AUTHORS
Ethan Thompson, PhD, Julia Piasecki, BA2, Hao Tang, PhD3, Nikita Sakhanenko, PhD4, Nathan Martin, PhD5, Shari Kaiser, PhD6, Debashree Basudhar, PhD7
AFFILIATIONS
1 Bristol Myers Squibb, Princeton, New Jersey, United States
2 Bristol Myers Squibb, Princeton, New Jersey, United States
3 Bristol Myers Squibb, Princeton, New Jersey, United States
4 Bristol Myers Squibb, Princeton, New Jersey, United States
5 Bristol Myers Squibb, Princeton, New Jersey, United States
6 Bristol Myers Squibb, Princeton, New Jersey, United States
7 Bristol Myers Squibb, Princeton, New Jersey, United States
ABSTRACT
Introduction: Idecabtagene vicleucel (ide-cel), a B-cell maturation antigen (BCMA)-directed CAR T-cell therapy, significantly improved progression-free survival (PFS) versus standard regimens (SOC) in the phase 3 KarMMa-3 trial (NCT03651128) in patients with triple-class–exposed relapsed/refractory multiple myeloma (RRMM) with 2–4 prior lines of therapy (mPFS 13.3 vs 4.4 mo; HR, 0.49; P<.001) [1]. PFS after ide-cel treatment was heterogeneous, underscoring the need for baseline biomarkers that describe disease biology, refine risk assessment, and identify features associated with durable response. We profiled baseline serum proteomics in KarMMa-3 to characterize this heterogeneity and evaluated associations with PFS, extramedullary disease (EMD), and durable response after ide-cel.
Methods: Serum proteins were profiled using the Olink Explore HT assay. Samples were collected at screening from 250 patients randomized to the ide-cel arm (Arm A) and 131 patients to the SOC arm (Arm B). Associations between protein abundance and PFS, either univariately or controlling for other risk factors, were evaluated using Cox Proportional Hazards models. Associations of protein abundance with EMD and PFS >48 mo were assessed using a Wilcoxon rank-sum test. False discovery rates (FDRs) < 0.05 were treated as significant, with the exception of the ide-cel arm PFS analysis, where a stricter threshold of FDR < .005 was used to focus analysis on the strongest associations.
Results: 253 of 5,416 serum analytes profiled at screening demonstrated significant correlation with PFS in the ide-cel arm (FDR <0.005). The majority (248 of 253) correlated with shorter PFS, whereas 5 of 253 correlated with longer PFS. sSDC1 (sCD138) was the strongest correlate. Ide-cel arm patients with lower than median sSDC1 had median PFS of 22.6 months, as compared to 7.4 months in patients with higher than median sSDC1. sBCMA also correlated with shorter PFS, consistent with previous reports. Predominant biological processes featured within the analytes correlating with shorter PFS include cell survival regulation, genomic stress, tumor burden and inflammation. B-cell–associated markers, including sCR2, sCD22, sFCER2, and sKIT, correlated with longer PFS.
Among the 253 analytes identified in the ide-cel arm, 14 also correlated with PFS in the SOC arm (FDR <0.05) with the same directionality as observed in the ide-cel arm, including sSDC1 and sBCMA. Study arm significantly correlated with improved PFS (p < 10-8) after controlling for all 14 of these proteins, indicating that ide-cel outperformed standard regimens in all identified patient segments.
Beyond sSDC1 and sBCMA, the remaining PFS-associated analytes may reflect additional biological mechanisms and provide independent prognostic value. Specifically, 248 analytes remained significantly associated with PFS after controlling for sBCMA, and 189 remained significant after controlling for sSDC1. Furthermore, these associations were not fully explained by conventional clinical risk factors, as all 253 analytes continued to correlate with PFS after adjustment for age, del(17p), t(4;14), and t(14;16). 216 of 253 PFS-associated analytes significantly correlated with the presence of EMD; however, similar to the other risk factors, all 253 analytes continued to correlate with PFS after controlling for EMD.
22 patients in the ide-cel arm achieved PFS >48 mo. Only 1 protein, MAP9, significantly correlated with PFS > 48 mo (FDR < .05).
Conclusions: In this large-scale serum proteomic analysis from KarMMa-3, baseline proteomics identified strong correlates of PFS after ide-cel treatment. The strongest was soluble SDC1, previously identified as a prognostic signal in multiple myeloma and shown to promote tumor invasion and angiogenesis [2]. The biomarkers associated with PFS in the ide-cel arm provided information beyond tumor burden and conventional clinical risk factors, indicating that serum proteomics captures disease biology not fully reflected by other measures. 14 analytes associated with PFS in both arms, suggesting that these signals are broadly prognostic of outcome in multiple myeloma. These findings support serum proteomics as a strategy to generate hypotheses for mechanisms of response and resistance and guide biomarker development in multiple myeloma.
[1] Rodríguez-Otero et al., N Engl J Med, 2023; [2] Jung, O. et al., Oncogenesis, 2016
DISCLOSURES
Ethan Thompson, PhD: BMS - Current equity holder in publicly-traded company; BMS - Current Employment