Generic framework for quantifying the influence of the mitral valve on ventricular blood flow

Abstract

Blood flow within the left ventricle provides important information regarding cardiac function in health and disease. The mitral valve strongly influences the formation of flow structures and there exist various approaches for the representation of the valve in numerical models of left ventricular blood flow. However, a systematic comparison of the various mitral valve models is missing, making a priori decisions considering the overall model's context of use impossible. Within this study, a benchmark setup to compare the influence of mitral valve modeling strategies on intraventricular flow features was developed. Then, five mitral valve models of increasing complexity: no modeling, static wall, 2D and 3D porous medium with time‐dependent porosity, and one‐way fluid–structure interaction (FSI) were compared with each other. The flow features velocity, kinetic energy, transmitral pressure drop, vortex formation, flow asymmetry as well as computational cost and ease‐of‐implementation were evaluated. The one‐way FSI approach provides the highest level of flow detail, which is accompanied by the highest numerical costs and challenges with the implementation. As an alternative, the porous medium approach with the expansion including time‐dependent porosity provides good results with up to 10% deviations in the flow features (except the transmitral pressure drop) in comparison to the FSI model and only a fraction (11%) of numerical costs. However, jet propagation speed is highly underestimated by all alternative approaches to the FSI model. Taken together, our benchmark setup allows a quantitative comparison of various mitral valve modeling approaches and is provided to the scientific community for further testing and expansion.

Inter‐model and inter‐modality analysis of left ventricular hemodynamics: Comparative study of two CFD approaches based on echocardiography and magnetic resonance imaging

“…The MV movement and LV contraction is modeled as fluid structure interaction (FSI) using an immersed boundary/finite element method and yields reasonable agreement with in vivo data. Yet, other studies report that reduced MV approaches are also capable to reproduce intraventricular flow features [9,45]. Concomitant challenges with modeling a moving MV as well as the large ventricular deformations are distortions in the computational mesh, possibly leading to negative cell volumes [7].…”

Cardiac hemodynamics computational modeling including chordae tendineae, papillaries, and valves dynamics

“…In order to investigate the interaction of the ventricular blood flow with MV, two different mathematical modeling can be carried out. The first approach consists in using lumped parameter models to simulate valve dynamics, representing the fluid dynamics effect of the valve through the relationship between its pressure drop and flowrate [85, 48, 12, 66]; moreover, more sophisticated diode type models can be used, in which a time-dependent permeability is associated to the MV plane, in order to model the opening/closure of the valve without considering the leaflets [76, 77, 90, 21, 87]. The second approach consists in the 3D (or at least 2D) representation of the valve geometry, obtained either by parametric modeling [52, 22] or by segmentation of patient-specific real-time imaging data [95, 58, 96], whose presence can be considered through mesh conforming techniques (such as the Resistive Immersed Surface method [5, 78]) or fully Eulerian methods (such as the Immersed Boundary Method [69, 55, 34, 26, 27, 28] or the Fictitious Domain method [35, 56]).…”