Image-Based, Whole System Hemodynamic Modeling for Valvular Heart Diseases
Please login to view abstract download link
About 2.5% of the U.S. population has valvular heart disease with aortic stenosis (AS) and mitral regurgitation (MR) being the two most common. These diseases can cause life-threatening complications such as cardiac arrhythmia, stroke, and heart failure (HF). Understanding the impact of these diseases on the cardiovascular system and the success of various treatments will allow clinicians to provide patients with the best care. Hemodynamics can be assessed with non-invasive and invasive clinical tools; however, non-invasive tools are limited in their accuracy and invasive tool have the risk of possible complications during or after the operation and are limited to gathering data in specific regions in the cardiovascular system. Advances in computational fluid dynamics (CFD) and reduced order modeling has enabled the development of patient-specific modeling pipelines that allow for a detailed evaluation of patient hemodynamics. We leverage advanced image processing techniques, CFD methods, and multi-scale modeling to capture the impact of valvular disease and treatments on local and global hemodynamics. Neural networks are a powerful tool that can be used to segment heart chambers and major vessels. Image registration techniques are used to track the motion of the dynamic cardiac images and apply it to the 3D model. Advanced CFD methods allow us to solve stabilized Arbitrary Lagrangian-Eulerian Navier-Stokes equations to solve for the flow in the heart and the great vessels. Advances in multi-scale modeling allow for the 3D models to be coupled to closed-loop 0D models. This allows us to solve for hemodynamics in the remaining cardiovascular system. We leverage advanced image processing techniques, CFD methods, and multi-scale modeling to capture the impact of valvular disease and treatments on local and global hemodynamics. In the first study, we modeled patient-specific cases of mitral regurgitation (MR) to better understand the effect of MR on the right ventricle (RV) function. In the second study, we modeled cases of aortic stenosis and the impact of a transcatheter aortic valve replacement (TAVR) on local hemodynamics. In our third study, we modeled the impact of an aortic root enlargement (ARE) on local hemodynamics in surgical aortic valve replacement (AVR) patients. To conclude, these methods use advanced CFD and multi-scale modeling methods to study how these various diseases and treatment options affect local and global hemodynamics.
