Sera from 51 ovarian cancer patients in clinical remission undergoing a second-look operation were studied. CA 125, CA 15-3, and 90K assays were performed in the week preceding surgery. Twenty-seven patients (53%) had no evidence of disease; 3 (6%) and 21 (41%) had microscopic and macroscopic disease. All patients but two with no evidence of disease had normal CA 125 serum levels, as did the three cases with microscopic disease and 6 patients with macroscopically detectable tumor. Of 17 patients with high CA 125 values, 15 (88%) were found to have persistent disease. Hence, the sensitivity and specificity of a CA 125 assay were 62 and 93%, respectively. CA 15-3 levels were low in all patients with no evidence of disease, in 2 cases with microscopic disease, and in 2 cases with residual tumor greater than 1 cm. All patients with high CA 15-3 titers had a positive second-look. The sensitivity and specificity of assay were 69 and 100%, respectively. With the 90K assay, 4 false-positive and 4 false-negative results were found and the sensitivity and specificity were 69 and 67%, respectively. However, 3 of 4 cases with a false-positive 90K assay had recurrence of disease with 7, 9, and 20 months later. Combination of the three markers increased sensitivity to 79%. These data suggest that more effective surveillance of disease status at second-look is obtained by using a combination of tumor marker assays.