Key Takeaways:
- AI-guided echocardiography enables novice users to obtain diagnostic-quality cardiac images comparable to expert sonographers, achieving perfect diagnostic-quality rates (100%) for primary cardiac parameters.
- AI-assisted echocardiography shows rapid adoption by novice users, potentially broadening access to critical diagnostic imaging, especially in underserved or resource-constrained healthcare environments.
A new clinical trial demonstrates that artificial intelligence (AI)-guided echocardiography allows novice users, such as nurses without prior ultrasound experience, to reliably acquire diagnostic-quality cardiac ultrasound images. The HeartFocus study, a prospective international trial (NCT05836259), evaluated AI-based software designed to guide novice users in two large teaching hospitals in the US and France in acquiring echocardiographic views essential for assessing cardiac function.
In the study, 240 adult patients (mean age 62.6 ± 15.1 years, 49% women) underwent two echocardiograms: one performed by a novice using AI guidance, and another performed by an expert without AI assistance. The AI software provided real-time visual guidance to novices, automatically recording diagnostic-quality images or enabling manual selection of the best-quality images captured. Remarkably, all novice-acquired echocardiograms (100%) achieved diagnostic quality for the primary echocardiographic parameters—left ventricular size and function, right ventricular size, and the presence of significant pericardial effusion—with equivalent quality compared to expert-acquired images. Furthermore, agreement among expert cardiologists evaluating these parameters exceeded 85% (range 87.1% – 99.6%), underscoring the reliability and consistency of AI-assisted exams.
The study also noted rapid proficiency, with novice users averaging only 23.6 minutes per exam and producing diagnostic-quality clips even from their earliest examinations. The AI system showed strong performance across various patient subgroups, including obese individuals and those with known cardiac disease.
The authors emphasized the potential clinical impact: “AI-based echocardiography guidance software is safe and reliable for clinical practice, significantly shortening the learning curve and expanding access to essential diagnostic imaging, particularly in resource-limited settings.”