MSACL 2025 Abstract

Performance Evaluation of Reflex Tg-MS testing in TgAb-positive Samples: A Retrospective Concordance Analysis with Immunoassay

Preejith P. Vachali, Matthew Dee, Delaney Ryan, Leah El Hage, Mario Skugor, Grace M. Kroner and Jessica M. Colón-Franco
Cleveland Clinic, Cleveland, OH, USA.

Preejith Vachali, MS PhD (Presenter) Cleveland Clinic >> POSTER (PDF)

Presenter Bio: Dr. Preejith Vachali is a Clinical Chemistry Fellow at the Cleveland Clinic with a PhD in bioanalytical chemistry. He brings a strong background in laboratory medicine and assay development, with prior experience in R&D roles across both clinical labs and academic settings. His interests lie in advancing diagnostic testing through innovation, with a focus on mass spectrometry-based assay development, lab automation, and biosensor technologies.

Relevant Financial Disclosures (within past 24 months, reported on Jun 10, 2026) No relevant financial relationship(s) to disclose. Conflict mitigation NOT REQUIRED.

BACKGROUND:
Thyroglobulin (Tg) is an essential biomarker for monitoring differentiated thyroid cancer. However, the presence of Tg antibodies (TgAb) can interfere with Tg measurements obtained by immunoassays. To address this, reflex testing using a mass spectrometry-based Tg assay (Tg-MS) is employed when TgAb levels exceed a defined threshold. But recent literature findings have raised concerns regarding the validity of using a fixed TgAb threshold to trigger reflex MS testing (1). Expert recommendations now suggest that MS test may not be needed if Tg is <0.1 µg/L with no evidence of structural disease (1). This study assesses the analytical agreement and correlation between Tg-MS (LC-MS/MS) and the initial Beckman Tg immunoassay in TgAb-positive samples.

OBJECTIVE:
To determine if reflex testing with the Tg- MS assay is necessary in TgAb-positive samples by evaluating its concordance with the initial Beckman Tg immunoassay, and whether strong agreement between the two methods supports limiting or omitting reflex testing in certain clinical scenarios.

METHODS:
Consecutive tests from March 2023 to February 2025 included 484 unique patients, and 928 total of Tg measurements. Tg values were assessed using two assays: an initial Beckman Tg immunoassay for all samples, with reflex Tg-MS testing conducted only for those with elevated TgAb (>4 IU/mL). Analytical focus was placed on values ≥0.5 ng/mL due to differences in detection thresholds (Beckman: ≥0.1 ng/mL, Tg-MS: ≥0.5 ng/mL). Tg values were categorized as <1, 1–2, and >2 ng/mL based on clinical cut-offs. Descriptive statistics were applied to demographic and assay data, using medians and interquartile ranges for non-normally distributed continuous variables, and frequencies and percentages for categorical variables. Non-normality was confirmed for both Tg assays, and log transformations failed to achieve normalization. Spear-man correlation analysis was performed to assess the strength and direction of association be-tween assays. To handle repeated measures, rank-transformed values were analyzed using the rmcorr package in R to produce within-subject correlation coefficients. Concordance correlation coefficient (CCC) was calculated based on a mixed-effects Poisson regression model incorporating random intercepts for patient ID, assay type, and their interaction, as per Tsai and Lin’s method (2) for non-normal distributions. The delta method was used to estimate 95% confidence intervals (CI). All analyses were conducted using SAS 9.4 and R version 4.3.

RESULTS:
The patient cohort had a median age of 57 years (IQR: 42–69), with 80.1% female. Across all 928 observations, the median Beckman Tg was 0.1 ng/mL and Tg-MS was 0.5 ng/mL. When restricted to values ≥0.5 ng/mL, Beckman Tg also had a median of 0.5 ng/mL, indicating baseline alignment with Tg-MS. The median absolute difference between assay values was 0.0 ng/mL, confirming minimal deviation. Among 624 observations with Beckman Tg levels <0.5 ng/mL, 93.7% also had Tg-MS below this threshold. Conversely, of the 262 observations with Beckman Tg >0.5 ng/mL, 95.8% showed corresponding Tg-MS values >0.5 ng/mL. The Spearman correlation coefficient between Tg-MS and Beckman Tg was 0.60 (95% CI: 0.53–0.65, p<0.001), indicating moderate positive correlation. The CCC was 0.98 (95% CI: 0.96–1.00, p<0.001), reflecting very strong agreement. The discrepancy between correlation and concordance is attributed to the non-linear relationship between values, suggesting that although trends are only moderately aligned, individual test values are nearly identical. Among the 484 unique patients, 57.2% had a single Tg measurements, while the remainder had multiple Tg measurements over time, enabling robust within-subject repeated measures correlation analysis.

CONCLUSIONS:
This analysis demonstrates a very high level of concordance between Tg-MS and Beckman Tg assays in TgAb-positive samples. While Tg-MS is currently performed reflexively when Tg anti-bodies are elevated, the strong agreement observed raises the question of whether reflex testing is always needed. In cases where Beckman Tg and Tg-MS values align closely, it may be reasonable to reconsider the routine necessity of reflex testing, particularly in settings where assay agreement is reliably high. However, further prospective testing and validation would be required before modifying clinical protocols.

REFERENCES:
1. Giovanella L., et al., hsTg&TgAb Consensus Working Group. Thyroglobulin and thyroglobulin antibody: an updated clinical and laboratory expert consensus. Eur J Endocrinol, 2023 189(2): R11-R27.
2. Miao-yu T and Chao-Chun L., Concordance correlation coefficients estimated by variance components for longitudinal normal and Poisson data. Comput Stat and Data An, 2018. 121 p. 57-70.