Brief Article
Copyright ©2013 Baishideng Publishing Group Co., Limited. All rights reserved.
World J Gastroenterol. Dec 14, 2013; 19(46): 8652-8658
Published online Dec 14, 2013. doi: 10.3748/wjg.v19.i46.8652
Assessment of the diagnostic performance and interobserver variability of endocytoscopy in Barrett’s esophagus: A pilot ex-vivo study
Yutaka Tomizawa, Prasad G Iyer, Louis M Wongkeesong, Navtej S Buttar, Lori S Lutzke, Tsung-Teh Wu, Kenneth K Wang
Yutaka Tomizawa, Prasad G Iyer, Louis M Wongkeesong, Navtej S Buttar, Lori S Lutzke, Kenneth K Wang, Barrett’s Esophagus Unit, Division of Gastroenterology and Hepatology, Mayo Clinic Rochester, MN 55905, United States
Tsung-Teh Wu, Department of Laboratory Medicine and Pathology, Mayo Clinic Rochester, MN 55905, United States
Author contributions: Tomizawa Y and Wang KK contributed to the Study design, data analysis, and manuscript preparation; Iyer PG, Wongkeesong LM and Buttar NS contributed to the endocytoscopy interpretation; Wang KK contributed to the endoscopic treatment and reviewer of the paper; Wu TT contributed to the histopathology assessment; Tomizawa Y and Lutzke LS contributed to the data collection.
Correspondence to: Kenneth K Wang, MD, Professor, Barrett’s Esophagus Unit, Division of Gastroenterology and Hepatology, Mayo Clinic Rochester, 1st Street SW, Rochester, MN 55905, United States. wang.kenneth@mayo.edu
Telephone: +1-507-2842174 Fax: +1-507-2557612
Received: May 1, 2013
Revised: October 7, 2013
Accepted: October 13, 2013
Published online: December 14, 2013
Abstract

AIM: To investigate a classification of endocytoscopy (ECS) images in Barrett’s esophagus (BE) and evaluate its diagnostic performance and interobserver variability.

METHODS: ECS was applied to surveillance endoscopic mucosal resection (EMR) specimens of BE ex-vivo. The mucosal surface of specimen was stained with 1% methylene blue and surveyed with a catheter-type endocytoscope. We selected still images that were most representative of the endoscopically suspect lesion and matched with the final histopathological diagnosis to accomplish accurate correlation. The diagnostic performance and inter-observer variability of the new classification scheme were assessed in a blinded fashion by physicians with expertise in both BE and ECS and inexperienced physicians with no prior exposure to ECS.

RESULTS: Three staff physicians and 22 gastroenterology fellows classified eight randomly assigned unknown still ECS pictures (two images per each classification) into one of four histopathologic categories as follows: (1) BEC1-squamous epithelium; (2) BEC2-BE without dysplasia; (3) BEC3-BE with dysplasia; and (4) BEC4-esophageal adenocarcinoma (EAC) in BE. Accuracy of diagnosis in staff physicians and clinical fellows were, respectively, 100% and 99.4% for BEC1, 95.8% and 83.0% for BEC2, 91.7% and 83.0% for BEC3, and 95.8% and 98.3% for BEC4. Interobserver agreement of the faculty physicians and fellows in classifying each category were 0.932 and 0.897, respectively.

CONCLUSION: This is the first study to investigate classification system of ECS in BE. This ex-vivo pilot study demonstrated acceptable diagnostic accuracy and excellent interobserver agreement.

Keywords: Endocytoscopy, Barrett’s esophagus, Dysplasia, Esophageal adenocarcinoma, Interobserver agreement

Core tip: The current gold standard for surveillance of esophageal adenocarcioma in Barretts’s esophagus (BE) is endoscopic random biopsy and pathological diagnosis. Endocytoscopy (ECS) has the potential to provide a virtual histological diagnosis in vivo and in real-time. However, a major issue relates to that interpretation of cellular and nuclear images may be subject to similar interobserver variability associated with conventional histopathological diagnosis, and there have been no reliable classification systems for the endocytoscopic diagnosis. We presented the first study to investigate classification system of ECS in BE. This ex-vivo pilot study demonstrated acceptable diagnostic accuracy and excellent interobserver agreement.