IRVINE, Ca. – There are more than a million colon cancer survivors in the United States, according to the American Cancer Society. That number is rising because doctors are finding and removing more pre-cancerous polyps during colonoscopies.
Every Friday, William Karnes, MD, clinical professor of medicine at UC Irvine has the help of an artificial intelligence system (AI) for colonoscopies at UC Irvine. He’s helping develop the system, which operates similarly to facial recognition.
“We painstakingly drew little boxes around all those polyps on tens of thousands of images and then trained the AI, tested the AI on its ability on a new set of images to find those polyps,” said Karnes.
Karnes said colonoscopists should find polyps in half of patients over 50, but sometimes that rate is as low as 10 percent.
“It’s that gap between the prevalence of polyps and our ability to find them which is responsible for interval colon cancers. Those are cancers we get despite being up to date with colonoscopy,” Karnes explained.
He said missed polyps result in seven percent of colon cancers.
John Gifford volunteered to get an AI-assisted colonoscopy. It found two pre-cancerous polyps that were removed.
“They were able to find something in its very formative stage, which was a relief because they found that, which means that they didn’t find anything else,” Gifford said.
Karnes says with the AI overlay, expert colonoscopists found 20 percent more polyps. He hopes upcoming clinical trials have similar results.
Karnes has tested the AI-assisted colonoscopies on about 100 patients so far. He hopes to begin multi-center clinical trials by next June.