S1025 Progression of Barrett’s Esophagus and Dysplasia Diagnosed by Wide-Area Transepithelial Sampling With 3-Dimensional Computer-Assisted Analysis (WATS3D): Results of Long-Term Follow-Up

Introduction: WATS3D, used as an adjunct to forceps biopsy (FB) for Barrett’s esophagus (BE) sampling, increases the yield of dysplasia and intestinal metaplasia (IM). However, the risk of neoplastic progression in WATS3D-diagnosed BE is poorly understood. We aimed to evaluate the risk of progression to high-grade dysplasia (HGD) or esophageal adenocarcinoma (EAC) in the largest cohort to date of WATS3D-diagnosed columnar-lined esophagus without IM (non-IM CLE), nondysplastic BE (NDBE), indefinite (IND)/crypt dysplasia (CD), and low-grade dysplasia (LGD). Methods: This was a retrospective study of patients undergoing WATS3D from 2013-2024. Data were collected from the WATS3D commercial database. Patients with ≥1 cm CLE and at least 1 follow-up endoscopy ≥12 months from index WATS3D were included. Patients with prior dysplasia or endoscopic therapy for BE were excluded. The primary outcome was the rate of progression to HGD/EAC (detected by either WATS3D or FB) stratified by baseline WATS3D histology. Progression rate was reported as percentage/patient-year for each baseline histologic category. Results: Of 337,206 patients undergoing WATS3D, 30,782 met inclusion criteria. Overall, 76 progressed to HGD/EAC over a mean of 3.35 years (range, 1.0-17.92). Progression was detected in 37 patients by FB and 66 by WATS3D. The progression rate to HGD/EAC, as detected by FB, was 0.19%/patient year in NDBE, 1.0% in IND/CD, and 3.41% in LGD. In comparison, the progression rate as detected by WATS3D was 0.10%, 0.32%, and 1.86% in NDBE, IND/CD, and LGD, respectively. Differences in progression rates as detected by WATS3D and FB were statistically significant in those with baseline NDBE (P = 0.004) and baseline IND/CD (P = 0.02) but not in those with LGD (P = 0.43) or non-IM CLE (P = 0.16). Increasing age, male sex, White race, and longer BE segment were more common in progressors vs nonprogressors. Conclusion: In this largest reported cohort to date, patients with LGD detected by WATS3D progressed to BE/EAC at similar rates whether detected by FB or WATS3D, with rates comparable to those previously reported for FB-diagnosed LGD. Progression rates were somewhat lower in WATS3D-diagnosed NDBE and IND/CD when the outcome was measured by WATS3D compared to FB. Progression rates in non-IM CLE were miniscule. These data suggest that LGD diagnosed by WATS3D should trigger the same treatment as that recommended for FB-detected LGD.