BE images with the corresponding targeted biopsy specimen were

retrieved. Images of the highest quality were selected by an investigator (who did not participate in the assessment); histology was available for all selected images. A 1-hour structured interactive teaching session was conducted with 6 endoscopists by using 8 AFI/NBI images (teaching set). All images for the interobserver agreement study were evaluated by 6 endoscopists, 3 experts (faculty members) and 3 nonexperts (trainees). Hormones antagonist The senior author of this article (P.S.), who performed all of the endoscopies of CX-5461 mouse the study, did not participate in the interobserver study. To assess interobserver agreement, an independent testing set of images was inserted in a PowerPoint presentation with a black background. Each image was numbered, and the slides were randomized in a computer-generated fashion with the image number displayed during the PowerPoint slide show. The 6 endoscopists were then given the testing set and

were asked to review each image. The following information was collected: 1 Patterns: AFI, abnormal purple fluorescence/normal rest of the areas; magnification NBI: regular/irregular All reviewers were blinded to the patient’s medical history, histological diagnosis, and other imaging data. Patient demographics, endoscopy findings (location of landmarks, BE length, length of hernia, AFI and NBI patterns), biopsy, and histology reports were collected and recorded

in a standardized case report form. Details of subjects enrolled in the study were then entered into a central database. The collected data included (1) demographics: age, sex, race, and body mass index; (2) endoscopic details: the date of procedure, presence of hiatal hernia, presence of visible not lesions, BE length by using Prague C&M8; (3) histological diagnosis of all of the areas analyzed in the BE mucosa; and (4) highest grade of histological diagnosis during the procedure (overall dysplasia). For the final analysis in this study, AFI and magnification NBI evaluation of only the flat mucosa away from the visible lesions was considered. All lesions detected with AFI, NBI, and random biopsies were regarded as abnormal areas with AFI and/or NBI and normal areas if detected with random biopsies only. Areas that were indefinite for dysplasia were included under LGD during data collection and analysis.