Comparative real-world outcomes of CD19-directed CAR T-cell therapies in large B-cell lymphoma

Abstract

Although 3 commercial CD19-targeted chimeric antigen receptor (CAR) T-cell therapies are available for large B-cell lymphomas (LBCLs), no randomized clinical trials have compared their efficacy and safety. In this retrospective multicenter cohort study, we evaluated real-world clinical outcomes of patients with relapsed/refractory LBCL treated with axicabtagene ciloleucel (axi-cel), tisagenlecleucel (tisa-cel), or lisocabtagene maraleucel (liso-cel). Between April 2016 and July 2024, 624 patients received CD19-targeted CAR T-cell therapies (344 axi-cel, 142 tisa-cel, and 138 liso-cel). At a median follow-up of 20.9 months, the respective estimated 2-year progression-free survival (PFS) and overall survival (OS) rates were 46% and 63% for axi-cel, 30% and 45% for tisa-cel, and 45% and 58% for liso-cel. After adjusting for potential confounders in multivariate analyses, tisa-cel was associated with inferior PFS and OS compared to axi-cel. No significant survival differences were found between liso-cel and axi-cel. Propensity score and subanalyses of patients treated in the second-line vs third-line or later settings yielded similar outcomes. Compared to axi-cel, the objective response rate at 100 days was higher for liso-cel and lower for tisa-cel. Rates of cytokine release syndrome, immune effector cell-associated neurotoxicity syndrome, and immune effector cell-associated hematotoxicity, and febrile neutropenia were significantly higher with axi-cel. However, no significant differences in the cumulative incidence of infections or nonrelapse mortality were found. Axi-cel was associated with faster vein-to-vein time (axi-cel, 35 days; tisa-cel, 43 days; liso-cel, 41 days) and fewer out-of-specification products (axi-cel, 2%; tisa-cel, 4%; liso-cel, 11%). These results provide insights into potential differential outcomes depending on product selection.

Graphical abstract

Figure 1.

Comparison of survival outcomes. Kaplan-Meier estimates of PFS (A) and OS (B) for patients treated with axi-cel, tisa-cel, or liso-cel. Differences in survival were assessed using the log-rank test. Median OS and PFS in months, HRs from the univariable Cox regression model, and 2-year survival probabilities are indicated on the graph.

Figure 2.

Comparison of survival outcomes after adjustment for confounders. (A) MVAs of factors potentially impacting survival. (B) PSMW sensitivity analysis for patients treated with axi-cel and liso-cel, adjusting for factors that may have influenced product selection. CR, complete response; DLBCL, diffuse large B-cell lymphoma; HGBL, high-grade B-cell lymphoma; LD, lymphodepletion; NOS, not otherwise specified; PD, progressive disease; PR, partial response; SD, stable disease.

Figure 3.

Survival outcomes for patients treated in the second-line vs third-line or later settings. (A-B) Kaplan-Meier estimates of PFS (A) and OS (B) for patients treated with axi-cel, tisa-cel, or liso-cel in the third-line or later setting. (C-D) Kaplan-Meier estimates of PFS (C) and OS (D) for patients treated in the second-line setting. Differences in survival were assessed using the log-rank test. Median OS and PFS in months and HRs of the univariable Cox regression model are superimposed. HRs after adjusting for confounders in MVAs and 1- or 2-year survival probabilities are shown on the graph. MVAs could not be performed for OS in the second-line setting due to insufficient events.

Figure 4.

Comparison of treatment response. (A) Distribution of best response at 100 days after CAR T-cell infusion. ORs from logistic regression comparing ORRs in tisa-cel and liso-cel to axi-cel are indicated. (B) Kaplan-Meier estimates of the DOR. Differences in DOR were evaluated using the log-rank test, and median DOR (in months) is shown. CR, complete response; PD, progressive disease; PR, partial response; SD, stable disease.

Figure 5.

Comparison of toxicities. (A-C) Grades of CRS (A), ICANS (B), and ICAHT (C). (D) Proportion of patients who had neutropenic fever (214 patients had missing data for neutropenic fever). Comparisons in panels A-D were performed using Pearson χ² test. (E) Cumulative incidence of infections (viral, bacterial, or fungal) from lymphodepletion to 30 days and 1 year after infusion. (F) Cumulative incidence of NRM, with incidence at 3 months and 2 years indicated. Comparisons in panels E-F were performed using Gray test.

Figure 6.

Resources for managing toxicities. Proportion of patients requiring steroids (A), tocilizumab (B), anakinra (C), or admission to ICU (D). Of note, data on ICU admission were missing for 20 patients. Comparisons were performed using Pearson χ² test. ICU, intensive care unit.