Comparison of circulating CA15-3 and carcinoembryonic antigen levels in patients with breast cancer.

Abstract

An immunoradiometric assay (IRMA) has been used to determine circulating levels of the breast cancer-associated antigen, CA15-3. Of 1,050 normal control subjects, serum from 99 (9.4%) had CA15-3 antigen levels > 22 U/mL, while that from 58 (5.5%) and 14 (1.3%) had levels > 25 U/mL and 30 U/mL, respectively. In contrast, 115 of 158 patients (73%) with metastatic breast cancer had CA15-3 levels > 22 U/mL. Thirteen of 26 patients (50%) with only local metastases, 27 of 34 (79%) of those with only bone metastases, and 20 of 24 (83%) with hepatic metastases had CA15-3 levels > 22 U/mL. Furthermore, nine of 31 patients (29%) with primary breast cancer had CA15-3 levels > 22 U/mL. CA15-3 and carcinoembryonic antigen (CEA) levels were compared for the same patient population. Significantly more patients with metastatic breast cancer had elevated CA15-3 levels than had elevated CEA levels (P < .001). Furthermore, the CA15-3 IRMA was more sensitive than the CEA assay in patients with only bone metastases, as well as those with only local metastases. Significantly more patients with primary carcinoma of the breast also had elevated CA15-3 than had elevated CEA levels (P < .02). CA15-3 levels were > 22 U/mL in patients with nonmalignant conditions, including five of 25 patients (20%) with benign breast diseases, and 23 of 52 patients (44%) with benign liver diseases. Furthermore, CA15-3 levels were also > 22 U/mL in 24 of 54 patients (44%) with gastrointestinal (GI) malignancies, 12 of 17 patients (71%) with bronchogenic carcinoma, and 29 of 44 patients (66%) with epithelial ovarian carcinoma. Serial CA15-3 levels correlated with clinical disease course. Nineteen of 21 patients (91%) with tumor progression had at least a 25% increase in CA15-3 levels. Conversely, seven of nine patients (78%) with tumor regression had at least a 50% decrease in CA15-3 levels. Among 27 patients with stable disease, 16 (59%) had levels that did not vary by more than .+-. 25% of the original CA15-3 levels. These results indicate that the CA15-3 antigen is a sensitive marker for the evaluation and monitoring of patients with breast cancer.