Determination of the Risk of Rupture of Intracranial Aneurysms Through Numerical Simulation and Data Classification

Determining the risk of rupture of intracranial aneurysms is a great challenge. In recent years, hemodynamics parameters, such as shear stress and oscillatory shear index, have gained attention to predict aneurysm rupture. Numerical flow simulations in complex biological structures have also gained considerable attention. However, the primary challenge is the validation of the numerical procedures due to the complexity of measuring (velocity and shear stress) and testing the computed results. The main goal of the present work is to classify geometrical and hemodynamic parameters to predict the possible rupture of the aneurysm.

Numerical simulations were performed using realistic geometries. Aneurysms 3D models were reconstructed from CT scans, using the 3D-Slicer software to perform the numerical simulations and to 3D print the models for validation purposes. The blood was modelled as a Newtonian fluid with constant properties. As a first approximation, the arterial walls were rigid, and no-slip boundary condition was taken into account.

The 1R machine learning algorithm was used for classifying the geometrical and hemodynamic parameters. The analysis used the following hemodynamic parameters: wall pressure and wall shear stress, oscillatory shear index (OSI), residence time, gradient oscillatory number (GON), vorticity, q criterion, stokes number for particles and enstrophy to classify the aneurysm in combination with the traditional parameters.

The performed numerical simulations showed good agreement with the experiments (qualitatively). The first classification presents a reliability of 82.8%. This is the first achievement of the project.