Impact of comprehensive optical diagnosis training using Workgroup serrAted polypS and Polyposis classification on detection of adenoma and sessile serrated lesion

New England Journal of Medicine, 370(14), 1298-1306

Objective The effect of colonoscopy on colorectal cancer mortality is limited by several factors, among them a certain miss rate, leading to limited adenoma detection rates (ADRs). We investigated the effect of an automatic polyp detection system based on deep learning on polyp detection rate and ADR. Design In an open, non-blinded trial, consecutive patients were prospectively randomised to undergo diagnostic colonoscopy with or without assistance of a real-time automatic polyp detection system providing a simultaneous visual notice and sound alarm on polyp detection. The primary outcome was ADR. Results Of 1058 patients included, 536 were randomised to standard colonoscopy, and 522 were randomised to colonoscopy with computer-aided diagnosis. The artificial intelligence (AI) system significantly increased ADR (29.1%vs20.3%, p<0.001) and the mean number of adenomas per patient (0.53vs0.31, p<0.001). This was due to a higher number of diminutive adenomas found (185vs102; p<0.001), while there was no statistical difference in larger adenomas (77vs58, p=0.075). In addition, the number of hyperplastic polyps was also significantly increased (114vs52, p<0.001). Conclusions In a low prevalent ADR population, an automatic polyp detection system during colonoscopy resulted in a significant increase in the number of diminutive adenomas detected, as well as an increase in the rate of hyperplastic polyps. The cost–benefit ratio of such effects has to be determined further. Trial registration number ChiCTR-DDD-17012221; Results.

Colonoscopy is commonly used to screen for neoplasia. To assess the performance of screening colonoscopy in everyday practice, we conducted a study of the rates of detection of adenomas and the amount of time taken to withdraw the colonoscope among endoscopists in a large community-based practice. During a 15-month period, 12 experienced gastroenterologists performed 7882 colonoscopies, of which 2053 were screening examinations in subjects who had not previously undergone colonoscopy. We recorded the numbers, sizes, and histologic features of the neoplastic lesions detected during screening, as well as the duration of insertion and of withdrawal of the colonoscope during the procedure. We compared rates of detection of neoplastic lesions among gastroenterologists who had mean colonoscopic withdrawal times of less than 6 minutes with the rates of those who had mean withdrawal times of 6 minutes or more. According to experts, 6 minutes is the minimum length of time to allow adequate inspection during instrument withdrawal. Neoplastic lesions (mostly adenomatous polyps) were detected in 23.5% of screened subjects. There were large differences among gastroenterologists in the rates of detection of adenomas (range of the mean number of lesions per subject screened, 0.10 to 1.05; range of the percentage of subjects with adenomas, 9.4 to 32.7%) and in their times of withdrawal of the colonoscope from the cecum to the anus (range, 3.1 to 16.8 minutes for procedures during which no polyps were removed). As compared with colonoscopists with mean withdrawal times of less than 6 minutes, those with mean withdrawal times of 6 minutes or more had higher rates of detection of any neoplasia (28.3% vs. 11.8%, P<0.001) and of advanced neoplasia (6.4% vs. 2.6%, P=0.005). In this large community-based gastroenterology practice, we observed greater rates of detection of adenomas among endoscopists who had longer mean times for withdrawal of the colonoscope. The effect of variation in withdrawal times on lesion detection and the prevention of colorectal cancer in the context of widespread colonoscopic screening is not known. Ours was a preliminary study, so the generalizability and implications for clinical practice need to be determined by future studies. 2006 Massachusetts Medical Society