Quality of optical diagnosis of diminutive polyps and associated factors

Abstract

The aim of the study was to identify endoscopist-related and procedural factors that may be associated with the quality of optical diagnosis of diminutive polyps using narrow-band imaging (NBI). All subjects who participated in a randomized trial on cap-assisted colonoscopy were eligible for the current study. Optical polyp diagnosis was an a priori outcome of the initial trial. Ten participating endoscopists used NBI to assess all of the diagnosed polyps as adenomatous or non-adenomatous in real-time and provided a degree of diagnostic certainty. The main outcome measures were quality benchmarks of optical diagnosis (negative predictive value [NPV] for diminutive rectosigmoid adenomas, agreement with pathology-based surveillance interval) and assessment of endoscopist-related and procedural factors potentially associated with the quality of optical diagnosis. A total of 1650 polyps were found in 607 patients, with 1311 polyps (79 %) being diminutive, of which 672 (53 %) were adenomatous. The NPV of optical diagnosis for rectosigmoid adenomas was 95 %. The optical diagnosis-based surveillance interval agreed with the pathology-based recommendation in 93 % of patients. Prior experience with image-enhanced endoscopy had no effect on optical diagnosis. Low and high adenoma detectors were not different in achieving the quality benchmarks. Cap-assisted colonoscopy was not associated with quality of optical diagnosis. Quality metrics of optical diagnosis remained similar during the first and second half of the study period. High quality optical diagnosis of diminutive polyps can be achieved and sustained by endoscopists previously inexperienced in this practice with minimal training. None of the examined factors appear to affect the quality of optical diagnosis; particularly, endoscopists' adenoma detection was not associated with optical diagnosis.