Description:

This example shows the generation of unstructured grids for finite element calculations of blood flow in carotid arteries reconstructed from contrast-enhanced MRA images. Each individual arterial branch is reconstructed using a deformable model which yields a surface triangulation for each branch. These branches are then merged using an adaptive voxelization technique. The finite element grid generation is then directly performed starting with the surface triangulation.

The following series of images show the merging of the surface triangulations corresponding to different branches of the carotid artery (internal, external and common) of a normal subject. The first image shows a portion of the external and common while the second figure shows a portion of the internal and common. The next two figures show the merged model and the surface of the generated finite element grid, respectively.

The second example shows the mergin of three branches of the carotid artery of a normal subject.

 
The next example shows the case of a patient with carotid artery stenosis in the internal carotid artery.

The following series of images show how the implantation of a stent in the stenosis can be simulated. The stenotic region of the internal carotid artery as well as the geometry of the stent are shown in the following two images.

The next two images show the relative position of the stent and the final geometrical model after implanting the stent, respectively.

 Observations:

• The segmentation of individual arterial branches has the advantage of avoiding fake intersections between close arteries or veins that do not really intersect. However, it is necessary to join or merge the surface triangulations representing each branch. These examples have shown that a voxelization technique combined with iso-surface extractions yield satisfactory surface models for unstructured grid generation.
• The segmantation, surface processing and grid generation algorithms work for both normal and diseased patient cases.
• "Virtual interventions" can also benefit from these techniques. Adding extra features such as stents, bypass grafts, etc., can be done at the triangulation level.