A Patient-Specific CFD Study of Carotid Webs: Hemodynamic Analysis and the Role of Blood Viscosity

This patient-specific CFD study reveals that while symptomatic carotid webs exhibit lower time-averaged wall shear stress and higher oscillatory shear index compared to asymptomatic cases, wall-based hemodynamic metrics alone cannot reliably distinguish carotid webs from normal carotid bifurcations.

The Gist

The Big Picture: The "Shelf" in the Artery

Imagine your carotid artery (the main highway carrying blood to your brain) is a smooth, wide river. In some people, a strange, thin shelf of tissue grows out of the riverbank. This is called a Carotid Web (CaW).

Think of it like a small, invisible diving board sticking out into a fast-flowing stream. Even though the river isn't blocked (there's no big dam), that little diving board messes up the water flow right behind it.

Doctors know these webs can cause strokes, especially in younger people who don't have typical risk factors like high cholesterol. But they didn't fully understand why or how to tell a dangerous web from a harmless one. This study used computer simulations to look at the "water flow" (blood flow) around these webs.

The Experiment: A Digital Twin Lab

The researchers took CT scans of 28 patients:

• 16 people with webs who had already had a stroke (Symptomatic).
• 6 people with webs who had no symptoms (Asymptomatic).
• 6 people with normal arteries (Controls).

They built digital twins of these arteries on a computer. They then simulated blood flowing through them to see exactly how the "river" behaved around the "diving board."

The Key Questions & Findings

1. Does the "thickness" of the blood matter?

Blood isn't just like water; it's a bit like ketchup. It gets thinner when it flows fast and thicker when it flows slow. The researchers tested three different ways to model this "ketchup effect" (Newtonian, Carreau-Yasuda, and Casson models).

• The Result: It didn't matter which model they used. The flow patterns looked almost identical.
• The Analogy: Imagine trying to predict how a leaf floats down a river. Whether you calculate the water as slightly thick or slightly thin, the leaf still gets stuck in the same eddy behind the rock. For these large arteries, the specific "thickness" of the blood doesn't change the big picture.

2. Can we spot the "bad" webs just by looking at the flow?

The team looked at three main metrics to see if the flow was "trouble":

• TAWSS (Time-Averaged Wall Shear Stress): How hard the blood rubs against the artery wall. (Low rubbing = bad).

• OSI (Oscillatory Shear Index): How much the blood flow wobbles back and forth. (Wobbling = bad).

• RRT (Relative Residence Time): How long blood particles hang around near the wall. (Hanging around = bad, because it lets clots form).

• The Result: They found a lot of overlap. The "bad" webs and the "normal" arteries had very similar numbers.

• The Analogy: It's like trying to tell the difference between a calm lake and a lake with a hidden whirlpool just by measuring the average temperature of the water. Sometimes, a normal river bend creates a small swirl that looks just like the swirl caused by a web. The computer couldn't reliably say, "This is a web, that is normal," just by looking at the average numbers.

3. What about the "Stroke" vs. "No Stroke" groups?

This is where it got interesting. While the numbers overlapped with normal arteries, there was a trend when comparing the two groups of web patients:

• Symptomatic (Stroke) patients: Their webs caused slower rubbing (lower shear), more wobbling (higher oscillation), and blood that stayed longer near the wall.

• Asymptomatic (No Stroke) patients: Their webs were slightly less chaotic.

• The Analogy: Think of two people with a diving board in their pool.

  • Person A (No Stroke): The diving board is small. The water swirls a little behind it, but it mostly clears out.
  • Person B (Stroke): The diving board is larger or angled differently. The water gets stuck in a deep, lazy circle behind it. If you threw a leaf in, it would spin there for a long time. That "stuck" water is where clots (thrombus) like to form.

The "Aha!" Moment

The study found that symptomatic webs create a "stagnation zone"—a pocket where blood gets lazy and hangs out. This is the perfect breeding ground for a clot to form and eventually break off, causing a stroke.

