Implementing Multilevel Network Meta-Regression for Time-To-Event Outcomes: A Case Study in Relapsed Refractory...

Area of Science:

• Biostatistics
• Clinical Trial Analysis
• Pharmacoeconomics

Background:

• Multilevel network meta-regression (ML-NMR) integrates individual patient data (IPD) and aggregate data from randomized controlled trials (RCTs).
• It enables robust comparative efficacy assessments of multiple treatments while accounting for between-study heterogeneity.
• This study focuses on applying ML-NMR to time-to-event outcomes.

Purpose of the Study:

• To provide an overview of ML-NMR for time-to-event outcomes.
• To apply ML-NMR in a case study for relapsed/refractory multiple myeloma.
• To demonstrate the implementation of ML-NMR with R code.

Main Methods:

• Evaluated idecabtagene vicleucel, selinexor+dexamethasone, belantamab mafodotin, and conventional care for overall survival.
• Combined single-arm trials and real-world data to form artificial RCTs (aRCTs).
• Utilized ML-NMR models adjusted for prior lines of therapy, triple-class refractory status, and age, comparing models with leave-one-out information criterion.

Main Results:

• The Weibull ML-NMR model demonstrated the best fit.
• Idecabtagene vicleucel showed superior overall survival compared to selinexor+dexamethasone, belantamab mafodotin, and conventional care.
• Triple-class refractory status was the only significant prognostic factor; other effect modifiers had minimal impact.

Conclusions:

• ML-NMR is a valuable method for time-to-event outcome analysis in complex treatment networks.
• The study provides practical R code to facilitate ML-NMR implementation.
• Practitioners are encouraged to use ML-NMR for treatment comparisons requiring population adjustment.