Background: Chimeric antigen receptor T-cell therapies tisa-cel and liso-cel are effective treatments for r/r DLBCL (Schuster 2019, Abramson 2020). This study compared efficacy outcomes of tisa-cel and liso-cel in r/r DLBCL using matching-adjusted indirect comparison (MAIC). Methods: Individual patient-level data (IPD) from JULIET (tisa-cel; NCT02445248; 02/2020 datacut) were weighted to match the patient population in TRANSCEND (liso-cel; NCT02631044; 08/2019 datacut). Baseline prognostic factors available in both trials were adjusted for age, sex, histology, ECOG performance status [ECOG PS], left ventricular ejection fraction, radiologic sum of product diameters, lactate dehydrogenase, prior stem cell transplantation [SCT], use of bridging therapy, and number of and refractoriness to prior therapies, in the MAIC. Overall survival (OS), progression-free survival (PFS), complete response (CR) rate, and overall response (OR) rate were compared. Primary analyses compared infused patients in JULIET (N=106, excluding 8 without lymphodepleting chemotherapy [LDC] and 1 large cell neuroendocrine carcinoma) with efficacy-evaluable set in TRANSCEND (N=256, infused patients). A scenario analysis compared JULIET infused to TRANSCEND primary analysis set (PAS) (N=133, dose level 2, excluding those with ECOG PS 2, prior allogeneic SCT, primary mediastinal B-cell lymphoma, follicular lymphoma [FL] 3B, or transformation from indolent lymphoma besides FL). Sensitivity analyses included JULIET patients with only fludarabine-based LDC or only adjusted significantly different baseline prognostic factors. Safety outcomes were not compared because adverse event management has evolved and differed between the two trials; MAIC is unable to adjust for such differences. Results: After adjusting for differences in baseline characteristics, OS, PFS, and CR were comparable between tisa-cel infused patients and the liso-cel efficacy-evaluable set (Table). The results were consistent across all scenario and sensitivity analyses. OR rate trended higher in the TRANSCEND efficacy-evaluable set (72.7% vs. 62.9%, p=0.07) and was higher in TRANSCEND PAS than in the respectively matched JULIET infused set (74.4% vs. 60.9%, p < 0.05). Conclusions: The MAIC results indicate there is no evidence suggesting differences in OS, PFS and CR between tisa-cel and liso-cel in r/r DLBCL. Analyses using IPD from both trials and/or real-world evidence are warranted to confirm these findings.