Real-world treatment patterns and survival outcomes in second and third line settings in large B-cell lymphoma (LBCL)

Abstract Introduction: Despite the proven efficacy of chimeric antigen receptor T-cell (CAR T) therapy for patients with relapsed or refractory (R/R) LBCL in clinical trials leading to regulatory approval and guideline recommendations, real-world uptake of CAR T in eligible patients remains low. The real-world outcomes of patients who receive CAR T in second line (2L) compared to those who receive alternative therapies, despite being eligible for CAR T, is not well elucidated. Additionally, for patients who have CAR T delayed until third line (3L), it is unclear if this delay affects their long-term outcomes. This study examines real-world treatment patterns and outcomes for patients with LBCL in the US across 2L and 3L settings and within the respective time periods of CAR T indication expansion. Methods: This retrospective study used the US-based Flatiron Health Research Database. Patients were eligible if they had chart confirmed pathological diagnosis of LBCL after Jan 1, 2011 and were considered R/R to first line (1L) therapy, and did not receive hematopoietic stem cell transplant (HSCT) or a clinical study drug before or as 1L therapy (data cutoff date: Mar 31, 2025). We analyzed treatment patterns and outcomes in 2L and 3L cohorts with treatment initiation between Jan 2022 and Dec 2024, when CAR T was available as a 2L treatment. Additional 2L and 3L cohorts were analyzed for historical context with treatment initiating between Jan 2018 and Dec 2021, when CAR T was available as a 3L option. Patients were included in 2L cohorts if they initiated an NCCN guideline-eligible 2L therapy within 12 months (mos) of ending a 1L Chemo+anti-CD20 treatment. Patients were included in 3L cohorts if both 2L and 3L treatments were initiated within the respective time period. For each cohort, treatment patterns were descriptively analyzed and real-world overall survival (rwOS) and real-world time to next treatment (rwTTNT) from the respective index dates (2L or 3L initiation) were assessed. Cox adjusted and unadjusted Kaplan-Meier analyses were conducted for rwOS and rwTTNT. Results: Among 481 patients who initiated 2L from 2018-21 (before 2L CAR T was approved), the largest subset (38.3%, n=184) received rituximab in combination with ifosfamide, carboplatin, etoposide (R-ICE) or dexamethasone, ara-C, and cisplatin (R-DHAP), of whom 20.7% (38/184) subsequently received HSCT. Among 365 patients who initiated 2L from 2022-24, only 41 (11.2%) received R-ICE/DHAP, of whom 7.3% (3/41) subsequently received HSCT. Overall from 2022-24, 15.6% (57/365) of patients received 2L CAR T and had the longest rwTTNT of all treatment groups, with a median of 15.9 mos; CAR T was significantly associated with improved rwTTNT compared to all non-CAR T therapies (hazard ratio [HR] = 0.28; 95% CI, 0.18-0.46; p <0.001). Patients who received 2L CAR T had a median follow-up time of 20 mos (95% CI, 15-24) and a 12-mo rwOS probability (SP) of 63% (95% CI, 51-79%). Among 339 patients considered eligible for 3L CAR T in 2018-21, CAR T (n=112) had the longest median rwOS compared to non-CAR T (n=227) at 30.8 vs 13.5 mos and was significantly associated with a lower risk of mortality (HR = 0.57 [95% CI, 0.38-0.84; p=0.005]). Among 202 patients eligible for 3L CAR T in 2022-24, median follow-up was 14 mos, and a lower risk of mortality with CAR T (n=75) was similarly observed (HR = 0.36 (95% CI, 0.20-0.64; p<0.001). Patients with 3L CAR T had a numerically higher 12-mo SP than non-CAR T in both time periods (2018-21: 73% vs 53%; 2022-24: 68% vs 48%). Additionally, patients with 3L CAR T had the longest rwTTNT vs non-CAR T before (16.9 vs 5.1 mos) and after (12.9 vs 4.6 mos) CAR T indication expansion into 2L. Conclusions: With the emergence of alternative 2L therapies like CAR T, the paradigm of 2L R-ICE/DHAP followed by HSCT may be losing relevance. CAR T therapy is extending treatment-free intervals, a proxy for prolonged disease control, and is showing promise in extending rwOS over less durable alternatives. These findings also reinforce the value of CAR T in improving survival outcomes in 2L and 3L settings and the potential impact of bringing CAR T into an upfront setting.