SUMMARY

Cardiovascular disease is the leading cause of death worldwide. Computational models may be used to study disease causes, progression, and treatments. This thesis presents methods of developing and validating patient-specific computational models of mitral valves (MV) from clinical computed-tomography (CT) images. Four patient-specific MV models were created from CT images of patients diagnosed with mitral regurgitation (MR). Each model contains both mitral valve leaflets with thickness, calcification, dynamic chordae tendineae origins, and a dynamic mitral annulus. The effect of calcification on MV biomechanics was compared via simulations with and without calcification for each model. The models suggest that the calcified mitral valves experienced higher loading than the non-calcified valves. High stress regions in the non-calcified models corresponded to calcification locations, indicating high stress may lead be linked to the development of calcification. The validated MV models may be used to investigate the effects of different MR treatments, such as transcatheter MV repair (TMVr). This thesis used finite-element simulations to provide a head-to-head comparison of the biomechanical effects of implanting two different TMVr devices, PASCAL and MitraClip, by simulating three different implantation configurations for each device. It was found that the PASCAL reduced leaflet stresses and strains compared to pre-operative leaflet loading conditions for the studied patient model, while the MitraClip caused them to increase, and the MitraClip reduced the regurgitant orifice area more so than the PASCAL. Analysis of results from the different implant configurations for each device may improve patient outcomes by providing scientific rationale for the clinical decisions of choosing the proper implant device, number of devices, implant location(s), and patient selection.