Medtronic's colonoscopy AI cuts number of missed polyps in half, study finds

In a standard colonoscopy, as many as one-third of colorectal polyps and adenomas—typically benign clumps of cells that form on the lining of the colon and may be precursors to cancer—can go by undetected.

Hoping to cut down on these inaccuracies is Medtronic, the global distributor of an artificial intelligence system originally developed by Cosmo Pharmaceuticals that aims to increase the number of precancerous polyps detected during a colonoscopy.

It’s well on its way to achieving that goal, as evidenced by the results of the first U.S. trial of the GI Genius system. Published Tuesday in the American Gastroenterological Association journal Gastroenterology, the study found that adding the AI to a standard colonoscopy did indeed vastly improve polyp detection.

GI Genius processes colonoscopy imaging data in real time, using its AI to automatically spot colorectal polyps of all shapes and sizes. The module is designed to connect to standard endoscopic equipment from all major brands.

In the study, just about 15% of polyps were missed when the AI system was used during the exam, compared to a miss rate of more than 32% in AI-less colonoscopies.

GI Genius had a significantly lower miss rate than standard methods on practically every comparison point: It spotted more polyps in each of the proximal and distal regions of the colon and more of those that were tiny in size and those less prominent on the colon’s lining.

The researchers also compared each method’s rate of false negatives based on whether a follow-up was able to find polyps that had gone undetected in an initial colonoscopy performed either with or without the use of GI Genius. They found that in follow-up exams performed after an AI-assisted colonoscopy, additional polyps were detected at a rate of less than 7%, compared to nearly 30% in follow-ups to non-AI exams.

“We know that colonoscopy is the gold standard for colon cancer screening, and this study unequivocally demonstrates that AI technology can help physicians better detect polyps during the procedure,” said Austin Chiang, M.D., the recently appointed chief medical officer of Medtronic’s gastrointestinal business.

“As a gastroenterologist, I worry about missed polyps because around half of all cases of post-colonoscopy colorectal cancer may be attributed to not catching them during the index colonoscopy,” Chiang said. “The impact of missed polyps could ultimately be the difference between life and death when we consider that 90% of patients with colon cancer can beat it when it’s caught early.”

The system has already passed the FDA’s muster, with a first-of-its-kind clearance that arrived in April 2021.

Hot on Medtronic’s tail in the field of AI-assisted colonoscopy are a crush of companies that includes Olympus, Docbot and Iterative Scopes, the latter of which kicked off the year by picking up $150 million to usher its own colorectal cancer-spotting tech through the FDA review process.