Artificial Intelligence Improves Detection, Regardless of Endoscopist Experience
Douglas K. Rex, MD, MASGE, reviewing Repici A, et al. Gut 2021 Jun 29.
Five randomized controlled trials (4 from China and 1 from Europe and the U.S.) have demonstrated that artificial intelligence (AI) increases the adenoma detection rate (ADR) by a mean of 11%. The previous controlled trial from the West was performed among expert endoscopists. The current study was initiated by the same Western group to examine the benefits of the same AI program (GI Genius; Medtronic, Minneapolis, Minn). Ten European endoscopists, each with less than 2000 completed colonoscopies, performed colonoscopy in 660 patients, aged 40 to 80 years, undergoing screening, surveillance, or diagnostic colonoscopy. The patients were randomized to a group with or without AI.
Compared to the non-AI group, the ADR for the AI group was 22% higher (53.3% vs 44.5%), and the rate of adenomas per colonoscopy was 21% higher. Most of the gains were in small and nonpolypoid lesions.
When the 2 studies were combined, the benefits of AI were the same, regardless of endoscopist experience.
COMMENTThese results indicate it would be good for endoscopists to incorporate AI for adenoma detection early in their career.
Note to readers: At the time we reviewed this paper, its publisher noted that it was not in final form and that subsequent changes might be made.
CITATION(S)
Repici A, Spadaccini M, Antonelli G, et al. Artificial intelligence and colonoscopy experience: lessons from two randomised trials. Gut 2021 Jun 29. (Epub ahead of print) (http://dx.doi.org/10.1136/gutjnl-2021-324471)
