Intrathecal chemotherapy for management of immune effector cell-associated neurotoxicity syndrome (ICANS): A multi-center, real-world analysis of clinical outcomes

Abstract Background Chimeric antigen receptor T-cell (CAR-T) therapy has transformed the management of large B-cell lymphoma (LBCL) but is limited by immune-related toxicities, including cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS). Initial treatment of ICANS primarily relies on high-dose corticosteroids (CS). For CS-refractory ICANS, additional agents such as anakinra are often utilized, though cost, infection risk, and morbidity limit use. Intrathecal (IT) therapy (steroids +/- chemotherapy) offers an alternative approach by targeting immune effector cells directly in the cerebrospinal fluid (CSF), thereby limiting systemic toxicity, with therapeutic efficacy demonstrated in several case series of patients (pts) with CS-refractory ICANS. We report outcomes from the first and largest multi-institutional retrospective series of CAR-T treated LBCL pts who received IT therapy for ICANS. Methods We conducted a multi-center retrospective study of adults with LBCL treated with commercially available CD19-directed CAR-T who developed ICANS. Baseline characteristics were summarized descriptively. Non-parametric tests assessed skewed variables; chi-square compared categorical variables. Progression-free survival (PFS) and overall survival (OS) were estimated using the Kaplan-Meier method, with log-rank tests comparing IT-treated and systemic-only cohorts. Results We identified 191 pts with ICANS; 51% (n=97) had grade 3-4 ICANS. Median age was 65 years (range, 23-86), 62% (n=118) were male. Axi-cel was the most frequently administered CAR-T product (83%, n=158), followed by Liso-cel (16%, n=30). CRS occurred in 93% (n=178) of pts, and 86% (n=164) received tocilizumab. ICANS developed a median of 5 days post-CAR-T (range, 0-24), and all pts received CS for management. Anakinra was used in 33% (n=62). Of the cohort, 42 pts (22%) received IT therapy, while 149 (78%) received systemic therapy alone. IT-treated pts were more likely to have grade 3-4 ICANS (69% vs 46%, p=0.01) and receive anakinra (48% vs 29%, p=0.02). IT therapy began a median of 2 days after ICANS onset (range, 0-37). 74% of pts received early IT within 5 days of onset and 26% received late IT (>5 days). IT regimens included hydrocortisone (HC) alone in 14% and HC plus methotrexate, cytarabine, or both in the remainder. ICANS recovery did not differ significantly between IT and systemic-only pts (83% vs 93%). Among grade 3-4 ICANS (n=97), recovery rates were similar (82% vs 84%). In grade 3-4 ICANS not treated with anakinra, recovery was 100% (n=11) with IT vs 89% (n=33) with systemic-only. The overall response rate (ORR) to CAR-T did not differ significantly between groups (69% vs 82%, p=0.07). In terms of safety, IT therapy was not associated with increased risk of CNS relapse or infectious complications. However, anakinra recipients had higher infection rates (53% vs 26%, p<0.01). PFS did not differ between IT and systemic-only groups overall (p=0.29) or in grade 3-4 ICANS (p=0.37). Not surprisingly, median OS was shorter in grade 3-4 vs grade 1-2 ICANS (11.1 vs 25.6 months, p<0.01). Among pts with grade 3-4 ICANS, median OS did not differ between those receiving early IT therapy vs systemic-only or late IT therapy (10 vs 12 months; p=0.41). However, in grade 3-4 ICANS pts, those treated with anakinra +/- IT therapy had a shorter median OS compared with those not receiving anakinra (7 vs 19 months; p=0.047). In a subgroup analysis of the most severe (grade 4) ICANS, IT therapy trended toward improved outcomes, though not statistically significant. In this group, IT-treated pts had lower median cumulative systemic steroid dose (907 mg vs 964 mg), fewer days with ICANS (10 vs 12.5), and fewer seizures (10% vs 25%). Conclusions To our knowledge, this is the largest retrospective series evaluating IT therapy for ICANS. Our findings demonstrate that IT therapy is feasible, well-tolerated, and may be an effective modality for ICANS management in LBCL pts treated with CAR-T. This data also reinforces the poor survival associated with grade 3-4 ICANS. The benefit of anakinra remains uncertain given increased infectious risk and worse survival in high-grade ICANS. Together, the findings of this study support IT therapy as a potential strategy for CS-refractory, high-grade ICANS and underscore the need for large-scale prospective trials to clarify the role of CSF-directed IT steroids and/or chemotherapy.