Tech & Innovation in Healthcare

Question: As a futurist and an oncologist, I’m fascinated by the ways researchers are using artificial intelligence (AI) to help diagnose their patients’ conditions. I would love to read more about studies involving AI and cancer diagnoses. Do you have any resources?

Michigan Subscriber

Answer: Using a novel AI blood testing technology, researchers at Johns Hopkins Medicine successfully detected more than 80 percent of liver cancers in a recent study. The study was published on Nov. 18, 2022, in Cancer Discovery. The blood testing technology was originally developed by Johns Hopkins Kimmel Cancer Center, and researchers used the technology to successfully detect lung cancer in a 2021 study.

Called DNA evaluation of fragments for early interception, or DELFI for short, the blood test detects the fragmentation DNA changes from cancer cells as they are released into the patient’s bloodstream. This DNA is known as cell-free DNA (cfDNA). The DELFI technology examines millions of cfDNA fragments for abnormal patterns.

Researchers obtained blood plasma samples from 724 individuals in the U.S., the European Union, and Hong Kong, and used the DELFI technology on the samples to test for hepatocellular cancer (HCC), which is a type of liver cancer. Of the 724 samples collected, 501 were collected in the U.S. and European Union, with samples from 75 people known to have HCC. Those samples helped train and validate a machine learning (ML) model. Researchers analyzed an additional 223 samples from individuals in Hong Kong, where 90 people were known to have HCC.

When the samples were analyzed, the DELFI technology accurately detected 88 percent of the liver cancers.