AI Tool Detects More than 80 Percent of Liver Cancers in New Study

John Hopkins Kimmel Cancer Center in Baltimore, Md., used artificial intelligence (AI) blood test technology to detect more than 80 percent of liver cancers in current liver cancer and at-risk patients.

Researchers used the AI tool to assess 724 individuals from the U.S., EU or Hong Kong with hepatocellular carcinoma (HCC) or who were at average or high-risk for HCC.

The blood test uses technology that detects cell-free DNA — fragments from DNA cancer cells that are shed into the bloodstream —  to identify abnormal patterns.

The AI tool’s sensitivity for accurately detecting cancer was 88 percent in the average risk population and its specificity for correctly detecting negative results was 98 percent. Among high-risk individuals the sensitivity was 85 percent, and the specificity was 80 percent.