However, the study also admitted a limitation: We can't just use a ruler to measure the flow. Because every person's artery is shaped differently (some are curvy, some are straight), the "average" numbers get messy. A normal artery can sometimes look just as "disturbed" as a web artery if you only look at the numbers.

The Takeaway for Everyone

1. Carotid Webs are tricky: They aren't always obvious blockages, but they create "traffic jams" for blood flow that can lead to strokes.
2. The "Stuck" Factor: The danger isn't just the web itself, but how much it makes the blood stall and swirl behind it. The more the blood stalls, the higher the risk of a stroke.
3. One Size Doesn't Fit All: Because every patient's anatomy is unique, doctors can't rely on a single "magic number" to diagnose risk. They need to look at the specific shape of the web and how it interacts with that specific person's blood flow.

In short: The researchers built a video game of blood flow to prove that while we can see the "swirls" caused by these webs, telling a dangerous web from a safe one requires looking at the whole picture, not just the numbers.

Technical Summary

1. Problem Statement

Carotid webs (CaWs) are shelf-like intimal protrusions in the internal carotid artery (ICA) increasingly recognized as a cause of ischemic stroke, particularly in younger patients without traditional risk factors. While clinical imaging (DSA) shows delayed contrast clearance downstream of CaWs, suggesting flow stasis, the specific hemodynamic mechanisms linking CaWs to thrombosis remain unclear.

Existing Computational Fluid Dynamics (CFD) studies face several limitations:

• Geometric Simplification: Many use artificial geometries that fail to capture patient-specific anatomical heterogeneity.
• Metric Sensitivity: Traditional methods for comparing hemodynamics (e.g., local extrema or fixed Region-of-Interest averaging) are either unstable (sensitive to mesh resolution) or dilute localized disturbances.
• Viscosity Uncertainty: It is unclear whether modeling blood as a non-Newtonian fluid (shear-thinning) significantly alters hemodynamic predictions in the disturbed flow fields created by CaWs compared to a standard Newtonian model.
• Differentiation Challenge: It is unknown if wall-based hemodynamic metrics (Time-Averaged Wall Shear Stress [TAWSS], Oscillatory Shear Index [OSI], Relative Residence Time [RRT]) can effectively distinguish CaW geometries from normal bifurcations or symptomatic from asymptomatic cases.

2. Methodology

The study employed a patient-specific CFD framework to analyze 28 subjects: 22 with CaWs (16 symptomatic, 6 asymptomatic) and 6 age-matched controls with normal bifurcations.

• Data Acquisition & Segmentation:
  • Geometries were reconstructed from CT Angiography (CTA) using 3D Slicer.
  • The domain included the common carotid artery (CCA), bifurcation, external carotid artery (ECA), and proximal ICA.
  • Digital Subtraction Angiography (DSA) was used in a subset of patients for qualitative validation of flow patterns.
• CFD Simulation Setup:
  • Solver: OpenFOAM (v2206) using the pimpleFOAM transient solver for incompressible Navier-Stokes equations.
  • Boundary Conditions: Rigid walls with no-slip conditions; pulsatile inlet flow based on published data (rescaled to patient heart rate); traction-free outlets.
  • Meshing: Unstructured tetrahedral meshes (>4 million elements) verified for independence via WSS convergence.
• Viscosity Models: Three models were tested to evaluate sensitivity:
  1. Newtonian: Constant viscosity ($3.5 \times 10^{-3}$ Pa·s).
  2. Carreau–Yasuda: Captures shear-thinning behavior.
  3. Casson: Incorporates yield stress and shear-thinning.
• Novel Analysis Framework:
  • Gaussian-Weighted Spatial Averaging: To overcome the limitations of local extrema or fixed ROIs, the authors introduced a continuous weighting function centered at the CaW throat. This method smoothly down-weighs distant contributions, providing a single, comparable scalar value for TAWSS, OSI, and RRT across different patient geometries.
  • Metrics: TAWSS (frictional force magnitude), OSI (directional change), and RRT (blood residence time near the wall).

