To determine the role of secondary cytoreductive surgery in the management of patients with ovarian cancer, 106 patients with advanced invasive epithelial malignancies seen at the UCLA Medical Center during the 5-year period from 1974 to 1979 were retrospectively reviewed. Of those, 73 underwent either primary or secondary surgery as part of their therapy. Thirty-two of the 73 patients underwent secondary cytoreductive surgery at the authors' institution. All of these patients had undergone primary cytoreductive surgery followed by a course of chemotherapy. Twelve patients (38%) underwent optimal secondary resection of the tumor (largest residual tumor mass less than 1.5 cm in diameter) and 20 underwent nonoptimal cytoreductive surgery. The operations consisted of resection of recurrent tumor (32), bowel (17), residual omentum (12), and lower urinary tract (1). There was no operative mortality and the complication rate was 34%, most of which (8 of 12) were prolonged ileus. The median survival of patients in the “optimal” group (20 months) was significantly longer than that of those in the “nonoptimal” group (5 months; P<.01). This difference was independent of the type of subsequent treatment, tumor grade, patient age, and the presence of bowel obstruction. Patients with clinically inapparent ascites (less than 1000 ml) had a median survival of 18 months, significantly longer than the 5-month median in patients who had ascites (at least 1000 ml; P<.01). Within the optimal group, patients who underwent optimal primary resection of the disease survived longer than those who had nonoptimal primary operations (P<.05). Patients with smaller maximum tumor size (less than 5 cm) before secondary resection survived longer (P<.05), as did those whose tumors were resected at the time of a “second look” rather than with clinical disease progression (P<.05). These data indicate that a secondary attempt at bulk removal of tumor appears justifiable in patients who have had incomplete responses to primary chemotherapy and who have no clinical evidence of ascites.