Automated Segmentation and Skin-Layer Thickness Estimation by Extracting the Optical Scattering Coefficient and Speckle Contrast Parameter from Optical Coherence Tomography Scans.

Abstract

Introduction: Knowing thicknesses of skin-tissue layers within the depths of several hundred micrometers from the surface is highly important for numerous biomedical applications - cosmetology, treatment of wounds and burns, characterization of various lesions, etc. In this regard, optical coherence tomography (OCT) with its ability to noninvasively enable visualization depth of the order of 1 mm with a resolution of the order of several micrometers offers exceptional diagnostic possibilities largely unavailable to other techniques. This explains high interest to OCT utilization in dermatology.

Methods: This study demonstrates that appropriate physics-based processing of OCT data allows one to objectively reveal and automatically estimate thicknesses of morphological skin layers, even if they are not visible as optical layers with differing intensity in initial structural OCT scans. To this end, we apply recently developed efficient methods of spatially resolved estimation of the scattering coefficient µ_s to locally characterize the scattering strength of scatterers in the tissue. Another useful parameter termed speckle contrast (SC_I) characterizes fluctuations of scattering strengths of scatterers and their clustering. Mapping parameters µ_s and SC_I enable differentiation of morphological skin layers, although in structural OCT scans they cannot be clearly delineated even by experts.

Results: Analysis of parameters µs and SC_I for OCT data in vivo acquired in various localizations for women and man of various ages confirmed the possibility to clearly segment three medically significant skin layers: (i) stratum corneum, (ii) living-cell layer of epidermis, and (iii) upper dermis layer even if in the initial OCT scans these layers are hardly discernible. The present study is intentionally limited to healthy facial skin, for which significant variations in the layer thicknesses were demonstrated for various age groups, as well as in quite close localizations for the same person.

Conclusion: These findings underscore high potential of in vivo OCT imaging supplemented with the analysis of speckle features and optical attenuation parameter for objective real-time differentiation of morphological skin layers, which is critically important for numerous applications (surgical reconstruction of skin wounds and/or burn injuries, controlled drug delivery in targeted skin layers, choice of therapy). Similar approach may be extended for diagnosing various skin diseases/pathologies.