Multispectral Imaging for Oral Cancer Detection

University Of Texas Md Anderson Can Ctr, Houston TX

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Oral cancer is a major health problem. Overall five-year-survival rates for all oral cancer patients are barely over 50%; for patients with advanced stage disease survival rates have remained at 30-40% or less for thirty years. Patients with early disease have better chances for cure and good functional outcome, but even in this country, most patients still present with advanced tumors. Early detection improves outcomes for oral cancer patients, yet clinical tools for early detection have limited sensitivity and specificity. Thus, identification of effective means of screening and early cancer diagnosis is needed to improve early detection efforts. The goal of this proposal is to develop a new technology, multispectral digital microscopic (MDM) imaging, for non-invasive detection of oral neoplasia. The MDM is a new type of digital microscope, which provides high-resolution color images of tissue. It can operate in two modes: (1) it can image the color of light reflected from tissue, and (2) it can be used to image the autofluorescence produced by tissue when illuminated with monochromatic light. Our group has shown that the process of carcinogenesis produces changes in the optical properties of epithelial cells and the supporting stroma. Our previous work has been limited to sampling small, biopsy-sized areas of tissue, which can be assessed with a pencil-sized fiber optic probe. In this proposal, we will test the hypothesis that changes in biochemistry and tissue morphology produced during carcinogenesis can be imaged in real time using multi-spectral digital microscopy. In Aim 1 we will compare the ability of MDM imaging to that of standard white light examination for assessment of oral lesions in a small clinical investigation. In Aim 2, we will assess the ability of MDM imaging compared to standard white light examination to delineate peripheral mucosal margins of oral carcinoma and dysplasia. In both trials, standard histopathologic evaluation will be used as the gold standard. In Aim 3, we will develop preliminary diagnostic algorithms based on our results using MDM imaging, and determine the sensitivity and specificity of these algorithms to non-invasively distinguish dysplasia and early carcinoma from benign lesions and normal mucosa. In addition, we will determine the optimal wavelength combinations to be tested in future large-scale clinical trials. Successful completion of this research will determine the feasibility for MDM imaging as a clinical tool to improve detection and monitoring of oral neoplasia.