The quantitative analysis of vessel wall thickness in the carotid bifurcation region based on blackblood MR imaging is a difficult problem because of partial volume effects, strong variations in surrounding tissue contrast, and flow artifacts, which frequently appear in the carotid bulbus. The abstract presents an automatic vessel wall thickness quantification approach based on a segmentation that integrates the information from a TOF-MRA sequence and a MSD-T2-weighted variable flip angle 3D MRI sequence without changing the image data.
The method is validated using a crossection-wise comparison with contours corrected by 6 different observers on 60 vessel crossections.
The results show a good overall agreement. Major deviations between observers and automatic segmentation occur in regions with strong artefact.