S945 A Community Center's Experience With WATS3D and the Use of Artificial Intelligence for Diagnosis and Surveillance of Barrett's Esophagus

Introduction: Barrett's esophagus (BE), the change of stratified squamous epithelium to columnar epithelium with intestinal metaplasia in the distal esophagus, is a recognized risk factor for the majority of patients diagnosed with esophageal adenocarcinoma (EAC). Current recommendations for BE surveillance suggest an upper endoscopy (EGD) with Seattle Protocol biopsies every 3-5 years for patients with no dysplasia, shorter intervals for those with low-grade (LGD) and high-grade dysplasia patients. Despite these guidelines, the incidence of EAC in the United States continues to increase. Traditional methods often miss areas of dysplasia due to limited sampling. This study evaluates wide-area transepithelial sampling with computer-assisted 3-dimensional analysis (WATS3D) in conjunction with standard tissue sampling for BE during EGD. Methods: All patients of this study gave informed consent and were enrolled at the Midwest Center for Day Surgery between May 2021 to July 2024. Patients were evaluated by 4 board-certified gastroenterologists with extensive experience in evaluation of BE. Of the 1,000 evaluated patients, 502 patients qualified and enrolled in the study (55% men, 45% women, mean age 61.3 years). All patients underwent standard Seattle Protocol surveillance of BE with 4 quadrant biopsies every 1-2 cm and WATS3D brushings. Biopsies were analyzed by experienced gastroenterology pathologists, and brush biopsies were sent to CDx Diagnostics’ central lab. Clinical, endoscopic, and pathological data were collected from electronic health records. The National Clinical Trial number is NCT04312633. Results: Results showed that 30.7% of patients (154) had non-dysplastic Barrett’s esophagus on biopsy alone, while WATS3D detected an additional 13.9% patients (70), improving diagnostic yield. WATS3D also identified 2 patients with LGD and 5 patients with indefinite dysplasia, significantly increasing detection (↑ yield 100%, ↑ yield 250%, respectively). Conclusion: The results of our study suggest WATS3D is an effective adjunct to standard Seattle Protocol biopsies, enhancing BE detection by 13.9% and identifying a subset of patients with advanced BE that might otherwise be missed. The use of forceps alone would have led to undetected BE, risking the patient not being appropriately screened, and possibly increasing risk for EAC. While further studies are needed, this study suggests that WATS3D can significantly increase BE detection and surveillance across practice settings.