Project

Back to overview

OPTAVI: adjoint-based OPTimization of deflection-based cerebral embolic protection devices for reducing stroke risk in Transcatheter Aortic Valve Implantation

Applicant Zolfaghari Hadi
Number 191786
Funding scheme Early Postdoc.Mobility
Research institution DAMTP Centre for Mathematical Sciences University of Cambridge
Institution of higher education Institution abroad - IACH
Main discipline Fluid Dynamics
Start/End 01.04.2020 - 30.09.2021
Show all

All Disciplines (2)

Discipline
Fluid Dynamics
Mathematics

Keywords (7)

TAVI; High-order computational fluid dynamics; Trasncatheter Aortic Valve Implantation (TAVI); Cerebral Embolic Protection Device; Nonlinear Adjoint-based Optimization; Flow Control; High-performance Computing

Lay Summary (French)

Lead
Bien que la chirurgie soit utilisée principalement pour remplacer une valve aortique malade, cette procédure peut être risquée pour le patient âgé. L’implantation transcathéter de valve aortique (ITVA), lors de laquelle la valve est remplacée par un tube depuis une artère offre une alternative à la chirurgie. Le risque significatif d’attaque cérébrale, dû, par exemple, à l’embolisation de tissu valvulaire calcifié, représente un inconvénient majeur. Des dispositifs de protection d’embolie cérébrale (DPEC) ont été récemment proposés afin de diminuer ce risque. Cependant, ils ne sont que partiellement efficaces. Ce projet utilise des techniques avancées de mécanique des fluides numérique afin de trouver des modèles de DPEC optimaux pour réduire encore le risque d’attaque cérébrale lors d’une ITVA.
Lay summary

Lead

Bien que la chirurgie soit utilisée principalement pour remplacer une valve aortique malade, cette procédure peut être risquée pour le patient âgé. L’implantation transcathéter de valve aortique (ITVA), lors de laquelle la valve est remplacée par un tube depuis une artère offre une alternative à la chirurgie. Le risque significatif d’attaque cérébrale, dû, par exemple, à l’embolisation de tissu valvulaire calcifié, représente un inconvénient majeur. Des dispositifs de protection d’embolie cérébrale (DPEC) ont été récemment proposés afin de diminuer ce risque. Cependant, ils ne sont que partiellement efficaces. Ce projet utilise des techniques avancées de mécanique des fluides numérique afin de trouver des modèles de DPEC optimaux pour réduire encore le risque d’attaque cérébrale lors d’une ITVA.

But du projet de recherche au début de la période de recherche et après l’achèvement du projet

Le projet utilise un solveur numérique basé sur GPU pour le calcul du flux sanguin et des techniques d’optimisations basé sur le gradient pour trouver des scenarii avec un moindre risque d’attaque cérébrale. Ces techniques ont été utilisées dans le passé avec succès pour l’optimisation d’aile d’avion. Leur application dans le domaine cardiovasculaire pourrait contribuer à une modélisation optimale en diminuant les coûts expérimentaux de manière substantielle. Après son achèvement, le projet fournira des outils d’optimisation concernant la mécanique du flux sanguin lors d’ITVA ainsi que d’un modèle optimal de scenarii afin de réduire le risque d’attaque cérébrale lors d’ITVA.

Contexte scientifique et sociétal du projet de recherche

Le projet inclut des aspects scientifiques de pointe, tels que l’optimisation non-linéaire, nécessitant des ressources avancées de hautes performances numériques. Outre l’aspect scientifique, le projet a pour but de minimiser le risque d’attaque cérébrale lors d’ITVA, constituant un impact clinique direct concret.

