Incidence and clinical predictors of cranial nerve palsies (CNPs) post ciltacabtagene-autoleucel (cilta-cel) in patients with relapsed or refractory multiple myeloma (RRMM).

Abstract

e19508 Background: Cilta-cel was FDA-approved for RRMM patients who have received 4 lines of therapy (LOT) including a PI, an IMID and a CD38 monoclonal antibody or 1 prior LOT, and are refractory to lenalidomide, based on the Cartitude 1 and 4 studies. (Martin et al., 2023, San Miguel et al., 2023) Neurotoxicity post cilta-cel is reported in up to 20% of patients, including CNPs. Factors associated with CNPs post cilta-cel have not been described. We aim to evaluate the incidence and clinical predictors of CNP after cilta-cel. Methods: We retrospectively reviewed all patients who received cilta-cel and developed CNPs (n=14) as well as a control cohort without CNP. In addition to reviewing demographics, known indicators of tumor burden and biology, we examined sequential absolute lymphocyte count measurements (ALC) from day of cilta-cel infusion to day +30 and compared to a control cohort (n=38), matched by age, gender, presence of extra-medullary or high-risk disease, high marrow burden (≥50%), and ferritin >400 at time of lymphodepletion. Max ALC, days to max ALC, max ALC slope (change in ALC/time) and ALC D7 to max ALC slope (Max ALC - Day +7 ALC/ Day max ALC -7) were calculated for all patients with CNP and matched controls. Results: Between May 2022 to October 2024, 140 patients received cilta-cel at our institution and 14 (10%) developed CNPs, at a median time of 17.5 days (range, 14-32) from infusion day. Patients with CNP had a median age of 67.5, 71.4% were males, 21.4% had high risk disease, 28.6 % had extra-medullary disease, and had received a median of 3 prior LOT. None had high marrow burden. Six patients had >1 CN involved, and CN 7 (n=11) was the most common. Patients were treated with steroids +/- IVIG. CNPs resolved in 71.4% of cases, within a median 61 days of its onset(range, 23-201). Patients who developed CNPs were older, had fewer prior LOT, higher max C-reactive protein value and more infections, when compared to control cohort ( p < 0.05 for all). There were no differences in incidence of CRS or ICANs or day 30 response. The median time to max ALC was 12 days (range, 10-30) in both cohorts, but max ALC was higher in the CNP cohort when compared to matched control cohort (4.98 vs 2.36, p = 0.011). Similarly, the max ALC slope and ALCD7 to max ALC slope were higher in the CNP cohort vs the matched control cohort. (2.92 vs 1.36, p =0.009 and 1.09 vs 0.43, p =0.011, respectively). Conclusions: This study noted an incidence of CNP of 10%. Patients who developed CNP were less heavily pretreated and had higher max ALC count and rate of increase in ALC than controls. Validation of these findings in a prospective cohort and primary prevention with a short course of dexamethasone in patients with elevated or rapidly rising ALC could be a consideration to mitigate this complication.