Fast automated image-based patient-specific finite element biomechanical modeling

This work proposes a novel method to create biomechanical patient-specific finite element model of the hemipelvis. The method is based on prior knowledge in the form of a template model (atlas) and on intensity-based image mapping. Direct approaches to create an accurate pelvic model combine image segmentation, surface reconstruction and mesh generation. Despite the active research in these areas the process to create quality pelvic model requires expert assistance in all three mentioned areas, and is not automated as per current state-of-the-art, requiring typically couple of weeks to obtain valid pelvis finite element model.; Method proposed in this work starts from the already defined template pelvic model and establishes the connection between the template and the patient data through image mapping of the corresponding CT scans. Once the mapping is established it can be used to transfer all information from the template to the patient including the finite element mesh, bone material properties, loading and boundary conditions. The need for mesh generation for each patient is eliminated as is a need for image segmentation and surface reconstruction. Instead of performing these costly steps for every patient, they are done only once for the template model and independently of surgical planning. To create patient-specific data, all that remains to be done is to determine the mapping between the template and patient CT data.; The mapping starts with global alignment, in which the rigid body translations and rotations, and uniform scalings are applied to reposition the template and bring it closer to the patient configuration. In the second stage the template is locally deformed to account for geometric variations between the template and the patient data. Optimization formulation has been adopted for both stages of the mapping process and has been implemented in a computer code. The process is fully automated and does not require expert involvement. The time to create a patient-specific finite element model has been reduced to the order of minutes, which meets the requirements of efficient real-time pre-operative planners.