the disease treated in a single institution under a multidisciplinary program by a homogeneously trained group of physicians. A recent paper by Panades et al. [25] also failed to show breast cancer specific survival (BCSS) differences when comparing IBC patients treated between 1980 and 1990 with patients treated between 1991 and The 10-year BCSS rates were 27.4% (95% confidence interval [CI], 18.8% 36.7%) and 28.6% (95% CI, 20.3% 37.5%), respectively (p.37) [25]. In univariate analysis, patients with HR-positive disease, absence of LVI, or nuclear grade 1 or 2 had a significantly longer RFS survival than patients with HR-negative

8 Gonzalez-Angulo, Hennessy, Broglio et al. 911 Table 3. Multivariate models Hazard ratio 95% CI p-value Overall survival Diagnosis year (continuous) Age (continuous) Nodes (positive versus negative) Nuclear grade (3 versus 1 and 2) LVI (yes versus no) pcr (yes versus no) HR status (positive versus negative) Recurrence-free survival Diagnosis year (continuous) Age (continuous) Nodes (positive versus negative) Nuclear grade (3 versus 1 and 2) LVI (yes versus no) pcr (yes versus no) HR status (positive versus negative) Abbreviations: CI, confidence interval; HR, hormonal receptor; LVI, lymphovascular invasion; pcr, pathologic complete response. disease, LVI, or nuclear grade 3 disease. Patients who achieved a pcr also had a longer RFS than those who did not. Interestingly, looking at the data we observed changes in the biological markers, mainly ER, over time. Explanations for these phenomena may include: (a) bias because of missing data, which could not be missing completely at random; (b) changes in population and referral patterns with time; and (c) changes in the method of evaluating ER and PgR. HR-negativity was the only independent predictor of RFS and OS in the model. pcr and age were independent predictors of RFS. The differences could not be demonstrated statistically for other known characteristics, most likely because of the lower numbers of cases studied with complete information in the early decades of diagnosis. The Panades et al. [25] study showed that, on multivariate analysis, ER status, menopausal status, and more intense chemotherapy were independently associated with BCSS and mastectomy was significantly associated with local RFS. pcr was a favorable prognostic factor for BCSS [25]. Other studies of patients with breast cancer have shown that pcr is associated with superior survival outcomes [26, 27], but not necessarily for IBC [3]. Interpreting trends over time is fraught with peril. Treatments and diagnostic criteria as well as patient and disease characteristics change over time. While some of these changes are dramatic and measurable, others are subtle and unmeasurable. Any conclusions made about trends in IBC must consider these changes and these limitations. A limitation of note in this analysis is the amount of missing data for patients diagnosed before Any inference about or comparison with these patients is extremely limited as there is little information available and the patients for whom characteristics are known are not likely to be representative of the entire group. Overall, it is quite worrisome that the survival of these patients has not improved much over these decades, and not only the development of clinical trials to treat this group of patients, but a better understanding of the first principles for IBC, acquired through molecular biology, etiologic epidemiology, and animal models, is needed to improve patient survival. IBC patients have been and still are excluded from the majority of clinical trials, probably because of the rareness of the disease, the lack of knowledge on how to treat these patients, and also the poor advances made against this type of breast cancer. Only through coordinated multidisciplinary, multicenter efforts and clinical trial participation will it be possible to develop more specific and targeted treatments able to significantly affect the prognosis of these patients. ACKNOWLEDGMENTS This study was supported by the Nellie B. Connally Breast Cancer Research Fund.

