preheader BJH 1

Matched propensity scoring suggests a higher efficacy, with increased toxicity for axicabtagene ciloleucel vs. tisagenlecleucel in the treatment of relapsed/refractory diffuse-large B-cell lymphoma

Results of a matched propensity score analysis of real-world data coming from the French DESCAR-T registry suggest that axicabtagene ciloleucel (axi-cel) induces a higher rate of (complete) responses than tisagenlecleucel (tisa-cel) in the treatment of patients with relapsed/refractory diffuse large B-cell lymphoma (r/r DLBCL) who previously received at least 2 lines of treatment. However, the analysis also revealed that axi-cel was associated with a higher incidence of CAR-T cell related toxicity compared to tisa-cel.


In recent years, both axi-cel and tisa-cel were shown to induce durable responses in patients with r/r DLBCL. In fact, the pivotal phase II ZUMA-1 trial demonstrated an objective response rate (ORR) of 83% for axi-cel, with a complete response rate of 58%. Corresponding estimates for tisa-cel in the JULIET study were 52% and 40%. However, in the absence of head-to-head trials, it is difficult to assess the comparative efficacy and safety of these two CAR-T cell therapies. To facilitate an indirect comparison of both therapies, propensity score matching (PSM), a widely used technique to reduce confounding biases in observational studies, can be used. At ASH 2021, Bachy and colleagues already presented the results of a PSM analysis comparing the outcome of r/r DLBCL patients in the French DESCAR-T registry who were treated with either axi-cel or tisa-cel. In this analysis, axi-cel was found to be associated with a longer prolonged progression-free survival (PFS) compared to tisa-cel, but this came at the cost of a more toxicity. However, follow-up in this analysis was short and as a result, no overall survival (OS) data were reported. During the 2022 annual EHA meeting, results were presented of a second PSM analysis of DESCAR-T data with a longer follow-up and a higher number of patients treated with axi-cel or tisa-cel.


All patients treated with either axi-cel or tisa-cel between July 1st 2018 and October 1st 2021 and who were included in the DESCAR-T registry were eligible for the presented analysis. PSM was used to create a balanced covariate distribution between a cohort of patients treated with tisa-cel and a cohort of patients treated with axi-cel. For this indirect comparison, an exhaustive list of covariates was used including age, sex, LDH level, C reactive protein (CRP) level, time between last treatment and infusion, ECOG performance status (PS), Ann Arbor stage, the number of prior treatment lines, the use and the response to bridging therapy, prior stem cell transplant (SCT) either autologous or allogeneic, bulk assessed at lymphodepletion, treating center and the histological diagnosis.


The presented analysis included a total of 809 r/r DLBCL patients who received at least 2 prior lines of therapy and were treated with either tisa-cel or axi-cel in one of the 25 French centers involved in DESCAR-T. Of these patients, 60 were not infused due to progression or death, while another 20 patients were not infused for other reasons. As such, this leaves 729 patients who proceeded to lymphodepletion and CAR-T infusion. Using the PSM algorithm, a matched population of 209 patients treated with tisa-cel and 209 patients treated with axi-cel was created. The best overall response rate (ORR) was 66% with tisa-cel and reached 80% with axi-cel (p< 0.001). The corresponding best complete response (CR) rates were 42% and 60% with tisa-cel and with axi-cel, respectively (p< 0.001). At the 1-year landmark, 33% of tisa-cel treated patients was alive and free of progression while this was the case for 47% of patients treated with axi-cel (HR[95%CI]: 1.65 [1.26-2.18]; p= 0.0003). Corresponding one-year OS rates reached 49% with tisa-cel and 63% with axi-cel (HR[95%CI]: 1.58[1.13-2.21]; p= 0.0072). These findings were subsequently also confirmed using an Inverse Probability Treatment Weighting analysis.

In terms of safety, patients treated with axi-cel were more likely to experience grade 1 or 2 cytokine release syndrome (CRS; p= 0.004). However, no significant difference was seen in terms of grade ≥3 CRS (9% with tisa-cel vs. 5% with axi-cel; p= 0.130). Patients treated with axi-cel were also more likely to experience immune effector cell-associated neurotoxicity syndrome (ICANS; all grade: 48% vs. 22%). In total, 29 patients (14%) treated with axi-cel experienced grade ≥3 ICANS, while this was only the case for 6 patients (3%) treated with tisa-cel.


The results of this real-world matched-comparison study suggest that axi-cel has a higher efficacy than tisa-cel in the third or further line treatment of patients with r/r DLBCL. However, this higher efficacy does come at the price of a higher incidence of all-grade CRS and ICANS (all-grade and grade ≥3).


Bachy E, et al. A matched comparison of tisagenlecleucel and axicabtageneciloleucel car t cells in relapsed or refractory diffuse large b-cell lymphoma: A real-life Lysa study from the French DESCAR-T registry. Presented at EHA 2022; Abstract S260

Speaker Emmanuel Bachy

Emmanuel Bachy

Emmanuel Bachy, MD, PhD, Hospices Civils de Lyon, France


See: Keyslides


Back to Top