To assess the impact of adopting an artificial intelligence computer-aided detection system, namely GI Genius™ by Medtronic, on the detection of polyps and adenomas during routine colonoscopy in real-world clinical practice

A retrospective analysis was conducted, including all colonoscopies performed over a period of three months before and three months after AI implementation at a private hospital. Demographic data, the indication of colonoscopy (screening, surveillance, diagnostic), and endoscopist experience (1 year, 3 years, >10 years) were recorded. The number, size, location, morphology according to Paris classification, and histology (adenoma, advanced adenoma, hyperplastic) of the detected polyps were documented for each examination. Categorical variables were analyzed using Chi-square testing, while continuous variables were analyzed using the Mann–Whitney U test.

A total of 240 patients were included, of whom 55.4% were male, with a mean age of 58.93 ± 13.88 years; 136 procedures (56.7%) were performed with AI. There were no differences in sex and age of patients between groups. In the AI-assisted cases, the presence of adenomas was significantly higher in screening colonoscopies (16.3% vs 0%, p=0.018). There were no differences between the two groups regarding total polyp count, size, or morphology. Among endoscopists with 3 years of experience, AI significantly increased the detection of Paris Is polyps (40% vs 18.9%, p=0.047) and right-sided polyps (27.7% vs 8.1%, p=0.022). In the overall cohort, no significant differences were observed in the total number of polyps, size, morphology, or histology.

AI significantly increased adenoma detection in screening colonoscopies and improved the performance of intermediate-experience endoscopists, especially for Paris Is and right-sided lesions. While overall polyp burden and histological findings were similar between the two groups, these results indicate the selective benefit of AI-assisted colonoscopy in enhancing examination quality.