Preoperative Chemotherapy Permits More Conservative Primary Surgery in Women with Breast Cancer


Synopsis: Preoperative chemotherapy with cyclophosphamide and doxorubicin allows more women to undergo lumpectomy for local tumor control in women with breast cancers larger than 2 cm. Disease-free and overall survival were not significantly different whether chemotherapy was given before or after primary surgery.

Sources: Fisher B, et al. J Clin Oncol 1997;15:2483-2493; Fisher B, et al. Proc Am Soc Clin Oncol 1997; 16:127a.

The data from fisher and colleagues describe the effects of preoperative chemotherapy on the size of the primary breast cancer, the status of axillary lymph nodes, and the rate of lumpectomy in women who received four cycles of doxorubicin 60 mg/m2 plus cyclophosphamide 600 mg/m2 (AC) compared to a randomized control group who received the same chemotherapy postoperatively. The objectives of the trial were to determine: 1) if preoperative chemotherapy improved disease-free (DFS) and overall survival (OS), 2) if the response of the primary tumor and axillary lymph nodes correlated with DFS and OS, 3) if preoperative chemotherapy permitted more breast-conserving surgery and decreased the incidence of ipsilateral breast tumor recurrence, and 4) if preoperative chemotherapy resulted in a downstaging of the axillary lymph node status and increased the proportion of node-negative patients. Eligibility requirements included having a palpable, operable breast cancer (T1-3, N0-1, M0); it could not be fixed to the underlying muscle or chest wall, and, if present, palpable axillary lymph nodes could not be fixed to each other or to the neurovascular bundle. Patients with inflammatory breast cancer were excluded. The diagnosis of breast cancer was made by either fine-needle aspiration or tru-cut biopsy only.

After the diagnosis was established, the investigators were required to specify the type of operation that they intended to perform before the patient was randomized; 759 patients received postoperative chemotherapy and 747 received preoperative chemotherapy at participating NSABP institutions from 1988 through 1993. The patient characteristics were similar in both randomized groups, with approximately 50% of patients being premenopausal and 75% clinically node negative. The clinical size of the breast tumors was 2 cm or smaller in 28% of patients, between 2 and 5 cm in 59%, and larger than 5 cm in 13%. Response to chemotherapy was measured by clinical tumor measurements before each cycle of chemotherapy and pathologically after surgery. All patients over 50 years of age received tamoxifen 10 mg bid for five years starting after chemotherapy. Patients who underwent lumpectomy began postoperative radiation therapy four weeks after their surgery or postoperative chemotherapy.

A clinical complete response in both breast and axilla was documented in 35% of patients randomly assigned to receive preoperative chemotherapy, and a partial response was noted in another 44%, for an overall clinical response rate of 79%. Only 4% of patients had clinical evidence of progressive disease during treatment. Initial tumor size influenced the likelihood of a clinical complete response: 57% of tumors 2 cm or smaller, 35% of tumors 2-5 cm, and 17% of tumors larger than 5 cm. Pathological examination revealed that only 26% of the clinical CRs had no evidence of cancer, and in another 11% only ductal carcinoma in situ could be identified. Importantly, the pathologic CR rate in patients who experienced a clinical CR with chemotherapy was the same regardless of the clinical tumor size or the presence of palpable lymph nodes at the start of treatment. Overall, a pathological CR was produced by chemotherapy in 9% of patients (14% in tumors < 2 cm, 9% in tumors 2-5 cm, and 3% in tumors > 5 cm). Of the 185 patients with clinically positive lymph nodes, 7% had a pathological CR in both the breast and axillary lymph nodes.

When the preoperative chemotherapy group was compared to the postoperative chemotherapy group, significantly more patients had pathologically negative nodes (59% vs 43%, respectively; P < 0.001). One-half of the patients in the group that received postoperative chemotherapy who were judged to be clinically node- negative had pathological evidence of tumor in the lymph nodes, compared to only one-third of the patients who received preoperative chemotherapy. At the time of randomization, lumpectomy was the planned operation for two-thirds of the patients in both groups. However, the operation that was actually performed differed between the two groups—67% of all patients who received preoperative chemotherapy underwent lumpectomy vs. 60% who had their surgery first (P = 0.002). The percentage of patients undergoing lumpectomy according to their initial tumor size is shown in the table.


