The article from New Atlas discusses a significant advancement in the field of gastroenterology, where researchers have found that AI-assisted colonoscopies can greatly improve the detection of polyps by inexperienced doctors. This development is particularly important in reducing the chances of missing polyps, which are potential precursors to colorectal cancer.
The use of AI in medical diagnostics has been expanding, with applications in mammography, ultrasound, and MRI. The introduction of AI in colonoscopy represents a significant step forward. Colonoscopy is a crucial diagnostic tool for reducing colorectal cancer-related deaths by detecting and removing premalignant polyps, known as adenomas. However, the effectiveness of colonoscopy can be limited due to factors like flat morphology of adenomas, poor bowel preparation, and the endoscopist’s experience.
In a study conducted by the Chinese University of Hong Kong’s Faculty of Medicine, researchers used a computer-aided polyp detection (CADe) system, a deep learning AI, to assist junior endoscopists with less than 500 endoscopies and less than three years of training. The study involved 766 patients, with half receiving AI-assisted colonoscopies and the other half undergoing conventional procedures. The results showed a significantly higher adenoma detection rate (ADR) in the AI-assisted group compared to the control group. This improvement was observed in both beginner and intermediate-level endoscopists.
The researchers noted that the benefit of CADe for large and advanced adenomas is still unclear and suggested further optimization of the algorithm and development of computer-aided adenoma diagnosis systems. They advocate for incorporating AI devices into endoscopy training curricula, given the positive outcomes observed in the study.
This study, published in the journal Clinical Gastroenterology and Hepatology, highlights the potential of AI in enhancing the diagnostic accuracy of colonoscopies, especially for less experienced doctors, thereby improving patient outcomes in colorectal cancer prevention.