Head-to-head longitudinal comparison of two mass spectrometry-based methods for monitoring monoclonal proteins in multiple myeloma: Analytical and clinical insights from the GEM-CESAR trial

Abstract

Abstract Background: Mass spectrometry (MS) is increasingly used to detect monoclonal proteins (MPs) in patients with monoclonal gammopathies. Two major intact light chain approaches—MALDI-TOF (Exent) and LC-Q/TOF—have shown superior sensitivity over conventional methods. However, direct comparisons between them are lacking, limiting their harmonized clinical application. Methods: We analyzed 55 serum samples from 18 patients with high-risk smoldering multiple myeloma (SMM) enrolled in the GEM-CESAR trial. Results from 25 patients will be presented at the congress. Samples were collected at diagnosis, post-induction, post-autologous stem cell transplantation (ASCT), post-consolidation, and after 2 years of maintenance (M2). All had been previously analyzed with Exent. For this study, the same samples were reanalyzed using an LC-Q/TOF workflow (LC/MS), involving affinity purification of serum immunoglobulins, liberation of intact light chains, and untargeted high-resolution detection. Measurable residual disease (MRD) in bone marrow (BM) samples was assessed using next-generation flow (NGF) following the recommendations of the IMWG. Results: At diagnosis, the isotype identified by both MS techniques was concordant in all but three patients: one showed IgA kappa by Exent but only kappa by LC/MS (with matching light chain masses); another showed two IgA Kappa peaks by Exent but only one by LC/MS; and a third showed two IgA lambda peaks by Exent versus one by LC/MS with evidence of glycosylation. The latter patient remained positive through M2, with both IgA lambda peaks consistently observed by Exent. Among the 55 paired serum samples, 40 (72.7%) showed concordant results: 22 were positive and 18 negative by both methods. Discordant results (n=15, 27.3%) were primarily due to LC/MS detecting residual disease not identified by Exent. LC/MS detected 36 positive samples versus 23 by Exent. Concordance varied by treatment phase: post-induction (12/14), post-ASCT (11/15), post-consolidation (8/15), and M2 (9/11); discordances were most frequent post-consolidation and predominantly Exent− / LC/MS+. The proportion of double-negative samples increased with treatment, reaching 73% at M2. Despite the limited sample size per timepoint and use of biochemical progression as a censoring event, both MS techniques stratified progression-free survival (PFS) across all phases, reaching statistical significance at M2 (Exent: p=0.0016, HR 0.08; LC/MS: p=0.007, HR 0.13). Overall, the results from both methodologies demonstrated statistically significant prognostic value for PFS: Exent, p=0.0112, HR 0.41 and LC/MS, p=0.032, HR 0.37. Combined analysis demonstrated significantly longer PFS in double-negative cases (mPFS not reached) compared to double-positive (mPFS 3.57 years) or Exent− / LC/MS+ (mPFS 4.25 years). Compared with MRD assessment in BM by NGF, Exent showed 74.6% concordance and LC/MS 69.1%. Discordances between Exent and NGF were bidirectional (6 Exent− / NGF+ and 8 Exent+ / NGF−), while those between LC/MS and NGF were predominantly LC/MS+ / NGF− (16; 29.1%). Analysis of PFS based on combined results from MS and NGF showed that patients negative in both serum (by either MS method) and BM had the best prognosis, with mPFS not reached. In contrast, patients positive in serum (by either MS method), BM, or both had significantly shorter PFS. Conclusions: This study provides the first longitudinal comparison of two MS-based techniques in MM. Both methods proved valuable for disease monitoring: LC/MS demonstrated greater sensitivity, while Exent showed slightly higher concordance with NGF. The discrepancies observed between the two MS methods may be partly attributable to the fact that the samples were analyzed at substantially different times and subjected to varying numbers of freeze-thaw cycles. Importantly, combining both MS techniques with each other or with NGF improved prognostic stratification, highlighting the complementary role of these approaches and supporting ongoing efforts to standardize MS-based monitoring in monoclonal gammopathies.