Validation of Artificial Intelligence-based Bowel Preparation Assessment in Screening Colonoscopy

Abstract

Background and Aims

Accurate bowel preparation assessment is essential for determining colonoscopy screening intervals. Patients with suboptimal bowel preparation are at a high risk of missing >5mm adenomas, and should undergo an early repeat colonoscopy. In this study, we employed artificial intelligence (AI) to evaluate bowel preparation and validated the ability of the system in accurately identifying patients who are at high risk of missing >5mm adenoma due to inadequate bowel preparation.

Patients and methods

This prospective, single-center, observational study was conducted at the Eighth Affiliated Hospital, Sun Yat-sen University from October 8, 2021, to November 9, 2022. Eligible patients underwent screening colonoscopy were consecutively enrolled. The AI assessed bowel preparation using e-Boston Bowel Preparation Scale (BBPS) while endoscopists evaluated using BBPS. If both BBPS and e-BBPS deemed preparation adequate, the patient immediately underwent a second colonoscopy, otherwise the patient underwent bowel re-cleansing before the second colonoscopy.

Results

Among the 393 patients, 72 >5mm adenomas were detected, while 27 >5mm adenomas were missed. In unqualified-AI patients, the >5mm AMR was significantly higher than in qualified-AI patients (35.71% vs 13.19%, p=0.0056, OR 0.2734, 95% CI 0.1139, 0.6565), as were the AMR (50.89% vs 20.79%, p<0.001, OR 0.2532, 95% CI 0.1583, 0.4052) and >5mm PMR (35.82% vs 19.48%, p=0.0152, OR 0.4335, 95% CI 0.2288, 0.8213).

Conclusions

This study confirmed that patients classified as inadequate by AI showed unacceptable >5mm AMR, provided key evidence for implementing AI in guiding the bowel re-cleansing, potentially standardizing the future colonoscopy screening; ClincialTrials.gov, NCT05145712.