Artificial intelligence-based assessments of colonoscopic withdrawal technique: a new method for measuring and enhancing the quality of fold examination

Abstract

Background This study aimed to develop an artificial intelligence (AI)-based system for measuring fold examination quality (FEQ) of colonoscopic withdrawal technique. We also examined the relationship between the system’s evaluation of FEQ and FEQ scores from experts, and adenoma detection rate (ADR) and withdrawal time of colonoscopists, and evaluated the system’s ability to improve FEQ during colonoscopy. Methods First, we developed an AI-based system for measuring FEQ. Next, 103 consecutive colonoscopies performed by 11 colonoscopists were collected for evaluation. Three experts graded FEQ of each colonoscopy, after which the recorded colonoscopies were evaluated by the system. We further assessed the system by correlating its evaluation of FEQ against expert scoring, historical ADR, and withdrawal time of each colonoscopist. We also conducted a prospective observational study to evaluate the systemʼs performance in enhancing fold examination. Results The system’s evaluations of FEQ of each endoscopist were significantly correlated with expertsʼ scores (r = 0.871, P < 0.001), historical ADR (r = 0.852, P = 0.001), and withdrawal time (r = 0.727, P = 0.01). For colonoscopies performed by colonoscopists with previously low ADRs (< 25 %), AI assistance significantly improved the FEQ, evaluated by both the AI system (0.29 [interquartile range (IQR) 0.27–0.30] vs. 0.23 [0.17–0.26]) and experts (14.00 [14.00–15.00] vs. 11.67 [10.00–13.33]) (both P < 0.001). Conclusion The system’s evaluation of FEQ was strongly correlated with FEQ scores from experts, historical ADR, and withdrawal time of each colonoscopist. The system has the potential to enhance FEQ.

Current and future implications of artificial intelligence in colonoscopy

“…Mucosal exposure is the most critical element in differentiating a high-quality colonoscopy, since only complete and accurate mucosal exposure can permit polyp detection, regardless of the use of a CADe system. Recently, a Chinese group developed an AI system for measuring “fold examination quality” during withdrawal in colonoscopy [ 46 ]. They compared the system’s evaluation of examination quality with the evaluation produced by expert endoscopists.…”

Withdrawal time in colonoscopy, past, present, and future, a narrative review

“…For example, studies have shown that finding an adenoma in a segment increases risk of synchronous or metachronous adenomas in the same segment (35)(36)(37). Therefore, finding an adenoma in a segment should lead to a more careful inspection of the same segment, a recommendation that can be reinforced with AI models similar to that used by Liu et al (38). The AI model will recommend relook in the right colon in high-risk population.…”

Targeting the low detector with artificial intelligence

“…Two studies in this issue of Endoscopy assessed the techniques of individual endoscopists by using AI as an objective benchmark [1,2]. In particular, the Liu et al study addressed the issue of mucosal exploration, and the Troya et al study investigated lesion spotting.…”

“…The circle of AI for quality of fold examination has been successfully closed by Liu et al by demonstrating that the assistance of the AI system was able to enhance the performance of low detectors but not of high detectors, underlining how AI can only improve what was defective in the first place, while its relevance for high detectors is likely to be marginal, if any [1].…”

“…By using AI, Liu at al. objectively measured the quality of fold examination of different endoscopists categorized according to their baseline ADR [1]. Briefly, the AI system measured the proportion of the examination in which the endoscopist achieved a wall view (i. e. exposure of colorectal mucosa to the lens of the scope) compared with a lumen view, which was considered clinically useless for the detection of colorectal neoplasia.…”