• aorta
• DTI
• fiber tracking
• MR
• python

Diffusion Tensor Imaging (DTI) is a magnetic resonance imaging technique which allows to obtain information on tissue microstructure non-invasively. The basic concept behind DTI is that water molecules tend to diffuse differently along a specific anatomical region depending on its type, architecture and integrity. In fact, in many biological tissues, the diffusion process can be modeled as an ellipsoid, which, mathematically, can be represented by a 3 x 3 symmetric matrix (six independent parameters), known as tensor. This tensor, called diffusion tensor, is estimated from diffusion weighted images (DWI) and describes the degree of anisotropy and the orientation of the water diffusion process at each voxel. On the basis of quantitative parameters provided by diffusion tensor, such as the fractional anisotropy (FA), the path of the fiber bundles can be reconstructed by using 3D tracking algorithms.
The objective of this thesis is to study the architecture of the human aortic wall by using diffusion-weighted images acquired from ex-vivo samples. To this end, a customized workflow of DTI processing of the aortic datasets was implemented.
Firstly, we tested different DTI software tools (Slicer, MedINRIA and FSL) which are commonly adopted for the DTI processing of brain images as DTI of the white matter is currently the unique application field in clinical setting. Secondly, a custom pipeline for DTI analysis of aorta datasets was developed in Python. The pipeline was composed by several processing steps: 1) conversion of the dataset from DICOM to NIfTI format; 2) eddy current corrections, 3) segmentation, 4) diffusion tensor estimation, 5) diffusion indices estimation, 6) fiber tractography, 7) quantitative analysis of the fiber distribution (number fibers, fiber angles, median FA). Finally, a graphical user interface was developed using PyQt in order to perform the entire workflow in a simple and fast way.