Epigenetic Liquid Biopsy Enables Universal Mutation-Agnostic Molecular Surveillance for High-Risk Neuroblastoma.
PURPOSE: Liquid biopsy monitoring in pediatric solid tumors is limited by low mutational burden and lack of trackable genomic drivers. We sought to develop a mutation-agnostic, methylation-based liquid biopsy framework enabling universal molecular surveillance of high-risk neuroblastoma. EXPERIMENTAL DESIGN: Using whole-genome Oxford Nanopore Technologies sequencing of high-risk neuroblastoma tumors, we compared tumor-derived methylation profiles with a comprehensive atlas of normal human cell types and identified 72 neuroblastoma-specific differentially methylated regions (meNBL) that were reliably detectable in cell-free DNA (cfDNA). Marker robustness and specificity were validated using independent neuroblastoma methylation datasets and assessed against methylation profiles from other cancer types. We established neuroblastoma as a distinct methylation entity within the reference atlas by integrating a panel of 25 meNBLs, enabling quantitative estimation of tumor-derived cfDNA. Assay performance was evaluated across diagnostic, remission, relapse, and healthy control samples and compared with mutation-based and copy number-based approaches. RESULTS: Neuroblastoma-derived cfDNA was consistently detected at diagnosis and relapse but was absent in healthy controls and during confirmed remission. Methylation-based deconvolution demonstrated high specificity, with no detectable background signal in controls, and improved performance relative to copy number-based tumor fraction estimation. Longitudinal profiling enabled early molecular detection of relapse and reliable disease monitoring. CONCLUSIONS: We establish a robust, mutation-independent methylation-based liquid biopsy strategy for neuroblastoma that enables accurate, quantitative disease monitoring across all high-risk patients, including those lacking trackable genomic alterations. This approach supports the clinical translation of methylation-based cfDNA deconvolution as a broadly applicable platform for pediatric precision oncology.