Primary chemotherapy for operable breast cancer: the NSABP experience

Introduction

Primary or neoadjuvant systemic chemotherapy, initially described for patients with locally advanced or borderline inoperable breast cancer, has been increasingly utilized for patients with less advanced or operable breast cancer. Theoretically, primary systemic therapy could inhibit the rapid growth of metastases after surgery and may decrease the emergence of chemoresistant clones of cells. On a practical level, primary systemic therapy has the potential to increase the use of breast conservation by decreasing tumor size. Beginning in 1988, the National Surgical Adjuvant Breast and Bowel Project (NSABP) cooperative group conducted two sequential trials to test the value of neoadjuvant chemotherapy and to optimize the treatment regimen for operable breast cancers.

Methods

NSABP Protocol B-18 was designed to compare pre-operative chemotherapy with doxorubicin (adriamycin) and cyclophosphamide (AC) given every 3 weeks for four cycles versus the same chemotherapy treatment given in the adjuvant setting. In protocol B-18, 1523 women with operable breast cancer were randomized to receive four cycles of AC followed by surgery or surgery followed by four cycles of AC. Women 50 years of age or older also received tamoxifen for 5 years, starting after chemotherapy.

Subsequently, NSABP Protocol B-27 was conducted with the intent to determine the effect of adding docetaxel (taxotere [T]) after four cycles of preoperative AC on disease-free survival (DFS) and overall survival (OS) of women with operable breast cancer. A total of 2411 women with operable primary breast cancer were randomized to receive either four cycles of preoperative AC followed by surgery (group I) or four cycles of AC followed by four cycles of T, followed by surgery (group II), or four cycles of AC followed by surgery and then four cycles of T (group III). Tamoxifen was given to all patients, starting concurrently with chemotherapy.

Results

In protocol B-18, mean tumor size was 3.5 cm. Preoperative AC produced objective clinical responses in 79% of the treated patients and clinical complete responses (cCR) in 36%. Pathologic complete responses (pCR, defined as no invasive cancer in the breast) were observed in 13%. OS and DFS were similar in the two randomized treatment groups. Preoperative chemotherapy resulted in a statistically significant increase in the rate of breast conserving therapy (BCT), from 60% to 68%. This was particularly notable in the patients with tumors >5 cm, in whom BCT was increased from 8% to 22% [1]. Although there was a trend toward increased ipsilateral breast tumor recurrence (IBTR) in preoperative chemotherapy patients who were downstaged to lumpectomy compared with patients treated preoperatively who were considered to be candidates for BCT at the outset (15.9% versus 9.9%), this difference was not statistically significant after controlling for patient age and tumor size [2]. Patients in the preoperative chemotherapy group who experienced a pCR had significantly improved DFS and OS compared with all other patients in the preoperative chemotherapy group (P < 0.0001). Clinical response was also associated with improved outcomes with long-term follow up [2, 3].

For protocol B-27, mean tumor size was 4.5 cm; this and other key characteristics were evenly balanced among the three treatment arms. The addition of docetaxel preoperatively resulted in significant increases in cCR and pCR at the time of surgery compared with AC alone (63.6% versus 40.1% and 26.1% versus 13.7%, respectively) [4]. Despite this, addition of docetaxel to AC did not significantly impact on survival in this cohort of patients [5]. There was a trend toward improved DFS in group II patients who received preoperative T, but this was not statistically significant (72% versus 67% DFS at 5 years; HR = 0.86, P = 0.10). In an analysis of relapse-free survival (RFS), which did not include second primary cancers, group II had a significantly better outcome compared with group I (74% versus 69% RFS at 5 years; HR = 0.81, P = 0.03). Group III RFS was not significantly different from group I (71% at 5 years; HR = 0.91, P = 0.32). Addition of docetaxel significantly reduced the incidence of local recurrences as first events, including IBTR in patients treated with breast conservation. There were no significant interactions between treatment and estrogen receptor status, age, tumor size, or clinical nodal status. An exploratory analysis of treatment effects in subsets of patients according to clinical response to AC suggests that preoperative T, but not postoperative T, significantly increased DFS in patients who had a partial clinical response after four cycles of AC (63%, 74%, 65% at 5 years for groups I, II, and III; HR = 0.68 for group II versus group I, P = 0.003). Addition of T did not appear to be beneficial in patients who were nonresponders after AC nor in those patients who had a cCR after AC. Pathologic complete response was a highly significant predictor of DFS and OS in all treatment groups (HR = 0.45, P < 0.0001, and HR = 0.33, P < 0.0001, respectively). In addition, pathologic nodal status after chemotherapy was a significant prognostic factor for survival, independent of pathologic response in the breast.

Conclusion

The B-18 trial did not demonstrate superiority of neoadjuvant over adjuvant chemotherapy for operable breast cancer, but with equivalent survival and increased BCT, the neoadjuvant approach can safely be used to offer BCT to more women with breast cancer. This trial also demonstrated a strong association between pCR and improved patient outcomes. In B-27, however, despite a doubling of the pathologic complete response rate in the breast with the addition of T preoperatively, we have not yet observed a significant improvement in DFS or OS for the study as a whole. Addition of preoperative or postoperative docetaxel decreased the incidence of local recurrences. There was a decrease in relapses with the addition of preoperative T, particularly in a subset of patients who had partial clinical responses to AC alone. Post-treatment pathologic response in the breast and nodal status remained powerful predictors of patient outcomes. Future studies will examine the value of additional drugs given with doctaxel after AC preoperatively and will also be designed to assess the ability of genomic and molecular profiles of pretreatment tumor to predict responsiveness to chemotherapy.