Artificial Intelligence Empowers Novice Users to Acquire Diagnostic-Quality Echocardiography

Area of Science:

• Cardiology
• Medical Imaging
• Artificial Intelligence

Background:

• Cardiac ultrasound (echocardiography) is vital for clinical decisions but requires expert sonographers.
• AI-driven deep learning tools can potentially train novices for diagnostic echocardiography.
• This could expand access to and improve the quality of cardiac care.

Purpose of the Study:

• To assess if nurses with no prior ultrasound experience could acquire diagnostic-quality echocardiographic views using AI software.
• To determine the efficacy of AI guidance in novice-performed cardiac ultrasounds.

Main Methods:

• A prospective, international, nonrandomized study involving 240 adult patients across two medical centers.
• Novices used AI software for one limited cardiac exam; experts performed another without AI guidance.
• Cardiologists evaluated primary endpoints (left/right ventricle size, pericardial effusion) for diagnostic quality.

Main Results:

• 100% of exams by novices using AI software met diagnostic quality standards for primary endpoints.
• Cardiac parameters from novice and expert exams showed strong correlation.
• AI software facilitated safe and effective echocardiogram acquisition by novices.

Conclusions:

• AI-based software offers a safe method for novices to achieve diagnostic-quality cardiac ultrasounds.
• Short training with AI tools empowers non-experts to perform essential echocardiographic assessments.
• AI holds promise for democratizing cardiac imaging expertise.