DP067 | ISOLATED CENTRAL NERVOUS SYSTEM RELAPSE IN LARGE B-CELL LYMPHOMAS AFTER CAR-T CELL INFUSION: AN ANALYSIS FROM THE ITALIAN CART-SIE STUDY

Large B-cell lymphomas (LBCL) patients with a high-risk CNS-International Prognostic Index (CNS-IPI), and/or testis, breast, kidneys, adrenal glands involvement, showed a 2-year rates of CNS disease of 10.2-12% (Schmitz, JCO 2016). Chimeric antigen receptor (CAR)-T cells improved outcome in relapsed/refractory (R/R) LBCL, but the impact on potential CNS relapse is still unknown. Aim of this analysis was to evaluate the cumulative incidence at 2-year (2-yr CI) of CNS relapse in R/R LBCL treated with CAR-T and enrolled in the observational CART-SIE study. An univariable Fine and Gray models for CNS relapse was conducted to test the impact of CNS-IPI, age, sex, histology, CAR-T product, disease status, Ann Arbor stage, bulky disease, number (N) of extranodal sites, N of previous therapies, bone marrow (BM)/kidney/testis/breast involvement. Since March 2019 to December 2024, 680 R/R LBCL were enrolled and evaluable for outcomes. Clinical characteristics were: median age 61 (IQR 53; 67); 80% DLBCL and 20% HGBCL; 74% were refractory; 76% had stage III-IV; 57% received axicabtagene ciloleucel (axi-cel) and 43% tisagenlecleucel (tisa-cel); median N of previous treatments were 2 (IQR 2;3). Regarding CNS-IPI, 30% had a low, 57% an intermediate and 13% a high risk; 16% had a BM infiltration, 22% > 2 extranodal sites, 9% kidney and 1% testicular involvement. At a median follow up of 17.93 months (IQR 6.55; 25.59), 2% (13 out of 680) patients had a CNS relapse: 3 HGBCL, 10 DLBCL (4 double expressor); 9 refractory to the last treatment; 6 received tisa-cel and 7 axi-cel; none of them had testicular or breast localization, one had a kidney involvement at diagnosis. The 2-yr CI estimates for CNS relapse was 4.1% (1.9%-7.6%); by CNS-IPI groups were 3.7% (0.88-10%) for low-risk, 4.9% (1.4%-12%) for intermediate-risk and 8% (0.92%-26%) for high-risk group, (Figure 1). In the univariable Fine and Gray model for CNS relapse analysis only age >65 years (HR 0.95, 95% CI: 0.92-0.99, p 0.008), BM involvement (HR 21.4, 95% CI 2.52-101, p 0.001) and N of extranodal sites > or = 2 (HR 3.5, CI 1.17-10.5, p 0.025) were correlated with CNS relapse. The median overall survival (OS) after occurrence of CNS relapse was 7.27 months (IQR 3.26; NA). In conclusions, in the CART-SIE study, the cumulative incidence of CNS relapse is lower than expected in a high-risk population. A longer follow-up is needed to confirm the impact of CAR-T cell in reducing the risk of CNS relapse.