Polarized light spatial frequency domain imaging for non-destructive quantification of soft tissue fibrous structures

Abstract

The measurement of soft tissue fiber orientation is fundamental to pathophysiology and biomechanical function in a multitude of biomedical applications. However, many existing techniques for quantifying fiber structure rely on transmitted light, limiting general applicability and often requiring tissue processing. Herein, we present a novel wide-field reflectance-based imaging modality, which combines polarized light imaging (PLI) and spatial frequency domain imaging (SFDI) to rapidly quantify preferred fiber orientation on soft collagenous tissues. PLI utilizes the polarization dependent scattering property of fibers to determine preferred fiber orientation; SFDI imaging at high spatial frequency is introduced to reject the highly diffuse photons and to control imaging depth. As a result, photons scattered from the superficial layer of a multi-layered sample are highlighted. Thus, fiber orientation quantification can be achieved for the superficial layer with optical sectioning. We demonstrated on aortic heart valve leaflet that, at spatial frequency of f = 1mm(-1) , the diffuse background can be effectively rejected and the imaging depth can be limited, thus improving quantification accuracy.

Keywords: (110.0113) Imaging through turbid media; (170.0110) Imaging systems; (170.6935) Tissue characterization; (290.5855) Scattering, polarization.

Fig. 1

Cylinder light scattering simulation geometry. The cylinders orient at 30 degree with respect to x axis. The polarized light illuminates a collection of cylinders with multiple diameters at normal incidence. The polarization angle θ of the incident light rotates anti-clockwise from 0 degree to 180 degrees.

Fig. 2

(a) Color picture of pSFDI imaging system. (b) Schematic of the pSFDI imaging system.

Fig. 3

SHG image shows the collagen fibers in porcine aortic valve leaflet tissue.

Fig. 4

Polarization dependent back-scattered light over 180-degree polarization range for (a) collagen fibers and (b) collagen fibrils. Two peaks can be identified. The higher peak indicates that illumination polarization is parallel to the fiber orientation, while lower peak indicates they are perpendicular.

Fig. 5

Polarization dependent back-scattered light over 180-degree polarization range for 6 fiber distributions. The higher peaks locate at 30 degree for normal distribution. With higher fiber orientation deviation, both higher peak and lower peak intensity decrease. No peak can be found for random distribution.

Fig. 6

Collagen fiber orientation mapping on bovine tendons. (a) Blue arrows indicate the gross fiber orientation of the bovine tendons. Red rectangle shows the imaged area. (b) Back-scattered intensity plot over a small region indicated by a black square using both DC and AC components. (c) Fiber orientation map extracted using on AC components. (d) Fiber orientation map extracted using DC component.

Fig. 7

Depth controlled pSFDI imaging of a two-layer bovine tendon sample (a). The collagen fiber orientation map retrieved using polarized light imaging without spatial pattern in map (b). The collagen fiber orientation maps (c) and (d) retrieved at spatial frequencies 0.09 and 0.5 mm-1, respectively. The intensity plot (e-g) for localized regions in (b-d), respectively.

Fig. 8

pSFDI imaing on PVL (a-1 and 2) and PLV-tendon combination (f-1 and 2). DC component based fiber orientaion maps (b and g) for PLV and PLV-tendon combination, respectively. AC component based fiber orientation maps (c and h) for PLV and PLV-tendon combination, respectively. A SALS based orientation map (d) for PLV was obtained for comparison purpose. Polar plots of both AC and DC components from a small region indicated by black sqares (b and g) for both PLV (e) and PLV-tendon combination (i) respectively. Black arrows in (b) and (d) indicate unreliable fiber orientation. Dashed black rectangle in (g) indicates the overlapping region of PLV-tendon combination.