Christopher Shuford (Presenter)
Laboratory Corporation of America
Authorship: Christopher M. Shuford (1), Jay S. Johnson (2), J. Will Thompson (3), Patricia L. Holland (1), James Murphy (2), Russell P. Grant (1)
(1) Laboratory Corporation of America, Burlington, NC (2) Waters Corporation, Milford, MA (3) Duke University, Durham, NC
Thyroglobulin is measured in thyroid cancer patients following total-thyroidectomy as a marker for recurrent disease given circulating concentrations should be effectively zero. Although LC-MS/MS assays are resistant to autoantibody interferences (via tryptic digestion) that inhibit immunoassays, recent publications have demonstrated approximately 40% of individuals with recurrent disease go undetected by LC-MS/MS assays (Netzel et al, JCEM, 2015), suggesting autoantibodies decrease the circulating concentrations below the LLOQ of the assays used (<0.5 ng/mL). Attempting to improve the clinical sensitivity, we developed of an assay with an LLOQ of 20 pg/mL and observed a significant fraction of individuals with previously undetectable thyroglobulin did indeed have detectable/quantifiable amounts circulating.
Thyroglobulin is measured in thyroid cancer patients following total thyroidectomy as a marker for recurrent disease given circulating concentrations should be effectively zero. Immunoassays measuring thyroglobulin provide inaccurate measurements due to the presence of autoantibodies, resulting in false negative results. Although LC-MS/MS assays are resistant to autoantibody interferences (via tryptic digestion), recent publications have demonstrated approximately 40% of individuals with recurrent disease go undetected by LC-MS/MS assays (Netzel et al, JCEM, 2015), suggesting autoantibodies decrease the circulating concentrations below the lower limits of quantification (LLOQ) of the assays used (<0.5 ng/mL). Attempting to improve the clinical sensitivity Tg monitoring, we developed an LC-MS/MS assay with an LLOQ of 20 pg/mL and observed a significant fraction of individuals with previously undetectable thyroglobulin did indeed have detectable/quantifiable amounts circulating.
LabCorp’s in-house LC-MS/MS assay for quantification of thyroglobulin in human serum, previously validated on a SCIEX Triple Quad™ 5500 LC-MS/MS system (1 mL/min, LLOQ 0.2 ng/mL), was re-deployed on an Waters ionKey/Xevo-TQS MS system (3 uL/min LLOQ, 20 pg/mL). Only minor modifications were made to the sample preparation, which involves tryptic digestion and subsequent enrichment of the signature peptide, FSPDDASAVLLR. Due to the added sensitivity at lower flow-rates, the concentration of the cleavable internal standard peptide was reduced 10-fold due and the wash stringency/volume during enrichment was increased to eliminate isobaric interferences. The ionKey/MS system was run in a dual-pump trap-elute configuration with an additional guard column up stream of the trap to prevent protein accumulation/precipitation and improve system ruggedness.
A head-to-head comparison of the two LC-MS/MS platforms was performed using a dilution series of the neat signature peptide containing a constant amount of labeled signature peptide. The detection limit for both was defined as the lowest amount loaded on column demonstrating a S/N>10, imprecision in the peak area ratio (PAR) <20%, and dilutional bias<20% across 4 replicate injections. The ionKey/MS system proved to be 40-fold more sensitive than the high-flow system (5 vs 200 amol on-column detection limits).
Initial attempts to re-deploy the assay on the ionKey/MS system produced spurious chromatographic pressure profiles consistent with on-column protein precipitation, resulting in poor column longevity. Addition of a protein-scavenging column up-stream of the trap column resolved this issue, allowing for >1200 injections on a single trap and analytical column during subsequent validation experiments. Additionally, the higher sensitivity afforded by the ionKey/MS system revealed contamination of the previously validated calibration matrix relative to the new target LLOQ of 20 pg/mL. Multiple surrogate matrices were screened for contamination and SigMatrix (Sigma-Aldrich), which is prepared from recombinant HSA, was observed to be free of contamination.
Following these changes, the ionKey/MS thyroglobulin assay was validated across 10 batches. Intra-assay (20x1) and inter-assay (2x10) imprecision and inaccuracy in calibrators and controls spanning from 20 pg/mL to 2000 pg/mL were all less than 15%, except at the 20 pg/mL LLOQ (19.6% CV). Additionally, the 3-sigma LOD was validated to be 5.7 pg/mL. The ionKey/MS assay showed excellent agreement with a commercially-available immunoassay in 50 autoantibody-negative samples (Deming slope = 0.998, R = 0.9914) and with the high-flow LC-MS/MS assay in 74 autoantibody-positive samples (Deming slope = 0.975, R = 0.9956). Finally, 57 autoantibody-positive and 53 autoantibody-negative samples previously measuring below the LLOQ of the high-flow assay (<0.2 ng/mL) and immunoassay (<0.1 ng/mL), respectively, were measured on the ionKey/MS assay. In both cohorts, over 70% of samples had detectable amounts of thyroglobulin (>5.7 pg/mL) and over 40% had quantifiable amounts (>20 pg/mL).
Conclusions & Discussion
Measurement of thyroglobulin on the ionKey/MS platform using microliter/minute flow-rates improved the analytical sensitivity of LC-MS/MS measurements by an order of magnitude. Feasibility to use this platform routinely was achieved by inclusion of a large inner diameter protein scavenging column up stream of the trapping and analytical columns, which dramatically improved system ruggedness. Subsequent detection and quantification of "sub-clinical" levels of Thyroglobulin in patients with levels below the reporting limits of current testing platforms, particularly in autoantibody positive patients, indicates this method has the potential to improve the clinical sensitivity of LC-MS/MS measurements - though further work is needed to substantiate this claim.
References & Acknowledgements:
|Salary||yes||Laboratory Corporation of America|
|Stock||yes||Laboratory Corporation of America|
IP Royalty: no
|Planning to mention or discuss specific products or technology of the company(ies) listed above:||