Background: The U.S. incidence of esophageal adenocarcinoma (EAC) has increased >500% in the last four decades and overall 5-year mortality is approximately 80%. Barrett’s esophagus (BE), the only known precursor for EAC, can be effectively treated with endoscopic eradication therapies with up to 90% success rates. Although guidelines, including those by the American College of Gastroenterology (ACG), recommend BE screening in at-risk patients to potentially reduce EAC deaths, <10% of recommended patients undergo screening endoscopy (EGD). Recent guideline updates recommend nonendoscopic biomarker testing as an acceptable alternative to EGD to increase compliance. The EsoGuard^® Esophageal DNA Test (EG, Lucid Diagnostics, Inc.) assesses methylation at 31 CPG islands across two genes, vimentin (VIM) and CCNA1. Two case-control studies of EG, on cell samples collected using the EsoCheck^® Esophageal Cell Collection Device (EC, Lucid Diagnostics, Inc), an FDA-cleared, nonendoscopic balloon cell collection device, have reported high sensitivity for detection of BE (81%-88%) and EAC (88%-100%).
Aim: Determine performance characteristics of EG/EC compared to EGD for detection of BE in patients who meet ACG BE screening criteria.
Methods: 144 patients without a prior history of BE/EAC who met ACG BE screening criteria were enrolled at 11 centers in ESOGUARD BE-1(NCT04293458), a multicenter, single-arm study comparing EG/EC to EGD with biopsy at detecting BE/EAC. All EC samples underwent EG testing at a single CAP-accredited CLIA-certified central laboratory (LucidDx Labs, CA).
Results: The mean age was 60.5 ±7.7 yrs. and 53% of patients were white. 130/144 (90%) successfully completed EC cell collection, of which 129/130 (99.2%) also completed an EGD. Of the 129 patients with both EG and EGD results, 9 (7%) were EGD-positive for BE and none for EAC. 26/129 (20%) of EC samples did not have sufficient DNA for EG, 2 were unevaluable, and 7 did not undergo EG for other reasons. The remaining 94/129 (72.9%) had binary EG results reported. EG’s BE sensitivity was 88% (95% CI: 47%-99%) and BE specificity was 81% (95% CI: 72%-89%). Short segment BE (SSBE, <3cm) sensitivity was 86% (95% CI: 42%-100%) and long segment BE (LSBE, ≥3cm) sensitivity was 100%. EG’s PPV was 30% (95% CI: 13%-53%), and NPV was 98.6% (95% CI: 92.4%-99.9%).
Conclusion: EG demonstrates high sensitivity and specificity at detecting BE in a screening population based on ACG guidelines, replicating its performance in prior case control studies. These results demonstrate that non-endoscopic biomarker testing using EG/EC has the potential to dramatically improve BE screening compliance and disease detection, by triaging at-risk appropriate patients to UE.

Brian deGuzman

Appearances