Researchers at the Johns Hopkins Kimmel Cancer Center have developed a groundbreaking artificial intelligence (AI)-driven liquid biopsy that analyzes genome-wide patterns of circulating cell-free DNA (cfDNA) fragments in the blood. This innovative test can identify early signs of liver conditions such as fibrosis and cirrhosis, which often remain undetected until significant damage has occurred. The potential to detect these silent diseases years before symptoms arise could fundamentally change how liver diseases are diagnosed and managed.
The study, which was published in Science Translational Medicine on March 4, 2026, represents the first systematic application of a DNA fragmentation analysis technique, known as fragmentome technology, to chronic diseases unrelated to cancer. Traditionally, this approach has been utilized primarily for cancer detection, but this new research opens avenues for diagnosing other critical health conditions.
In their research, the team performed whole genome sequencing on cfDNA samples from 1,576 individuals with liver disease and other health issues. By examining the patterns of DNA fragments across the entire genome, they sought to uncover signals that could indicate the presence of disease. The analysis encompassed approximately 40 million fragments distributed across thousands of genomic regions, producing a substantial dataset that surpasses most existing liquid biopsy tests.
Understanding Fragmentome Technology
The fragmentome approach diverges from conventional liquid biopsy methods that typically focus on identifying specific gene mutations associated with cancer. Instead, this new technique studies how DNA fragments are fragmented, organized, and distributed throughout the genome. This broader focus allows for application beyond cancer, potentially encompassing other chronic diseases that could elevate cancer risk.
Akshaya Annapragada, a lead author and M.D./Ph.D. Student in the Velculescu lab, highlighted the power of analyzing the entire fragmentome, stating, “We are analyzing the entire fragmentome, which contains a tremendous amount of information about a person’s physiologic state.” This large-scale data analysis, combined with advanced machine learning algorithms, enables the development of classifiers specific to various health conditions.
Early Detection Benefits and Implications
Victor Velculescu, M.D., Ph.D., co-director of the cancer genetics and epigenetics program at Johns Hopkins, emphasized the significance of early detection for liver diseases. He noted that approximately 100 million people in the United States are affected by liver conditions that increase their risk of cirrhosis and liver cancer. Current blood tests for fibrosis often lack sensitivity, particularly in the disease’s early stages, with standard markers failing to identify early fibrosis in many cases.
“Many individuals at risk don’t know they have liver disease,” Velculescu stated. “If we can intervene earlier—before fibrosis progresses to cirrhosis or cancer—the impact could be substantial.” Early identification of precursor conditions could enable healthcare providers to treat underlying issues sooner and potentially prevent the progression to more severe diseases, including cancer.
Future Directions and Broader Applications
The current study is an extension of previous research focused on the fragmentome of liver cancer, where subtle DNA signals linked to disease were identified even in patients with seemingly normal fragmentation profiles. This prompted further investigation into fragmentome patterns specifically related to liver fibrosis and cirrhosis.
In a related analysis involving individuals with suspected serious illnesses, researchers developed a fragmentation comorbidity index, which distinguished between individuals with varying comorbidity scores. This index was able to predict overall survival and, in some instances, proved more specific than traditional inflammatory markers.
While the liver fibrosis assay remains a prototype and has not yet been introduced as a clinical test, researchers are keen to refine and validate the classifier for liver disease and explore fragmentome signatures associated with other chronic illnesses. This research suggests the technology may eventually have broader medical applications, potentially aiding in the detection of cardiovascular, inflammatory, and neurodegenerative disorders.
The implications of this research are far-reaching, as early detection could significantly alter the landscape of chronic disease management. As the team continues to refine their techniques and validate their findings, the potential to develop disease-specific classifiers could pave the way for innovative diagnostic tools in the future.
As the research progresses, the medical community will be closely watching for updates on the adoption of this technology in clinical settings. The hope is that such advancements will lead to improved patient outcomes and a better understanding of chronic diseases.
For those interested in the intersection of technology and healthcare, the ongoing developments in AI-driven diagnostics promise to offer exciting new possibilities. Your thoughts and comments on this breakthrough are welcome.
Disclaimer: This article is for informational purposes only and does not constitute medical advice.
