In a ground-breaking study, Yale University and Mass General Brigham, Harvard School of Medicine, have demonstrated  that artificial intelligence (AI)-assisted colonoscopy greatly improves the detection of colonic lesions over conventional colonoscopy. Analysing randomised controlled trials (RCTs), the study sought to find out whether artificial intelligence-powered polyp detection (CADe) systems would be more effective than conventional techniques in spotting polyps and precancerous growths during colonoscopy exams.

To evaluate important parameters between artificial intelligence-assisted and conventional colonoscopy, the researchers searched extensively across big scientific databases. These included secondary variables such as the adenoma detection rate (ADR), adenoma miss rate (AMR), and ACN detection rate, in addition to the average adenoma per colonoscopy (APC). While ACNs are more severe, advanced lesions with a larger chance of advancing to colorectal cancer, adenomas are polyps in the colon that might develop into cancer.

The study indicated that in identifying polyps and precancerous growths, AI-assisted colonoscopy systems beat conventional colonoscopy. Using artificial intelligence made it possible to screen more thoroughly and find polyps that could have escaped notice with more traditional approaches. However, the AI-assisted method only marginally improved when it came to identifying ACNs. The researchers came to the conclusion that AI-assisted colonoscopy did not significantly outperform the traditional method when it came to finding ACNs per colonoscopy. This means that both methods are almost equally good at finding these more serious lesions.

Another important finding from the study is that AI-assisted colonoscopy was most helpful for doctors who were not very good at finding adenomas or for patients who had not had a previous faecal immunochemical test. Fascinatingly, the researchers also discovered no appreciable difference in adenoma diagnosis among several artificial intelligence systems.

To better grasp the long-term advantages of artificial intelligence in colonoscopy, the research team underlined the need for more investigation. The research team recommended focusing future studies on interval post-colonoscopy colorectal cancer and implementing a study design that randomly assigns colonoscopists instead of patients to achieve more precise findings.

This study adds to the mounting body of data showing how artificial intelligence can raise early detection rates and improve the quality of cancer screenings—qualities essential for lowering colorectal cancer death.