PURPOSE To evaluate the use of an alternate, non-cross-resistant adjuvant chemotherapy regimen in women with a poor pathologic response to a preoperative doxorubicin-based regimen. PATIENTS AND METHODS Patients with locally advanced breast cancer received three cycles of vincristine, doxorubicin, cyclophosphamide, and prednisone (VACP) every 21 days followed by surgery. Patients with less than 1 cm(3) residual tumor at mastectomy received an additional five cycles of VACP. Those with more than 1 cm(3) residual tumor were randomly assigned to receive an additional five cycles of VACP or five cycles of vinblastine, methotrexate with calcium leucovorin rescue, and fluorouracil (VbMF). RESULTS One hundred ninety-three patients were evaluable. Overall clinical response was seen in 83.4% after three cycles of VACP, whereas the pathologic complete response was 12.2%. One hundred six patients were randomly assigned to VACP or VbMF. Those receiving VbMF achieved higher relapse-free survival (RFS) and overall survival (OS) than those who received additional VACP, although the differences did not reach statistical significance. Initial stage of tumor, clinical complete response, and pathologic complete response were all associated with statistically superior survival rates. CONCLUSION Clinical and pathologic response to preoperative doxorubicin-based chemotherapy predicted for improved survival in women with operable breast cancer. For those with a poor response to initial neoadjuvant chemotherapy, treatment with VbMF was associated with a trend toward improved RFS and OS compared with those continuing with the doxorubicin regimen.