Comparative analysis of three thyroglobulin immunoassays: analytical performance and clinical implications.

[OBJECTIVES] Serum thyroglobulin (Tg) is a key biomarker in the post-surgical monitoring of differentiated thyroid cancer (DTC). However, inter-assay variability among different immunoassay platforms can impact clinical interpretation, particularly at low Tg concentrations. This study aimed to compare the analytical performance and concordance of three widely used Tg immunoassays, Access (Beckman Coulter, Tg-B), Atellica (Siemens, Tg-A), and Liaison (Diasorin, Tg-L), with a focus on their agreement across clinically relevant Tg ranges.

[METHODS] A total of 103 residual serum samples from subjects with or without thyroid pathology were analyzed using Tg-B, Tg-A, and Tg-L. Correlation analysis, Bland-Altman plots, and concordance rates were evaluated across three Tg concentration intervals: <2 ng/mL, 2-50 ng/mL, and > 50 ng/mL. Tg-B was used as the reference method for comparison.

[RESULTS] All three assays demonstrated strong overall correlations. Tg-L showed a very strong correlation with Tg-B (ρ = 0.89), with moderate agreement at Tg < 2 ng/mL. Tg-A also correlated well with Tg-B (ρ = 0.92), though agreement declined slightly at higher concentrations (> 50 ng/mL). The concordance rate for detecting undetectable Tg (< 0.2 ng/mL) was 96% for Tg-L and 98% for Tg-A when compared to Tg-B. Bland-Altman analysis revealed a significant negative bias for Tg-L versus Tg-B, while Tg-A and Tg-B showed no significant difference. A significant discrepancy was also observed between Tg-L and Tg-A.

[CONCLUSIONS] Although the three Tg immunoassays demonstrated high correlation, notable differences emerged at lower and higher Tg levels, likely due to assay-specific antibody characteristics and calibrator variability. Our findings underscore the need for re-baselining when switching methods in longitudinal follow-up. However, the use of residual serum samples from a heterogeneous population, including individuals with and without thyroid pathology limits the direct clinical generalizability of the results and underscores the need for further validation in well-characterized post-thyroidectomy DTC cohorts.