Computational Analysis of Fluid Dynamics in the Transcatheter Aortic Valve Replacement

Keywords:
Flow Mechanics; Transcatheter Aortic Valve Replacement/methods; Hemodynamics; Regional Flood Flow

Transcatheter aortic valve replacement (TAVR), a minimally invasive heart surgery, was introduced by Cribier et al.¹ as an alternative to the traditional open-heart surgery in the treatment of individuals with severe aortic valve stenosis and at high surgical risk due to advanced age or the presence of multiple comorbidities.² After the first pioneering efforts, the advent of innovative prosthetic valves, and more technologically refined approaches and devices, the use of TAVR for patients with intermediate surgical risk has been a worldwide trend.³ However, variation in the prosthetic valve positioning and orientation post TAVR procedure can produce significant changes in the aortic hemodynamics and the corresponding stresses in the vessel wall.⁴

Within the aorta, there are two categories of vessel wall stress. The first category of stress is the result of the friction between the moving blood and the vessel wall, which is proportional to the blood speed, moving away from the intimate layer of the vessel wall. This kind of stress is known as wall shear stress (WSS). The second category of stress is due to the variation in pulse pressure generated during the cardiac cycle. In this category, there are circumferential, axial and radial stress transferred to all vessel wall layers. With advancing age, the aorta enlarges, the arch changes shape from a near-perfect semicircle, and the vessel generally becomes more tortuous.⁵ Moreover, the change in the natural curvature of the aorta introduces secondary flow dynamics and flow asymmetry, which directly influence WSS distribution and magnitude over the vessel wall.

Among the available imaging modalities, computed tomography (CT) is widely considered the gold standard method for studying and analyzing the aorta, coronary and femoral arteries. Recent developments using a wide coverage detector design (256 or 320 slices) or high-frequency dual-source CT have made it possible to use less contrast and a lower radiation dose. Although CT can present the geometrical and functional complexities of the aorta, it is currently limited to capture a snapshot of the blood flow at a defined instant of time during the cardiac cycle.

On the other hand, four-dimensional (4D) flow magnetic resonance imaging (MRI) is a novel technique with the capability of assessing aortic blood flow in three-dimensional space as a function of time, which permits the quantification of aortic hemodynamics.⁶ This new imaging acquisition technique may improve our understanding of the inherent dynamicity of aortic blood flow. However, CT can be improved with computational fluid dynamics (CFD) modeling, which can compute previously unmeasurable hemodynamic parameters to understand the biomechanical behavior of blood flow in both normal and diseased vessels.

In the absence of a readily applicable means to directly measure WSS, CFD has been applied in CT and MRI images to understand both the spatial and temporal patterns of WSS and the influence of aortic flow dynamics on this parameter.^7–9 Using CT images as the input of a CFD model, Celis et al.¹⁰ demonstrated that small variations of the aortic valve tilt angle could modify the nature of the flow and produce changes in the distribution of the WSS over the aorta wall.

CFD is a feasible method that has been used for ages¹¹ in determining fluid flow and 3D model of coronary arteries and can simulate an accurate vessel flow based on a set of given parameters. For incompressible fluids, most CFD analysis solve the Navier-Stokes and continuity equations that govern fluid motion. This set of equations includes non-linear and partial differential equations based on the principle of conservation of mass and momentum. Navier-Stokes equation describes the viscous motion of fluids¹² and, according to Newton’s law of viscosity, the relationship between the shear stress and shear rate of a fluid, subjected to mechanical stress, is a constant for a given temperature and pressure, and is defined as the viscosity or coefficient of viscosity. Physiologically, this means that the blood flow in the cardiovascular system is equal to the change of blood pressure divided by the system resistance.¹³

Despite the availability of powerful CFD software packages to model fluid flow, such as ANSYS FLUENT, OpenFOAM, SIMVascular, and ADINA,¹⁴ the current CFD methods have large computational time cost, which prevents them from being used in large patient cohorts. This time cost basically comes from the complexity of the models, which need patient anatomic geometries, tissue properties, hemodynamics loading conditions, and proper selection of modeling techniques. A potential paradigm-changing solution to the bottlenecks in current CFD methods is to incorporate machine learning (ML) algorithms¹⁵ to expedite computational analysis, starting from geometry modeling to computational model setup, and simulation completion.

Liang et al.¹⁶ have recently developed a novel machine learning approach that demonstrated the feasibility of using ML as a fast and accurate surrogate of CFD to estimate steady-state hemodynamic fields of human thoracic aorta. In their approach, CFD is treated as a black box, and the ML algorithm learns the nonlinear relationship between CFD input and output. On average, the proposed method took minutes to run a CFD simulation for each aorta model, which seems to be fast enough for clinical applications.

