From 1984 to 1988, thirty-nine untreated patients with epithelial ovarian cancer received Cisplatin 50 mg/m2, Doxorubicin 50 mg/m2, and Cyclophosphamide 750 mg/m2 (CAP), at 3 weekly intervals. All patients had FIGO stage III or IV tumors except 2 patients with stage IIb and IIc, respectively. After initial surgery 23 patients had residual disease > 2 cm in diameter. Twenty-five patients (64%) were evaluable for response to chemotherapy. Objective responses were observed in 13 out of 25 patients (52%, 95% confidence intervals (CI), 32.42% to 71.58%), 6 patients had a cCR (24%) and 7 had a cPR (28%). Seventeen out of the 39 patients (44%) had a second-look laparotomy. A pCR was achieved in 5 out of 17 patients (29%); a pPR was obtained in 8 patients (47%). Median duration of survival was 41,5 months (range 2-107+); median duration of time to failure was 21 months (range 2-107+). Median disease-free survival was 86 months (range 3,5-107+). Eleven patients (28%) are alive and 9 patients (23%) are free of recurrence at median follow-up of 86 months. Only 4 of 11 long-term survivors had a pCR. In univariate analysis, histology, clinical response to chemotherapy, and the presence of ascites at the time of diagnosis, achieved a significant correlation with survival and time to failure (TTF); in addition, TTF was significantly affected by pathological response to induction chemotherapy. The only important predictors of disease-free survival (DFS) were tumor grade and stage of disease. In multivariate analysis, the presence of ascites was the only significant prognostic factor with respect to survival and TTF. Our study confirms the effectiveness of CAP in the treatment of epithelial ovarian cancer and the relatively poor long term prognosis.