A Magnetic Resonance Imaging-based Computational Protocol for Analysis of Plaque Morphology and Hemodynamics in Patients with Carotid Artery Stenosis

Our research assesses physiologically relevant risk factors for plaque embolism and stroke in patients with internal carotid artery stenosis. Specifically, we look at how plaque morphology and its hemodynamic environment differ amongst patients with carotid artery stenosis. In the field of cerebrovascular research, recent efforts have aimed to identify predictors of stroke in asymptomatic carotid artery stenosis patients with a specific focus on patient-specific risk factors, imaging characteristics, and hemodynamic parameters, which can be linked to increased stroke risk.

Computational fluid dynamics enables non-invasive patient-specific analysis of blood flow, and it's increasingly used alongside magnetic resonance imaging and photon counting CTA to assess carotid plaque structure and composition. Current magnetic resonance imaging is limited by long scan times, complex image interpretation, and repositioning errors. And on the other hand, computational fluid dynamics suffers from limited patient-specific data and poor model tuning, thus reducing its accuracy.

Our group has demonstrated that patients with similar degrees of internal carotid artery stenosis narrowing exhibit distinct hemodynamic profiles and that in patients with bilateral stenosis, the degree of stenosis severity influences each side's hemodynamics and flow characteristics. And this really underscores the complex cerebrovascular and hemodynamic interactions in these patients. To begin, launch the CRIMSON software on a computer system.

Import de-identified DICOM image data for patient-specific anatomy into CRIMSON by using the import button in the data manager. Using the geometry modeling window, select vessel path editing and create a vessel tree containing the common carotid artery, external carotid artery, and internal carotid artery. Begin the common carotid artery center line at the level of C5 where the PC MRI flow waveform was obtained.

Place the internal carotid artery center line so it ends one to two centimeters distal to the stenosis, matching the PC MRI waveform location. Then place the external carotid artery center line endpoint proximal to the first order branches, matching the PC MRI waveform acquisition location. Now use the vessel re-slice window to visualize the center line and cross-sectional views perpendicular to the center line after adding at least two points along each vessel.

Using the same window, add vessel contours to specify the boundaries of the vessel wall. The left side of the vessel re-slice window displays the original image while the right side displays the image gradient for defining contours. Place contours frequently enough to capture vessel curvature and changing geometry without overfitting.

After defining all contours, use the loft button in the vessel contour modeling window to generate a combined 3-dimensional solid model via lofting. Then use the vessel blending window to select the fillet algorithm for blending the vessel model into a single solid geometry. Open the meshing and solver setup window and click the meshing button to view meshing options to configure mesh parameters.

In the global options window, set the global element size to an absolute value between 0.5 millimeter and 0.75 millimeter. Then set boundary layer type to geometric growth, the total number of layers to three, the first layer thickness to 0.2 millimeter and total layer thickness of one millimeter. Then apply curvature refinement to add mesh elements at high curvature regions such as the stenosis.

Right-click on the mesh and click on the mesh information button to review mesh metrics including element count, aspect ratios, and distribution. To specify boundary conditions, click on the meshing and solver setup window. Select the solver setup icon, then add a boundary condition set using the BC icon.

View the boundary conditions currently available in CRIMSON. Click the BC icon again, select no slip to implement rigid, non-deformable walls, and apply this to all walls using the apply to all walls option. Then select prescribed velocity.

Import the previously defined inflow waveform and map a parabolic velocity profile to the CCA inlet. Similarly, import the ECA pulsatile outflow waveform mapping a parabolic velocity profile to the ECA outlet. Now click the BC icon, choose RCR, and populate a three element Windkessel model, consisting of proximal resistance, distal resistance, and capacitor.

Map the RCR to the ICA outlet based on patient-specific calculations. To prepare solver parameters, navigate to the meshing and solver window, click on the solver setup icon, then choose solver parameters. Set the step size to 0.1 millisecond for four cardiac cycles, requires residual to 10 to the power of minus four, and blood density to 1, 060 kilograms per cubic meter.

Use solver setup to generate all simulation input files, including the flow data, inlet flow at each time step, the mesh and boundary conditions, face on which each boundary condition is applied, the first time step number of the simulation, three element Windkessel data, pressure and velocity at every point in the mesh, and the instructions for the flow solver. Add the Carreau-Yasuda model into the solver. inp and add to the simulation files to allow for blood to be modeled as a non-Newtonian fluid.

Now run simulation in the steady pane of the solver setup window to run the CRIMSON Navier-Stokes flow solver. Specify the number of processors in the command window. When the solver starts running, the output file histor.

dat will be printed in the command line and saved in a new directory, n-procs-case. Use the Linux prompt tail f histor. dat to observe the file in real time.

The first column of the file corresponds to time step. The second column is the elapsed time. The third column is the non-linear residual, and the fourth column is the log residual value.

A high quality mesh with low aspect ratio elements was generated to accurately represent the carotid artery bifurcation geometry. A representative velocity profile across the carotid bifurcation and ICA stenosis was simulated, showing a maximum flow velocity of approximately 275 centimeters per second at peak systole across the stenosis. Pressure mapping showed negligible pressure gradient across the stenosis in one case with proximal and distal pressures nearly overlapping throughout the cardiac cycle.

In a contrasting case, pressure proximal to the stenosis was significantly higher than distal pressure, revealing a prominent pressure drop. Wall shear stress was low across the bifurcation in the non-stenotic model, especially in the outer walls of the internal and external carotid arteries. In the stenotic model, high wall shear stress was concentrated at the internal carotid artery stenosis.

Oscillatory shear index mapping showed lesion-associated OSI values before surgery were lower compared to postoperative values. HU match imaging identified intraplaque hemorrhage through hyperintense signal on T1 weighted imaging and lowered values on the T1 map. Calcified plaque was identified by consistently hypointense signal on dark blood, T1 weighted, and T2 weighted sequences.