Prediction of cardiac remodeling and myocardial fibrosis in athletes based on IVIM-DWI images

Myocardial microcirculation in athletes and its relationship with cardiac remodeling (CR) and myocardial fibrosis (MF) are not fully understood. We prospectively enrolled 174 athletes and 54 healthy sedentary controls for intravoxel incoherent motion (IVIM) diffusion-weighted imaging of cardiac magnetic resonance imaging. Athletes exhibited significantly lower fast apparent diffusion coefficient (ADC_fast) and perfusion fraction (f) in 16 myocardial segments and each blood supply area compared to controls (<i>p</i> < 0.05). Athletes with CR and/or MF had lower myocardial slow apparent diffusion coefficient (ADC_slow) values than those without (<i>p</i> < 0.05). A gradient boosting machine (GBM) effectively predicted CR and/or MF based on these hypoperfusion parameters, with an area under the receiver operating characteristic curve of 0.947 in the training set and 0.841 in the test set. The GBM model, leveraging IVIM parameters, could predict the occurrence of CR and/or MF, offering a potential tool for monitoring and managing the athletes' health.