Hemodynamic effects of bifurcation and stenosis geometry on carotid arteries with different degrees of stenosis

Carotid artery stenosis (CAS) is a key factor in pathological conditions, such as thrombosis, which is closely linked to hemodynamic parameters. Existing research often focuses on analyzing the influence of geometric characteristics at the stenosis site, making it difficult to predict the effects of overall vascular geometry on hemodynamic parameters. The objective of this study is to comprehensively examine the influence of geometric morphology at different degrees of CAS and at bifurcation sites on hemodynamic parameters. A three-dimensional model is established using computed tomography angiography images, and eight geometric parameters of each patient are measured by MIMICS. Then, computational fluid dynamics is utilized to investigate 60 patients with varying degrees of stenosis (10%-95%). Time and grid tests are conducted to optimize settings, and results are validated through comparison with reference calculations. Subsequently, correlation analysis using SPSS is performed to examine the relationship between the eight geometric parameters and four hemodynamic parameters. In MATLAB, prediction models for the four hemodynamic parameters are developed using back propagation neural networks (BPNN) and multiple linear regression. The BPNN model significantly outperforms the multiple linear regression model, reducing mean absolute error, mean squared error, and root mean squared error by 91.7%, 93.9%, and 75.5%, respectively, and increasing from 19.0% to 88.0%. This greatly improves fitting accuracy and reduces errors. This study elucidates the correlation and patterns of geometric parameters of vascular stenosis and bifurcation in evaluating hemodynamic parameters of CAS. This study opens up new avenues for improving the diagnosis, treatment, and clinical management strategies of CAS.