Directional Bias in Polyp Detection during Colonoscopy: A Prospective Observational Study.

INTRODUCTION

Colonoscopy is essential for colorectal cancer prevention, yet a proportion of lesions remain undetected, contributing to interval cancers. Although adjunctive technologies can improve adenoma detection, the role of inherent visibility differences across the endoscopic field has not been fully clarified. This study evaluated whether polyp detection varies according to the short-axis direction of the colonic lumen (SADCO) during colonoscopic withdrawal.

METHODS

In this prospective observational study conducted at a tertiary center between June 2024 and February 2025, the short-axis direction of initial lesion visualization was recorded using the short-axis directional classification (SADCO). The primary outcome was the distribution of lesion detection between the upper-left direction (ULD) and lower-right direction (LRD). Subgroup analyses, multivariable logistic regression, and post hoc quadrant analysis were performed to assess factors associated with ULD detection.

RESULTS

A total of 1,185 lesions from 432 patients were analyzed. Detection in the ULD was significantly lower than in the LRD (42.4% vs. 57.6%; p < 0.001), and this pattern was consistent across lesion size, bowel segment, macroscopic type, and endoscopist experience. Detection of advanced adenomas tended to be lower in the ULD (41.4% vs. 58.6%; p = 0.071), and detection of advanced neoplasia was significantly reduced (40.7% vs. 59.3%; p = 0.043). On multivariable analysis, a history of colonoscopy within 3 years was independently associated with increased ULD detection (p = 0.040). Quadrant analysis confirmed the lowest detection rate in the upper-left quadrant.

CONCLUSION

Polyp detection during colonoscopic withdrawal demonstrates a consistent directional bias, with the ULD functioning as a region of relatively reduced detection frequency. Awareness of this intrinsic limitation and deliberate inspection of the ULD may help optimize visual inspection patterns during colonoscopy.