High-resolution 3D visualization of ductular proliferation of bile duct ligation-induced liver fibrosis in rats using x-ray phase contrast computed tomography

Abstract

X-ray phase-contrast computed tomography (PCCT) can provide excellent image contrast for soft tissues with small density differences, and it is particularly appropriate for three-dimensional (3D) visualization of accurate microstructures inside biological samples. In this study, the morphological structures of proliferative bile ductules (BDs) were visualized without contrast agents via PCCT with liver fibrosis samples induced by bile duct ligation (BDL) in rats. Adult male Sprague-Dawley rats were randomly divided into three groups: sham operation group, 2-week and 6-week post-BDL groups. All livers were removed after euthanasia for a subsequent imaging. The verification of the ductular structures captured by PCCT was achieved by a careful head-to-head comparison with their corresponding histological images. Our experimental results demonstrated that PCCT images corresponded very well to the proliferative BDs shown by histological staining using cytokeratin 19 (CK19). Furthermore, the 3D density of proliferative BDs increased with the progression of liver fibrosis. In addition, PCCT accurately revealed the architecture of proliferative BDs in a 3D fashion, including the ductular ramification, the elongation and tortuosity of the branches, and the corrugations of the luminal duct surface. Thus, the high-resolution PCCT technique can improve our understanding of the characteristics of ductular proliferation from a new 3D perspective.

Figure 1

Slices and corresponding 3D microvasculature rendering images from different groups. A hepatic vascular tree is presented with regular shapes in the control group (a,d). Tortuous and abnormal microvascular structure caused by fibrosis tissues occured at week 2 post-BDL (b,e). The disordered microvascular structure and abnormal microvascular morphology become more apparent at week 6 post-BDL as the fibrosis stage increased (c,f). The ductular proliferation marked by a yellow cuboid is revealed in (f).

Figure 2

CT image and corresponding pathological sections of the BDL rats. In the immunohistochemical staining for CK19 (b), CK19-expressing cells appear to form larger clusters that are stained brown, which is in accordance with the CT findings (a), which are marked by the white arrows. In the histological slice stained with Sirius Red (c), fibrous bands are clearly shown within the regions marked by black arrows.

Figure 3

Illustration and quantification of fibrosis and proliferative BDs in liver fibrosis progression. Histological sections stained by Sirius Red showing representative images of the control group, the 2-week and 6-week post-BDL groups are shown in (a) to (c). The volume rendering images of three groups are shown in (d) to (f). The amplified images of the upper surface in (d), (e) and (f) are provided in (g) to (i), respectively. Fibrosis area (j) and the density of proliferative BDs (k) are summarized and shown for different groups. The black and white arrows indicate the fibrous bands and proliferative BDs, respectively. For all comparisons, P was less than 0.01. Note that the ductular proliferations are not detected in the livers of the control group.

Figure 4

3D reconstruction of ductular proliferation and VME to 3D track targeted BD. (a) Local 3D reconstruction of proliferative BDs. (b) A single segmented BD. Bifurcations in (b) and (d) (red arrows), and the corrugations in (b), (e) and (f) (yellow arrows) are discernible. (c) Centre line of the segmented BD in (b). The branching points and breakpoints are marked by green dots. Successive pathway tracing is shown in (d) to (f).

Figure 5

Schematic illustration of the experimental station at BL13W1. A monochromatized x-ray is projected on a liver sample that was mounted on a rotary stage. The scanning images can be acquired at a proper distance by adjusting the SDD that can be varied from 0.05 m to a maximum of 8 m. In this study, the transmitted beam was recorded by two different image detectors with spatial resolution of 9 μm (detector A) and 3.25 μm (detector B), respectively. During the CT imaging, the projections were collected over 180° of rotation.