Objective: To compare the sensitivity of CA-15.3 with cytologic diagnosis and carcinoembryonic antigen (CEA) and CA-125 content and to determine if elevated CA-15.3 suggests a breast primary.
Study design: A retrospective study of 111 consecutive effusions. CEA was measured by enzyme immunoassay (cutoff > 5 ng/mL), CA-125 by radioimmunoassay (cutoff > 5,000 micron/mL) and CA-15.3 by radioimmunoassay (cutoff > 15 micron/mL). Results were correlated with cytologic diagnosis, histology (when available), clinical and radiologic data, and follow-up.
Results: The results were: benign, 39; malignant, 72; sensitivities and specificities: cytology, 76%/100%; CEA, 80%/97%; CA-125 22%/100%; and CA-15.3, 69%/89%. The sensitivity of CEA and cytology was 97% and of CA-15.3 and cytology, 87%. Neither mean (293.4 micron/mL) nor range of CA-15.3 predicted a breast primary. Falsely elevated CEA and CA-15.3 were noted in ascites with fecal contents from colonic perforation. High CA-125 and low CEA predicted an ovarian/endometrial primary.
Conclusion: CA-15.3 assay in effusions is not recommended since it neither enhances the sensitivity of cytologic diagnosis nor predicts a breast primary.