Imaging-based prediction of histological clot composition from admission CT imaging.

Machine learning based analysis of admission imaging allows for prediction of clot composition. Perspectively, such an approach could allow selection of clot-specific devices and retrieval procedures for personalized thrombectomy strategies.