Background: Axicabtagene ciloleucel (axi-cel) and tisagenlecleucel (tisa-cel) have both demonstrated impressive clinical activity in relapsed/refractory (R/R) diffuse large B-cell lymphoma (DLBCL). In a previous propensity score matching (PSM) analysis (Bachy et al., ASH 2021) allowing for balanced comparison between axi-cel and tisa-cel outcomes, we reported on a prolonged progression-free survival (PFS) but a higher toxicity associated with axi-cel compared with tisa-cel. No OS difference was observed but the follow-up was short (6 months). Aims: We aim at reporting on PSM analysis with longer follow-up and with additionnal patients treated with axi-cel or tisa-cel. Methods: All patients treated in France with axi-cel or tisa-cel from the 1st July 2018 to the 1st October 2021 and included in the DESCAR-T registry were considered. Propensity score matching (PSM) was used to create a balanced covariate distribution between a cohort of patients treated with tisa-cel and a cohort of patients treated with axi-cel. An exhaustive list of covariates was used for PSM: age, sex, LDH level, C reactive protein (CRP), time between last treatment and infusion, Eastern Cooperative Oncology Group (ECOG) performance status (PS), Ann Arbor stage, number of prior lines of treatment before CAR-T, bridging and response to bridging, prior stem cell transplant (SCT) either autologous or allogeneic, bulk assessed at lymphodepletion, centre, histological diagnosis. PSM was performed considering a 1:1 matching without replacement and with optimal matching applying a calliper width of the propensity score set at 0.1. Inverse probability of treatment weighting (IPTW) was used as another approach to further validate PSM analysis. Results: 809 patients from 25 French centres with R/R DLBCL after at least 2 lines of previous therapy had a commercial CAR-T order with axi-cel or tisa-cel and were registered in DESCAR-T. Out of 809 patients with a CAR-T order, 60 were not infused due to progression or death between leukapheresis and lymphodepletion and 20 did not proceed to lymphodepletion for other reasons. Finally, 729 proceeded to lymphodepletion and CAR-T infusion. In the 1:1 matched population (N=418, 209 patients treated with tisa-cel and 209 patients treated with axi-cel), the best ORR/CRR was 66/42% versus 80/60% for patients treated with tisa-cel compared with axi-cel, respectively (P<0.001 for both ORR and CRR comparisons). After a median FU of 11.7 months (95% CI, 10.5-12.0 months) the 1-yr PFS was 33% for tisa-cel and 47% for axi-cel (HR=1.65, 95% CI 1.26-2.18, P=0.0003). OS was also significantly longer following axi-cel infusion than following tisa-cel infusion (1-yr OS 63% versus 49%; HR=1.58, 95% CI, 1.13-2.21; P=0.0072). Similar findings were found using IPTW statistical approach. Grade 1-2 CRS were significantly more frequent with axi-cel than tisa-cel (P=0.004) but no significant difference was observed for grade 3 or more CRS (9% versus 5% for tisa-cel and axi-cel respectively, P=0.130). Regarding ICANS, both all grades and severe (i.e. grade ≥3) ICANS were significantly more frequent with axi-cel than tisa-cel. 48% of patients experienced ICANS after axi-cel infusion compared to 22% after tisa-cel infusion. 29 patients (14%) presented a grade ≥3 ICANS with axi-cel compared with 6 (3%) only with tisa-cel. Image:Summary/Conclusion: In conclusion, our matched-comparison study supports a higher efficacy but also a higher toxicity of axi-cel compared with tisa-cel in third or more treatment line for R/R DLBCL.