Image processing and advanced analysis of contrast-enhanced microCT data for the microstructural characterisation of cardiovascular tissues

Nowadays, cardiovascular diseases (CVDs) are considered as the main cause of death worldwide and many of these diseases are affecting the microstructure (i.e. the specific fibrous organisation at the micros scale) of cardiovascular tissues. To better diagnose and treat CVDs, it is essential to gain more knowledge on this microstructure, but for now, its characterisation is mainly based on qualitative observation and lacks of quantitative information. Imaging techniques, such as contrast-enhanced 3D microfocus X-ray computed tomography (CECT) are allowing to visualise in 3D the microstructure of cardiovascular tissues, which makes it a good tool to study it, but there is a need to develop robust protocols to analyse the images obtained. This study had two objectives. The first one was to develop a filtration and segmentation protocol to extract the collagen and elastin fibers, that are the main constituents of the microstructure, from tissues imaged by CECT. To do so, different filters and their parameters have been investigated to smooth the images, remove the noise, and increase the contrast of collagen and elastin fibers in a rat aorta dataset. Then, a segmentation protocol was described, to separate the elastin and collagen components of the aortic wall. These methods have been extended to datasets of images representing a porcine aorta and a porcine mitral valve. In both cases, the protocols provided good results but were limited by the image quality, and the density of fibers visible on the images. The second objective was to extract structural quantitative data from the segmented datasets. Three parameters have been investigated: the volume fraction of elastin and collagen, the thickness of the elastin sheets, and the tortuosity of these sheets. The three analysis were made on the rat aorta dataset, and the values have been compared to the literature. This comparison assessed that the tools used were efficient, and could be used for further analysis on other datasets, if they have been successfully segmented. To conclude, in this study, an image processing and analysis protocol was developed, allowing to extract structural data from the microstructure of cardiovascular tissues imaged by CECT. Further perspectives are to assess the impact of the filtration method by having a more objective criterion for the selection of optimal filters, to extend the data processing analysis protocols to other datasets than the rat aorta, and to extract other structural parameters that are characterising the fibrous microstructure.