Hemodynamic Modeling of Biological Aortic Valve Replacement Using Preoperative Data Only

Abstract

Objectives: Prediction of aortic hemodynamics after aortic valve replacement (AVR) could help optimize treatment planning and improve outcomes. This study aims to demonstrate an approach to predict postoperative maximum velocity, maximum pressure gradient, secondary flow degree (SFD), and normalized flow displacement (NFD) in patients receiving biological AVR.Methods: Virtual AVR was performed for 10 patients, who received actual AVR with a biological prosthesis. The virtual AVRs used only preoperative anatomical and 4D flow MRI data. Subsequently, computational fluid dynamics (CFD) simulations were performed and the abovementioned hemodynamic parameters compared between postoperative 4D flow MRI data and CFD results.Results: For maximum velocities and pressure gradients, postoperative 4D flow MRI data and CFD results were strongly correlated (R2 = 0.75 and R2 = 0.81) with low root mean square error (0.21 m/s and 3.8 mmHg). SFD and NFD were moderately and weakly correlated at R2 = 0.44 and R2 = 0.20, respectively. Flow visualization through streamlines indicates good qualitative agreement between 4D flow MRI data and CFD results in most cases.Conclusion: The approach presented here seems suitable to estimate postoperative maximum velocity and pressure gradient in patients receiving biological AVR, using only preoperative MRI data. The workflow can be performed in a reasonable time frame and offers a method to estimate postoperative valve prosthesis performance and to identify patients at risk of patient-prosthesis mismatch preoperatively. Novel parameters, such as SFD and NFD, appear to be more sensitive, and estimation seems harder. Further workflow optimization and validation of results seems warranted.

Comparison of hemodynamics in biological surgical aortic valve replacement and transcatheter aortic valve implantation: An in‐silico study

“…Since only one SAPIEN 3 prosthesis was available as a model for reconstruction, the other prostheses and sizes used in the intervention were estimated by scaling the SAPIEN's geometry according to the dimensions given by the manufacturers. Representations of the Carpentier‐Edwards Magna Ease (Figure 1) were generated as described in Hellmeier et al 15 For virtual AVR‐procedure, two free edge loops were added to the annular base of the valve, projecting above and below the valve as cylinder‐shaped extensions, enabling a merging to the aorta's geometry and achieving the correct mesh topology.…”

“…This contrasts with findings of Pibarot et al stating that in patients with comparable anulus diameter, the orifice area of TAVI and SAVR does not differ. 8 Since selection of SAVR size based on MDCT usually results in an oversizing of at least one size, 15,18 we decided to virtually implant each patient with a SAVR prosthesis one size smaller than the one originally selected, except for cases 7 and 15, since the narrowest prosthesis was already selected here.…”

CT-Based Analysis of Left Ventricular Hemodynamics Using Statistical Shape Modeling and Computational Fluid Dynamics

“…Detailed reviews on different modeling approaches in cardiovascular medicine are provided by Quarteroni et al ( 12 ), Doost et al ( 13 ), and Hirschhorn et al ( 14 ). Such image-based CFD frameworks can also be employed to investigate post-operative outcomes after virtual treatment ( 5 , 15 , 16 ). Furthermore, CFD can complement cardiac computed tomography (CCT) by functional analysis of patient-specific intraventricular hemodynamics via so-called CCT-based CFD ( 17 ).…”

Effect of transcatheter edge-to-edge repair device position on diastolic hemodynamic parameters: An echocardiography-based simulation study

“…Computational methods and image-based modeling provide tools to investigate several hemodynamic parameters on the basis of patient specific input data. Such approaches have, for instance, been applied to investigate left ventricular hemodynamic flow structures ( 22 – 24 ), the outcome after implantation of biological and mechanical aortic valve prostheses ( 25 ), MV tissue properties ( 26 ), as well as mitral hemodynamics with and without simulation of diseases and treatment ( 27 ). Caballero et al ( 18 ) and Errthum et al ( 28 ) were the first to use advanced computational methods to systematically investigate post-interventional hemodynamic characteristics for one case and several TEER strategies, as well as for specific devices.…”