AI-Based Liquid Biopsy Detects Liver Fibrosis, Cirrhosis and Chronic Disease Signals

Detecting these conditions earlier could allow treatment before irreversible damage occurs and reduce the risk of liver cancer. Researchers have now developed an artificial intelligence (AI)–driven liquid biopsy that analyzes cell-free DNA patterns in blood to detect early liver disease and potentially identify broader chronic health conditions.

The approach, developed by scientists at the Johns Hopkins Kimmel Cancer Center (Baltimore, MD, USA), analyzes cell-free DNA (cfDNA) fragmentomes—patterns describing how DNA fragments are cut, distributed, and packaged across the genome. Using AI and genome-wide sequencing, the method identifies disease-specific fragmentation signatures associated with liver fibrosis and cirrhosis.

In the study published in Science Translational Medicine, researchers analyzed blood samples from 1,576 individuals with liver disease and other medical conditions using whole-genome sequencing. Approximately 40 million cfDNA fragments were evaluated in each analysis across thousands of genomic regions, including repetitive sequences that are rarely studied in conventional diagnostics. Machine-learning algorithms examined the size, distribution, and genomic locations of these DNA fragments to detect patterns linked to liver disease.

The resulting classification system was able to identify early liver fibrosis, advanced fibrosis and cirrhosis with high sensitivity. Unlike traditional liquid biopsy methods that focus on specific genetic mutations, this approach evaluates the entire DNA fragmentation landscape, providing broader insight into physiological changes occurring throughout the body. The researchers also developed a fragmentation comorbidity index, which distinguished individuals with higher overall disease burden using cfDNA patterns.

In a cohort of 570 individuals with suspected serious illness, the index correlated with the widely used Charlson Comorbidity Index and was able to predict overall survival in some analyses. Beyond liver disease, fragmentomic signals associated with cardiovascular, inflammatory and neurodegenerative conditions were also detected, suggesting potential for broader diagnostic applications. Early liver fibrosis can often be reversed with treatment, but it frequently goes undetected until it progresses to cirrhosis or liver cancer.

Current blood markers lack sensitivity for early disease, while imaging methods require specialized equipment that may not always be available. The researchers emphasize that the liver fibrosis test remains a prototype and requires further validation before clinical use. Future work will focus on refining disease-specific classifiers and exploring fragmentome signatures for additional chronic conditions.

“Many individuals at risk don’t know they have liver disease,” said Victor Velculescu, M.D., Ph.D., co-senior author of the study. “If we can intervene earlier — before fibrosis progresses to cirrhosis or cancer — the impact could be substantial.”

Related Links:
Johns Hopkins Kimmel Cancer Center