Direct link to Lay Summary Last update: 10.03.2020

Responsible applicant and co-applicants

Abstract

Transcatheter aortic valve implantation (TAVI) has quickly become the clinical standard for patients with medium to high risk for surgery. TAVI usually involves guiding two or more catheters through an artery to the aorta, typically one for carrying the crimped aortic valve prosthesis, and one for contrast injection for X-ray angiography. Despite being a promising treatment for aortic valve disease in the elderly, stroke remains a major complication of TAVI. Large randomized clinical trials reported stroke within 30 days in 5-7% of the patients undergoing TAVI [1, 2, 3, 4]. Transcranial Doppler ultrasound imaging has showed cerebral embolic signals in 100% of the TAVI patients, mainly during valve deployment [5]. Diffusion-weighted MRI has also shown ischemic brain lesions in 68-100% of patients after the TAVI [6, 7, 8]. While the implications of such new brain lesions are not fully understood, the accumulation of 1.5-4.3cm 3 [9] of these “silent” cerebral lesions has been attributed to neurological complications such cognitive decline, high risk of dementia [10], and more than three-fold increase in risk of subsequent stroke [11]. Cerebral embolic protection devices (CEPD) such as Sentinel[12], or TriGaurd HDH[13] have been developed to prevent the migration of debris (emboli) to the brain during TAVI, and resulted in 44-46% decrease in brain lesions [12, 13]. Despite this progress, the risk still remains significant, and so it is imperative to develop new approaches, including improving the design of the new devices so that brain lesions can be reduced if not eliminated. This is of crucial impact particularly in the western aging societies, where transcatheter procedures have become the treatment of choice for aortic valve disease.?When the catheters are guided through the aorta to the valve position during the TAVI procedure, they may scratch fragile atherocelerotic plaques on the aortic wall. These embolic debris can find their way to the brain while being advected in an unsteady and high Reynolds number flow in the aorta. A significant amount of debris also enters the bloodstream when the TAVI valve is deployed: by radially squeezing the diseased and calcified valve, small chunks of calcification or valve tissue can detach from the valve and enter the bloodstream. The captured debris during TAVI procedures using the Sentinel device has is made up of a variety of materials, e.g., calcification, valve tissue, aortic wall pieces and thrombus [14]. Current CEPDs are based on fairly simple-minded ideas, such?as placing filters inside brachiocephalic and left common cartoid arteries (Sentinel), or simply covering the arteries in the aortic arch with a filter to deflect the debris downstream (e.g. TriGaurd HDH and Trigaurd 3). Because the flow here is turbulent and laden with solid particles with nearly random size, shape and density, more advanced physical understanding is needed to examine and/or enhance the hydrodynamic efficacy of CEPDs.?This project aims at creating a 3D high-order nonlinear adjoint-based [15, 16] framework for optimization of the CEPDs performance when exposed to the particle-laden turbulent flow in the aorta. This employs a massively parallel GPU-based turbulent flow simulation tool [17, 18] for the nonlinear direct and adjoint computations. The Lagrangian multiplier in the adjoint formulation will accommodate the incompressible Navier-Stokes equations?together with particle transport equations as constraints. First, the flow through a model of a prosthetic heart valve (PhD project of the applicant) will be extended to include the full patient specific geometry of the thoracic aorta using a sharp-interface immersed boundary method. Second, a Lagrangian model for finite-sized particles representing embolic debris will be coupled into the flow solver. Third, a deflection-based CEPD geometry will be?introduced into the model. As a fourth and the most significant step, nonlinear adjoint-based variational capabilities will be added on top of the particle-laden turbulent flow solver. The researcher will extend his preliminary experience with nonlinear adjoint-based methods with the theoretical expertise that the host group in Cambridge has to offer, to design and deploy a gradient-based method for minimization of particle transport to the brain. Upon development of the adjoint equations and code suite, iterative direct-adjoint looping simulations will be performed to produce a design shape with maximum cerebral protection.References:[1] David H Adams, et al., Transcatheter aortic-valve replacement with a self-expanding prosthesis. New England Journal of Medicine, 370(19):1790-1798, 2014.?[2] Craig R Smith, et al., Transcatheter versus surgical aortic-valve replacement in high-risk patients. New England Journal of Medicine, 364(23):2187-2198, 2011.?[3] Jeffrey J Popma , et al., Transcatheter aortic valve replacement using a self-expanding bioprosthesis in patients with severe aortic stenosis at extreme risk for surgery. Journal of the American College of Cardiology, 63(19):1972-1981, 2014.?[4] Blase A Carabello. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. Current Cardiology Reports, 13(3):173-174, 2011.?[5] Gabor Erdoes, et al., Transcranial doppler-detected cerebral embolic load during transcatheter aortic valve implantation. European Journal of Cardio-thoracic Surgery, 41(4):778-784, 2011.?[6] Philipp Kahlert, et al., Silent and apparent cerebral ischemia after percutaneous transfemoral aortic valve implantation: a diffusion-weighted magnetic resonance imaging study. Circulation, 121(7):778-784, 2010.?[7] Martin Arnold, et al., Embolic cerebral insults after transapical aortic valve implantation detected by magnetic resonance imaging. JACC: Cardiovascular Interventions, 3(11):1126-1132, 2010.?[8] Timothy A Fairbairn, et al., Diffusion-weighted mri determined cerebral embolic infarction following transcatheter aortic valve implantation: assessment of predictive risk factors and the relationship to subsequent health status. Heart, 98(1):18-23, 2012.?[9] Thomas R Insel, et al., The NIH brain initiative. Science, 340(6133):687-688, 2013.?[10] Sarah E Vermeer, et al., Silent brain infarcts: a systematic review. The Lancet Neurology, 6(7):611-619, 2007.?[11] C Bernick, et al., Silent MRI infarcts and the risk of future stroke: the cardiovascular health study. Neurology, 57(7):1222-1229, 2001.?[12] Julia Seeger, et al., Significant differences in debris captured by the sentinel dual-filter cerebral embolic protection during transcatheter aortic valve replacement among different valve types. JACC: Cardiovascular Interventions, 11(17):1683-1693, 2018.?[13] Alexandra J Lansky, et al., A prospective randomized evaluation of the TriGuard HDH embolic deflection device during transcatheter aortic valve implantation: results from the deflect III trial. European Heart Journal, 36(31):2070-2078, 2015.?[14] Samir R Kapadia, et al., Protection against cerebral embolism during transcatheter aortic valve replacement. Journal of the American College of Cardiology, 69(4):367-377, 2017.?[15] Rich R Kerswell, Nonlinear nonmodal stability theory. Annual Review of Fluid Mechanics, 50:319-345, 2018.?[16] Thomas R Bewley, et al., DNS-based predictive control of turbulence: an optimal benchmark for feedback algorithms. Journal of Fluid Mechanics, 447:179-225, 2001.?[17] Hadi Zolfaghari, et al., High-order accurate simulation of incompressible turbulent flows on many parallel GPUs of a hybrid-node supercomputer. Computer Physics Communications, 244:132-142, 2019.?[18] Rolf Henniger, et al., High-order accurate solution of the incompressible Navier-Stokes equations on massively parallel computers. Journal of Computational Physics, 229(10):3543-3572, 2010.
-