T-cell engaging bispecific antibody shows potent MRD negativity with variable remission and survival outcomes: A bayesian meta-analysis of randomized controlled trials

Abstract Introduction B-cell acute lymphoblastic leukemia (B-ALL) remains a challenging malignancy, particularly in relapsed/refractory (R/R) settings where prognosis is poor despite intensive chemotherapy. Blinatumomab, a T-cell engaging bispecific antibody targeting CD19 on B-cells and CD3 on T-cells, has shown promise in inducing deep remissions and improving survival across various disease stages. However, trial results vary, with differences in patient populations (R/R vs. frontline) and endpoints, necessitating a comprehensive synthesis. This Bayesian meta-analysis aggregates evidence from pivotal trials to evaluate blinatumomab's efficacy on complete remission (CR), measurable residual disease (MRD) negativity, and overall survival (OS), providing clinically relevant insights for optimizing B-ALL therapy. Methods We conducted a Bayesian meta-analysis of three key trials—TOWER (R/R B-ALL), BLAST (MRD+ B-ALL), and E1910 (newly diagnosed B-ALL)—comparing blinatumomab to standard chemotherapy or historical controls. Outcomes included CR rate (log relative risk, positive favoring blinatumomab), MRD negativity (log relative risk, positive favoring blinatumomab), and OS (log hazard ratio, negative favoring blinatumomab). A random-effects hierarchical Bayesian model with shrinkage estimation adjusted study effects toward the pooled mean. Non-informative priors were used for pooled effects and between-study variance (tau), generating 95% credible intervals (CrI) and prediction intervals via Markov Chain Monte Carlo simulations. Heterogeneity was assessed with tau, and sensitivity analyses examined influences of disease stage and trial design. Data were extracted from published trials as of August 06, 2025. Analysis was done using R 4.5.1 Outcomes The analysis included 1,009 patients across the trials. For CR rate, the pooled log relative risk was 1.23 [95% CrI: -0.46, 3.04], with shrinkage-adjusted estimates of -0.24 [95% CrI: -0.48, -0.00] for TOWER, 5.45 [95% CrI: 2.67, 8.23] for BLAST, and 1.46 [95% CrI: 1.04, 1.88] for E1910, indicating a potential benefit but with uncertainty. The prediction interval was 1.22 [-2.42, 5.02], and heterogeneity was high (tau = 1.45 [0.56, 2.55]). For MRD negativity, the pooled log relative risk was 1.52 [95% CrI: 0.34, 2.61], with estimates of 1.13 [95% CrI: 0.71, 1.55] for TOWER, 1.24 [95% CrI: 0.81, 1.68] for BLAST, and 2.59 [95% CrI: 1.95, 3.23] for E1910, demonstrating a significant advantage for blinatumomab. The prediction interval was 1.52 [-0.87, 3.81], with moderate heterogeneity (tau = 0.80 [0.19, 1.78]). For OS, the pooled log hazard ratio was -0.32 [95% CrI: -0.87, 0.23], with estimates of -0.34 [95% CrI: -0.61, -0.08] for TOWER, -0.12 [95% CrI: -0.48, 0.24] for BLAST, and -0.53 [95% CrI: -0.96, -0.09] for E1910, suggesting a survival trend without definitive evidence. The prediction interval was -0.32 [-1.40, 0.77], with low heterogeneity (tau = 0.22 [0.00, 0.95]). Conclusion This Bayesian meta-analysis highlights blinatumomab's robust efficacy in achieving MRD negativity in B-ALL, a critical prognostic factor that may enable curative allogeneic transplantation and long-term remission. While CR and OS benefits show promise, particularly in frontline settings, heterogeneity underscores the influence of disease stage on outcomes. Clinically, these findings support integrating blinatumomab into standard regimens for R/R and MRD+ patients to enhance deep responses and potentially reduce relapse, urging further trials in pediatric and high-risk cohorts to refine its role in transforming B-ALL care.