ECCOMAS 2024

A Comparison of Image-Driven, Patient-Specific Direct Numerical Simulations of the Right Ventricle to 4D Flow MRI

  • Yildiran, Ibrahim (George Washington University)
  • Capuano, Francesco (Universitat Politècnica de Catalunya)
  • Loke, Yue-Hin (Children’s National Hospital)
  • Olivieri, Laura (UPMC Children's Hospital of Pittsburgh)
  • Balaras, Elias (George Washington University)

Please login to view abstract download link

The analysis of intracardiac blood flow patterns has the potential to provide reliable biomarkers that can improve our understanding and treatment of many cardiovascular diseases. In this work we focus on the right ventricle (RV) where currently there is a significant lack of knowledge on the blood flow dynamics, in healthy as well as diseased conditions. We have developed a workflow for patient specific, image-based Direct Numerical Simulations (DNS) that provide detailed spatio/temporaly resolved blood flow patterns. The kinematics of the RV are reconstructed from multi-slice MRI datasets (i.e. short-axis and longitudinal contours of the RV walls) with temporal resolution of approximately 30 phases per cardiac cycle. Feature-tracking software extracts the time history of the endocardium contours, which are then the input to image-registration algorithms resulting in a time-continuous representation of the entire RV geometry. The latter is accomplished using the Large Deformation Diffeomorphic Geometric Mapping (LDDMM) framework combined with an atlas-based approach. Overall the method results in a good representation of the complex RV contraction pattern, including the clinically relevant longitudinal shortening. The simulations driven by the resulting wall motion, provide comprehensive intracardiac flow visualization and quantification at high temporal/spatial resolution. There are, however, uncertainties attributed in part to kinematic reconstruction, boundary conditions, and geometric simplifications. Therefore, rigorous validation of DNS results is essential. In this study, we will compare DNS results of the flow inside the RV with 4D-flow MRI results. The comparison is based on healthy and congenital heart disease (CHD) cases, with particular reference to Tetralogy of Fallot (ToF). We will show that the velocity and vorticity fields are in good agreement with the 4D-MRI. The presence of the atrium in the model improves ventricular filling. Similarly, a new strategy to reconstruct the tricuspid valve kinematics yielded a clear, organized ring-vortex as in the clinical data.