Systems and methods are disclosed for predicting the progression of disease from factors like vessel geometry, physiology, and hemodynamics, said progression in disease comprising the onset or change over time of lesions in a blood vessel of a patient. One method includes: training a machine learning system based on longitudinal data of a plurality of individuals, that is corresponding data taken from the same individuals at different points in time, of vessel geometry, physiology, and hemodynamics, the data used to train the machine learning algorithm comprising multiple time-variant scans of the same individual taken at different times, to learn vessel characteristics in a location at an earlier time point that are correlated with the progression of disease in the same location at a later time point; acquiring an image of a patient; and for imaged locations in the patient, using the machine learning system's training data of local disease progression to predict the change in disease at said locations.
展开▼
