408 WATS Brushings Are More Sensitive for Detecting Intestinal Metaplasia in Barrett’s Esophagus Than Concurrent Forceps Biopsies: Results From a National Registry

Abstract

INTRODUCTION: Forceps biopsies (FB) are used to sample the esophageal mucosa for intestinal metaplasia (IM) to diagnose BE, confirm eradication of BE, and detect recurrence after treatment. FB are limited by their focal field of sampling. Wide area transepithelial sample biopsy with three-dimensional computer-assisted analysis (WATS3D, CDx Diagnostics, Suffern, NY) samples a wider area of mucosa, but its sensitivity in routine clinical settings is incompletely understood. Our aim was to determine the sensitivity of WATS compared to FB for IM in patients undergoing surveillance of non-dysplastic BE (NDBE). METHODS: The WATS Registry is an ongoing multicenter U.S. study enrolling subjects undergoing WATS as part of usual care at 13 sites. Patients are aged 18 to 80 and are enrolled in one of three categories: screening endoscopy for gastroesophageal reflux symptoms without a history of BE, surveillance endoscopy for already-diagnosed BE without prior endoscopic eradication therapy, or surveillance endoscopy for BE following eradication therapy. We analyzed all endoscopy visits where subjects underwent both WATS and FB for surveillance of prevalent, treatment naïve NDBE with not less than 1 cm of columnar mucosa. Given that all these subjects previously had both endoscopic and histologic evidence of BE, FB or WATS not demonstrating IM was assumed to represent sampling error. We calculated the sensitivity of WATS and FB for IM, manifest by goblet cell metaplasia, with 95% asymptotic confidence limits. RESULTS: This analysis included 454 endoscopies of patients with a previous diagnosis of BE, presenting for surveillance and undergoing both FB and WATS. We observed a higher (p = 0.005) overall sensitivity of WATS (63.0% (95% CL 58.6 – 67.4%) than FB (53.7% (95% CL 49.2 – 58.3%)) of diagnosing IM (Figure). In addition to the patients with IM on FB, greater than 1/3 more subjects (36%, 87/244) had IM on WATS. The sensitivity of WATS for IM was similar across demographic groups (Table). Sensitivity of both FB and WATS increased with the length of BE being sampled. CONCLUSION: WATS has higher sensitivity for detecting IM compared to concurrent FB in patients with a previous history of BE. The addition of WATS to FB in these patients results in an overall additional 36% detection of IM. Given the included patients had previously demonstrated IM in visible segments of columnar epithelium in the esophagus, the incremental yield of WATS likely represents sampling error by FB.