Eighty-eight women with epithelial ovarian carcinoma, treated by first-line chemotherapy, achieved a complete clinical response and underwent second-look laparotomy to assess the true pathologic status of their disease. Persistent tumor was found in 50 patients (57%). Thirty-two of these (36%) had obvious gross tumor, whereas, 16 (18%) had microscopic disease. Thirty-eight patients (43%) had no pathologic evidence of persistent ovarian carcinoma. With a follow-up of 6 to 60 months, 30 of these patients (79%) remain without evidence of recurrence. Multivariate logistic regression analysis revealed three covariates that were independently significant in predicting continued disease-free status. These included: the greatest diameter of the largest residual tumor left at the primary operation; histologic features of the tumor; and the diameter of the largest tumor aggregate found at initial operation. A mathematical model based on the most significant covariates was designed to assess the relative risk of any patient having persistent tumor at second-look laparotomy. A comparison of the predicted to actual outcome revealed a sensitivity of the model of 88%, a specificity of 71%, and an accuracy of 77%. Second-look laparotomy represents the basis on which potentially curative second-line salvage therapy can be initiated. With an increasing period of follow-up and with greater numbers of patients, it can potentially document a complete pathologic response to first-line therapy administered with curative intent, and identify patients for additional, adjunctive therapy, who are at risk of recurrence.