The present invention provides a method for determining thromboembolic risk in a patient. The method includes processing functional images of a patient's heart in order to create a computational fluid dynamic (CFD) modeling of the patient's heart. Once a CFD model is obtained, various metrics can be determined to estimate the patient's risk of left ventricular thrombosis. This method is particularly suited for determining thromboembolic risk in patients having suffered a myocardial infarction. However, the method can also be applied to a broader population at risk of cardioembolic and cryptogenic stroke.
