CAR T-Cell Therapy for DLBCL with CNS Involvement: Overall Safe and Effective with Limited Efficacy in Patients with Leptomeningeal Disease

Introduction: Relapsed/refractory primary or secondary central nervous system lymphoma (r/r P/SCNSL) presents an unmet need with poor survival outcomes. Anti-CD19 chimeric antigen receptor (CAR) T cell therapy has transformed the management of diffuse large B cell lymphoma (DLBCL) and is now established as a standard of care for r/r disease. While CAR-T cell have demonstrated modest efficacy in PCNSL and SCNSL, data remains limited. Here, we conducted a comprehensive analysis of patients (pts) with DLBCL involving the CNS, treated with CAR-T cell therapy sourced from an international multi-center registry, aiming to provide detailed insights into treatment outcomes. Methods: We performed a multicenter retrospective analysis of pts with r/r de novo or transformed DLBCL with CNS involvement at any point during their disease course treated with available CD19-directed CAR T cell between 2016 and 2024 at four academic institutions in the US, and Israel. Detailed data, including the timing and localization of CNS involvement both pre- and post-CAR-T relapse, were collected. Descriptive statistics, logistic regression, Kaplan-Meier estimates, and log-rank tests were used. Leptomeningeal disease was defined as any cerebrospinal fluid involvement, including positive flow cytometry, MRI findings of leptomeningeal enhancement, or cranial nerve involvement. Results: A total of 49 pts, 44 with SCNSL and 5 pts with PCNSL were identified. Baseline characteristics included a median age of 61 years (IQR 49-71), with 49% male. The Karnofsky performance status was < 90 in 85% of pts. LDH levels were elevated in 55% of pts. Of the pts, 37% received ≥ 3 lines of therapy, and 27% had undergone prior high-dose chemotherapy and ASCT. CNS involvement at any point during the disease course included 47% of pts (22) with exclusively parenchymal (PAR) disease, 37% (18) with exclusively leptomeningeal (LEP) disease, and 16% (8) with both types of disease. Of the 22 with PAR involvement, 50% had only 1 lesion, 33% had 2 lesions and 17% ≥ 3 lesions. At apheresis, 22 pts (45%) had both CNS and systemic disease, 14 had only CNS disease and 8 had systemic disease only. For bridging, thirteen pts (33%) received CNS-directed radiation therapy, 15 (35%) were treated with high dose MTX, 12(31%) had intra thecal therapy and 10 pts (26%) received a BTK inhibitor. Pts received CAR-T cell therapy with axicabtagene ciloleucel (31%), tisagenlecleucel (29%), lisocabtagene maraleucel (24%), or point-of-care anti-CD19 therapy (16%) (Kedmi et al, Tansplant Cell Ther 2022). Among all pts, the overall response rate (ORR) was 65%, with a complete response (CR) rate of 58%. With a median follow-up of 11 months (IQR: 7, 16) the median overall survival (OS) from CAR-T was 15 months (95% CI: 9, NE), and the 1-year OS estimate was 57% (95% CI:44%, 75%). Median progression-free survival (PFS) was 7.5 months (95% CI: 3.7, NE), with a 1-year PFS estimate of 34% (95% CI: 21%, 55%). In univariable analysis, previous ASCT was associated with better OS (HR 0.22, 95% CI 0.05-0.95), while elevated LDH pre-lymphodepletion was associated with poorer OS (HR 2.96, 95%CI 1.12-7.84). Neither previous ASCT nor elevated LDH pre lymphodepletion were associated with PFS. When analyzing pt with exclusively PARE vs LEPT disease, pts with LEPT involvement had significantly inferior OS (median 8.6 vs. 19 months, p=0.038) and PFS (median 4.7 vs. 19 months, p=0.034). Grade >2 CRS developed in 45% (22) of pts, and 41% (20) experienced > grade 2 ICANS. Six pts experienced grade 4 ICANS. No patients experienced ICANS grade 5. The toxicity profile was comparable in isolated PARE vs LEPT. During the follow-up period, 27 pts relapsed or progressed following CAR-T cell. Of them, 15 pts had CNS relapse (LEPT=7, PARE=6, both LEPT/PARE=2). Active CNS involvement was present in 87% (13/15) of these pts before apheresis, with only 1 patient with CNS CR at the time of CAR-T cell infusion. Systemic relapse occurred in 11 pts without CNS involvement. Conclusions: In pts with S/PCNSL treated with CAR-T cell, we observed a safety and efficacy profile of CAR-T cell therapy comparable to that seen in recent CNSL series. Furthermore, we confirmed inferior PFS and OS in pts with isolated LEPT disease compared to those with PARE disease, suggesting the need for novel therapeutic to better control LEPT disease pre-CAR-T cell or consolidation approaches tailored specifically for pts with LEPT involvement.