Comparison of image driven direct numerical simulations to 4D flow MRI in congenital heart disease*

The advent of advanced medical imaging and computational techniques has significantly enhanced our ability to analyze and understand complex intracardiac flow phenomena, especially within the right ventricle (RV), which has a more intricate geometry and wall motion. The high temporal and spatial resolution in image-driven direct numerical simulations (DNS) provide more detailed insights into intracardiac flow patterns compared to four-dimensional magnetic resonance imaging (4D-MRI), which offers lower resolution imaging of blood flow. These insights can eventually provide reliable biomarkers for monitoring the evolution of postoperative morbidities and the timing of surgeries, particularly in congenital heart disease (CHD). Despite these advantages, computational models may suffer from various uncertainties, including reconstructed motion and boundary conditions. These uncertainties often limit DNS's ability to accurately capture the diastolic filling of the RV, a critical aspect of understanding CHD's fluid dynamics. To address this, we propose a computational workflow that adjusts the reconstructed motion of the RV to maintain mass conservation, thereby improving the representation of diastolic filling. We compare qualitative and quantitative results based on healthy cases against both 4D-MRI and DNS results obtained with non-adjusted RV reconstruction, demonstrating the potential for enhanced accuracy.