Presentation (CTI)

OBJECTIVES: To perform a comparative evidence synthesis for generalized myasthenia gravis (gMG) using model-based network meta-analysis (MBNMA) with time-adjusted treatment effects, and to characterize time-dependent responses for individual treatments where data permitted.
METHODS: A fractional polynomial time-course MB-NMA was conducted by using longitudinal trial data in adults with gMG. Treatments were compared versus placebo/standard care using parametric time-course functions to model outcomes across multiple follow-up time-points. Analyses were performed within a Bayesian framework using Markov Chain Monte Carlo (MCMC) methods, accounting for repeated measures within studies and allowing treatment effects to vary over time.
RESULTS: All evaluated treatments demonstrated efficacy benefits versus placebo/standard care; however, robust indirect comparative assessment remained challenging. Based on Surface Under the Cumulative Ranking Curve (SUCRA) rankings from the time-course MB-NMA, rozanolixizumab (ROZA) and efgartigimod (EFG) consistently ranked among the top-performing treatments across multiple efficacy outcomes (Myasthenia Gravis Activities of Daily Living [MG-ADL] responder, Quantitative Myasthenia Gravis Score [QMG], Myasthenia Gravis Composite [MGC], Myasthenia Gravis Quality of Life [MG-QoL]). Eculizumab showed strong relative performance for MG-QoL, while batoclimab ranked favorably for MG-ADL and MGC. Zilucoplan also demonstrated favorable rankings across several efficacy domains, although results were less consistent across endpoints. Overall, SUCRA rankings indicated that ROZA and EFG demonstrated the most consistently favorable efficacy profiles across assessed outcomes, while other therapies showed end-point specific strengths.
CONCLUSIONS: Time-course MB-NMA offers a valuable framework for evaluating dynamic treatment effects in gMG while maximizing the use of longitudinal data across heterogeneous follow-ups. Specific subgroup analyses by antibody status or MGFA class were not feasible in this iteration due to data limitations. Clinical heterogeneity and indirect evidence warrant cautious interpretation of findings.