|author:||Peter E. Hammer|
|adviser:||Robert D. Howe|
Heart valves are functionally complex, making surgical repair difficult. Simulation-based surgical planning could facilitate repair, but current finite element (FE) studies are prohibitively slow for rapid, clinically-oriented simulations. An anisotropic, nonlinear mass-spring (M-S) model is presented to approximate the membrane behavior of heart valve leaflet tissue, and it is coupled with a fast method for simulating valve dynamics. An efficient FE model is also described for simulating valve leaflets. The speed-accuracy tradeoff between the FE and M-S models is quantified so that the strength of each method can be leveraged where appropriate. The FE model is applied to study a generalized aortic valve repair technique that incorporates graft material into the native valve, where the graft has significantly different mechanical properties than native leaflets. Results show that the graft must be larger than the native leaflets and predicts optimal graft height and width. The M-S method is applied to fully image-based models of the mitral valve to simulate valve closure and loading for fast applications like intraoperative surgical planning. This model is used to simulate a technique used in valve repair and to assess the importance of chordae in determining the closed configuration of the valve. Direct image-based comparison was used for validation. Results of M-S model simulations showed that it is possible to build fully image-based models of the mitral valve and to rapidly simulate closure with sub-millimeter accuracy. Chordae, which are presently difficult to image, are shown to be strong determinants of the closed valve shape.