The mucin glycoprotein-detecting assay CA 15-3 is a valuable tool for monitoring the course of disease in breast cancer patients. Assays of CA 15-3 are based on the use of two MAbs to polymorphic epithelial mucin (PEM). We evaluated the technical and clinical performance of the Chiron ACS BR, an automated competitive chemiluminescence assay using a single MAb, B27.29, and compared the assay's results with those of the Centocor CA 15-3 RIA, the Abbott IMx CA 15-3, and the Boehringer Mannheim Enzymun-Test CA 15-3. The study population consisted of 253 healthy women, 66 patients with benign breast disease, 168 breast cancer patients, and 76 patients with other carcinomas. In the technical evaluation, we assessed the precision and linearity on dilution of the ACS BR assay. Cutoff values (upper limits of values seen in healthy subjects) were determined for all four assays. Agreement between the assays was studied by linear regression analysis. The ACS BR assay gave within- and between-assay CVs of 2.2% and 3.9%, respectively. Three samples from healthy women gave discordant values by ACS BR and were not included in the calculations. All four assays exhibit a highly similar pattern when monitoring breast cancer disease; the closest agreement of values was obtained between ACS BR and Centocor CA 15-3. We conclude that the ACS BR assay is a fast and reliable immunoassay for measuring PEM in serum. Although it detects a slightly different epitope on the PEM molecule than is targeted in other assays, for cancer serum samples it agreed better with the original Centocor CA 15-3 assay than did the other two CA 15-3 assays tested.