Cost-effective Mueller matrix microscopy for label-free differentiation of melanoma from normal skin and common skin cancer types.

A major goal in the management of skin melanoma is to optimize early detection of the disease. The current gold standard method for diagnosing skin cancer relies on pathologists' interpretation of dermoscopy images and on histologic analysis, but this approach has low accuracy for melanoma detection and is time-consuming. Though advanced optical imaging technologies can increase the detection accuracy for non-melanoma skin cancer, they are still unreliable for melanoma detection and are associated with high costs for the equipment and training. In this study, a low-cost wide-field transmission microscope powered by Mueller matrix formalism and decomposition methods is developed to image collagen birefringence in normal human skin, melanoma, and common types of skin cancer (basal cell carcinoma or BCC and squamous cell carcinoma or SCC). The results show that two-dimensional images of retardance can highlight clusters of collagen fibers in tumorous skin. In addition, analyzing orientation as a function of retardance is useful to differentiate normal skin from tumorous skin, while analyzing orientation as a function of depolarization is useful in categorizing types of skin cancer.