A Geometry Modeling and Optimization Pipeline for the Atrioventricular Heart Valves

Abstract

This research aims to develop a pipeline for modeling the tricuspid heart valve that can be used as an adaptable tool for furthering research in the field, including treatment options for heart valve disease such as functional tricuspid regurgitation. We first gathered data from micro-computed tomography scans of porcine heart valves to extract the valve shape and identify the annulus. The data were transformed and sent to a CAD modeling software in a streamlined process. We then combined the initial shape of the valve with a set of input parameters to define a leaflet surface and chordae tendineae using non-uniform rational B-splines (NURBS). The resulting model was used to represent the valve's shape, which we provide examples of using multiple patient data sets. We also combined the model with a nonlinear isotropic constitutive model for the leaflets to directly use isogeometric analysis (IGA) to evaluate the closure of the valve. This valve model creation, using only a set of initial data and input parameters, was combined with a genetic algorithm search pattern to demonstrate optimization capabilities of the modeling pipeline. The pipeline was used to minimize an objective function for the coaptation area of the valve model, which affects the quality of the valve's closure and reduces chances of tricuspid regurgitation. The presented modeling pipeline provides the next step in streamlining the process from data acquisition to improving biomechanical understanding of the tricuspid valve, bridging the research gap currently present with the tricuspid valve.