Building upon the foundation of liquid biopsy utility for the early detection of cancer, analysis of genome-wide cell-free DNA fragmentation with machine learning classification and modeling can also extend to the identification of liver cirrhosis and other chronic diseases, according to findings published in Science Translational Medicine.
“This builds directly on our earlier fragmentome work in cancer, but now using AI and genome-wide fragmentation profiles of cell-free DNA to focus on chronic diseases,” said co-senior study author Victor Velculescu, MD, PhD, Co-Director of the Cancer Genetics and Epigenetics Program at the Johns Hopkins Kimmel Cancer Center. “For many of these illnesses, early detection could make a profound difference, and liver fibrosis and cirrhosis are important examples. Liver fibrosis is reversible in early its stages, but if left undetected, it can progress to cirrhosis and ultimately increase the risk of liver cancer.”
Study Methods and Key Findings
Liquid biopsy use in other diseases has been largely unexplored, until now.
The researchers used whole-genome sequencing to examine cell-free DNA fragmentomes in 1,576 individuals. The cohort included patients with liver disease as well as patients with vascular, autoimmune, and neurodegenerative conditions. Analysis focused on fragment size as well as distribution of the fragments within the entire genome.
They also developed a machine learning classifier to detect signatures of early liver disease, advanced fibrosis, and cirrhosis across the vast amount of fragmentomes. The classifier was tested in a discovery (n = 423) and validation cohort (n = 221), and demonstrated limited cross-reactivity for other diseases and high sensitivity.
Analysis of genome-wide fragmentomes and methylomes found liver-derived and immune-mediated changes in the cell-free DNA of patients with liver disease. Changes in fragmentomes were also found in individuals with other diseases, and reflected disease-specific changes in their circulation.
Then, a machine learning model was created using cell-free DNA fragmentomes to predict the survival of patients with several of the identified diseases. The model was tested on separate discovery (n = 571) and validation (n = 231) cohorts.
“The fact that we are not looking for individual mutations is what makes this study so powerful,” said first author Akshaya Annapragada, an MD/PhD student working in the Velculescu lab. “We are analyzing the entire fragmentome, which contains a tremendous amount of information about a person’s physiologic state. The scale of these data, coupled with machine learning, enables development of specific classifiers for many different health conditions.”
“The fragmentome can serve as a foundation for building different classifiers for different diseases, and importantly, these classifiers are disease-specific and do not cross-react,” she added. “A liver fibrosis classifier is distinct from a cancer classifier. This is a unique, disease-specific test built from the same underlying platform.”
DISCLOSURE: The research was supported in part by the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation, SU2C in-Time Lung Cancer Interception Dream Team Grant, Stand Up to Cancer-Dutch Cancer Society International Translational Cancer Research Dream Team Grant, the Gray Foundation, The Honorable Tina Brozman Foundation, the Commonwealth Foundation, the Mark Foundation for Cancer Research, the Danaher Foundation and ARCS Metro Washington Chapter, the Family of Dan Y. Zhang AACR Scholar in Training Award, the Cole Foundation, and National Institutes of Health grants. For full disclosures of the study authors, visit science.org.
