Label-free in vivo pathology of human epithelia with a high-speed handheld dual-axis confocal microscope

Abstract

Oral cancer is one of the most common malignancies worldwide, with an approximate global prevalence of 300,000 individuals in 2012. Roughly 50,000 new cases and over 10,000 deaths are attributed to this disease each year in the United States. Dentists and head & neck physicians are trained to recognize aberrations in gross morphology that are indicative of malignancy, as well as to utilize a number of wide-field imaging techniques to improve the visualization of lesions, some of which can detect disease with high sensitivity but poor specificity. As a result of poor diagnostic specificity, the majority of suspected lesions, when biopsied and analyzed via histopathology, reveal benign conditions rather than premalignant or malignant lesions. In addition to the high costs and time associated with obtaining large numbers of unnecessary tissue samples for pathological analysis, the need for an invasive biopsy often results in patient discomfort, noncompliance, complications and/or diagnostic delays. Reflectance confocal microscopy can potentially provide a real-time non-invasive method to triage and guide excisional biopsy. However, miniaturization is a challenge and previous in vivo systems have necessitated trade-offs between critical parameters such as device size, imaging speed (frame rate), imaging depth, resolution, and field-of-view. In this thesis project, a fully packaged miniature line-scanned reflectance microscope, which combines a dual-axis confocal architecture with MEMS-based beam scanning, was built to achieve an optimized clinical device for real-time micropathological detection of oral lesions. This handheld device, with a tip diameter of 12 mm, enables high-contrast sub-cellular imaging (lateral resolution: 1µm, axial resolution: 1.72 µm) deep within tissues of the oral cavity. Furthermore, high-speed (>20 Hz) optical sectioning was achieved, which greatly enhanced the clinical usability of this device by minimizing motion artifacts and enabling high quality real-time video mosaicking. However, due to the size limitation of the miniature device (12-mm diameter tip), the field of view limits the capability of the device for imaging a large extent of tissue with high resolution. To extend the imaging field during the clinical use of such devices, we have developed a real-time video mosaicking method based on a previous post-processing mosaicking method, which allows users to efficiently survey larger areas of tissue. For the images that are 400 x 400 pixels in size, real-time mosaicking is possible at >30 frames/sec. Unlike other real-time mosaicking methods, our strategy can accommodate image rotations and deformations that often occur during clinical use of a handheld microscope. We perform a feasibility study to demonstrate that the use of real-time mosaicking enables significantly more efficient sampling of a desired imaging field when using a handheld dual-axis confocal microscope.