Several randomized studies in patients with advanced ovarian cancer have dealt with the comparison of cisplatin and carboplatin when given as either a single drug or in combination. The German Ovarian Cancer Study Group (GOCA) performed a prospective randomized trial in a subgroup of patients specified by a successful cytoreductive operation before the start of chemotherapy. From February 1987 to May 1990, 173 previously untreated patients with stage III and IV disease and limited tumor bulk of <2 cm postoperatively received either cyclophosphamide 600 mg/m2plus carboplatin 350 mg/m2or cyclophosphamide 1000 mg/m2plus cisplatin 80 mg/m2. The drugs had to be administered on Day 1 every 28 days for six subsequent courses.Results.In 158 assessable patients no significant differences in pathologically confirmed CR rates (pCR: 14% vs 16%), median time to progression (PFI: 19 months vs 26 months), and overall survival (OS: 35 months vs 37 months) were observed. Refusal of therapy due to toxicity was more frequent in the cisplatin arm, whereas patients with progressive disease predominated in the carboplatin arm. Nonhematologic adverse effects were more likely to occur with cisplatin whereas carboplatin patients experienced more myelosuppression. As prognostic factors associated with an increased risk of progressive disease and shorter overall survival time, stage of disease and amount of residual tumor after first surgery were determined.Conclusions.Carboplatin proved to be effective in patients with optimally debulked ovarian cancer. However, neither regimen used in this trial is sufficiently active to prevent tumor progression in the majority of patients.