Enhancing the Positive Predictive Value of EGD for Diagnosis of Barrett's Esophagus Through EsoGuard® Triage

Background: Guidelines support Barrett's esophagus (BE) screening, but most eligible patients do not undergo endoscopic evaluation; non-endoscopic strategies are now supported as a reasonable alternative by U.S gastroenterology societies. EsoGuard (EG) is a DNA assay used with EsoCheck, a non-endoscopic cell collection device for detection of BE, which can be utilized as a triage to esophagogastroduodenoscopy (EGD) in patients meeting screening criteria. In doing so, EG may serve to enrich the population undergoing EGD, resulting in more BE diagnoses while potentially reducing utilization of already-limited endoscopy resources. Aim: To test the hypothesis that BE detection in EGDs performed on EG positive patients will be significantly higher than the positive predictive value (PPV) of screening EGD alone. Methods: Real-world data was retrospectively collected from EG positive patients for whom EGD diagnoses were available. Baseline patient characteristics, risk factors, and EGD results were obtained from the treating physicians. PPV of screening EGDs was the comparator and estimated by literature-established disease prevalence of BE, which in the U.S gastroesophageal reflux disease population is ~10.6%. The hypothesis was tested using t-tests for single proportions at a one-sided 5% significance level. Results: Data from 209 patients found 60 (28.7%) subjects with salmon-colored mucosa on EGD and specialized intestinal metaplasia on histopathology. However, 10 (4.8%) had < 1cm of disease on visual inspection, therefore, did not meet the American College of Gastroenterology definition of BE so was excluded from the analysis. Of the remaining 199 patients, 50 (25.1%) had BE on EGD. In the cohort of patients meeting ACG screening criteria, 28.9% (33/114) had BE. Overall, a 2.4-fold increase in BE detection was observed compared to the PPV of screening EGD, and in the ACG cohort this increase was 2.7-fold. Among ACG patients ≥65 years old, the increase was nearly 2.5-fold (25.9% detection rate). Conclusions: Our data suggests EG and EC used as a triage test enriches the population undergoing EGD for BE, and compared to screening EGD alone, can help direct more efficient use of endoscopy resources to unburden the system without reducing the number of eligible patients screened and diagnosed.