Abstract
The success of CAR T is typically assessed by disease response and rates of immune complications like Cytokine Release Syndrome (CRS) and Immune Effector Cell-Associated Neurotoxicity Syndrome (ICANS). Given the variability in both efficacy andof toxicity among available CD19 CAR T products in relapsed/refractory large B-cell lymphoma (LBCL), traditional outcomes may not fully capture the overall clinical impact of this therapy. Composite end-points may yield novel insights into net clinical benefit and optimal therapeutic index.
To evaluate the combined contribution of efficacy and toxicity in the first 100 days post CAR T, we defined novel composite end-points: toxicity-free complete response at day 100 (tfCR100) and toxicity-free, progression free survival at day 100 (tfPFS100). Toxicity was characterized as experiencing grade ≥3 CRS or ≥3 ICANS. Specifically, tfCR100 was defined as the proportion of patients (pts) achieving a complete response (CR) at day 100 post-infusion and without toxicity; tfPFS100 was defined as the proportion of pts alive, free of lymphoma progression at day 100 post-infusion, and without toxicity. CR100wt pts were defined as CR at day 100 with prior ≥gr 3 CRS or ≥gr 3 ICANS. We compared outcomes between tfCR100 and CR100wt and patients without CR at day 100. We then compared 2-yr PFS and OS by tfCR100 overall and by product. Relapse and non-relapse mortality at 2 years were estimated using cumulative incidence function using competing risks.
We included 627 consecutive CD19 CAR T recipients with LBCL reported to the Cell Therapy Consortium registry between 4/2016-7/2024; median age was 64.5 yrs (range 18.5-86.2), 83% had ECOG 0-1, 64% were males, most had >2 lines of therapy. Products used included lisocabtagene maraleucel (liso-cel; n=83 pts), tisagenlecleucel (tisa-cel; n=207 pts) and axicabtagene ciloleucel (axi-cel; n=337 pts). 2-year PFS for 3 products were 50% (liso-cel), 30% (tisa-cel) and 42% (axi-cel; p<0.01) and respective 2-yr OS were 69% vs 47% vs 58% (p<0.01).
At day 100 post CAR T, tfPFS100 in the entire cohort was 53% (95%CI 49-57%). Product specific tfPFS100 was highest for liso-cel (76%; 95%CI 65-84) compared to 53% (95% CI 46-59) for tisa-cel and 47% for axi-cel; 95% CI 42-52; p=0.01).
By day 100, 242 pts (38.6%) achieved tfCR100, 65 pts (10.4%) experienced CR100-wt, and 320 (51%) did not achieve CR. Rates of tfCR100 were similar by age yet differed significantly by CAR T product: liso-cel 65.1%, tisa-cel 47.2% and axi-cel 32.9% (p<0.01). CR100wt was more frequent with axi-cel (17.2%) compared to tisa-cel (2%) and liso-cel (3.6%).
At a median follow-up of 22 months (range 1.1-80.7), tfCR100 pts yield 2-year OS 86% (95% 80-91) compared to 63% (95% CI 49-75) for the CR100-wt pts (p<0.01). Among patients without CR by day 100, the 2-year OS was estimated at 31% (25-36%; p<0.01). The 2-year PFS was higher in tfCR100 patients (69 vs 53%; p=0.02). Notably, within the tfCR100 group (n=242), 2 yr OS and PFS did not differ by the product (liso-cel 84% vs tisa-cel 85% vs axi-cel 87% and 67% vs 63% vs 73%; NS). To evaluate the contributors for survival, we noted that non-relapse mortality was significantly lower in tfCR100 group (5%; 95%CI 2-8) compared to 25%; 95%CI 13-37 in the CR100-wt group (p<0.01). NRM was mostly attributable to ICANS (19% vs no >gr 3 ICANS 4%; p<0.01). Remarkably, 2-year cumulative incidence of relapse did not differ between the tfCR100 group (26%) and the CR100-wt group (21%;p=0.77). Univariate and multivariate analysis for factors associated with novel endpoints will be presented at the meeting.
Our data demonstrate that about half of CAR-T recipients experience CR without serious toxicity and they yield excellent outcomes regardless of product used. However, products differ significantly in early endpoint tfPFS100 with superiority of liso-cel compared to others. NRM was high in CR100-wt patients predominantly due to ICANS. Importantly, the relapse rate in pts who achieve CR at day 100 with and without toxicity is similar, suggesting toxicity- directed steroids treatment does not enhance relapse risk. Our data reveal the emergence of novel tools that may be valuable in benchmarking cell therapies, health economic modeling, and guiding future strategies to optimize both safety and efficacy of CAR-T therapy.
