Artificial Intelligence and Computational Methods in Cardiovascular Ultrasound.
Authors
Affiliations (5)
Affiliations (5)
- Department of Medicine and Health Sciences "Vincenzo Tiberio", University of Molise, via Francesco De Sanctis 1, Campobasso 86100, Italy; Department of Cardiovascular Sciences, Responsible Research Hospital, Campobasso, Italy. Electronic address: [email protected].
- Department of Medicine, Surgery and Dentistry, University Study of Salerno, Baronissi, Salerno, Italy.
- Department of Cardiovascular Sciences, Responsible Research Hospital, Campobasso, Italy.
- Heart Department, University Hospital "San Giovanni di Dio e Ruggi d'Aragona" Salerno, Italy.
- Department of Medicine and Health Sciences "Vincenzo Tiberio", University of Molise, via Francesco De Sanctis 1, Campobasso 86100, Italy.
Abstract
Cardiovascular ultrasound is particularly suitable for artificial intelligence: it is widely available, information-rich, dynamic, and strongly dependent on acquisition quality and operator interpretation. Contemporary computational methods can guide image acquisition, recognize views, segment chambers, quantify ventricular function, characterize myocardial mechanics, and detect disease signatures. In heart failure, multimodal architectures can combine echocardiographic images and Doppler data with clinical variables, biomarkers, electrocardiograms, and omics to support patient-level phenotyping and prognostic assessment, while emerging foundation models enable adaptation across multiple tasks. The most credible near-term model is augmented echocardiography: automation of repetitive measurements combined with transparent quality control and expert clinical interpretation.