Morphological Variation in the Human Mitral Valve Using Statistical Shape Modelling

Statistical shape modelling (SSM) offers a robust framework for quantifying anatomical variability and constructing representative virtual patient cohorts of 3D anatomies that can be used as the foundation of biomechanical in silico clinical trials. In this study, we developed a SSM of the mitral valve using 72 contrast-enhanced computed tomography angiography (CTA) scans of the heart. Principal component analysis revealed dominant modes of shape variation that align with previously reported anatomical patterns in the literature, validating the model's physiological relevance. The resulting shape model effectively captures the geometric diversity of the mitral valve without making any presuppositions about the importance of landmarks or linear measurements. Our results demonstrate the utility of SSMs in generating virtual patient populations from existing scan data. These findings support the integration of SSMs into computational modelling pipelines for preclinical testing, device design, and personalised medicine.