SIGNATURE WORK
CONFERENCE & EXHIBITION 2024

This project aims to leverage machine learning methodologies to develop predictive models for estimating the risk of hemorrhagic transformation (HT) using only readily available patient information typically accessible from medical imaging. By analyzing patient status and MRI images, the project seeks to furnish risk estimations for various interventions, thereby assisting clinicians in making informed decisions to mitigate the risk of HT. Through statistical analysis and classification performed by decision tree, support vector machine (SVM), and logistic regression, atrial fibrillation, NIHSS score, and Magnetic Resonance Imaging (MRI) parameters are positively related to HT incidence. At the same time, the time from symptom onset to the first MRI is negatively associated with the occurrence of HT.