Artificial intelligence matched and in some cases outperformed gastroenterologists in assessing Crohn’s disease during endoscopy, according to a study published Aug. 18 in Clinical Gastroenterology and Hepatology.
Researchers at Ann Arbor-based University of Michigan developed an AI model that spotted mucosal ulceration in Crohn’s disease patients as accurately as physicians. The tool also closely matched the Simple Endoscopic Score for Crohn’s Disease, the most common scoring system, according to an Aug. 27 news release from Michigan Medicine.
The AI reached a higher similarity score (.591) than the agreement between two gastroenterologists (.462).
AI can offer more objective and reproducible metrics than current manual tools, senior author Ryan Stidham, MD, said in the release.
Study authors also said AI-assisted endoscopy could improve consistency among providers, help with treatment decisions where IBD specialists are scarce and support education and drug development.
“It’s very early, but computer vision tools like this may provide an important component for the future of automated care, where AI and experts work together in treating patients,” Dr. Stidham said in the release
