Presentation pdf 2.1MB

Development of a Patient-Specific Fluid-Structure Interaction Model for Acute Ischemic Stroke

An estimated 700,000 acute ischemic strokes (AIS) occur annually in the United States [1] due to embolic occlusion of a cerebral artery. Despite improved recanalization rates with new stent retriever devices, over 15% of patients cannot be recanalized and another 17% die within 90 days despite successful recanalization1. In order to investigate the underlying mechanics of a lodged embolus and determine its optimal surgical removal procedure, we are developing a continuum-based fluid-structure interaction (FSI) model to simulate dislodgement under applied aspiration pressures. The model will allow for patient-specific geometries, blood and thromboemboli properties, and hemodynamic conditions to be investigated and to determine their individual contributions to thrombus removal. The modeling framework combines established viscoelastic blood and embolus models with a cohesive zone (CZ) model for the interface between the embolus and artery wall. The CZ is a physical surface with negligible inertia that can provide a traction distribution as a function of the opening displacement and rate[2]. All constitutive models were implemented in OpenFOAM in a tightly coupled finite-volume FSI framework. The general FSI framework in OpenFOAM was further modified to include two-way FSI coupling on both the upstream and downstream surfaces of the thromboembolus. To account for the variable viscosity of our viscoelastic blood model, the near-wall blood viscosity was found at each FSI interface and used to calculate the fluid traction forces that were applied to the solid mesh as boundary conditions. Dynamic mesh motion of the fluid domains was implemented through OpenFOAM’s dynamic mesh library and governed by a Laplace equation with variability diffusivity. Aitken under-relaxation was also employed within the FSI solver to improve stability of the partitioned approach. To validate the FSI computational solver, an experimental flow loop was developed to track the displacement of an embolus under controlled cyclic aspiration in a fluid filled chamber. The flow loop consisted of a medical aspirator, solenoid valve, Arduino controller, 3D-printed middle cerebral artery chamber, and a rubber embolus analog. Cyclic pressures were applied near the embolus surface via an aspiration catheter at 1 Hz and the embolus analog’s motion was tracked and analyzed with in-house Matlab code. Following the solver’s validation in the simplified geometry, an AIS patient’s Circle of Willis anatomy was reconstructed from CT images and separate fluid and solid domain meshes were created with ‘snappyHexMesh’. AIS patient internal carotid and basilar artery flow rates were prescribed at the model inlets [4] and 3-element Windkessel model boundary conditions were prescribed at the six CoW outlets. Baseline simulations were performed under pulsatile cardiac conditions and without surgical intervention to determine the embolus motion and flows and pressures local to the lodged embolus. Future simulations will incorporate an AIS aspiration catheter and clinical aspiration protocols to determine the FSI model’s ability to predict clot removal.

1. Grech et al. The Neuroradiology Journal. 2015. 2. Costanzo. International Journal of Engineering Science. 1998.