Dynamic three-dimensional reconstruction and modeling of cardiovascular anatomy in children with congenital heart disease using biplane angiography.

Modeling and simulation of cardiovascular biomechanics and fluid dynamics from patient-specific data is a continuing topic of research investigation. Several methodologies utilizing CT, MRI and ultrasound to re-create the three-dimensional anatomy of the cardiovascular system have been examined. Adaptation of these models to pediatric applications has not been studied as extensively. There is significant need for such techniques in pediatric congenital heart disease since local anatomy may exhibit highly unusual geometry, and three-dimensional information would be of significant use for surgical and interventional planning, biomechanical and fluid dynamic simulation, and patient counseling. We report here on the adaptation and application of a three-dimensional reconstruction technique that utilizes bi-plane angiographic images as the base data sets. The method has been validated in a variety of adult imaging situations including coronary artery imaging and intervention. The method uses a skeletonization approach whereby local centerline, diameter, branching and tortuosity of the vasculature are obtained to create the three-dimensional model. Ten patients with a variety of etiology were imaged and 3D reconstructions were obtained. Excellent images were obtained of complex anatomy including the highly branched pulmonary vasculature and Fontan surgical connections. The data were then translated into solid and surface models to facilitate viewing, export into computational fluid dynamic grids, and into files suitable for stereo lithography fabrication (STL). This method appears promising for the dynamic study of complex cardiovascular anatomy found in congenital heart disease. Optimization of the method to facilitate on-line reconstruction and simulation are currently ongoing.