Particle.news
Download on the App Store

AI Liquid Biopsy Reads cfDNA Fragmentome to Detect Early Liver Fibrosis and Cirrhosis

Machine learning on genome-wide cfDNA fragmentation provides a noninvasive readout of liver damage often missed by standard blood tests.

Overview

  • Johns Hopkins researchers reported the results March 4 in Science Translational Medicine, with partial support from the National Institutes of Health.
  • Analyzing cfDNA fragmentomes from 1,576 people, the team used whole‑genome sequencing and AI to identify disease‑specific fragmentation signatures.
  • The classifier distinguished early liver disease, advanced fibrosis and cirrhosis with high sensitivity in separate discovery and validation cohorts.
  • A fragmentation comorbidity index correlated with Charlson Comorbidity Index scores and independently predicted overall survival in tested cohorts.
  • Investigators observed fragmentomic signals tied to other chronic conditions, but the liver assay remains a prototype requiring further development and validation before clinical use.