Scientists on the Johns Hopkins Kimmel Most cancers Heart have created an AI-driven liquid biopsy check that examines genome-wide cell-free DNA (cfDNA) fragmentation patterns and repeat landscapes to determine early liver fibrosis and cirrhosis. This innovation additionally uncovers indicators of wider persistent illness burdens.
The analysis, backed partly by the Nationwide Institutes of Well being and revealed March 4 in Science Translational Medication, pioneers the usage of fragmentome expertise—initially developed for most cancers—in persistent noncancer ailments.
Fragmentome Evaluation in Motion
Groups carried out whole-genome sequencing on cfDNA samples from 1,576 sufferers with liver situations and associated well being points. They scrutinized fragment sizes, genomic distributions, and patterns in repetitive areas, processing about 40 million fragments throughout 1000’s of areas—way more information than typical liquid biopsies.
Machine-learning fashions sifted by way of this huge dataset to pinpoint disease-specific signatures, attaining excessive sensitivity for early liver illness, superior fibrosis, and cirrhosis.
This strategy extends prior fragmentome analysis from most cancers to persistent situations utilizing AI and complete cfDNA profiles, acknowledged Victor Velculescu, M.D., Ph.D., co-director of the most cancers genetics and epigenetics program at Johns Hopkins Kimmel Most cancers Heart and co-senior writer. For a lot of such ailments, early detection transforms outcomes; liver fibrosis and cirrhosis exemplify this, as early fibrosis reverses however advances to cirrhosis and liver most cancers threat if missed.
Differing from mutation-focused biopsies, fragmentome expertise research DNA fragment cleavage, packaging, and unfold, suiting it for noncancer illnesses that will precede tumors. Robert Scharpf, Ph.D., professor of oncology, and Jill Phallen, Ph.D., assistant professor of oncology, co-led the trouble.
We acquire energy by skipping particular person mutations to research the complete fragmentome, which reveals huge physiologic insights, defined lead writer Akshaya Annapragada, M.D./Ph.D. pupil in Velculescu’s lab. Huge information scales plus machine studying yield tailor-made classifiers for various well being states.
Tackling Widespread Liver Dangers
Practically 100 million Individuals stay with liver points elevating cirrhosis and most cancers risks. Present blood checks miss early fibrosis and catch cirrhosis solely half the time; imaging calls for specialised ultrasound or MRI, usually unavailable.
Excessive-risk folks usually stay unaware of liver illness, Velculescu noticed. Early intervention earlier than development to cirrhosis or most cancers guarantees main positive aspects, probably averting most cancers through treatable precursors.
From Most cancers Insights to Comorbidity Instruments
The work stems from a 2023 liver most cancers research the place fibrosis sufferers confirmed faint fragmentome shifts regardless of regular profiles, spurring focused evaluation. In 570 critically in poor health sufferers, a brand new fragmentation comorbidity index outperformed some irritation markers in predicting survival and Charlson Comorbidity Index ranges.
Particular signatures linked to poorer outcomes; the strategy additionally flagged cardiovascular, inflammatory, and neurodegenerative alerts, hinting at broad utility pending additional research.
Fragmentomes allow disease-tailored classifiers that keep away from overlap—a liver fibrosis detector differs sharply from most cancers ones on this unified platform, Annapragada famous.
Subsequent Steps for Scientific Impression
This liver fibrosis prototype awaits refinement, validation, and extension to extra persistent ailments for real-world use.
