Patients with Barrett's esophagus are at risk of developing esophageal adenocarcinoma, a cancer that is rapidly increasing in incidence in the United States and other countries. Practice guidelines recommend periodic endoscopic surveillance with multiple random biopsies of the Barrett's segment primarily to detect critical lesions of high-grade dysplasia and early adenocarcinoma at a time when intervention can be curative. However, this approach can be labor-intensive, time-consuming, and hindered by low sampling yield, inconsistent histological interpretation, pathology-related costs and delay in diagnosis. As an adjunct to gastrointestinal (GI) endoscopy, optical spectroscopy offers the opportunity to enhance the surveillance of Barrett's esophagus by identifying tissue pathology in a rapid, non-intrusive and real-time manner in vivo. Among the optical spectroscopic techniques being evaluated for tissue diagnosis, near-infrared Raman spectroscopy (NIRS) is particularly promising because of the wealth of molecular information contained in its spectra. Recently, a technological breakthrough was achieved and the in vivo feasibility of measuring Raman spectra of GI tissues during endoscopy was demonstrated, using a fiber-optic NIRS system developed by our group. Our long-term objective is to develop NIRS as a novel complementary technique to endoscopy for enhancing the surveillance of patients with Barrett's esophagus. In this proposal, we plan to assess the diagnostic performance of NIRS, coupled with multivariate statistical methods of spectral analysis, for identifying Barrett's esophagus with high-grade dysplasia and early adenocarcinoma. Patients scheduled to undergo endoscopy in our specialized Barrett's Esophagus Unit will be offered participation in the study. Barrett's sites that are interrogated with the fiber-optic Raman probe will immediately be biopsied for corresponding histologic diagnoses. The first specific aim will consist of analyzing a set of in vivo Raman spectra using multivariate statistical techniques in order to develop robust diagnostic algorithms (spectral classifiers) that separate high-risk Barrett's tissues (high-grade dysplasia and early adenocarcinoma) from low-risk Barrett's tissues (no dysplasia, indefinite for dysplasia, and low-grade dysplasia). The second specific aim will consist of evaluating the developed algorithms in a prospective fashion in order to assess sensitivity and specificity of the Raman system and spectral classifiers for identifying high-risk Barrett's esophagus. Should NIRS prove to be reliable in assessing the neoplastic potential of Barrett's mucosa, the technique would represent a welcome addition to endoscopy and a significant improvement in the management of patients with Barrett's esophagus.
