Increased Detection of Barrett’s Esophagus and Dysplasia In Community Gastroenterology Practices Resulting From the Addition of Computer-Assisted Transepithelial Brush Biopsy to Forceps Biopsy

Abstract

Introduction: Barrett’s esophagus (BE) is a known precursor to esophageal cancer. Current surveillance strategies include random four quadrant forceps biopsies (FB) of a known segment of BE every 1-2 cm. Wide Area Transepithelial Sampling with computer-assisted 3-dimensional tissue analysis (WATS3D) previously has been shown as an effective adjunctive tool to FB for detection of BE and dysplasia. A previous study in community practices demonstrated that adding WATS3D to FB increased BE detection in all patients from 15% to 25% and dysplasia from 0.68% to 1.12%, for adjunctive yields of 68% and 65%, respectively. The aim of this study is to further investigate the added benefit of WATS3D in routine use in community gastroenterology practices. Methods: A prospective registry involving 19 gastroenterologists was established to follow patients undergoing screening or surveillance endoscopy with both FB and WATS3D. Data points collected included demographic information, detection of BE and dysplasia, and the order FB and WATS3D were performed during endoscopy. FB and WATS3D samples were sent together to a central laboratory. WATS3D samples were analyzed using a neural network to identify goblet cell metaplasia and dysplasia. FB samples underwent standard histologic review. Both sample sets were evaluated by trained GI pathologists. De-identified data was aggregated for analysis. Results: A total of 1655 patients were enrolled in the study. The average age was 59 years (range 16-92), and 43% of patients were male. The indication for endoscopy was GERD in over 75% of patients. In cases with suspected BE, segment length was less than 3 cm. FB alone identified BE in 203 cases (12%), and dysplasia in 9 cases (0.54%). Adjunctive use of WATS3D identified an additional 203 cases of BE, increasing the yield to 25%. WATS3D also detected another 12 cases of dysplasia, increasing the yield to 1.26%. These data demonstrate added yields of 110% for BE and 133% for dysplasia. In a subset of 551 patients, the normal sequence of WATS3D being performed prior to FB was reversed. There was no statistically significant difference in added yield found as a result of biopsy technique order. Conclusion: The increased detection yields from 12% to 25% for BE and from 0.54% to 1.26% for dysplasia found in this large study are highly consistent with results of previous WATS3D studies performed in community-based gastroenterology practices. These data further demonstrate that widespread adjunctive use of this sampling technique during upper endoscopy can be expected to significantly increase the detection of both BE and dysplasia, leading to improved patient care. Disclosure - Seth Gross - advisory board; Vivek Kaul - advisory board/research support; Michael Smithadvisory board/research support.