Paper Details

ABSTRACT

Introduction: ZUMA-7 (Locke, NEJM 2022) and TRANSFORM (Kamdar, Lancet 2022) established axicabtagene ciloleucel (axi-cel) and lisocabtagene maraleucel (liso-cel) as second-line (2L) standard of care (SOC) for relapsed/refractory large B-cell lymphoma (LBCL). Real-world 2L populations differ from trial cohorts, as most patients receive holding (HT) and/or bridging therapy (BT), while CAR-T is arranged. Herein, we report 2L CAR-T (axi-cel and liso-cel) outcomes with a median follow-up exceeding 2.5 years, focusing on the impact of intermediary therapy.

Methods: This multicenter retrospective study across 7 U.S. academic centers analyzed patients with LBCL who received 2L CAR-T from their respective approval dates (axi-cel: 4/1/22, liso-cel: 6/24/22) through 12/30/25. CAR-T-associated toxicities were graded by ASTCT consensus criteria (Lee, TCT 2019), and investigator-assessed responses were reported per Lugano 2014 criteria. Intermediary therapies included HT and BT. HT was given between the time of documented relapse and leukapheresis, whereas BT was given between leukapheresis and lymphodepleting chemotherapy (LDC). Real-world outcomes for all treated patients were estimated using Chi-square tests, Kaplan-Meier methods, and log-rank tests. Multivariable analysis was used to evaluate predictors associated with outcomes. Cumulative incidence of non-relapse mortality (NRM) was evaluated using competing-risk analyses and compared using Gray’s test.

Results: Of 157 infused patients, 146 (93%) received axi-cel and 11 (7%) liso-cel; median follow-up was 32.5 months (95% CI 31.2-33.6). Median age was 66 years (range, 20-86), 63% were male, 11% had ECOG PS ≥2, and 63% had primary refractory disease at the time of referral. Histology was DLBCL, NOS in 68%, transformed follicular lymphoma in 17%, and high-grade B-cell lymphoma in 11%; 58% were germinal-center subtype. Stage III/IV disease was present in 75%, IPI 3-5 in 50%, and bulky disease (≥10 cm) in 19%. The median duration between the last provided dose of frontline chemotherapy and documented disease progression was 43.5 days (IQR 19.3-150.5). Median time between patient evaluation (Eval) for CAR-T and infusion was 47.5 days (IQR 36.3-67), including 17 days (IQR 9-29.2) from Eval to leukapheresis, and 28 days (IQR 26-34.8) from leukapheresis to infusion.

Overall, 79% (N=124) of patients received intermediary therapy: 38% (N=59) HT, 69% (N=108) BT, and 27% (N=43) both. Among patients who received HT, 44% were treated with chemotherapy-based regimens, 19% received radiation (RT), and 19% received steroids. Of those who received BT, 51% were treated with chemotherapy-based regimens, 23% with RT, and 18% with steroids. Polatuzumab vedotin-containing regimens were used in 8% (N=5) of HT patients and 16% (N=17) of BT patients. Median time from Eval to infusion was longer among patients receiving HT than among those who did not (69 vs 46 days). LDH at leukapheresis was highest among patients receiving BT and HT and did not significantly decline between leukapheresis and LDC regardless of whether HT, BT, or both were given (all p>0.05).

Cytokine release syndrome occurred in 89% (N=138), with grade 3-4 events in 5% (N=7). Immune effector cell-associated neurotoxicity occurred in 57% (N=89), with 30% (N=48) of patients having grade 3-4. NRM was 5% and 8% in the first and second years, respectively, predominantly due to infection, most commonly bacterial pneumonia, influenza A, and COVID-19.

In patients receiving 2L CAR-T, the best ORR and CR were 85% and 73%, respectively. At 2 years, OS was 68.2% (95% CI 60.1-75), and PFS was 54% (95% CI 45.8-61.6). Median OS and PFS were not reached. The Median OS after 2L CAR-T relapse was approximately 11 months.

In multivariable Cox models among axi-cel recipients (N=136), elevated LDH at leukapheresis was the only independent predictor of PFS (HR 2.63, 95% CI 1.19-5.80, p=0.016). Receipt of intermediary therapy was not associated with PFS (HR 1.03, 0.50-2.11) or OS (HR 1.22, 0.50-3.03), nor with cumulative incidence of relapse (Gray's p=0.14) or NRM (p=0.76).

Conclusion: At a median follow-up of 32.5 months, over half of 2L LBCL patients treated with CAR-T in the real world remain alive and in remission. However, we have identified a knowledge gap that warrants prospective clinical trials: nearly 80% of patients receive intermediary holding and/or bridging therapies in SOC practice despite a lack of prospective data in the CAR-T setting.