Radiomic Analysis of CMR for Prediction of Left Ventricular Mass Index Changes in Subsequent Year for Fabry Disease

Motivation: Cardiovascular diseases are the primary cause of mortality in Fabry disease (FD) patients, even with enzyme replacement therapy (ERT). Early detection of cardiac involvement can improve outcomes. Goal(s): This study aims to predict potential increases in left ventricular mass index (LVMi) in FD patients using radiomics analysis of CMR cine images in the subsequent follow-up year. Approach: Radiomic features were extracted from 2D and 3D CMR images, and a linear SVM model was developed to differentiate increased LVMi patients from stable LVMi patients. Results: The 2D model with Chi2 achieved the best performance (AUC=0.79), highlighting the potential of radiomics in predicting FD prognosis. Impact: Radiomic analysis of CMR cine images demonstrated good performance in distinguishing increased LVMi from stable LVMi patients, providing a novel approach for predicting the changes of LVMi in patients with FD in the subsequent follow-up year.