Serum tetranectin (Se-TN) and CA-125 were measured in 315 patients with a pelvic tumor and 458 healthy females. At a false-positive rate of 0.7% the sensitivity for ovarian cancer stage 1 and 2 was 33% for Se-TN and 76% for both CA 125 and an index based on Se-TN and CA 125 (Index 1). At a false-positive rate of 0.4% the sensitivity was 29% for Se-TN, 62% for CA 125 (35 U/ml), and 76% for Index 1. A negative correlation was found between the Se-TN level and the stage of cancer. The sensitivity for benign tumors was 6% for Se-TN, 17% for CA 125, and 21% for Index 1 at a false-positive rate of 0.4%. In the present study the sensitivity and specificity levels of Se-TN were not sufficiently high to warrant the use of Se-TN as an individual marker for ovarian cancer. The sensitivity rose with the index-based Se-TN and CA 125 without causing a concomitant increase in the rate of false-positive results. None of the markers rose to levels that would allow their use in clinical diagnosis to discriminate between localized cancer and benign tumors.