Characterization of arterial stenosis using 3D imaging: Comparison between three imaging techniques (MRA, spiral CTA and 3D DSA) and four display methods (MIP, SR, MPVR, VA) in a phantom study

Introduction. Accurate assessment of arterial stenosis is a major public health issue for the diagnosis and treatment of cardiovascular diseases. The number of imaging techniques and types of software for display of imaging data is increasing. Few studies that compare these different techniques are available in the literature. Materials and methods. Using phantoms to reproduce the main types of arterial stenosis, the authors compared three 3D acquisition techniques (MRA, CTA, and 30 DSA) and four types of display methods (MIP, SR, MPVR, and VA). The degree, the shape, and the location of different types of stenoses were analyzed by three experienced observers during two successive readings. Intra- and inter-observer reproducibility were assessed. The results of the various acquisition techniques and display methods also were compared to the digital reference data (CFAO) of the physical phantoms. Results. The degree of intra- and inter-observer reproducibility for the assessment of shape and location of the stenoses was good. Visual assessment of the degree of stenosis showed significant differences between two observers as well as in two readings by one observer. The 3D DSA was the most accurate technique for assessing the degree of stenosis. CTA provided better results than MRA. MPVR provided an accurate assessment of the degree of the stenosis. 3D DSA and CTA assessed stenosis form and localization adequately, with no significant difference; both methods appeared to be more accurate than MRA. SR provided the best information on the eccentric nature of the stenosis. The shape was very well assessed by VA and MPVR. Conclusions. Even though 3D DSA is the most accurate acquisition technique for visualization, the combined use of SR and MPVR appears to be the best compromise to describe the morphology and degree of stenosis. Further improvements in automatic 30 image processing could offer a better understanding and increased possibilities for assessing arterial stenosis.