Disease response at apheresis and association with long-term outcomes following CAR-T cells for relapsed/refractory multiple myeloma (RRMM).

Abstract

7518 Background: The two approved BCMA-targeted CAR-T products, cilta-cel and ide-cel, have significant efficacy in RRMM, but are not considered curative. Initial studies in ≥4 th line RRMM required progressive disease (PD) at time of enrollment and T cell apheresis. We hypothesized that using CAR-T cells as a planned consolidation strategy (i.e. in patients (pts) with stable or responsive disease on their current therapy) may lead to lower toxicity and better long-term disease control. Methods: We conducted a retrospective review of all RRMM pts receiving commercial CAR-T cells at the University of Pennsylvania from 6/1/21 to 4/30/24, with at least 6 months of follow-up. Intent for consolidation was retroactively assigned by chart review. Kaplan-Meier methodology was used to determine PFS and OS. Results: We identified 149 pts for analysis, with a median follow-up of 14.4 months (mos). Median prior lines was 6 and 81% of pts were triple class-refractory; 46% had high-risk cytogenetics, 26% had extramedullary disease, and 17% had prior BCMA-directed therapy. Pts received either cilta-cel (54%) or ide-cel (46%), and 95% received bridging therapy. CAR-T cells were intended as planned consolidation in 51 pts (34%); of these, 36 (71%) had ≥PR at time of apheresis. For consolidation vs non-consolidation groups, this translated into greater depth of response post-CAR-T cells (≥VGPR, 86% vs. 66%, p=0.01), lower rates of ≥grade 3 CRS (1.9% vs. 9.1%, p=0.16), and longer PFS (median not reached vs. 10 mos, p=0.001), respectively. The PFS improvement was seen for both cilta-cel (p=0.01) and ide-cel (p=0.04). No differences in neurotoxicity were noted. We also performed analyses based on response at apheresis, regardless of intent (8% ≥VGPR, 23% PR, 27% stable disease (SD), and 42% PD). PFS at 20 mos was 88%, 47%, 55%, and 31% for ≥VGPR, PR, SD, and PD at apheresis, respectively (p=0.015). Median PFS of pts with at least SD (≥SD) at apheresis was not reached vs. 9.4 mos in those with PD (p= 0.003), with 20-month OS of 87% in the ≥SD group and 68% in the PD group (p=0.015). Subgroup analysis confirmed this PFS difference for both cilta-cel and ide-cel, while the OS impact was only seen for cilta-cel. On multivariate analysis, having ≥SD at apheresis was an independent predictor for PFS. No statistically significant differences in CRS and ICANS were observed based on response at apheresis. Pts with ≥SD at apheresis had higher absolute lymphocyte counts at days 7 and 14 post-CAR-T infusion than those with PD, indicating disease status at apheresis may be associated with CAR-T product quality. Conclusions: Our data suggest that disease control (≥SD) at time of T-cell collection is associated with more durable responses, supporting use of CAR-T cells as a consolidation strategy in RRMM. We cannot conclude these associations are causal. Further analyses of apheresed T cell characteristics are planned.