Outcomes and salvage strategies for large B‐cell lymphoma progressing after second‐line CAR T‐cell therapy: A DESCAR‐T study from the LYSA group

Abstract Relapse after chimeric antigen receptor (CAR) T‐cell therapy in large B‐cell lymphoma (LBCL) is associated with a dismal prognosis. Although Phase 3 trials have established CAR T‐cells as standard second‐line (2L) therapy, outcomes and optimal management after relapse in this setting remain unknown because no real‐world data are currently available. We conducted a multicenter retrospective study using the French DESCAR‐T registry, including patients with LBCL who relapsed after 2L CAR T‐cell therapy with axicabtagene ciloleucel or lisocabtagene maraleucel. The objective was to describe postrelapse treatments, outcomes, and prognostic factors. Among the 893 patients treated with 2L CAR T‐cells, 297 (33%) relapsed and were analyzed. Median time to relapse was 2.7 months, with 35% of these relapses occurring within 2 months. Among the 231 treated patients, bispecific antibody (BsAb)‐based regimens were the most common type of salvage therapy (65%). The overall response rate after relapse was 39.1%, with a complete response rate of 27.6%. Notably, despite the use of BsAb‐based therapy in 65% of patients, the median overall survival (OS2) after relapse was only 6.5 months (median progression‐free survival PFS2: 3.4 months). Patients treated with BsAb monotherapy achieved a median OS of 7.1 months. Multivariate analysis revealed that early relapse (<6 months), an Eastern Cooperative Oncology Group (ECOG) ≥ 2, and an elevated C‐reactive protein (≥30 mg/L) were independently associated with poor OS. This study represents the first real‐world analysis of outcomes after relapse following 2L CAR T‐cell therapy. Despite the more frequent incorporation of BsAbs into the therapeutic landscape, overall prognosis and treatment efficacy remain dismal, highlighting the urgent need for innovative therapeutic strategies and dedicated prospective trials in this emerging double‐refractory population.