Personalized 3D Myocardial Infarct Geometry Reconstruction from Cine MRI for Cardiac Digital Twins
Researchers have developed a new method to reconstruct 3D models of myocardial infarction (MI) from cine MRI scans, bypassing the need for contrast agents. This approach utilizes a novel explicit geometry-motion embedded model that extracts and fuses geometry and motion features from multi-view cine MRIs. The system achieved an average Dice score of 0.678 on 225 scans and demonstrated high consistency with contrast-enhanced MRI-derived data in downstream electrophysiological simulations, showing promise for contrast-free cardiac digital twin modeling. AI
IMPACT Enables more accessible and frequent cardiac scar characterization for personalized digital twin simulations.