Orientations of collagen fibers in aortic histological section

Histological sections of male abdominal aorta were analyzed with the aim to obtain collagen fibrils and smooth muscle cells nuclei (SMCN) orientations. The sections were stained with van Gieson and digitalized. Digital images of the sections were reduced to binary pixel maps with target component represented in white (collagen or SMCN). One image was used to assess the sensitivity to threshold in binary conversion. Consecutively images were processed by in house developed software BinaryDirections, which works with the algorithm of the rotating line segment to determine significant directions in digital images. The algorithm operates in the way that in each target pixel a line segment is rotated step by step to explore neighborhood of the pixel. Exploring the neighborhood, the number of unity pixels coincident with the line segment is determined in each rotating step. This operation is conducted based on matrix formulation of the neighborhood and the line segment. The distribution of orientations in the entire image is obtained either as averaged density distribution from all pixels or as histogram of the most frequent directions. It was found that both collagen fibrils and SMC nuclei analyses give unimodal distributions. The peak ranges from -50° to -20° (defined as declination from circumferential direction of an artery) depending on the method. It implies that preferred direction in aortic medial layer was oriented circumferentially rather than longitudinally. However, aligning with circumferential direction is not as close as commonly referred in the literature. This conclusion was independent of the threshold setup. Results suggest that the orientations of SMC nuclei and collagen fibrils are mutually correlated and determination of collagen fibril orientations, which may be stained in insignificant manner, could be supported with SMC nuclei orientations to obtain more realistic models.