Comprehensive Analysis of Pre-Treatment Biomarkers Associated with Toxicity and Durable Responses in Idecabtagene Vicleucel Therapy for Relapsed/Refractory Multiple Myeloma

Introduction. Idecabtagene vicleucel (ide-cel) has shown significant efficacy in relapsed/refractory multiple myeloma. A better understanding of factors that predispose to the development of cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) or associate with durable responses are paramount. We conducted an analysis of pre-treatment characteristics and key biomarkers in patients treated with ide-cel in standard of care.

Methods. Patients were included if they received a therapeutic dose of ide-cel and consented to a biospecimen protocol. Cytokines (GM-CSF, IL2, IL6, IL15, IFNγ, ANG1, ANG2, TNFα) were measured at apheresis, baseline (BL) prior to lymphodepleting chemotherapy, day of infusion, and weekly until day 28. Peripheral blood mononuclear cells were isolated at serial timepoints to capture CAR peak expansion. DNA was extracted and quantified, followed by qPCR to determine CAR copy numbers, with correction for batch effects, and calculation of CAR copies per µg of DNA. Pre-treatment bone marrow (BM) aspirates were evaluated for high plasma cell burden (PCB) defined as ≥50% involvement and phenotyping performed for the relative composition of T-, myeloid, and natural killer (NK) cells. CRS and ICANS were graded per ASTCT criteria. Patients were classified as having a durable response (DR) if alive without disease progression at 9 months post-treatment, while non-durable response (NDR) was defined by death or disease progression prior to 9 months.

Results. A total of 108 patients treated between May 5, 2021 and May 18, 2024 were included. Median age was 69 years (range 43-83) and the cohort included 79 (71%) non-Hispanic White, 20 (19%) non-Hispanic Black, and 9 (8%) Hispanic patients. Patients had received a median of 5 prior lines of therapy (range 3-13) and median follow-up was 12.5 months. Toxicity was similar to published data with any grade CRS in 82% (≥Gr2 20%) and any grade ICANS in 17% (all ≥ Gr2). At the time of follow-up, 51/94 (54%) patients had a DR.

The risk of developing ICANS associated with high disease burden: advanced R-ISS stage (p=0.04), triple-class refractoriness (p=0.008), PCB ≥50% (p=0.02), and higher inflammation: higher BL ferritin (p=0.004), CRP (p=0.03), IL6 (p=0.03), IL15 (p=0.003), and ANG2:ANG1 (p=0.07), especially in younger patients (p=0.04 in <70 years). Interestingly, the risk of developing higher grade CRS was higher in those patients that received a higher cell dose (median 440 vs 411x10⁶, p=0.01). As expected, higher peaks in IL6 (p=0.007) and ferritin (p=0.004) occurred in those with ≥grade 2 CRS whereas a more extensive inflammatory response occurred in those with ICANS with significantly higher peaks in IL6, IL15, IFNγ, TNFα, and ANG2:ANG1 (p<0.05). Higher peak IL6, IFNγ, TNFα, and GM-CSF (p<0.05) were more frequently noted in patients with higher peak expansion (≥ median 250x10⁶ CAR transgene copies/µg).

NDR was higher in patients with EMD (p=0.019), high PCB (p=0.019), prior BCMA exposure (p=0.019), and with higher BL ferritin (p=0.019). In pre-treatment BM aspirates, a higher proportion of monocytic myeloid-derived suppressor cells (M-MDSCs) [CD11b+CD33+CD14+HLA-DR^neg/low] were noted in those with reduced CAR-T peak expansion (p=0.0025), in particular co-expressing CD116 (p=0.0045) and iNOS (p=0.056). Higher proportions of regulatory T cells in pre-treatment BM did not appear to influence outcomes but did associate with lower peak ferritin. Pre-treatment BM associations with DR included higher pre-treatment CD4:CD8 ratios and more CD8 stem central memory cells. Higher proportions of BM CD16^dimCD107a+ NK cells were also noted in patients with DR. Lower CAR-T expansion was noted in patients with BM CD8 co-expressing multiple checkpoint ligands (LAG/PD1/TIGIT, p=0.04) and was also less in those that received alternative LD chemotherapy (median 186 vs 250k) although not statistically significant (p=0.6).

Conclusion. In a large cohort of real-world patients, we have identified that pre-treatment burden and BL inflammation pose a high risk of ICANS, whereas higher cell doses associated with higher grade CRS. Pre-treatment burden of M-MDSCs may reduce expansion and DR was linked to less disease and favorable BM profiles: higher CD4:CD8 ratios, the presence of cytotoxic NK cells, and more stem central memory CD8+ T cells.

*DKH, CLF, and MM are co-first authors

*FLL and LCP are co-last authors