Comparison of two free light chain assays: performance of the free light chain ratio as risk factor for MGUS progression

Abstract

Abstract Monoclonal gammopathy of undetermined significance (MGUS) is the most common plasma cell dyscrasia, present in ~3% of the population over 50yo. MGUS carries a 1% per year risk of malignant transformation for the rest of the affected individual’s life; long-term followup is recommended for all patients with MGUS. Different risk stratification models exist to evaluate potential for progression to malignancy. The Mayo risk model, published in 2005, uses the presence of 3 risk factors (abnormal free light chain (FLC) kappa/lambda ratio (KLr), M-spike ≥1.5 g/dL, non-IgG isotype) to confer a 3% per year risk of progression instead of the average 1%. For many years, the only FDA-approved measurement of FLC was the FreeLite assay (Binding Site), which uses polyclonal sheep antibodies. Recently, other FLC assays have become available in the US, including the Sebia FLC ELISA, which employs rabbit polyclonal antibodies. The assays have similar reference intervals (RI), however differences in analytical performance between the FreeLite and Sebia were previously reported. The objective of this study was to determine the performance of the Sebia FLC assay in the Mayo risk stratification model of MGUS patients for progression in comparison to FreeLite. Cryopreserved serum samples from a cohort of 923 MGUS patients with available long-term clinical followup (median follow-up time: 7.7years, range: 0.0 to 46.9) was used to measure FLC using Sebia ELISA on a DS2 automated platform (Dynex) and FreeLite on a Siemens BNII nephelometer. Passing-Bablok (P-B) regression and Spearman correlation were used to compare the KLr between assays. Cox proportional hazards models for progression to multiple myeloma or a related plasma cell malignancy were used to estimate the prognostic effect and ability to discriminate low and high risk of abnormal KLr for each assay. Analyses were carried out in R software. Correlation between Sebia and FreeLite KLr had a Spearman r = 0.796; P-B fit: Sebia KLr = 0.825*(FreeLite KLr) – 0.048, 95% CI of slope: 0.810-0.839). With FreeLite, 32.9% of patients had a KLr outside RI for the method. For Sebia, 48.0% of patients had a KLr outside the method specific RI. Despite the individual samples analytical disagreement, both assays showed similar clinical performance for risk stratification: Sebia abnormal KLr was associated with a hazard ratio (HR) of 2.77 (95%CI: 1.79-4.28 c-statistic=0.6505) for progression, while FreeLite abnormal KLr was associated with a HR of 3.31 (95%CI: 2.20-5.00, c-statistic=0.6529), and the difference in c-statistics was not statistically significant (p=0.93). The risk assessment is performed only once per patient, at the initial diagnosis of MGUS. The similarity in clinical performance between FreeLite and Sebia assays support the use of either method as a predictor for risk of progression for MGUS patients.