Wide Area Transepithelial Sampling With Computer-Assisted 3-Dimensional Analysis (WATS3D) and Forceps Biopsies Should Both Be Obtained to Maximize Detection of Intestinal Metaplasia Following Ablation of Barrett’s Esophagus

Abstract

Introduction: Barrett’s esophagus (BE) is a precancerous condition characterized by esophageal intestinal metaplasia (IM). Endoscopic ablation is performed to decrease risk of progression to adenocarcinoma. Post-ablation surveillance is accomplished using the modified Seattle protocol, where 4-quadrant forceps biopsies (FB) are taken at least every 2 cm throughout the original BE segment. Prior studies have demonstrated the benefit of adjunctive use of WATS3D with FB to improve detection of residual or recurrent IM following endoscopic ablation. Our aim was to investigate discordance between post-ablation WATS3D and FB and determine demographic and endoscopic factors that predict disparate results. Methods: Patients undergoing follow-up endoscopy after BE ablation between June 2012 and May 2014 were evaluated. If there was no visual evidence of BE, WATS3D biopsies were obtained using the standard 2-brush technique (CDx Diagnostics, Suffern, NY). The squamocolumnar junction was sampled with its own brush kit, followed by separate brushings from the tubular esophagus for all patients with an original BE length of at least 2 cm. Samples were analyzed by a central laboratory using a neural network to identify IM. Four-quadrant FB were obtained during the same endoscopy and reviewed by an expert GI pathologist. Results from both data sets were compared. Statistical analysis using two-tailed T-tests was performed using patient demographic and endoscopic data. Results: Post-ablation biopsies were taken in 114 cases (78% male, mean age 62 years). Diagnostic agreement between WATS3D and FB was seen in 89 cases (78%). Of the 31 cases in which IM was detected, only 6 had positive results on both WATS3D and FB. Fifteen cases were detected by FB alone and 10 cases by WATS3D alone. Therefore, the adjunctive yield of WATS3D was 47.6% (10/21). Analysis demonstrated no statistically significant difference between patients with diagnostic agreement versus discordance with respect to age, gender, smoking status or mean number of ablative treatments to visually eradicate BE. However, mean hiatal hernia size was substantially larger when IM was detected only with WATS3D (3.6 cm versus 2.6 cm, p=0.08). Conclusion: Residual or recurrent IM detected following BE ablation was found with both WATS3D and FB in only 19.4% of cases. This high degree of discrepancy when IM persists underscores the importance of using both techniques to survey normal-appearing post-ablation mucosa. WATS3D seems to be most helpful when sampling in the presence of larger hiatal hernias. This may be due to a greater luminal circumference and/or reduced inspection capability. Further improvements in WATS3D brush technology and technique will only augment its benefit in post-ablation surveillance. Disclosure - Dr. Smith - Advisory Board Member and Research Support: CDx Diagnostics.