Tetralogy of Fallot is a congenital heart disease that typically requires childhood surgical repair. Mortality risk increases substantially in the years following surgery due to deterioration in ventricular function; however, previous studies using regression models have failed to predict this deterioration effectively. This study presented by Samad et al. attempts to use machine learning to predict which patients with repaired Tetralogy of Fallot (rTOF) are likely to experience deterioration in ventricular size and function. Researchers selected 153 patients with rTOF who had undergone two cardiac MRI scans taken at least six months apart to assess changes in ventricular function. Utilizing clinical, electrocardiogram, and imaging data from the first scan, they trained a support vector machine (SVM) to predict future deterioration in ventricular size and function and categorize patients into three groups: no deterioration, minor deterioration, or major deterioration. The success of the machine learning model was measured with a mean area under the curve (AUC) incorporating various combinations of 24 variables in a predictive scale from 0.5 to 1.0 (0.5 representing random chance and 0.9–1.0 being excellent). This was done in four experimental scenarios based on differentiating the three deterioration categories. It was found that the machine learning models were able to predict who would experience deterioration with relative accuracy. The SVM model achieved a high AUC of 0.87 when distinguishing major deterioration from no deterioration and a high AUC of 0.82 when distinguishing any deterioration from no deterioration. The most important predictors of deterioration included left ventricular (LV) ejection fraction, pulmonary regurgitation fraction, LV circumferential strain, and the patient’s age at the time of surgical repair. The authors highlight that machine learning can help identify important predictors of heart deterioration, allowing for more personalized approaches to monitoring patients with rTOF.