Imaging local optic axis of multilayered tissue using Jones-matrix optical coherence tomography

Polarization-sensitive Optical Coherence Tomography (PS-OCT) is a label-free and volumetric modality for birefringence and optic axis imaging, capable of visualizing the organization and orientation of fiber tissues. Measuring the optic axis of fiber tissues has been challenging for PS-OCT due to the polarization effects accumulated from overlying birefringent layers. This paper proposes an efficient PS-OCT algorithm for reconstructing the local optic axis of multi-layered birefringent tissues by exploiting the vector and matrix products of Stokes vectors. When measuring multiple layers of retarders using PS-OCT, the measured Stokes vector rotates around the optic axis of each layer at an angle determined by its phase retardation on the Poincaré sphere. This causes the measured optic axis to be accumulatively rotated by the overlaying retarders. Our algorithm effectively geometrically reverses the accumulated rotations of the measured Stokes vector caused by the overlaying retarders, enabling the reconstruction of the optic axis local to each layer. The reconstruction algorithm was applied to datasets of a retarder phantom and an in vivo fingernail bed measured by Jones-matrix based PS-OCT. The phantom consisted of three 3D printer filaments stacked as a triangle. The inter-filament angles were computed from the structural OCT image for reference (denoted as reference angles). The optic axis orientation of each filament was computed, at first, at the regions not covered by other filaments. The maximum difference of the measured orientation to the reference angles was as small as 4.6°. The local axis orientations were then computed at two regions covered by two different filaments. The maximum difference of the axis orientations at the covered and non-covered regions of the same filament was as reasonably small as 14.8°. The nail and the skin at the nailbed region exhibited different optic axes.