Use of AI in detection of colorectal neoplasia during colonoscopy
29 Aug 2023 • This study aimed to assess the effects of using artificial intelligence computer-aided detection (CADe) during colonoscopy for detecting colorectal neoplasia. The goal was to determine whether CADe increases adenoma detection rates (ADRs) and reduces adenoma miss rates, but also whether it leads to overdiagnosis and overtreatment of nonneoplastic polyps.
The researchers conducted a systematic review and meta-analysis of randomized trials. They included 21 trials involving 18,232 patients. The main findings were as follows:
- The CADe group had a higher ADR compared to the standard colonoscopy group (44.0% vs. 35.9%), resulting in a relative risk of 1.24. This suggests that CADe increases the detection of adenomas.
- The use of CADe led to a 55% relative reduction in adenoma miss rates, indicating its effectiveness in reducing missed adenomas during colonoscopy.
- More nonneoplastic polyps were removed in the CADe group (0.52 per colonoscopy) compared to the standard group (0.34 per colonoscopy), resulting in a mean difference of 0.18 polypectomy.
- There was a marginal increase in mean inspection time with CADe (0.47 minute).
In conclusion, the utilization of CADe during colonoscopy enhances the detection of adenomas and decreases the miss rate for adenomas. However, it also leads to increased removal of nonneoplastic polyps, which might not necessarily be beneficial. The study highlights both the benefits and potential drawbacks of using CADe in colonoscopy for polyp detection. Source: Annals of Internal Medicine | Read full story