A Clinical Trial of Cancer Prevention by Biomarker Based Detections of Barrett's Esophagus and Its Progression

This Project aims to reduce mortality from Esophageal Adenocarcinoma (EAC) by pioneering two paradigm changing approaches. First, we will demonstrate the efficacy of non-endoscopic biomarker-based early detection of Barrett's esophagus (BE), in a population without symptoms of gastroesophageal reflux disease (GERD). Second, we will establish a novel molecular technology, dubbed “BAD”, for early detection and interception of BE progression toward EAC. BE is the precursor lesion of EAC, a cancer with 80% 5-year mortality. BE is currently detected only when individuals with GERD undergo endoscopic (EGD) screening. Once detected, BE patients undergo triennial surveillance EGD with random biopsies to catch progression to high-grade dysplasia (HGD), that can be eradicated endoscopically to prevent cancer. Weaknesses of the approach include: i) low acceptance of EGD screening among GERD patients; ii) the absence of any recommended screening among non-GERD patients, who fall completely outside of BE screening guidelines, but who account for 40% of EAC; frequent failure of random surveillance biopsies to detect early EAC, with many cancers arising in between surveillance exams and others already metastatic when detected. Our team developed a novel swallowable balloon-based device to enable targeted screening of the distal esophagus (the site of BE origin) in a simple 5minute office procedure. We mated this device with a methylated DNA biomarker panel that sensitively detects BE. In the GERD population we showed this approach detects non-dysplastic BE (NDBE) with 90% sensitivity and 92% specificity. We now find that non-GERD patients with 3 or more BE risk factors (age, obesity, smoking, male sex, white race) have BE risk similar to GERD patients. Aim 1 of this proposal will conduct a human trial demonstrating that our non-endoscopic biomarker-based technology will enable BE detection in this non-GERD population, with positive predictive value greater than that of current guidelines for EGD screening of GERD patients. Second, our group applied methods of deep DNA sequencing and AI analysis, developed for detecting cancer DNA in liquid biopsies, to instead detect abnormal DNA from nascent clones of progressed BE captured in esophageal brushings that comprehensively surveille the full BE disease segment. We showed this method, “BAD”, detects as “Very-BAD” 97% of EAC and 68% of HGD. We also showed Very-BAD identifies a 7% subset of non-dysplastic BE, that on retrospective review showed high risk of early progression to HGD or EAC. Aim 2 of this proposal will: i) implement improvements to the BAD methodology aimed at increasing sensitivity for HGD to 85%, while preserving specificity; ii) prospectively demonstrate that Very-BAD NDBE defines a population with high 3-year risk of BE progression to HGD; iii) implement BAD as a method to enable frequent non-endoscopic BE surveillance by adapting BAD to work with samples from our non-endoscopic balloon device. Last, molecular studies will determine the basis of false positive and false negative BE calls in non-GERD subjects, and will also visualize and molecularly interrogate the early progressed BE cells that are detected as Very-BAD.