Patients undergoing lumpectomy

Lumpectomy Performed (%)

Postoperative Preoperative Tumor Size (cm) Chemotherapy Chemotherapy

All patients 60% 67% £ 2 cm 79% 81% 2-5 cm 63% 71% > 5 cm 8% 22%

Lumpectomy was performed more frequently in clinically node-negative women in both groups; however, the percentage of patients who underwent lumpectomy was greater after chemotherapy (59%) than that proposed at the time of randomization (49%). There was no increase in postoperative complications in patients who received chemotherapy preoperatively.

The article does not provide any information on the comparative results of the disease-free survival (DFS) and overall survival (OS) of the two groups of women; however, preliminary results of that important end point were described at the recent ASCO meeting. Patients who received preoperative chemotherapy had the same DFS (67%) and OS (80%) as those who received postoperative chemotherapy.


Although B-18 is not the first, it is by far the largest study to use preoperative chemotherapy in the treatment of breast cancer. Bonadonna previously demonstrated that preoperative chemotherapy could shrink tumors more than 3 cm, making them suitable (by their criteria) for lumpectomy and radiation therapy.1 What wasn’t known from that trial was whether such an approach compromised long-term outcome. The importance of B-18 is that it randomly assigned patients to preoperative or postoperative chemotherapy and demonstrated that the DFS and OS were the same.

How might the results of B-18 influence the way that breast cancer is treated? It could make us more sanguine about recommending preoperative chemotherapy to women who want breast-conserving therapy; but, will it actually result in the performance of significantly more lumpectomies? The rate of lumpectomy was identical in both arms of the study for tumors that initially measured less than 2 cm and was only 8% greater in patients with tumors measuring 2-5 cm. Although the lumpectomy rate was increased to 22% from 8% in patients with tumors larger than 5 cm (stage IIIA disease), most oncologists already offer these patients treatment similar to what is given to patients with inflammatory breast cancer, namely preoperative chemotherapy followed by surgery, more chemotherapy, and, finally, radiation therapy. Therefore, it is unlikely that we will see marked changes in the management of newly diagnosed breast cancer patients.

The trial was designed to detect a 33% proportional reduction in the relapse rate and death rate, and it failed to show that preoperative chemotherapy was superior. This result could be interpreted as a rejection of the argument that earlier treatment of micrometastases will translate into improved long-term, disease-free, and overall survival. Statisticians could also argue that to prove equivalence of two treatments, many more patients would have had to have been enrolled in this trial. Clinicians will conclude that they are equal because the results are consistent with several other smaller studies, all of which show no detrimental effects of preoperative chemotherapy2-5 (in fact, one shows a trend toward a more favorable outcome4).

The greatest effect of the study by Fisher et al will likely be on the design of future adjuvant therapy trials. Patients who experienced a clinical CR had a significantly better DFS and a trend toward improved OS. The overall pathological CR rate was 9% plus an additional 4% who had DCIS alone, and these patients had improved DFS and OS. All efforts should now be directed toward increasing this pathological CR rate with the expectation that it will correlate with the eradication of micrometastatic disease and improve long-term DFS and OS. The NSABP is currently proceeding along this path with its ongoing B-27 study that tests the value of the addition of docetaxel to the AC regimen used in the B-18 trial. In addition to the taxanes, other drugs such as vinorelbine or gemcitabine could be added to the adjuvant chemotherapy for breast cancer. It is conceivable that improvements in the pathological CR rate can serve as a surrogate for DFS and OS and, therefore, speed the progress in the adjuvant therapy of breast cancer.


1. Bonadonna G, et al. J Natl Cancer Inst 1990;82: 1539-1545.

2. Makris A, et al. Proc Am Soc Clin Oncol 1996;15:96 (abstr).

3. Dilhuydy JM, et al. Radiother Oncol 1992;24:S4 (suppl; abstr).

4. Scholl SM, et al. Proc Am Soc Clin Oncol 1995;14:125 (abstr).

5. Tubiana-Hulin M, et al. Eur J Cancer 1994;30A:S37 (suppl 2; abstr).