| One of every seven Americans, over one million people per year, is afflicted with skin cancer in one of its various forms. Currently the only available methods to diagnose suspected cancerous lesions are by visual inspection and subsequent biopsy of the lesion. Using only visual inspection, nearly one-third of all melanomas are misdiagnosed, subsequently, the only other alternative, biopsy, is often performed unnecessarily on benign lesions. However, biopsy, which is the surgical removal of a tissue for a microscopic histological examination, is an invasive, expensive, and time-consuming process. Therefore, there is a need for the development of accurate, non-invasive skin cancer detection techniques. This study was focused on the development and testing of a novel Automated Mueller Matrix Polarization Imaging System that has the potential for non-invasive determination of cancerous lesions from their benign counterparts. As part of this research, the system was precisely calibrated, tested for known samples and the experimental results, in form of the full 16-element Mueller matrix (16-EMM), quantified. A 16-EMM is extremely powerful because it completely describes the polarization altering properties of a sample, however a raw 16-EMM is not easy to interpret. A method on how to interpret the Mueller matrix imaging for optical differentiation between cancerous and non-cancerous lesions based on polar decomposition was performed. A polar decomposition of the 16-EMM yielded the more familiar quantities of retardance, diattenuation, and depolarization to differentiate between cancerous and non-cancerous lesions. In addition, the extracted features were analyzed using statistical processing to develop an accurate classification scheme and to identify the importance of each parameter in distinguishing cancerous melanoma, non-cancerous mole, and normal tissue in a Sinclair swine animal model. |
