| dc.description.abstract | The early diagnosis of oral carcinoma has proven to be a difficult task. Fully developed oral squamous cell carcinomas can be easily recognized by trained clinicians due to the appearance of aggressive exogenous tumors and other visual and textural changes. However, pre-malignancies that can lead to oral squamous cell carcinoma are often difficult to diagnose due to their similar appearance to normal or benign oral tissue. The gold standard for diagnosis continues to be visual and palpatory inspection by a clinician which may be followed by an exploratory biopsy on suspicious lesions. This excised lesion is then sent for histopathological analysis and a recommendation is made based on the results. One of the flaws with this methodology is that the clinical diagnosis of oral carcinomas and pre-malignancies is subjective, and opinions can vary between clinicians. Additionally, achieving a histopathological diagnosis of a tissue sample and acting on that information takes a significant amount of time. Automation of the clinical diagnosis of oral cancers by utilizing optical imaging and image recognition techniques could drastically reduce this time.
Two promising modalities for cancer imaging are structured illumination and fluorescence lifetime imaging microscopy. Oral epithelial tissue undergoing carcinogenesis is subject to biochemical changes caused by fundamental shifts in the epithelial metabolism. This shift in the metabolism and biochemical state causes a change in the cellular microenvironment that can affect how long certain endogenous fluorophores emit light. This phenomenon can be probed by fluorescence lifetime
imaging. Although fluorescence lifetime imaging offers an exceptional ability to distinguish between different fluorophores within the focal plane of an imaging system, it suffers from the inability to accurately resolve multiple spectrally overlapping lifetimes that exist within a single pixel. The fluorophores of interest in oral carcinoma, NADH and FAD, share significant spectral overlap with collagen, an endogenous fluorophore in the submucosal connective tissue. Because collagen generally overlaps laterally with NADH and FAD, its fluorescence emission is usually found in both the NADH and, to a lesser extent, FAD channels. Luckily, the collagen containing connective tissue is exclusively deep relative to the oral epithelium. This axial spatial separation allows for the utilization of optical sectioning to spatially isolate NADH and FAD from the effects of collagen rather than spectrally. This thesis presents the development and testing of a fluorescence lifetime imaging microscope with optical sectioning capability provided by structured illumination with the goal of enhancing the early detection of oral carcinoma. | en |
