Computational Hemodynamics-Based Growth Prediction for Small Abdominal Aortic Aneurysms: Laminar Simulations Versus...

Area of Science:

• Biomedical Engineering
• Computational Science
• Cardiovascular Research

Background:

• Computational fluid dynamics (CFD) simulations are crucial for assessing patient-specific hemodynamics in abdominal aortic aneurysms (AAAs).
• Hemodynamic stressors derived from CFD are frequently used to predict AAA growth, but the impact of different simulation models remains unclear.

Purpose of the Study:

• To investigate the impact of different CFD simulation models (laminar vs. large eddy simulation) on the predictive modeling of AAA growth.
• To compare the ability of hemodynamic parameters from laminar and LES models to predict AAA growth status.

Main Methods:

• CFD simulations were performed on 70 AAA models from patient CTA data with known growth status (fast vs. slow).
• Laminar and large eddy simulation (LES) models were used to obtain hemodynamic parameters like wall shear stress (WSS) and flow stability.
• Machine learning classifiers (SVM, KNN, GLM) were employed using morphological, hemodynamic, and patient health data to predict AAA growth status.

Main Results:

• Aneurysmal flow stability and WSS were comparable between laminar and LES simulations.
• Hemodynamic variables from both models demonstrated similar capabilities in differentiating AAA growth status.
• The predictive performance of machine learning classifiers showed less than a 2% difference between the two simulation models (p > 0.05).

Conclusions:

• The choice between laminar and LES CFD simulation models did not significantly impact the outcomes of computational hemodynamics for AAA growth prediction.
• Both investigated flow simulation models provided comparable hemodynamic parameters for predicting AAA growth status.
• These findings suggest that simpler laminar models may suffice for predicting AAA growth, simplifying complex simulations.