Comparative Analytical Performance and Clinical Agreement of Chemiluminescence Microparticle Immunoassay and Electrochemiluminescence Immunoassay for CA19-9: A Diagnostic Accuracy Study

Background: Carbohydrate antigen 19-9 (CA19-9) is widely used as a tumor marker for pancreatic and hepatobiliary malignancies. However, inter-assay variability may affect the interpretation of results. This study compared the analytical performance, reference values, diagnostic accuracy, and clinical concordance of CA19-9 measured by chemiluminescent microparticle immunoassay (CMIA) and electrochemiluminescence immunoassay (ECLIA) in a Thai population. Methods: Serum CA19-9 concentrations were measured using both CMIA and ECLIA in 398 subjects comprising 145 healthy individuals, 115 patients with benign conditions, and 138 patients with malignancies. Method comparison, reference interval assessment, receiver operating characteristic (ROC) analysis, and concordance analysis were performed. Population-specific cut-off values were established and compared with manufacturer-recommended thresholds. Results: CMIA produced modestly but significantly higher CA19-9 concentrations than ECLIA in healthy individuals, resulting in higher upper reference limits. The two assays showed strong correlations across the study population (R2 = 0.922). Among malignant conditions, CMIA generally yielded higher CA19-9 values than ECLIA, with significant differences observed in pancreatic cancer and cholangiocarcinoma (p < 0.001). Values in benign conditions were largely comparable, except in cirrhosis, where ECLIA produced higher values. Diagnostic performance was similar between CMIA and ECLIA (AUC 0.856 vs. 0.840). Population-derived cut-off values (34.945 kU/L for CMIA and 34.395 kU/L for ECLIA) showed diagnostic performance comparable to manufacturer-recommended cut-offs and improved inter-assay concordance, particularly in cholangiocarcinoma. Conclusions: CMIA and ECLIA demonstrated strong analytical correlation and comparable diagnostic performance for CA19-9. Nevertheless, systematic inter-assay differences indicate that results should be interpreted using assay-specific reference intervals and clinical decision thresholds. Population-specific cut-off values may further improve clinical interpretation, particularly in regions with a high prevalence of hepatobiliary malignancies.