Medical imaging and patient-specific modeling of women pelvic system: application to magnetic resonance images

Common female pelvic floor dysfunctions, such as genital prolapse, have a large social impact and are strongly correlated with pelvic mobility. Medical imaging techniques, such as magnetic resonance imaging (MRI), have become a powerful tool to help clinicians gain structural and cinematic information on soft tissue, such as pelvic organs. On the other hand, numerical finite element simulation can provide clinicians with new insights into the mechanism of pathology through mobility, especially when anatomical structures are hard to observe in MRI (fascia and ligaments) and when organ motion is hard to predict at early stage of pathology. However, numerical simulation relies on consistent geometric modeling of the pelvic system. 3D virtual image correlation method is a new approach to efficiently build readily usable geometric models for simulation from image data. It generates directly parametric surfaces (NURBS) without surface remeshing and is more suitable for applications as complex as the pelvic system that contains multiple organs and suspension devices. Overall, the geometric modeling helps fill the gap between medical images and numerical simulation.