REMODEL: A modular framework for longitudinal ctDNA methylation profiling reveals treatment response and relapse-associated epigenetic remodeling

We developed REMODEL (Recovery-based Epigenetic Monitoring of Dynamic methylation for Longitudinal analysis), an integrated framework combining coverage-aware binning, count-based longitudinal statistics, and a novel Epigenetic Recovery Score (ERS) to quantify post-treatment methylation recovery.

Applied to 17 serial plasma samples from two T-cell lymphoma patients (2,025 genomic bins), REMODEL distinguished durable remission, characterized by coordinated demethylation (86.4% of bins) and sustained positive ERS, from relapse, marked by widespread hypermethylation (64.9% of bins) and persistently negative ERS despite clinical remission. By integrating dynamic methylation detection with recovery modelling, REMODEL provides an interpretable framework for longitudinal ctDNA methylation analysis and minimal residual disease monitoring.