Comparison of LC-MS/MS to immunoassay for measurement of thyroglobulin in fine-needle aspiration samples.

The measurement of thyroglobulin (Tg) by immunoassay (IA) is affected by the presence of anti-Tg autoantibodies (Tg-AAbs), which may cause false-negative or falsely low results (1 ). It has been shown that Tg quantification by LCMS/MS overcomes interference of Tg-AAbs in measurement of Tg (2 ). We previously observed that 20% of Tg-AAb–positive serum samples that tested negative for the presence of Tg by IA ( 0.1 ng/mL) had Tg concentrations 0.5 ng/mL when tested by LC-MS/MS (2). Here we discuss the potential utility of applying this LC-MS/MS method to testing Tg in fine-needle aspiration (FNA) samples. Ultrasound-guided FNA cytology and measurement of Tg concentrations in FNA samples are commonly used in the evaluation of suspicious lymph nodes in the diagnosis of thyroid carcinoma. Tg measurement in needle washout fluids (FNA-Tg) allows improved diagnostic accuracy of cytology (3 ) and is useful in patients with aggressive tumors, low serum thyroglobulin concentrations, or circulating Tg-AAbs. FNA-Tg concentrations 10 ng/mL are considered abnormal (4). Jeon et al. (5 ) proposed that Tg-AAbs could interfere with FNA-Tg measurements and cause falsely low FNA-Tg concentrations. Thus, we evaluated whether our LC-MS/MS method intended for serum Tg (2 ) could be used for measurement of FNA-Tg, and whether any of the FNA samples that tested negative by IA would have measurable Tg when analyzed by LC-MS/MS, suggesting Tg-AAb interference with IA. Samples used in the study were deidentified FNA patient samples submitted to ARUP Laboratories (Salt Lake City, UT) for routine testing. The study was approved by the University of Utah Institutional Review Board. We adapted the LC-MS/MS method for measuring serum Tg (2) for analysis of FNA-Tg. In brief, the samples were diluted with Tgnegative serum (ratio of 1 part to 9 parts, respectively), and Tg was enriched using rabbit polyclonal anti-Tg antibody and protein precipitation. The enriched proteins were then denatured and reduced and, after addition of an internal standard, digested with trypsin. A Tg-specific tryptic peptide was purified by immunoaffinity extraction and analyzed by LC-MS/MS (AB Sciex 5500) operated in positive ion, electrospray ionization mode. Two mass transitions were monitored for the targeted peptide (m/z 636.9/ 1060.5; 636.9/913.5) and the internal standard (m/z 639.9/1066.5; 639.9/ 919.5). Instrument cycle time was 6.5 min per sample. Imprecision of triplicate analysis of FNA samples containing 30, 2030, and 80 500 ng/mL was 15.2%, 11.2%, and 2.5%, respectively; between-run imprecision at 2 ng/mL (analyzed once per day for 15 days) was 13.9%. We evaluated performance of the methods in the vicinity of cutoff required for Tg-FNA samples by analysis of Tgnegative FNA samples (n 4) spiked with Tg to 15 ng/mL using IA and LC-MS/MS. The mean (CV) measured concentrations for IA and LC-MS/MS were 14.6 ng/mL (2.2%) and 18.4 ng/mL (16.0%), respectively. Using the LC-MS/MS method, we analyzed 73 FNA samples previously analyzed for Tg with the Beckman Coulter AccessTM Tg-IA. FNA samples (1 part diluted with 9 parts Tg-AAb negative serum) were tested for the presence of Tg-AAb using the Thyroglobulin Antibody IITM assay (Beckman Coulter). Serum samples of 7 individuals whose FNA samples were analyzed in this study were positive for Tg-AAb. None of the analyzed FNA samples had detectable Tg-AAb (cutoff 4 IU/mL). The range of Tg concentrations in the samples was 0 –160 000 ng/mL; 18 samples tested negative for Tg (cutoff 5 ng/ mL) by both methods. The median concentrations of Tg by LC-MS/ MS and immunoassay were 58 and 63 ng/mL, respectively. The Deming regression for the entire set was IA 1.14 * LC-MS/MS 473, r 0.994, Sy|x 2220; for samples with Tg 70 ng/mL, it was IA 1.63 * LC-MS/MS 1.2, r 0.879, Sy|x 4.8 (Fig. 1). For the 7 FNA samples with corresponding Tg-AAb– positive serum samples, the Deming regression equation for Tg was IA 1.10 * LC-MS/MS 34.3, r 0.976, Sy|x 162; the FNA-Tg concentration range was 0 –2400 ng/ mL. In 1 sample, the Tg concentration was substantially underestimated by the IA, and in 2 samples concentrations measured by IA were 2.6 and 2.9 times greater than by LC-MS/MS. The data showed reasonably good agreement between LCMS/MS and Beckman IA at Tg concentrations 70 ng/mL; the concentrations determined by LC-MS/MS were lower than by IA in samples with Tg 70 ng/mL. The lower observed concentrations could be related to Tg degradation during storage of the FNA samples; degradation could be related to the use of saline as a solvent for dilution of the needle biopsy samples. Alternatives to the saline diluents used for FNA samples should be evaluated to find a matrix that allows for improved stability of Tg. Future studies in which clinical information on the partici1 Nonstandard abbreviations: TG, thyroglobulin; IA, immunoassay; AAb, autoantibody; FNA, fineneedle aspiration. © 2014 American Association for Clinical Chemistry Clinical Chemistry 60:11 1452–1456 (2014) Letters to the Editor