Towards a Robust Automated Technique to Construct Aortic Finite Element Meshes Directly from Medical Images

Abstract

Three-dimensional (3-D) medical imaging creates new opportunities for engineering analyses, whether toward basic understanding of the physiology of an organ, or for the simulation of virtual surgical procedures. Finite element models constructed from patient- specific 3-D medical images can also be used to more fully understand a patient’s pathology. However, creation of such models can be labour-intensive as it typically relies on various segmentation techniques of the organ contour in 2-D cross-sections, then the reconstruction of a 3-D surface or volume, and finally the meshing of the resulting structure. In addition, image processing and mesh generation are often operator-dependent, lack robustness and may be of suboptimal quality.

The objective of this work is to use the information contained in 3-D images to deform a structured mesh approximating an organ of interest so that it matches the geometry represented in the images. For the specific case of the aorta, our proposed methodology relies on the gradient vector flow (GVF) field and a ray-tracing algorithm to detect image boundaries and locate the center of gravity of a local cross-section; it also relies on a marching method to repeat the process between two consecutive cross-sections and generate a structured FE mesh. Testing of the methodology on simplified 3-D synthetic images yielded high quality structured meshes fast with little computational effort and minimal input from the operator. Work is underway to expand the application of the proposed methodology for robust use with images of the aorta obtained from CT or MRI scanners.