Polarimetric analysis of anisotropic tissue using polarization-sensitive optical coherence tomography (PS-OCT)

Polarimetric properties of tissues including form-birefringence, form-biattenuance and optic axis orientation correlate with the size, density and orientation of fibrous structures. These polarimetric properties are physiologically significant for diagnosing several diseases by observing their variations. Polarization-sensitive Optical Coherence Tomography (PS-OCT) is an optical interferometric imaging modality that noninvasively determines depth-resolved, cross-sectional reflectivity and polarimetric properties of anisotropic tissue. Polarimetric properties in anisotropic tissue were geometrically studied by observing and analyzing trajectory of normalized Stokes vectors on the Poincaré sphere corresponding to backreflected light from the tissue. Analytic expressions for various derivatives of the trajectory were derived and employed to analyze characteristics of polarimetric properties in anisotropic tissue. A novel nonlinear fitting algorithm utilizing multiple incident polarization states was developed for high-sensitive determination of polarimetric properties in weak form-birefringent tissue recorded by PS-OCT. After verification of robustness by simulated PS-OCT data, polarimetric properties of form-birefringent tissues were determined by the algorithm. Complex polarization ratio (CPR) was adapted to determine polarimetric properties of anisotropic tissue. Another nonlinear fitting algorithm using the trajectory of the CPR representing anisotropic tissue was developed with multiple incident polarization states. Polarimetric properties of form-birefringent tissues were determined by the nonlinear fitting algorithm with the trajectory of the CPR. Complex representation and two-dimensional visualization of the trajectory of the CPR provides advantages in computation and efficiency over Jones, Mueller and Stokes vector representations. Two boundary detection techniques were performed to identify each layer in multilayered anisotropic tissue. These techniques used residuals in a nonlinear fitting algorithm and discontinuity of curvatures in the trajectory of normalized Stokes vectors. Slope change of the residuals at the boundary was observed as an indicator to detect a boundary using the residual technique. High-valued curvatures at the boundary were selected by thresholding speckle-noise curvatures with hypothesis testing based on probability density functions at and within boundaries between layers.