Table 1.
Description of an array of studies that have leveraged machine learning applications in echocardiography, computed tomography angiography, nuclear imaging, and cardiac magnetic resonance imaging.20–22,28,29,31–34,37,38,41,43 CTA: computed tomography angiography; CAD: coronary artery disease; TPD: total perfusion deficit; MACE: major adverse cardiac events
| STUDY | MACHINE LEARNING ALGORITHM | TYPE OF IMAGING | BRIEF STUDY DESCRIPTION |
|---|---|---|---|
| Samad et al. 20 | Supervised learning | Echocardiography | Utilizing clinical and echocardiographic variables to predict complications |
| Khamis et al. 21 | Supervised learning | Echocardiography | To automatically detect correct views on echocardiography |
| Knackstedt et al. 22 | Machine learning | Echocardiography | To automatically calculate ejection fraction and longitudinal strain |
| Casaclang-Verzosa et al. 28 | Unsupervised learning | Echocardiography | To detect unique phenotypes of aortic stenosis |
| Todoki et al. 29 | Unsupervised learning | Echocardiography | Utilizing echocardiographic variables to predict MACE complications |
| Motwani et al. 31 | Multiple machine learning algorithms | Computed Tomography | To estimate 5 year mortality in patients with CAD |
| Santini et al. 32 | Deep learning algorithm | Computed Tomography | To classify and segment lesions on CTA |
| Baskaran et al. 34 | Machine learning algorithm | Computed Tomography | To perform automatic segmentation of structures on CTA |
| Baskaran et al. 33 | Deep learning algorithm | Computed Tomography | To identify CTA cardiovascular structures |
| Arasajani et al. | Supervised learning algorithm | Nuclear Cardiology | Integrated echocardiographic and clinical characteristics to predict CAD. |
| Betancur et al. 38 | Deep learning | Nuclear Cardiology | To predict CAD and compare to TPD |
| Betancur et al. 37 | Supervised learning | Nuclear Cardiology | To estimate the occurrence of MACE events |
| Winter et al. 41 | Deep learning | Magnetic Resonance Imaging | To automatically calculate right and left ventricular mass and volumes |
| Leng et al. 43 | Deep learning | Magnetic Resonance Imaging | To identify ventricular contours and compare with manual tracing |