Pre-operative discrimination between benign and malignant ovarian tumors using a combination of CA125 and CA15.3...

Serum levels of CA125 and CA15.3 were measured in 70 patients presenting with an ovarian neoplasm, of whom 38 had an ovarian malignancy and 32 a benign ovarian tumor. CA125 levels exceeded 35 U/ml in 71% of ovarian carcinomas and in 25% of benign ovarian tumors. In the entire group of 70 patients, CA125 levels (greater than 35 U/ml) were elevated in 35 patients, of whom 27 had ovarian cancer. CA15.3 levels were found to be elevated (greater than 30 U/ml) in 9% of benign ovarian tumors and in 50% of ovarian malignancies. Of 8 patients with a false positive CA125 (greater than 35) elevation, only one had an elevated CA15.3 level whereas in 27 correct positive patients 19 also had elevated CA15.3 levels. Of all 20 patients with both markers elevated, 19 patients (95%) had ovarian cancer. When a cut-off level of 65 U/ml was used for the tumor marker CA125, all patients with simultaneous elevation of both markers were found to have an ovarian malignancy. Using a panel of CA125 (greater than 35 U/ml) and CA15.3 (greater than 30 U/ml) and requiring a simultaneous marker elevation, the sensitivity of the test decreased from 71% to 50% but the corresponding specificity of the test rose from 75% to 97%. Specificity was as high as 100% if in the same panel of tests a 65 U/ml cut-off for CA125 was taken. A comparison of early stage I-II ovarian cancer with benign ovarian tumors failed to demonstrate a discriminatory capacity of any test or test combination. We conclude that the use of a panel of tumor markers is advantageous in the pre-operative discrimination of benign and malignant ovarian tumors, since the predictive value for malignancy of a combined marker elevation was as high as 100% in the population studied.