In vivo measurements of parameters hemodynamics and the corresponding stress in the aorta are not practical. Therefore, CFD is widely used to estimate these parameters, but it is time consuming and computationally expensive. ML models can be a promising alternative for CFD simulations to aid clinical decisions and treatment based on specific patients. This can lead to better clinical results in many studies, such as the identification of the best position and orientation of the prosthetic valve in the TAVR procedure.

• Short Editorial related to the article: Prediction of Stress Map in Ascending Aorta - Optimization of the Coaxial Position in Transcatheter Aortic Valve Replacemento

Referências

• ¹
  Cribier A, Eltchaninoff H, Bash A, Borenstein N, Tron C, Bauer F, et al. Percutaneous Transcatheter Implantation of an Aortic Valve Prosthesis for Calcific Aortic Stenosis. Circulation. 2002;106(24):3006–8.
• ²
  Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson LG, et al. Transcatheter Aortic-Valve Implantation for Aortic Stenosis in Patients Who Cannot Undergo Surgery. N Engl J Med. 2010;363(17):1597–607.
• ³
  Morello A, Corcione N, Ferraro P, Cimmino M, Pepe M, Cassese M, et al. The best way to transcatheter aortic valve implantation: From standard to new approaches. Int J Cardiol. 2020 [Internet]. [Cited in 2020 Aug 09]. Avaiable from: internationaljpurnalofcardiology.com/action/showPdfpii=501’67-5273%2820%2933563-4
  » internationaljpurnalofcardiology.com/action/showPdfpii=501’67-5273%2820%2933563-4
• ⁴
  Groves EM, Falahatpisheh A, Su JL, Kheradvar A. The Effects of Positioning of Transcatheter Aortic Valves on Fluid Dynamics of the Aortic Root. ASAIO J [Internet]. 2014;60(5):545-602.
• ⁵
  Farag ES, Vendrik J, van Ooij P, Poortvliet QL, van Kesteren F, Wollersheim LW, et al. Transcatheter aortic valve replacement alters ascending aortic blood flow and wall shear stress patterns: A 4D flow MRI comparison with age-matched, elderly controls. Eur Radiol. 2019;29(3):1444–51.
• ⁶
  Dyverfeldt P, Bissell M, Barker AJ, Bolger AF, Carlhäll C-J, Ebbers T, et al. 4D flow cardiovascular magnetic resonance consensus statement. J Cardiovasc Magn Reson. 2015;17(1):72.
• ⁷
  Biasetti J, Hussain F, Gasser TC. Blood flow and coherent vortices in the normal and aneurysmatic aortas: a fluid dynamical approach to intra-luminal thrombus formation. J R Soc Interface. 2011;8(63):1449–61.
• ⁸
  Jarral OA, Tan MKH, Salmasi MY, Pirola S, Pepper JR, O’Regan DP, et al. Phase-contrast magnetic resonance imaging and computational fluid dynamics assessment of thoracic aorta blood flow: A literature review. Eur J Cardio-thoracic Surg. 2020;57(3):438–46.
• ⁹
  Callaghan FM, Grieve SM. Translational Physiology: Normal patterns of thoracic aortic wall shear stress measured using four-dimensional flow MRI in a large population. Am J Physiol - Hear Circ Physiol. 2018;315(5):H1174–81.
• ¹⁰
  Celis D, Alvares B, Gomes DA, Ibanez I, Azevedo PN, et al. Predição do Mapa de Estresse em Aorta Ascendente : Otimização da Posição Coaxial no Implante Valvar Aórtico Percutâneo. Arq Bras Cardiol. 2020; 115(4):680-687.
• ¹¹
  Papadopoulos KP, Gavaises M, Pantos I, Katritsis DG, Mitroglou N. Derivation of flow related risk indices for stenosed left anterior descending coronary arteries with the use of computer simulations. Med Eng Phys. 2016;38(9):929–39.
• ¹²
  Schneiderbauer S, Krieger M. What do the Navier{\textendash}Stokes equations mean? Eur J Phys. 2013;35(1):15020.
• ¹³
  Doutel E, Pinto SIS, Campos JBLM, Miranda JM. Link between deviations from Murray’s Law and occurrence of low wall shear stress regions in the left coronary artery. J Theor Biol. 2016;402:89–99.
• ¹⁴
  Ong CW, Wee I, Syn N, Ng S, Leo HL, Richards AM, et al. Computational Fluid Dynamics Modeling of Hemodynamic Parameters in the Human Diseased Aorta: A Systematic Review. Ann Vasc Surg [Internet]. 2020;63:336–81. Available from: http://www.sciencedirect.com/science/article/pii/S089050961930487X
  » http://www.sciencedirect.com/science/article/pii/S089050961930487X
• ¹⁵
  LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015 May 27;521(7553):436-44.
• ¹⁶
  Liang L, Mao W, Sun W. A feasibility study of deep learning for predicting hemodynamics of human thoracic aorta. J Biomech. 2020;99:109544.

Publication Dates

• Publication in this collection
  23 Oct 2020
• Date of issue
  Oct 2020