Site Editor



Detecting Liver Cancer Via Genome-Wide Cell-Free Fragmentome Features

By: Sarah Lynch
Posted: Thursday, December 1, 2022

Zachariah Foda, MD, of Johns Hopkins University, and colleagues have developed a method for screening for hepatocellular carcinoma by examining cell-free DNA fragmentome features. Examining fragmentation profiles allowed them to spot specific genomic and chromatin characteristics, particularly those known to impact the development of hepatocellular carcinoma. Overall, according to the investigators, both the success of the models and their cost-efficiency suggest future improvements in screening for hepatocellular carcinoma. These findings were presented at the 2022 American Association for the Study of Liver Diseases (AASLD) Annual Meeting (Abstract 201).

“Screening individuals at high risk for hepatocellular carcinoma, including those with cirrhosis and viral hepatitis, provides an avenue for improved survival, but current screening methods are inadequate,” the investigators noted. “Overall adherence to international guidelines remains low at 20%, and current screening tests have also shown limited sensitivity, especially for early-stage disease.”

The researchers used a cross-validated machine learning model that evaluated genome-wide fragmentome features from the blood samples of 391 patients. Of the patients selected, 47 had hepatocellular carcinoma and 344 were cancer-free. Among the cancer-free individuals, 52 were considered to be at high risk for hepatocellular carcinoma. The researchers examined genome sequencing data from a group of 223 patients, including 80 with early-stage hepatocellular carcinoma, 101 high-risk patients without liver cancer, and 32 healthy patients.

After analyzing the models, the researchers found an AUC of 0.97, including all patient data. For patients with hepatocellular carcinoma, the AUC was 0.98. For high-risk patients, the AUC was 0.92. At 90% specificity, the overall sensitivity of the model for detecting cancer was 94% among all patients and 79% among high-risk patients. The model was able to distinguish patients with hepatocellular carcinoma with an AUC of 0.87 and high-risk patients with an AUC of 0.86.

Disclosure: For full disclosures of the study authors, visit

By continuing to browse this site you permit us and our partners to place identification cookies on your browser and agree to our use of cookies to identify you for marketing. Read our Privacy Policy to learn more.