Correlation Between Artificial Intelligence-Based Fecal Coverage Area-Quantified Bowel Preparation Quality Score (FQ-BBPS) and Adenoma Detection Rate

Overview

A single-center prospective observational study evaluates the correlation between an AI-based bowel preparation quality score (FQ-BBPS\[25%\]), which quantifies fecal coverage area during colonoscopy withdrawal, and the adenoma detection rate (ADR) in screening/surveillance patients.

The FQ-BBPS system uses image segmentation to automatically calculate the percentage of mucosa obscured by feces/fluid, providing a continuous, objective alternative to the conventional Boston Bowel Preparation Scale (BBPS). Preliminary retrospective data showed a strong correlation (Spearman ρ = -0.829) between FQ-BBPS(25%) and ADR, and adding this metric improved adenoma prediction over e-BBPS alone. This prospective study (October 2026 - September 2027) will consecutively enroll patients aged 45-75 years undergoing colonoscopy for screening or polyp surveillance, excluding those with prior colorectal surgery, cancer, alarm symptoms, anticoagulant use, or recent trial participation. With a conservative assumed correlation of -0.70, a sample size of 560 (allowing 10% dropout) is planned. The primary outcome is ADR (proportion with ≥1 histologically confirmed adenoma), and secondary outcomes include detection rates for serrated lesions, advanced adenomas, and polyp detection by size/location; safety endpoints (perforation, immediate/delayed bleeding) will also be recorded. The FQ-BBPS(25%) score will be computed automatically from withdrawal videos, while endoscopists record BBPS and polyp characteristics. Spearman correlation will assess the association, and ROC analysis will explore the optimal cutoff against the 35% ADR reference. The study adheres to the Declaration of Helsinki, with ethics approval and written informed consent from all participants.

Conditions

Eligibility

Locations (1)

Outcomes

Central Contacts