Thirty-eight patients with stage III ovarian carcinoma were treated with a protocol consisting of an initial phase of induction of remission with cyclophosphamide, hexamethylmelamine, doxorubicin, and cisplatin (CHAD) combination chemotherapy and a second laparotomy for resection of residual tumors, followed by a consolidation phase with curative doses of whole abdominal radiation. Six patients (16%) had stage IIIA disease, ten (26%) IIIB, and 22 (58%) had stage IIIC disease. All patients received three to 14 courses of CHAD chemotherapy with a clinical response rate (complete [CR] and partial [PR]) of 91%. Thirty-three patients underwent the second operation. In 14 patients no residual tumor was found, and in another 11 residual tumors found were totally resected. Thus, 25 of 33 (76%) were classified as in pathological complete remission (PCR) after this operation. Whole abdominal irradiation was well tolerated, although 12 of 29 (42%) of the irradiated patients required more than a 2-week interruption of the treatment course because of leukopenia and/or thrombocytopenia. The actuarial 5-year survival and disease-free survival rates for the whole group were 27% and 17%, respectively, and for the 29 patients who received the complete sequence of the prescribed protocol treatments, 35% and 20%, respectively. A univariate analysis of clinical parameters showed that inherent biological features, such as histology and grade, were the most dominant factors affecting prognosis, and that neither the aggressive surgical approach employed, nor the high-dose whole abdominal irradiation, significantly affected the outcome. The long-term results suggest that although our combined modality protocol was well tolerated, it failed to enhance the cure of stage III ovarian carcinoma. The possible biological and therapeutic vectors affecting this outcome are discussed.