Mining data from hemodynamic simulations for generating prediction and explanation models

Arterial geometry variability is present both within and across individuals. To analyze the influence of geometric parameters on maximal wall shear stress (MWSS) in the human carotid artery bifurcation, the computer simulations were run to generate the data pertaining to this phenomenon. In our work we evaluate various prediction models for modeling relationship between geometric parameters of the carotid bifurcation and the MWSS. The results revealed the highest potential of using the neural network model for this prediction task. The achieved results and generated explanations of the prediction model represent progress in assessment of stroke risk for a given patient's geometry in real time.