Endocrine Abstracts

Correlation of three thyroglobulin assays in the management of differentiated thyroid cancer (DTC) patients

Introduction: Thyroglobulin(Tg) is the main biochemical marker of DTC. Usual immunometric assays(IMA) are prone to interference by autoantibodies which can lead to erroneous results. LC-MS/MS has been described as an assay that may provide accurate results, however the literature reported up to 40% of undetectable rates in patients with positive thyroglobulin antibodies(TgAb) and structural disease.

Objective: To compare three different Tg assays in patients with positive and negative TgAb.

Methods: 129 patients with negative (97) and positive TgAb (32), had Tg measured by LC-MS/MS assay(Tg-MS) with functional sensitivity of 0.7 ng/mL and at least one of the 2 IMAs (by Beckman^R and Roche^R) with FS of 0.1 ng/mL. Negative TgAb samples were considered when TgAb were negative at least by two Methods Roche ^R; Siemens^R and AntiTgII Siemens^R. The correlation between assays was assessed using Spearman’s correlation coefficient and Deming linear regression. Strength of the correlations were determined by the McBride scale: r < 0.90 poor; r = 0.90 to 0.95 moderate correlation; r = 0.95 to 0.99 is substantial; r > 0.99 almost perfect

Results: Comparing the performance of Tg-MS and Tg-IMA in samples with negative TgAb, the overall concordance of positive and negative results between Tg-IMAs and Tg-MS was 91.8% (89/97) when the FS of each assay was used. The concordance between Tg-Roche and Tg-Beckman was 99% (just one sample discordant). Method comparison between the Tg-IMAs and Tg-MS in TgAb negative specimens correlated well: correlation coefficient between Tg-MS and Tg-Beckman was r 0.950; Tg-MS and Tg-Roche were r 0.956; and the best correlation coefficient was seen between Tg-Beckman and Tg-Roche (r 0.982). On the other hand, comparing the performance of Tg-MS and Tg-IMA in samples with positive TgAb (n =55) undetectable results of Tg were observed in 20/32 patients (62.5%) in IMAs and in 22/32 (68.8%) in the LC-MS/MS assay. The overall concordance in TgAb positive samples between Tg-IMAs and Tg-MS was 87.5% (28/32). Tg-IMAs agreement was 96.4% (27/28). Correlations between Tg-IMAs and Tg-MS were worse than observed in negativeTgAb patients. The correlation coefficient between Tg-MS and Tg-Beckman was r 0.875 and between Tg-MS and Tg-Roche was r 0.878. Correlation between Tg-Beckman and Tg Roche was better (r 0.965)ConclusionThe 3 assays correlated well in negative TgAb samples but had poor correlation between IMAs and LC-MS/MS assays in the TgAb positive samples. Therefore, efforts should be done to develop assays able to measure Tg in TgAb positive samples.