C011 | CAR-T CELLS SPECIFIC TOXICITIES AS MAJOR DETERMINANTS OF NON-RELAPSE MORTALITY: AN ANALYSIS FROM THE ITALIAN CART-SIE STUDY

Chimeric antigen receptor (CAR)-T cell therapy has revolutionized outcomes of patients (pts) with relapsed/refractory lymphomas, yet non-relapse mortality (NRM) has emerged as a notable concern, with ECOG, older age and infections identified as major drivers in real-life studies. Within the prospective multicenter CART‑SIE study, we analyzed incidence and predictors of NRM in lymphoma pts treated with on‑label CAR‑T cells; CAR-T eligibility relies on Italian Drug Agency criteria, which include ECOG < 2, adequate organ function and bone marrow reserve and age ≤ 75 years (yrs) for large B‑cell lymphomas. From 2019 to 2025, 932 out of 1132 enrolled pts were infused and evaluable for outcomes: at median follow-up of 17.8 months (IQR 6.4–25.4), 47 (5%) had a NRM event. CAR-T product was Axi-cel, Brexu-cel or Tisa-cel in 55%, 21% and 24% pts, respectively. Median age was 65 yrs (IQR 56-68); median number of prior lines was 2 (IQR 2-3); CAR-HEMATOTOX was high in 81% pts. NRM was early (≤28 days) and late (>28 days) in 19 (40%) and 28 (60%) pts and mainly due to infections (47%), CRS (13%) and ICANS (13%) (Figure 1). The cumulative incidence of NRM was 5.5% and 8.8% at 1- and 2-years. In univariable Fine and Gray models, a higher risk for NRM was found in case of age >60 yrs (HR 2.24 p=.01), CAR-HEMATOTOX high (HR 2.25, p=.04), G≥3 CRS (HR 4.34, p<.01), ICANS any grade (HR 4.11, p<.01) and G≥3 (HR 6.44, p<.01), infections (HR 3.65, p<.01), SPM (HR 2.32, p=.04), cardiotoxicity (HR 5.6, p<.01), and use of steroids for management of CAR-T toxicities (HR 2.8, p<.01); neither histology, CAR-T product, disease burden, number of previous lines and hematotoxicity correlated with higher NRM. In multivariable analysis, age >60 yrs (HR 2.05, p=.029), G≥3 CRS (HR 2.45, p=.007), G≥3 ICANS (HR 4.75, p<.001), use of steroids (HR 1.88, p=.036) and infections (HR 2.92, p<.001) retained significance as independent predictors of NRM. In our cohort, older age, CRS, ICANS and infections were major drivers of NRM. The Italian Drug Agency criteria have excluded baseline patients’ characteristics that might otherwise affect NRM, highlighting the predominant impact of CAR-T toxicities; as late NRM exceeds early NRM and CAR‑T is moving into earlier lines, infectious prophylaxis, granulocyte colony‑stimulating factor support and long‑term oncologic surveillance are critical to successfully address late complications, especially in pts requiring steroids for management of early toxicities.