This study (Alabama Breast Cancer Project) reports the ten-year surgical results of a prospective randomized trial comparing Halsted radical mastectomy (RM) with modified radical mastectomy (MRM) for breast cancer. We entered 311 patients in the study between 1975 and 1978. Patients with histologically positive axillary lymph nodes were randomized after operation to receive melphalan or intermittent intravenous cyclophosphamide, methotrexate, and fluorouracil for one year. After a median follow-up of ten years, there was no significant difference in the survival of the two groups (RM, 71%; MRM, 64%). Local recurrence after RM was significantly lower than after MRM. A subset of patients with more advanced cancers (T3 and T2 with clinically positive axillary nodes) experienced significantly better survival at ten years following RM compared with MRM (59% vs 38%, respectively). These results indicate that overall survival is similar for patients treated by either RM or MRM. However, there is subset of patients with more advanced cancers whose ultimate survival can be favorably influenced by RM. © 1987, American Medical Association. All rights reserved.