This randomized study was designed to compare the efficacy and toxicity of MMV chemotherapy (mitomycin C(MMC), methotrexate (MTX) and vincristine(VCR)), MMVM (MMV+medroxyprogesterone acetate(MPA)) and MMVP (MMV+prednisolone(P)), and to evaluate potential additional effects of MPA or P to the chemotherapy in doxorubicin (ADR)-refractory cancer patients. A total of 108 advanced breast cancer patients who had been resistant to ADR or who had relapsed after response were randomized to the three treatment arms: MMC 4mg/m2, MTX 35mg/m2 and VCR 0.7mg/m2, i.v., on days 1 and 8; repeated every 3 weeks. MPA 1,200 mg/day or P 10mg/day was given orally in the MMVM and MMVP arms, respectively. An interim analysis showed that the response to MMV was marginally significantly lower than the other groups; entry to the arm was thus interrupted. 102 patients were evaluable. Totally, response rates of 9.5%(2/21) in MMV, 37.5%(15/40) in MMVM and 29.2%(12/41) in MMVP were obtained. There was a significant difference in the response between the MMVM and MMV arms (p = 0.0206), and a marginal difference between MMVP and MMV (p = 0.0784). Although overall survivals in the 3 groups were equivalent, the time to progression of patients treated with MMVM and MMVP was shown to be significantly longer than MMV patients. Hematological adverse effects, especially thrombocytopenia, were significantly diminished by the addition of hormones, particularly MPA, to the chemotherapy, thus increasing total doses of chemotherapeutic agents. It was shown that in ADR-resistant cancers MMVM had an effect equivalent to that in ADR-sensitive breast cancers. The authors conclude that MMVM may be a candidate for a noncross resistant regimen for ADR-resistant cancers, as well as an effective 2nd line chemoendocrine treatment of advanced breast cancer.