ABSTRACT
BACKGROUND AND AIM: It is unclear whether second-generation narrow-band imaging (NBI) improves colorectal adenoma detection in clinical practice. We aimed to evaluate the ability of NBI to detect adenomas in academic and community hospitals. METHODS: This observational, multicenter study was conducted in four academic and four community hospitals between July 2018 and April 2019. We enrolled patients aged ≥ 20 years who underwent colonoscopy for screening, polyp surveillance, or diagnostic workup. The primary endpoint was the adenoma detection rate (ADR) between NBI (NBI group) and white-light imaging colonoscopies (WLI group) after propensity score (PS) matching. RESULTS: Of 1831 patients analyzed before PS matching, the NBI and WLI groups included 742 and 1089 patients, respectively. After PS matching, 711 pairs from both groups were analyzed. ADR and the mean number of adenomas per patient did not differ significantly between the NBI and WLI groups (43.5% vs 44.4%, P = 0.71; 0.90 ± 1.38 vs 0.91 ± 1.40, P = 0.95, respectively). Academic hospitals showed higher ADR in the NBI group (60.5% vs 53.8%), whereas community hospitals showed higher ADR in the WLI group (35.8% vs 40.5%). In the NBI group, ADR was significantly higher among NBI-screening-experienced endoscopists than among NBI-screening-inexperienced endoscopists (63.2% vs 39.2%, P < 0.001). The mean number of flat and depressed lesions detected per patient was significantly higher with NBI than with WLI (0.62 ± 1.34 vs 0.44 ± 1.01, P = 0.035). CONCLUSIONS: Second-generation NBI could not surpass WLI in terms of ADR based on patient recruitment from both academic and community hospitals but improved the detection of easily overlooked flat and depressed lesions.