3. Key Contributions

1. Methodological Innovation: Development of a Gaussian-weighted spatial averaging technique that enables robust cross-patient comparison of hemodynamic metrics without relying on arbitrary region definitions or unstable local maxima.
2. Viscosity Model Validation: A systematic evaluation demonstrating that the choice of blood viscosity model (Newtonian vs. non-Newtonian) has negligible impact on hemodynamic predictions in carotid webs.
3. Hemodynamic Phenotyping: Detailed characterization of flow disturbances in "common" CaWs (web at bifurcation) versus "atypical" CaWs (web within ICA), revealing distinct flow patterns based on web location.
4. Clinical Correlation: Quantitative comparison of hemodynamic metrics between symptomatic and asymptomatic CaWs, and between CaWs and normal bifurcations.

4. Key Results

• Flow Validation: CFD simulations successfully reproduced the low-velocity recirculation zones and delayed contrast clearance observed in clinical DSA images, validating the model's ability to capture CaW-induced flow disturbances.
• Viscosity Impact: The three viscosity models yielded nearly identical results. Differences in peak WSS were less than 2%, and flow patterns remained unchanged. This suggests that for large artery hemodynamics with CaWs, the computationally expensive non-Newtonian models are unnecessary; a Newtonian assumption is sufficient.
• CaW vs. Normal Bifurcation:
  • While CaW cases showed greater inter-patient variability, the mean and median hemodynamic metrics (TAWSS, OSI, RRT) overlapped significantly with the normal group.
  • Statistical analysis showed no significant difference between symptomatic CaWs and normal bifurcations (e.g., TAWSS $p=0.858$).
  • Interpretation: The natural curvature and expansion of the carotid bulb generate low shear and oscillatory flow even in normal arteries, masking the specific signal of the CaW when using averaged metrics.
• Symptomatic vs. Asymptomatic CaWs:
  • Symptomatic CaWs exhibited a trend toward lower TAWSS (3.39 Pa vs. 6.63 Pa), higher OSI, and higher RRT compared to asymptomatic cases.
  • While these trends were not statistically significant (likely due to sample size), they suggest that symptomatic lesions create more severe stasis and flow oscillation.
• Location Matters:
  • Common CaWs: Generated pronounced downstream recirculation and low TAWSS distal to the web.
  • Atypical CaWs (ICA): Generated dominant upstream disturbances within the carotid bulb, with only minor downstream effects.

5. Significance and Implications

• Clinical Utility: The study suggests that wall-based hemodynamic metrics alone (TAWSS, OSI, RRT) may be insufficient to distinguish CaWs from normal carotid anatomy or to definitively predict stroke risk based solely on averaged values.
• Modeling Efficiency: The finding that Newtonian models are sufficient for CaW analysis simplifies future CFD studies, reducing computational costs without sacrificing accuracy.
• Future Directions: The results indicate that future risk stratification may require:
  • Volumetric/Transport Metrics: Moving beyond wall shear stress to Lagrangian particle tracking and residence time analysis to better capture stasis.
  • Morphological Integration: Combining hemodynamics with detailed geometric characterization (web thickness, angle, pocket volume) to understand how specific angioarchitectures drive thrombosis.
  • Larger Cohorts: The lack of statistical significance highlights the need for larger, multi-center datasets to detect subtle but clinically relevant hemodynamic differences.

Conclusion: While CaWs create distinct local flow disturbances (recirculation and stasis) that correlate with clinical DSA findings, these disturbances do not shift the averaged wall shear stress metrics enough to clearly separate CaW patients from normal controls. However, symptomatic CaWs do exhibit a trend toward more disturbed flow (lower shear, higher oscillation) than asymptomatic ones, suggesting that local flow dynamics play a role in thrombus formation that requires more sophisticated, non-averaged analysis methods to fully quantify.