Systematic review of non-icans neurological toxicities associated with BCMA and GPRC5D-directed CAR-T cell therapy in plasma cell dyscrasia Free

Introduction

CAR T-cell therapy targeting BCMA and GPRC5D has demonstrated remarkable efficacy in relapsed/refractory Multiple Myeloma. However, there are numerous reports of these CAR T-cells being associated with novel toxicities- particularly intriguing of which is Non-ICANS Neurotoxicity (NINT). There is a broad spectrum of NINTs, ranging from cranial and peripheral neuropathy to parkinson like movement and neurocognitive toxicities (MNTs). Given the novel and atypical nature of these toxicities, their reporting from different trials and real-world datasets is often limited and non-uniform. We conducted a systematic review to describe NINTs in adults with plasma cell dyscrasias who received BCMA or GPRC5D-directed CAR T-cell therapy.

Methods

The study was registered on PROSPERO (CRD420251074357) and followed the PRISMA reporting guidelines. Two independent authors performed a systematic search using the search strategy (“Multiple Myeloma”[Mesh] OR “Smouldering Multiple Myeloma”[Mesh] OR myeloma OR “Immunoglobulin Light-chain Amyloidosis”[Mesh] OR “AL Amyloidosis” OR “Light-chain Amyloidosis” OR “primary amyloidosis”) AND (“Receptors, Chimeric Antigen”[Mesh] OR “car t” OR “Chimeric Antigen receptor therapy”) over the EMBASE, Web of Science and Pubmed databases. The primary outcome was the pooled incidence of NINTs and specific neurotoxicity adverse events. The secondary outcomes included examining the time of onset, reversibility, and outcomes of patients with NINTs. R software (version 4.4.0) was used for analysis, and the pooled incidence was reported as a percentage with 95% confidence intervals using a random-effects model.

Results

Our initial search yielded 8719 studies, which were screened for eligibility and inclusion in the study. After removing studies that reported on the same dataset, nine studies were included for analysis, including 4 retrospective analysis and 5 prospective clinical trials. Five studies had data on patients who received cilta-cel, 2 studies had data on patients who received ide-cel and 2 studies had data on patients receiving ide-cel and cilta-cel respectively. We found no studies using GPRC5D CAR T-cells associated with NINTs. The studies by Cohen et al. and Martin et al. reported on the CARTITUDE-1 trial at different time points, and hence the information was combined for analysis purposes. One study (Ellithi et al) did not report on total incidence, and just the incidence of individual NINTs, and was hence was only included in the pooled analysis for specific NINTs. The total number of patients across studies was 2919. The pooled incidence of NINTs across studies was 5.9% (95% CI 0.00-0.16; I2=95.3%, p < 0.001, n = 1423). The most common individual NINT reported was Cranial Nerve palsy with a pooled incidence of 2.34% (95% CI 0.00-0.75, I2=94.4%, p<0.001, n=2919), followed by MNTs with a pooled incidence of 1.21% (95% CI 0.00-0.04, I2= 88%, p<0.001, n=2919). The weighted mean of the median time to onset of NINT was reported in only five studies, and was 24.5 days (Range 11-387 days) . Treatment for NINTs varied, ranging from steroids, IVIg, anakinra and cyclophosphamide. Information regarding reversal (partial or complete) of NINTs was available in only five studies with a pooled incidence of 69.2% (95% CI 0.48-0.88; I2=23.5%, p-0.26, n=58), while the pooled incidence of reversal of MNTs was 57.1% (95% CI 0.23-0.88, I2=0%, p-0.44, n=16).

The pooled incidence of NINTs was significantly higher in patients receiving cilta-cel compared to ide-cel (13.6% vs 0.04% p<0.001). The pooled incidence of MNTs (3.25% vs 0.03% p=0.068) and cranial nerve abnormalities (5.88% vs 0% p<0.001) was also significantly higher in patients who received cilta-cel compared to those who received ide-cel. Seven patients (12.1%) with NINTs died during different study follow-up periods- 2 due to progressive neurotoxicity and five due to infections, possibly secondary to immunosuppression given for the NINT.

Conclusion Our systematic review demonstrated a pooled incidence of 5.9% for NINTs across studies, with adverse events being more common in patients receiving cilta-cel compared to ide-cel. These adverse events lead directly or indirectly to death in 12% of patients with NINTs. The overall reversibility of these toxicities was 69%, but was lower for MNTs at 57%. This review highlights the need for more systematic reporting of these adverse events across clinical trials and real-world studies.