Parallel Comparison of Cancer Antigen 15.3 and Cancer Antigen 27.29 in Breast Cancer

Abstract

Abstract Objectives Mucins are large and complex glycoproteins expressed by several epithelial tissues. The expression of MUC1 mucin is upregulated in breast cancer and the protein is released in high amounts in patients’ blood. Cancer antigens (CA) 15.3 and CA 27.29, which measure different epitopes of MUC1 protein, are useful biomarkers for postoperative surveillance and therapy monitoring of advanced breast cancer. The aim of this study was to compare CA 15.3 assay and CA 27.29 assay by using paired patient samples. Methods The study was performed in 523 patient samples. CA 15.3 was measured by a sandwich assay on Beckman Coulter UniCel® DxI 800. CA 27. 29 was measured by a competitive immunoassay on Siemens ADVIA Centaur System. Results On Deming regression, the slope was 0.310 with an intercept of 8.6 and correlation coefficient of 0.951. The mean bias between the two assays was –39.2 (–56.6%). The agreement of the two methods was 89% (n = 467), among which, 12% (n = 62) were positive for both CA 15.3 and CA 27.29 (above the upper limit of the reference range), and 77% (n = 405) were negative for both assays (within the reference range). The disagreement was 11% (n = 56), which showed positive CA 27.29 and negative CA 15.3. The correlation coefficient between CA 15.3 and CA 27.29 is 0.9085 when both are positive and 0.6315 when both are negative. The bias is –238.9 (–84.7%) when both are positive and –8.5 (–52.4%) when both are negative. Paired t tests showed significant difference (P < .001) between the CA 15.3 and CA 27.29 in both positive samples and both negative samples. Conclusion In summary, there is a good correlation between CA 15.3 tested on Beckman Coulter UniCel® DxI 800 and CA 27.29 on ADVIA Centaur System. However, with the negative bias, they should not be used interchangeably.