Prediction of Adverse Events From Non-PCI Coronary Plaques Using Machine Learning

The aim of our project is to predict, with the help of artificial intelligence (AI) and machine learning techniques, the natural history of non-flow-limiting coronary stenoses and to develop and validate a machine learning risk score able to estimate the risk of MACE during follow-up based on OCT findings observed on coronary plaques not subjected to percutaneous coronary intervention (PCI) and on the clinical characteristics of the patients. We will collect information on all non-PCI coronary lesions assessed by performing OCT in non-culprit vessels of patients presenting with ACS or Chronic Coronary Syndrome (CCS) at the index procedure. All OCT runs will be digitized and analyzed with the help of AI. The OCT appearance of coronary plaques will be analyzed with AI and correlated with the risk of MACE during follow-up (a composite of cardiac death, myocardial infarction, or target lesion revascularization). An AI predictive risk model based on coronary plaque OCT images and patient clinical characteristics will be developed to estimate the risk of MACE incidence at follow-up. Cutting-edge machine learning algorithms, including convolutional neural networks, random forests, and support vector machines, will be used.