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

Introduction: CAR-T therapy has changed relapsed/refractory (R/R) Large B Cell Lymphoma (LBCL) treatment, yet nearly 60% of patients (pts) relapse. Several therapeutic options exist post-CAR-T, but their optimal use and accessibility remain unclear. This study evaluates the incidence, treatment strategies and outcomes for LBCL pts relapsing after CAR-T. Methods: We retrospectively analyzed LBCL pts treated post-CAR-T failure across 15 centers of the Italian Lymphoma Foundation. All received commercial CAR-T cells (axicabtagene ciloleucel, axi-cel, or tisagenlecleucel, tisa-cel) by December 31, 2023. Pts provided EBMT informed consent. Results: Among 241 CAR-T recipients, 119 (61 axi-cel, 58 tisa-cel) relapsed or were refractory. Most had Diffuse Large B-cell Lymphoma (72%), advanced stage (76%) and IPI ≥ 3 (59%). CRS occurred in 83% (severe 12%), ICANS in 23% (severe 8%). Relapse/progression was histologically confirmed in 50%, with CD19 loss in 48%. At relapse, 77% (n = 92) received curative-intent therapy. Although 67% were intended for bispecific antibodies (BiAB), only 19% accessed them due to limited availability. Other therapies included polatuzumab-based regimens (27%), chemotherapy (22%), immunomodulatory agents (15%), radiotherapy (6%) and experimental drugs (11%). Outcomes were poor, with a median overall survival (OS) of 6.5 months in treated pts and 0.7 months in palliative care (Figure 1). OS was longest with BiAB (12.5 months) and polatuzumab regimens (8.7 months), outperforming chemotherapy (3.4 months, p = 0.003). Progression free survival (PFS) was also superior with BiAB (6 months) versus polatuzumab (3.9 months), Immunomodulants (2.1 months), or chemotherapy (1.2 months; p < 0.001). Best response rates were seen with BiAB (72.5% ORR, 50% CR) and polatuzumab (37% ORR, 33% CR). At CAR-T infusion, bridging therapy response, IPI, CARHEMATOTOX score, stage, CRS, and ICANS correlated with post-CAR-T outcomes, but multivariate analysis confirmed only PET positivity at day 30. At CAR-T failure, ECOG > 1, LDH > 2× ULN, > 2 extranodal sites, relapse < 4 months, platelets < 100 × 109/L, and high EASIX score predicted poor prognosis. Multivariate analysis identified LDH, extranodal sites, and early relapse as OS predictors, while LDH and relapse timing were linked to PFS. Developed prognostic models for OS and PFS stratified pts into 3 risk categories (p < 0.001, Figure 1). The PC-PI score proposed by Iacoboni was also validated (Figure 1). Pts in CR after salvage therapy had better PFS when consolidated with allogeneic stem-cell transplantation (allo-SCT; n = 16) (not reached vs. 15.9 months, p = 0.016), although OS was not significantly improved. Conclusion: Post-CAR-T relapse in LBCL is associated with poor outcomes. BiAB show the most promise but have been difficult to access. Prognostic factors such as LDH, extranodal involvement, and early relapse can guide therapy. While allo-SCT may improve PFS in selected pts, its impact on OS requires further investigation.