Feasibility of improving sLe(a)/CA 19-9 harmonization by recalibration: a simulation study using three EQA programs

Objectives: The CA 19-9 tumor marker measures the sialylated Lewis a blood group antigen (sLe(a)) expressed on circulating proteins. CA 19-9 is primarily used to monitor pancreatic and gastrointestinal cancers. Studies have demonstrated the lack of comparability of CA 19-9 assays. Methods: Data from three EQA programs were used to simulate the potential for increasing the CA 19-9 harmonization. The simulation was based on method comparison studies of each pair of measurement procedures (MPs), selection of comparable MPs for value assignment of EQA samples, and mathematical recalibration of MPs. Harmonization status was determined pre- and post-harmonization, using consensus mean of all MPs, as reference, and method comparison studies. Results: Seventy-five EQA samples were included and four MPs (Abbott, Beckman, Roche and Siemens (Atellica)) were present in all EQA programs and five others from Siemens (Immulite), Ortho, Fujirebio, Biomerieux and Thermo Fisher, in at least one EQA program. The obtained harmonization procedure required recalibration of all MPs. Post-harmonization mean difference improved from -50.3 to 92.0 % into -0.48-1.25 %. The range of slopes from individual MP method comparisons improved from 0.090 to 2.56 into 0.93 to 1.10. Post-harmonization CV(%) of individual EQA samples remained at 22.0-49.2 % within individual method comparisons and alignment of upper limits of normal diversified after application of the harmonization procedure. Conclusions: Overall, CA 19-9 harmonization can be improved by recalibration, though a significant lack of correlation remains. The feasibility of achieving CA 19-9 harmonization of today's clinical assays is questionable.