Gastroenterology Research and Practice (Jan 2018)
Interobserver Agreement in the Diagnosis of Inflammatory Bowel Disease-Associated Neoplasia in China in Comparison to Subspecialized American Gastrointestinal Pathologists
Abstract
Background. The aim of this study was to evaluate the interobserver variability in diagnosing inflammatory bowel disease (IBD)-associated neoplasia among practicing pathologists from China using telepathology, a practice of remote diagnostic consultation increasingly used nationally and internationally, and its comparison with the interpretation of subspecialized gastrointestinal (GI) pathologists from the United States (US). Methods. Eight GI pathologists from the US and 4 pathologists from China with an interest in GI pathology participated in this study. A total of 50 colonic biopsies from patients with a clinical history of IBD from 8 medical centers in China were included. All microscopic slides in each case were digitized using an Aperio system. One pathologist (XL) reviewed the digitized full-slide images, and selected areas of interest were captured at low, medium, and high magnifications at a resolution of 1712 × 1072 pixels and saved as tagged image file format (TIFF) files on read-only DVD. Each pathologist evaluated the images and selected the most appropriate diagnostic category for each case (negative, indefinite, low-grade dysplasia [LGD], high-grade dysplasia [HGD], and carcinoma). A Fleiss’ kappa coefficient (K) analysis was performed to determine interobserver agreement and the agreement of each pathologist from China with the consensus diagnosis (defined as diagnostic agreement by at least 4 participating US GI pathologists). Results. There was substantial interobserver agreement among 4 pathologists from China on the interpretation of IBD-associated neoplasia (kappa value 0.68, 95% confidence interval: 0.56–0.78). A consensus diagnosis included negative (n=22), LGD (n=22), HGD (n=3), carcinoma (n=2), and indefinite for dysplasia (n=1). Using consensus diagnoses as references, the agreement between each pathologist from China and the consensus diagnosis was substantial with kappa values ranging from 0.75 to 0.80. Conclusions. This study reveals substantial interobserver agreement for the interpretation of colonic neoplasia in IBD using digitized images among Chinese pathologists as well as between each Chinese pathologist and a consensus diagnosis generated by US GI pathologists.