A NEW artificial intelligence system, PanEcho, has demonstrated the ability to fully automate echocardiogram interpretation with high accuracy, potentially transforming cardiovascular diagnostics and expanding access to expert-level heart imaging.
Echocardiography is a fundamental tool in cardiovascular care, yet it traditionally depends on expert interpretation and manual reporting, which can be time-consuming and limit access in resource-constrained settings. The development of AI-driven solutions offers the promise of faster, standardised, and more widely accessible cardiac assessment.
In this retrospective, multisite study, researchers developed and validated PanEcho, a multitask deep learning system trained on 1.2 million echocardiographic videos from 32,265 transthoracic echocardiography (TTE) studies involving 24,405 patients across Yale New Haven Health System (YNHHS) hospitals and clinics. The AI was evaluated on its ability to perform 18 diagnostic classification tasks and 21 echocardiographic parameter estimations. In internal validation, PanEcho achieved a median area under the receiver operating characteristic curve (AUC) of 0.91 for diagnostic tasks and a median normalised mean absolute error of 0.13 for parameter estimation. The system accurately estimated left ventricular ejection fraction (mean absolute error: 4.2% internally and 4.5% externally) and reliably detected moderate or worse left ventricular systolic dysfunction (AUC: 0.98 internal, 0.99 external), right ventricular systolic dysfunction (AUC: 0.93 internal, 0.94 external), and severe aortic stenosis (AUC: 0.98 internal, 1.00 external). The AI maintained robust performance even with limited imaging protocols, achieving a median AUC of 0.91 in abbreviated TTE studies and 0.85 in real-world point-of-care ultrasonography from emergency departments.
These findings suggest that PanEcho could serve as an adjunct reader in echocardiography laboratories or as a rapid screening tool in point-of-care settings, particularly where access to trained cardiologists is limited. For clinical practice, the integration of AI-enabled echocardiogram interpretation could accelerate diagnostic workflows, reduce reporting delays, and support earlier detection of cardiac conditions. However, prospective evaluation in real-world clinical workflows remains essential to ensure safety, reliability, and appropriate integration alongside clinician expertise. Continued research and careful implementation will be crucial to harness the full potential of AI in cardiovascular imaging, ultimately improving patient outcomes and healthcare efficiency.
Reference
Holste G et al. Complete AI-enabled echocardiography interpretation with multitask deep learning. JAMA. 2025;DOI:10.1001/jama.2025.8731.
