ABSTRACT

Tissue contrast in optical coherence tomography (OCT) images is generated by the intrinsic characteristics that are proportional to the density, size, and shape of tissue microstructures. Extraction of tissue optical properties from OCT images is therefore a generic method for obtaining quantitative information about tissue microstructure. The sensitivity and specificity of conventional OCT by assessing morphology only for melanoma detection is lower than anticipated. The aggregation of the predominant optical properties that contribute to OCT image formation diminishes the specificity of melanoma detection due to interrelationships of these properties. Importantly, the optical properties come at no additional cost as they are embedded in the image data and can readily be extracted via post-processing applicable to virtually all OCT systems. OCT images by providing morphology of the skin and angiography of the vessels, along with soma image processing and machine-learning methods can highlight the relevant diagnostic information, yielding unprecedented sensitivity/specificity.