Abstract Introduction Relapsed or refractory (r/r) DLBCL remains a clinical challenge after first-line chemoimmunotherapy. Chimeric antigen receptor T (CAR-T) therapy induces high response rates in DLBCL but often causes severe, prolonged cytopenias. Recent analyses report grade≥3 neutropenia in up to 60% of patients, and persistent (≥30 days) cytopenia in a substantial subset. Early cytopenias are largely attributed to intensive lymphodepleting regimens or bridging chemotherapy, whereas late cytopenias correlate with prior therapy and immune-mediated toxicities. Strategies to mitigate these hematologic toxicities are needed. Methods A retrospective analysis was conducted on 12 DLBCL patients, who underwent CAR-T cell therapy following autologous stem cell support at our center from December 2023 to April 2025. The lymphodepletion regime of these patients was modified BEAC regime, include carmustine (125mg/m2 on Day −8 to Day −7), etoposide (125 mg/m2, Day −6 to Day −3), cytarabine (250 mg/m2, Day −6 to Day −3), and cyclophosphamide (45 mg/kg, Day −7 to Day −6). The median time from leukapheresis to CAR-T cell infusion was 46 days (range, 21–76). Patients received the modified BEAC regimen starting on day −8. Hematopoietic stem cell (HSC) infusion was performed on day 0, followed by CAR-T cell infusion on day 3 (range, 1–4).The median MNC cell dose infused was 6.715×108/kg (range, 1.68–16.8). The CD34+ cell dose of ASCT was 3.355×106/kg (range,1.33–15.1). The CAR-T cell infusion dose was 5×106cells/kg (range,4–6). Treatment outcomes were assessed by overall response rate (ORR) and complete remission(CR) rate. Survival outcomes, including overall survival (OS) and progression-free survival (PFS), were analyzed using Kaplan-Meier curves. Adverse events were assessed according to CTCAE v5.0, while ICANS and CRS were evaluated using the Penn grading scale. Results At the time of analysis, all 12 patients were evaluable. The median age was 55 years (range, 33–67), and 9(75%) patients were female. CD19 was the primary CAR target in 11(91.7%) patients. 6 patients were classified as having primary refractory disease; 4 relapsed 12 months from initial therapy, and 2 relapsed 12 months from initial therapy. One patient had double-hit lymphoma and achieved a durable CR after CAR-T infusion; 4 patients were identified as double expressors. The median number of prior lines of therapy was 2 (range, 1–3). The median time to neutrophil engraftment after HSC infusion was 9.5 days (range, 8–15), and the median time to platelet recovery was 15 days (range, 8–23). At 1 month post-CAR-T cell infusion, 7 patients achieved CR, 4 had partial remission (PR), and 1 had stable disease (SD). At 3 months post-infusion, 10 patients achieved CR, 1 had progressive disease (PD), and 1 remained with SD. Among these PR patients, 3 achieved CR and 1 progressed to PD. The best ORR was 91.7% (11/12), with a CR rate of 83.3% (10/12). After a median follow-up of 9.3 months (range, 3–18.5 months), the OS rate was 100%, and the PFS rate was 91.7%. One patient remained in SD at the last follow-up. Regarding hematologic toxicities, all patients experienced grade 4 leukopenia, neutropenia, thrombocytopenia, and lymphopenia. Delayed hematologic toxicities beyond day 30 were infrequent, with grade ≥3 cytopenias observed in a minority of patients: leukopenia in 3 (25%), anemia in 2 (16.7%), and thrombocytopenia in 2 (16.7%). For non-hematologic toxicities, all patients developed CRS, with one patient experiencing grade ≥3 CRS. Only one patient experienced grade 3 ICANS. Elevated transaminases were observed in 4 patients, infections in 5 patients, and one patient developed cardiac dysfunction with BNP levels increasing from 93 to 6700 pg/mL. All toxicities were controllable and resolved with appropriate treatment. Conclusion The combination of CAR-T cell therapy with HSC infusion following a modified BEAC regimen demonstrated promising efficacy and manageable toxicity in patients with r/r DLBCL. The modified BEAC regimen enabled timely hematopoietic recovery. These findings support the feasibility of this approach and warrant further validation in larger cohorts.
You et al. (Mon,) studied this question.