A Real-World Weighted Comparison of Tisagenlecleucel and Axicabtagene Ciloleucel CAR T Cells in Relapsed or Refractory Diffuse Large B Cell Lymphoma Aged 75 Years and Older : A Lysa Study from the Descar-T Registry

Background:

CD19 CAR T cells have transformed the therapeutic landscape of relapse or refractory large B cell lymphoma (R/R LBCL). A first analysis of the French DESCAR-T registry (NCT04328298), on a population treated indiscriminately with axicabtagene ciloleucel (axi-cel) and tisagenlecleucel (tisa-cel), showed similar efficacy and survival in patients aged 75 and over (pts ≥75 yo) compared to younger patients (Guffroy, EHA 2024). However, non-relapse mortality (NRM) was higher in this group of elderly patients. Matched comparison of axi-cel and tisa-cel for 809 R/R adults LBCL in the 3d line setting has demonstrated a higher efficacy and toxicity of axi-cel over tisa-cel. Axi-cel superiority was observed regardless of age, although sub-groups analyses were carried out on pts ≥70 yo (Bachy, Nat Medicine 2022). Given the higher NRM observed in pts ≥75 yo, we postulated that the comparison of these 2 products in this group of age could lead to different results.

Aims:

To compare outcomes of R/R LBCL pts ≥75 yo treated in 3rd line or more with axi-cel versus tisa-cel.

Methods:

We retrospectively analyzed pts ≥75 yo registered in the French DESCAR-T registry infused with axi-cel or tisa-cel, and compared the outcomes using a stabilized weight (SW) method entering 13 baseline covariates. The aim of propensity score weighting is to balance covariates between axi and tisa-cel to account for all possible measured confounding variables. Four patients over 80 yo, treated with tisa-cel were removed from the analysis because they obtained extreme propensity scores. The primary endpoint was OS. The secondary endpoints were PFS, DOR, best ORR and CRR (Lugano 2014, local assessment), grade ≥3 CRS, ICANS and infection rates and non-relapse mortality (NRM) defined as patients who died of causes unrelated to lymphoma relapse/progression (death of unknown origin were excluded). All time-to-event analyses used time of CAR T cell infusion as the origin.

Results:

Between April 2018 and September 2023, 125 consecutive pts ≥75 yo (median age 76 yo) were infused with a CD19 CAR T cells commercial product for a R/R LBCL after at least 2 lines and registered in the DESCAR-T registry. Stabilized-IPTW defined 2 groups of 53 pts treated with tisa-cel and 67 pts treated with axi-cel, balanced in term of age, gender, LBCL subset, number of prior lines, age-adjusted International Prognostic Index, HCT-CI score, response to the bridging therapy, CRP and ferritin level. Standardized mean differences between both groups for all weighting variables were <0.15.

Median follow-up was 12.3 months (95% CI, 7.4-23.2 months) in the axi-cel group and 17.6 months (95% CI, 8.5-23.8 months) in the tisa-cel group. Best ORR/CRR was significantly higher at 77.0%/68.0% in the axi-cel group versus 54.1/45.6% in the tisa-cel group (p=0.008 and 0.014, respectively). Median DOR was not different in the two groups (14.8 months for axi-cel versus 10.0 months for tisa-cel, p=0.67). The estimated median OS was 21.4 months (95% CI; 8.2 - NR) with axi-cel and 19.8 months (95% CI; 7.9-NR) with tisa-cel (p=0.97) and the estimated median PFS was 9.2 months (95% CI; 6.0 – 18.5) with axi-cel and 4.1 months (95% CI; 1.0 – 11.3) with tisa-cel (p=0.19). Toxicity was higher in the axi-cel group versus tisa-cel with grade ≥ 3 CRS, ICANS and bacterial infection at 5.11 %, 10.91% and 16.6% versus 2.27%, 1.15% and 2.1%, with p-values at 0.032, <0.001 and 0.018, respectively. This higher toxicity resulted in a higher rate of ICU admittance with axi-cel in 45.4% of infused patients compared to 12.8% with tisa-cel (p=0.040).

NRM was not different between the 2 groups, observed in 19.7% in the axi-cel group compared to 12.3% in the tisa-cel group (p=0.0719) with no differences in terms of early NRM (death before day 28 post-infusion) occurring in 2.04% in the axi-cel group and 2.4% in the tisa-cel group (p=0.58). Infection was the main contributor of NRM in each group, representing 51.4% and 68.3% of non-relapse deaths in axi-cel and tisa-cel group, respectively.

Conclusion:

Although our patient sample was too limited to derive definite conclusions, our findings confirm a higher efficacy but also higher toxicity of axi-cel compared to tisa-cel in LBCL pts ≥75 yo, which however did not translate into PFS, OS or NRM significant differences. Our study supports the clinical use of both products in elderly patients, with patient selection for each product based on the trade-off between toxicity and efficacy.