Deep Learning and Machine Learning for Differentiation Between Contrast Extravasation and Hemorrhagic Transformation in Post-Thrombectomy Stroke CT

Key Points

• This research aims to evaluate the effectiveness of machine learning models in distinguishing between contrast extravasation and hemorrhagic transformation in post-thrombectomy stroke imaging.
• Applied machine learning models to non-contrast CT (NCCT) imaging data
• Compared U-Net deep learning with traditional machine learning models
• Recommended exploring larger studies and hybrid approaches combining radiomics and deep learning
• Machine learning models accurately differentiate contrast extravasation from hemorrhagic transformation
• U-Net demonstrated advantages in learning from raw imaging data
• Traditional models achieved comparable performance to deeper learning approaches