C074 | CRS-PSS TO IDENTIFY MANTLE CELL LYMPHOMA PATIENTS AT HIGHER RISK FOR TOXICITIES BURDEN OF CAR T CELLS: A VALIDATION ANALYSIS FROM THE ITALIAN CAR-T SIE STUDY

Chimeric antigen receptor T-cells have revolutionized treatment of B-cell lymphomas, including mantle cell lymphoma (MCL), with the limit of toxicities as cytokine release syndrome (CRS) and neurotoxicity (ICANS). Two prognostic scoring systems (PSS), CRS-PSS and ICANS-PSS, were developed to predict toxicities in lymphomas, though excluding MCL. We aimed to validate them in MCL treated with brexucabtagene autoleucel (brexu-cel). We analyzed 125 MCL patients (pts) treated with brexu-cel and enrolled in the CAR-T SIE study as of March 2025 (table1). CRS-PSS included bulky disease, platelets (PLT) <150k/µL, CRP >30 mg/L, and no bridge/bridge failure. We did not calculate ICANS-PSS (including sex, PLT <150k/µL, axi-cel product, and no bridge/bridge failure) as it would largely overlap with CRS-PSS after removing product type. Logistic regression was used to identify associations with outcomes: CRS, ICANS, steroids and tocilizumab use. CRS occurred in 96% pts, with grade ≥3 in 22%. ICANS occurred in 49% pts, with grade ≥3 in 18%. Tocilizumab and steroids were used in 85% and 70% pts respectively. CRS-PSS was available in 86% (108/125) and was high (>2) in 26% (28/108): grade ≥3 CRS occurred in 25% (7/28) of high-risk vs 17% (14/80) of low-risk pts, while grade ≥3 ICANS occurred in 25% (7/28) of high-risk vs 26% (21/80) of low-risk pts. Predictors of severe CRS included HEMATOTOX (OR=3.49, p=0.044), bone marrow disease (OR=2.83, p=0.045) and exposure to bridge therapy (OR=3.25, p=0.044), while response to bridge was protective (OR=0.31, p=0.029). Severe ICANS correlated only with exposure to bridge therapy (OR=3.97, p=0.039). While CRS-PSS was not associated with grade ≥3 CRS (OR=1.55, p=0.42) and ICANS (OR=1.69, p=0.33), correlation was seen with grade ≥2 ICANS (OR=2.56, p=0.042). To account for the possible impact of sample size, we analyzed predictors of steroids use as surrogate marker of severe toxicity. Significant variables were PLT <150k/µL (OR=0.39, p=0.017), high CRP (OR=4.28, p=0.002), HEMATOTOX (OR=3.67, p=0.008), and bridging therapy (OR=2.4, p=0.044). CRS-PSS was significantly associated with steroids use (OR=5.56, p=0.002) with 89% (25/28) of high-risk receiving steroids vs 60% (48/80) in low-risk pts. Our findings suggest that while CRS-PSS did not predict severe CRS, it could identify pts at risk of grade ≥2 ICANS and requiring steroids as a surrogate of toxicity burden. Larger studies including steroids cumulative doses could refine its clinical utility.