St. Jude Children’s Research Hospital and international collaborators introduced M‑PACT (Methylation‑based Predictive Algorithm for CNS Tumors), an AI platform that uses cerebrospinal fluid circulating tumor DNA (ctDNA) methylation profiling to molecularly classify pediatric central nervous system tumors noninvasively. The method, reported by the team, enables diagnostic resolution when tumor tissue is inaccessible or high‑risk to obtain. M‑PACT combines targeted methylation sequencing with machine learning to assign tumor class and subtype from CSF‑derived ctDNA. The platform could reduce the need for invasive biopsies and speed diagnostic subtyping that informs treatment and clinical trial eligibility. Clinicians should note that ctDNA methylation classification is a rapidly maturing approach: methylation signatures capture tumor‑specific epigenetic patterns, but assay sensitivity, standardized pipelines and validation across centers will determine clinical deployment timelines.