DP073 | A MULTI-CENTER REAL-LIFE ANALYSIS ON PATIENTS WITH LARGE B-CELL LYMPHOMA AFTER FAILURE OF CAR-T CELLS THERAPY.

Background CAR-T has changed treatment of relapsed/refractory (R/R) Large B Cell Lymphoma (LBCL), yet nearly 60% of patients (pts) relapse. While several options exist after CAR-T failure, their optimal use and accessibility remain unclear. We evaluate treatment strategies and outcomes in LBCL R/R after CAR-T. Methods We retrospectively analyzed LBCL pts treated post-CAR-T across 16 Fondazione Italiana Linfomi (FIL) centers. All received commercial CAR-T (axi-cel or tisa-cel) by Dec 31, 2023, with EBMT informed consent. Results Of 255 CAR-T recipients, 125 (65 axi-cel, 60 tisa-cel) relapsed or were refractory. Most had advanced stage (75%) and IPI≥3 (58%) disease. CRS occurred in 84% (12% severe), ICANS in 22% (9% severe). Relapse/progression was histologically confirmed in 50%, with CD19 loss in 51%. At relapse, 78% (n=97) received curative-intent therapy. Bispecific antibodies (BiAB) were planned in 65%, but only 19% eventually accessed them due to limited availability. Other therapies: polatuzumab-based regimens (27%), chemotherapy (22%), immunomodulatory agents (17%), radiotherapy (5%) and experimental drugs (10%). Median overall survival (OS) was 6.9 months in treated pts vs. 0.7 with palliative care (Fig1). BiAB (12 months OS) and polatuzumab regimens (8.7 m) outperformed chemotherapy (3.4 m, p=0.002). PFS was longest with BiAB (6 m) vs. polatuzumab (3.3), immunomodulants (1.9) or chemotherapy (1; p<0.001). Best responses were seen with BiAB (ORR 73%, CR 50%) and polatuzumab (ORR 37%, CR 33%). At infusion, bridging response, IPI, CARHEMATOTOX, stage, CRS, and ICANS impacted outcomes, but multivariate analysis confirmed only PET positivity at day 30. At failure, ECOG>1, LDH>2×ULN, >2 extranodal sites, relapse <4 months, platelets <100×10^9/L, and high EASIX predicted poor outcome. Multivariate analysis identified LDH, extranodal sites, and early relapse as OS predictors; LDH and early relapse for PFS. These factors defined prognostic models stratifying pts into 3 groups. The PC-PI score (based on ECOG, Hb, LDH, extranodal sites, early relapse) was also validated. Pts achieving CR post-salvage (n=39) had better PFS when consolidated with allotransplant (n=18; 22 vs. 12 months, p=0.004), though OS benefit was not significant. Conclusions CAR-T failure in R/R LBCL leads to poor outcomes. BiAB show best efficacy but limited access. Prognostic markers (LDH, extranodal disease, early relapse) may guide therapy. In responders, allotransplant improves PFS.