Efficacy and safety of CAR-T cell therapy versus bispecific antibodies in patients with relapsed/refractory B-cell non-Hodgkin lymphoma: A single-arm meta-analysis

Abstract Background Relapsed/refractory B-cell non-Hodgkin's lymphoma (R/R B-NHL) remains challenging to treat, and durable remission is often difficult to achieve with conventional therapies. Chimeric antigen receptor T-cell (CAR-T) therapy and bispecific antibodies (BsAbs), as emerging immunotherapeutic strategies, have shown promising therapeutic prospects in several early-stage clinical trials in recent years, mainly targeting B-cell antigens such as CD19 or CD20. However, a systematic comparison of their relative efficacy and safety is lacking. Objective The aim of this study was to systematically evaluate and compare the efficacy and safety of CAR-T cell therapy with bispecific antibodies in the treatment of patients with R/R B-NHL. Methods This meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.A systematic search for single-arm phase I/II clinical trials was performed to indirectly compare efficacy and safety using a random-effects single-arm meta-analysis.We searched PubMed, Embase, Web of Science, and Cochrane databases up to February 8, 2025. The primary outcome was the objective response rate (ORR), and the secondary outcome was the complete response rate (CR); meta-regression analyses were performed to adjust for relevant covariates. Results Among pre-established search criteria, 59 trials involving 2,914 pts met the eligibility criteria and were included in the analysis. A total of 59 clinical studies were included, comprising 46 CAR-T and 13 BsAbs trials (ratio ≈2:1). Among CAR-T studies, 32 targeted CD19, 7 dual CD19/CD20, 3 CD20, and 4 dual CD19/CD22; among BsAbs studies, 7 used a CD3×CD20 configuration and 6 a CD3×CD19 configuration. In terms of study design, 17 were multicenter trials, and 43 (71.7%) were single-center studies. CAR-T cell therapy was significantly more efficacious than BsAbs, with an ORR of 72% (95% CI: 67%–77%) and a CR of 54% (49%–59%), which were significantly higher than those of BsAbs in terms of ORR 50% (38%–62%) and CR 33% (23%–46%) (both p<0.01). Among these, CD20 targeted CAR-T therapies demonstrated a higher ORR 85% (95% CI: 5%–100%) than CD3×CD20 BsAbs 54% (37%–71%) (p<0.01), although the wide confidence interval reflects limited available data. Among grade ≥3 adverse events, the CAR-T group had an overall higher incidence than the BsAbs group, including cytokine release syndrome (CRS, 8% vs. 4%; p<0.01), neurotoxicity (8% vs. 6%; p<0.01), and infection (14% vs. 15%; p<0.01). Specifically, grade ≥3 immune effector cell-associated neurotoxicity syndrome (ICANS) was more common in the CAR-T group (12%, 95% CI: 9%–16%) vs. the BsAbs group (6%, 2%–18%; p<0.01). In addition, dual-target CAR-T constructs may be associated with a higher risk of adverse events. Among different generations of CAR-T products, the ORR of third-generation products was 79% (49%–93%), slightly higher than that of second-generation products 72% (67%–77%). In second-generation CAR-T, the 4-1BB co-stimulation conformation 74% (69%–79%) was slightly more efficacious than the CD28 conformation 67% (54%–78%). Among the FDA approved CAR-T products, Axicabtagene ciloleucel demonstrated a high ORR 78% (71%–83%); among the BsAbs, Epcoritamab had the highest ORR 76% (wide CI due to limited studies). Conclusion CAR-T cell therapy demonstrates superior efficacy, while BsAbs exhibit a more favorable safety profile. Both modalities offer distinct advantages that provide valuable guidance for clinical decision-making in R/R B-NHL.