Artificial intelligence blood test identifies silent liver disease years before physical symptoms appear

Johns Hopkins researchers use AI and DNA fragment patterns to identify liver fibrosis and cirrhosis years before symptoms, potentially preventing liver cancer.

By: AXL Media

Published: Mar 6, 2026, 6:24 AM EST

Source: The information in this article was sourced from Johns Hopkins Medicine

Revolutionizing Early Detection via Fragmentome Technology

A team at the Johns Hopkins Kimmel Cancer Center has pioneered an artificial intelligence-driven liquid biopsy capable of identifying liver damage long before clinical symptoms emerge. Unlike traditional liquid biopsies that search for specific genetic mutations associated with cancer, this new "fragmentome" technology analyzes genome-wide patterns of cell-free DNA (cfDNA) fragments circulating in the blood. Published in Science Translational Medicine on March 4, 2026, the study demonstrates how AI can read the specific ways DNA breaks apart to signal the presence of early-stage fibrosis and cirrhosis.

Decoding Metabolic and Genomic Signals

The research involved whole-genome sequencing of cfDNA samples from 1,576 individuals. Machine learning algorithms processed data from approximately 40 million DNA fragments per sample, examining their size and distribution across thousands of genomic regions—including repetitive areas typically ignored in genetic testing. Co-senior author Victor Velculescu, M.D., Ph.D., noted that while liver fibrosis is reversible in its early stages, it frequently goes undetected until it progresses to irreversible cirrhosis, significantly increasing the risk of liver cancer.

Advancing Beyond Cancer Diagnostics

The power of the fragmentome approach lies in its ability to capture a person’s overall physiologic state without needing to find a "needle in a haystack" mutation. First author Akshaya Annapragada explained that because the test analyzes how DNA is packaged and cut throughout the entire genome, it can develop specific classifiers for various health conditions. In one analysis of 570 people, the researchers created a "fragmentation comorbidity index" that successfully predicted overall survival and distinguished between different levels of chronic illness, sometimes outperforming traditional inflammatory markers.