Mo1523 A Japanese Multicenter Randomized Controlled Trial Using Narrow-Band Imaging for Detection of Colorectal Adenomas

Narrow-band imaging (NBI) provides unique views especially of the mucosal vascular network and facilitates improved visualization of neoplasia by enhancing contrast. One reason for the development of NBI was to increase colorectal adenoma detection compared to conventional white light imaging (WLI). Previous studies have yielded various conflicting results regarding improvement in colorectal adenoma detection rates using NBI, but most reports were single center studies or used a variety of different NBI settings. The aim of this multicenter randomized controlled trial was to investigate the potential of NBI for detecting colorectal adenomatous lesions. Our randomized controlled trial using back-to-back colonoscopy examinations and recommended NBI settings to resolve this issue was performed exclusively in six multi-referral centers in Japan. Back-to-back colonoscopies were conducted in the right side of the colon including the cecum, ascending and transverse colon of patients whose first colonoscopy was with either NBI or WLI followed immediately by a second colonoscopy using the other imaging method (Group A, NBI-WLI; Group B, WLI-NBI). We used the sequential EVIS LUCERA SPECTRUM (Olympus Co., Tokyo, Japan) video endoscopic system and a CF-H260AZI (Olympus) conventional high definition colonoscope with enhanced image function settings (surface structure enhancement, A-5; adaptive index of hemoglobin color enhancement, 3) for all examinations. Times were recorded for both insertion and withdrawal examination. All detected polyps were removed endoscopically for histopathological analysis. The primary outcome measure was the adenoma detection rate at the time of first observation and the secondary outcome measure was the adenoma missed rate using either NBI or WLI. A total of 406 patients were randomly assigned to Group A and 407 patients to Group B with 17 Group A patients and 14 Group B patients excluded for several reasons including inadequate bowel preparation and colons that were too dark because of melanosis. There were no significant differences between the two groups in terms of patient age, gender or withdrawal time. The total number of adenomatous lesions detected by NBI and WLI at the time of first observation was 165 (42.4% of 389 Group A patients) and 167 (42.5% of 393 Group B patients), respectively (p=0.98). The missed rate for all adenomatous lesions of 21.1% using NBI was significantly lower than the missed rate of 27.8% with WLI (p=0.03). This randomized controlled trial in six Japanese multicenter referral centers did not indicate any objective advantage of NBI over WLI in terms of improved adenomatous lesion detection, but use of NBI in the proximal colon significantly reduced the missed rate for adenomatous lesions in comparison to WLI.