Abstract
BACKGROUND & AIMS: Guidelines differ on the definition of Barrett's esophagus (BE). We aimed to estimate the detection rate of esophagogastroduodenoscopy (EGD) with biopsy for BE, using longitudinal outcome of adenocarcinoma of the esophagus or esophagogastric junction (EAC, EGJAC) as the gold standard. METHODS: We performed retrospective analyses of US Veterans with EAC/EGJAC between 2017 and 2021 who had an EGD before cancer diagnosis. We reviewed a random sample of 200 cases of EAC and 100 cases of EGJAC to tabulate the detection rate of EGD with biopsy, stratified by Siewert classification (S1 or S2). RESULTS: After manual review, there were 136 S1 and 108 S2 cases. Among patients with endoscopic suspicion of BE who had biopsies, 96.8% of S1 and 97.9% of S2 had intestinal metaplasia (IM). BE defined as ≥1 cm with IM had a detection rate of 66.2% for S1 and 39.8% for S2. When restricted to EGDs with appropriate biopsies, the detection rate of ≥1 cm with IM improved to 70.3% in S1 and 41.4% in S2. Seventy-three percent of S1 cases and 90% of S2 cases without BE diagnosed were related to lack of endoscopic suspicion for BE. BE or a visible lesion was present in 70% of photographs from EGDs not suspected of having BE by the original endoscopist. CONCLUSIONS: IM seems required for development of EAC and likely EGJAC. EGD with biopsy had only modest rate for detecting a precursor state for subsequent adenocarcinoma and was lower for S2 compared with S1. Variability in endoscopic suspicion, including segments <1 cm, are potential areas for improvement in detection of BE by EGD.