AI-Assisted Colonoscopy Detects More Polyps but Has Modest Effect on Cancer Risk

Currently, screening is performed using a test that detects blood in the stool (FIT screening). If the test identifies a certain amount of blood, the patient is referred for a colonoscopy. During these colonoscopies, artificial intelligence (AI) is commonly used as an assistant to help identify polyps. However, the clinical benefits of AI in this context need to be documented, much like other diagnostic and treatment tools. AI applications in colonoscopy are among the most studied areas of clinical medicine. A recent systematic review conducted during the guideline development process analyzed 44 randomized studies involving over 30,000 patients. AI was found to detect more polyps, but long-term effects such as cancer incidence and mortality were not assessed. A new study has now revealed that AI use during colonoscopy increased polyp detection by approximately 8%, but had minimal impact on cancer risk.

This large microsimulation study, led by the University of Oslo (Oslo, Norway), aimed to explore whether AI could help reduce colorectal cancer incidence and mortality. The study evaluated the effects of AI on cancer occurrence, mortality, and potential drawbacks such as increased follow-up colonoscopies when AI is used in colorectal cancer screening. The researchers reviewed data from about 15 studies in addition to the systematic review. The results showed that AI’s use in colonoscopy had little to no clinically significant effect on colorectal cancer risk or mortality after screening. Specifically, AI reduced cancer risk over 10 years from 0.82% without AI to 0.71% with AI. For those undergoing colonoscopy after FIT screening, the risk decreased slightly from 5.82% to 5.77%. The study, published in BMJ Medicine, also found that AI use led to a 20% increase in follow-up colonoscopies over 10 years, due to the enhanced detection of polyps. This increase results in more examinations for individual patients and additional strain on healthcare systems. The study's findings suggest that while AI improves polyp detection and can reduce cancer risk, it must be considered alongside the drawbacks of increased follow-up examinations, as the effect on cancer risk is minimal.

“The project shows that while artificial intelligence can reduce cancer risk by improving the detection of polyps, it must be weighed against the drawbacks of increased follow-up burden. The risk of cancer, which is the most important for patients, is minimally affected,” said physician and Ph.D. candidate Natalie Halvorsen from the Clinical Effectiveness Research Group at UiO.

Related Links:
University of Oslo