From July, 1978 to September, 1981, 184 patients with localy advanced breast cancer (T3; T4a-b; any N; M0) regardless of their hormonal receptor status, entered a trial to evaluate the contribution of radiotherapy when added to an intensive preoperative chemoendocrine regimen. Seventy-eight patients were ultimately disqualified. All patients underwent sequentially: (1) two cycles of chemotherapy: Day 1--Oncovin 1.4 mg/m2, cyclophosphamide 350 mg/m2, Adriamycin 30 mg/m2; Day 2--methotrexate 20 mg/m2, 5-fluorouracil 350 mg/m2 (in addition, antiestrogens were given to postmenopausal patients); (2) mastectomy with complete axillary dissection combined with oophorectomy in patients before and one year after menopause; (3) radiotherapy randomly to one-half of the patients; and (4) ten additional chemotherapy cycles as above, with antiestrogens to all patients. No serious local sequellae were encountered from mastectomy or radiotherapy, but complications of chemotherapy were numerous, particularly in irradiated patients. One death due to toxicity occurred after preoperative chemotherapy. The results to date suggest that in irradiated patients metastases may become enhanced and that their local disease is not more effectively controlled than in patients not having radiotherapy. Two factors may have been largely responsible for the differences observed between the two groups: the delay of chemotherapy in irradiated patients and the sustained immunosuppression known to occur after mediastinal radiotherapy.