Blandine GUFFROY

Chimeric antigen receptor (CAR) T-cells targeting CD19 have been approved for the treatment of relapse/refractory large B-cell lymphoma. Hematotoxicity is the most frequent CAR T-cell–related adverse event. Transfusion support is a surrogate marker of severe cytopenias. Transfusion affects patients’ quality of life, presents specific toxicities, and is known to affect immunity through the so-called transfusion-related immunomodulation that may affect CAR T-cell efficacy. We analyzed data from 671 patients from the French DESCAR-T registry for whom exhaustive transfusion data were available. Overall, 401 (59.8%) and 378 (56.3%) patients received transfusion in the 6-month period before and after CAR T-cell infusion, respectively. The number of patients receiving transfusion and the mean number of transfused products increased during the 6-month period before CAR T-cell infusion, peaked during the first month after infusion (early phase), and decreased over time. Predictive factors for transfusion at the early phase were age >60 years, ECOG PS ≥2, treatment with axicabtagene ciloleucel, pre–CAR T-cell transfusions, and CAR-HEMATOTOX score ≥2. Predictive factors for late transfusion (between 1 and 6 months after infusion) were pre–CAR T-cell transfusions, CAR-HEMATOTOX score ≥2, ICANS ≥3 (for red blood cells [RBC] transfusion), and tocilizumab use (for platelets transfusion). Early transfusions and late platelets (but not RBC) transfusions were associated with a shorter progression-free survival and overall survival. Lymphoma-related mortality and nonrelapse mortality were both increased in the transfused population. Our data shed light on the mechanisms of early and late cytopenia and on the potential impact of transfusions on CAR T-cell efficacy and toxicity.

The prognosis of relapsed primary central nervous system lymphoma (PCNSL) remains dismal. CAR T‐cells are a major contributor to systemic lymphomas, but their use in PCNSL is limited. From the LOC network database, we retrospectively selected PCNSL who had leukapheresis for CAR‐T cells from the third line of treatment, and, as controls, PCNSL treated with any treatment, at least in the third line and considered not eligible for ASCT. Twenty‐seven patients (median age: 68, median of three previous lines, including ASCT in 14/27) had leukapheresis, of whom 25 received CAR T‐cells (tisa‐cel: N = 16, axi‐cel: N = 9) between 2020 and 2023. All but one received a bridging therapy. The median follow‐up after leukapheresis was 20.8 months. The best response after CAR‐T cells was complete response in 16 patients (64%). One‐year progression‐free survival from leukapheresis was 43% with a plateau afterward. One‐year relapse‐free survival was 79% for patients in complete or partial response at CAR T‐cell infusion. The median overall survival was 21.2 months. Twenty‐three patients experienced a cytokine release syndrome and 17/25 patients (68%) a neurotoxicity (five grade ≥3). The efficacy endpoints were significantly better in the CAR T‐cell group than in the control group ( N = 247) (median PFS: 3 months; median OS: 4.7 months; p < 0.001). This series represents the largest cohort of PCNSL treated with CAR T‐cells reported worldwide. CAR T‐cells are effective in relapsed PCNSL, with a high rate of long‐term remission and a reassuring tolerance profile. The results seem clearly superior to those usually observed in this setting.