Predicting the outcome of transcatheter mitral valve implantation using image-based computational models

Abstract

Background: The appropriate placement and size selection of mitral prostheses in transcatheter mitral valve implantation (TMVI) is critical, as encroachment on the left ventricular outflow tract (LVOT) may lead to flow obstruction. Recent advances in computed tomography (CT) can be employed for pre-procedural planning of mitral prosthetic valve placement. This study aims to develop patient-specific computational fluid dynamics models of the left ventricle (LV) in the presence of a mitral valve prosthesis to investigate blood flow and LVOT pressure gradient during systole. Methods: Patient-specific computational fluid dynamics simulations of TMVI with varied cardiac anatomy and insertion angles were performed (n = 30). Wide-volume full cycle cardiovascular CT images prior to TMVI were used as source anatomical data (n = 6 patients). Blood movement was governed by Navier-Stokes equations and the LV endocardial wall deformation was derived from each patient's CT images. Results: The computed pressure gradients in the presence of the mitral prosthesis compared well with clinically measured gradients. Analysis of the effects of prosthetic valve angulation, aorto-mitral annular angle, ejection fraction, LV size and new LVOT area (neo-LVOT) after TMVI in silico revealed that the neo-LVOT area (p < 0.001) was the most significant factor affecting LVOT pressure gradient. Angulation of the mitral valve can substantially mitigate LVOT gradient. Conclusions: Computational fluid dynamics simulation is a promising method to aid in pre-TMVI planning and understanding the factors underlying LVOT obstruction.