9 912 Inflammatory Breast Cancer Trends DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST The authors indicate no potential conflicts of interest. REFERENCES 1 Anderson WF, Schairer C, Chen BE et al. Epidemiology of inflammatory breast cancer (IBC). Breast Dis ;22: Hance KW, Anderson WF, Devessa SS et al. Trends in inflammatory breast carcinoma incidence and survival: The Surveillance, Epidemiology, and End Results program at the National Cancer Institute. J Natl Cancer Inst 2005;97: Low JA, Berman AW, Steinberg SM et al. Long-term follow-up for locally advanced and inflammatory breast cancer patients treated with multimodality therapy. J Clin Oncol 2004;22: Jaiyesimi IA, Buzdar AU, Hortobagyi GN. Inflammatory breast cancer: A review. J Clin Oncol 1992;10: Tabbane F, Bahi J, Rahal K et al. Inflammatory symptoms in breast cancer. Correlations with growth rate, clinicopathologic variables, and evolution. Cancer 1989;64: Taylor GMA. Inflammatory carcinoma of the breast. Am J Cancer 1938; 33: Robbins GF, Shah J, Rosen P et al. Inflammatory carcinoma of the breast. Surg Clin North Am 1974;54: Zucali R, Uslenghi C, Kenda R et al. Natural history and survival of inoperable breast cancer treated with radiotherapy and radiotherapy followed by radical mastectomy. Cancer 1976;37: Stocks LH, Patterson FM. Inflammatory carcinoma of the breast. Surg Gynecol Obstet 1976;143: Krutchik AN, Buzdar AU, Blumenschein GR et al. Combined chemoimmunotherapy and radiation therapy of inflammatory breast carcinoma. J Surg Oncol 1979;11: Rouesse J, Friedman S, Sarrazin D et al. Primary chemotherapy in the treatment of inflammatory breast carcinoma: A study of 230 cases from the Institut Gustave-Roussy. J Clin Oncol 1986;4: Buzdar AU, Singletary SE, Booser DJ et al. Combined modality treatment of stage III and inflammatory breast cancer. M.D. Anderson Cancer Center experience. Surg Oncol Clin N Am 1995;4: Rouzier R, Extra JM, Klijanienko J et al. Incidence and prognostic significance of complete axillary downstaging after primary chemotherapy in breast cancer patients with T1 to T3 tumors and cytologically proven axillary metastatic lymph nodes. J Clin Oncol 2002;20: Early Breast Cancer Trialists Collaborative Group. Effects of adjuvant tamoxifen and of cytotoxic therapy on mortality in early breast cancer. An overview of 61 randomized trials among 28,896 women. N Engl J Med 1988;319: Early Breast Cancer Trialists Collaborative Group. Polychemotherapy for early breast cancer: An overview of the randomised trials. Lancet 1988; 352: Henderson IC, Berry DA, Demetri GD et al. Improved outcomes from adding sequential paclitaxel but not from escalating doxorubicin dose in an adjuvant chemotherapy regimen for patients with node-positive primary breast cancer. J Clin Oncol 2003;21: Mamounas EP, Bryant J, Lembersky B et al. Paclitaxel after doxorubicin plus cyclophosphamide as adjuvant chemotherapy for node-positive breast cancer: Results from NSABP B-28. J Clin Oncol 2005;23: Martin M, Pienkowski T, Mackey J et al. Adjuvant docetaxel for nodepositive breast cancer. N Engl J Med 2005;352: Cristofanilli M, Gonzalez-Angulo AM, Buzdar AU et al. Paclitaxel improves the prognosis in estrogen receptor negative inflammatory breast cancer: The M. D. Anderson Cancer Center experience. Clin Breast Cancer 2004;4: Buzdar AU, Montague ED, Barker JL et al. Management of inflammatory carcinoma of breast with combined modality approach - an update. Cancer 1981;47: Ueno NT, Buzdar AU, Singletary SE et al. Combined-modality treatment of inflammatory breast carcinoma: Twenty years of experience at M. D. Anderson Cancer Center Cancer. Chemother Pharmacol 1997;40: Cristofanilli M, Buzdar AU, Sneige N et al. Paclitaxel in the multimodality treatment for inflammatory breast carcinoma. Cancer 2001;92: The World Health Organization Histological Typing of Breast Tumors Second Edition. The World Organization. Am J Clin Pathol 1982;78: Black MM, Speer FD. Nuclear structure in cancer tissues. Surg Gynecol Obstet 1957;105: Panades M, Olivotto IA, Speers CH et al. Evolving treatment strategies for inflammatory breast cancer: A population-based survival study. J Clin Oncol 2005;23: Fisher B, Bryant J, Wolmark N et al. Effect of preoperative chemotherapy on the outcome of women with operable breast cancer. J Clin Oncol 1998; 16: Kuerer HM, Newman LA, Smith TL et al. Clinical course of breast cancer patients with complete pathologic primary tumor and axillary lymph node response to doxorubicin-based neoadjuvant chemotherapy. J Clin Oncol 1999;17: