Fig. 1

Recruitment Diagram. Patients (n = 208) were retrospectively identified over a 6 month timeline who were admitted for AMI. Of these patients, 155 patients were included in the study who had at least two of three apical echocardiographic views available for analysis. Using ultrasomics features from the images, topological data analysis was used to cluster patients into three groups. These three groups were assessed in a supervised machine learning algorithm to develop class labels for the external validation group. Ultimately, groups clustered using ultrasomics features were assessed for prediction of all-cause mortality and left ventricular infarct size