Poster T20 - Assessing Improvements in LV Hemodynamics Post-mitral Valve Repair Surgery via an Integrated CMR-PC-MRI-in-silico Approach

Introduction: Mitral valve prolapse (MVP) is a widespread heart valve issue characterized by the abnormal movement of one or both mitral valve leaflets more than 2 mm into the left atrium during left ventricular (LV) contraction, with the major complication of mitral regurgitation[1]. Affecting around 2-3% of the adult population in the United States, the sole definitive remedy for MVP involves surgery, either mitral valve repair (MVRe) or replacement (MVRp) using mechanical or biological prosthetic valves[2]. Consequently, the accurate characterization of intracardiac flow is crucial in assessing the effectiveness of the intervention. In this study, we leveraged a recently introduced patient-specific moving wall-computational fluid dynamics (CFD) simulation framework using phase-contrast magnetic resonance imaging (PC-MRI) data. The patient-specific CFD method was used to simulate the LV flow patterns of five patients with MVP and to characterize the LV hemodynamic and structural changes pre- and post-MVRe surgery.

Materials and

Methods: Standard cine and PC-MRI scans were performed on five MVP patients (n=5) before the MVRe surgery and post-operation. For each case, imaging resulted in a set of nine short-axis (SA) slices with 20 time frames over one cardiac cycle. The LV endocardial surface was segmented at all timepoints. The three-dimensional (3-D) LV endocardial surfaces were reconstructed from standard cardiac magnetic resonance imaging (CMR). Endocardial displacements and myocardial strains over the whole cardiac cycle were calculated through an in-house non-rigid image registration (NRIR) framework[3]. Calculated displacements were applied to the reconstructed endocardial surface to simulate its movement throughout the cardiac cycle. Additionally, using PC-MRI data for the same patients, the blood flow velocity field was calculated at the 2D area of the mitral inflow track, and the integrated mitral flow rate was applied as an inflow boundary condition in the FSI models. The FSI domains were solved using the incompressible Navier–Stokes equations with the arbitrary Lagrangian-Eulerian (ALE) formulation. The flow was considered to be laminar and incompressible, with blood viscosity and density of 0.0035 Pa·s and 1040 kg/m³, respectively.

Results, Conclusions, and Discussions: The CFD modeling provided simulated hemodynamic behavior before and after MVRe in three patients. For each patient, the LV vorticity field was computed in the diastolic phase, and changes in vorticity magnitude and the averaged LV kinetic energy during diastole were examined before and after MVRe surgery, as illustrated in Figure 1. In addition to flow characteristics, the NRIR framework was used to acquire myocardial strains. The correlation between flow characteristics and structural changes was explored, with the correlations of systolic longitudinal, radial, and circumferential strain with kinetic energy at the basal level depicted for one case in Figure 2.
While previous in-vivo studies have demonstrated the correlation between LV vortex structures and LV kinetic energy, as well as alterations in LV vortices post-MVRp surgery[4], our in-silico CFD modeling explores the flow patterns before and after MVRe. Additionally, statistical analysis examines the link between changes in LV structural kinematics—i.e., endocardial displacements and myocardial strains—and LV hemodynamic alterations. The proposed method holds the promise of better understanding the implications of LV remodeling following MVRe surgeries, ultimately improving surgical design in patients with MVP.

Acknowledgements (Optional):