Automated, Augmented Reality High Resolution Microendoscopy for the Management of Esophageal Neoplasia

点击翻译按钮获取中文摘要

Project Summary: Despite advances in chemoradiation therapy, the five-year survival rate for esophageal squamous cell neoplasia (ESCN) remains a dismal 15% due to diagnosis at a late, incurable stage. Endoscopic screening is typically performed in high-risk populations with Lugol’s iodine staining of the mucosa and targeted biopsy of abnormal (unstained) areas. While Lugol’s significantly increases the sensitivity of white light endoscopy (>95%), specificity remains poor (<65%) as inflammation and other benign mucosal changes mimic neoplasia. While confocal microendoscopy has been shown to dramatically enhance the diagnostic accuracy and yield of Lugol’s chromoendoscopy, existing platforms are costly (>$150,000) and only available in a handful of tertiary centers worldwide. In an ongoing RO1 (2014-2019) of 1300 subjects in China and the USA, our group is evaluating a portable, battery-powered, high-resolution microendoscope that provides subcellular, 1100x-magnified images of the esophageal epithelium.This low-cost (< $1,500) device uses a Samsung Galaxy tablet computer to display microscopic epithelial images and image-analysis software to highlight abnormal nuclei allowing a more objective, real-time means of differentiating benign from neoplastic epithelium. In our current prospective, randomized, controlled trial of ESCN screening/surveillance, our goal is to determine whether this portable device increases the accuracy, diagnostic yield and clinical impact of routine endoscopic screening by providing a more objective, accurate, in vivo diagnosis that can direct a clinician’s decision to biopsy, not biopsy, or perform endoscopic therapy. The goal of this Revision R01 is to refine and validate a low-cost (HRME + glasses < $2,100) mobile, augmented reality (AR) platform that facilitates accurate, rapid, real-time endoscopic screening by non-expert clinicians worldwide. Using AR glasses and a small tablet computer (or cell phone), microendoscopic images with a quantitative, software interpretation (neoplasia vs. no neoplasia) will be displayed in real-time, colocalized with white light endoscopic images, allowing easier, more accurate biopsy and treatment.