Artificial Intelligence in Cardiac Point-of-Care Ultrasound: A Narrative Review

Area of Science:

• Cardiology
• Medical Imaging
• Artificial Intelligence

Background:

• Cardiac point-of-care ultrasound (POCUS) is crucial in acute care but limited by operator variability.
• Artificial intelligence (AI), machine learning, and deep learning offer solutions to improve POCUS consistency.
• AI can aid in image acquisition, quantitative analysis, and interpretation of cardiac ultrasound.

Purpose of the Study:

• To review current applications of AI-assisted cardiac POCUS.
• To evaluate the clinical relevance and performance of AI tools in cardiac ultrasound.
• To identify areas of AI application and their validation status.

Main Methods:

• A narrative review of AI-assisted cardiac POCUS was conducted.
• Literature search from 2018-2026 in PubMed and Google Scholar.
• Included studies focused on clinical, educational, and image acquisition settings.

Main Results:

• Automated left ventricular ejection fraction (LVEF) assessment is the most developed AI application.
• AI tools show agreement with expert interpretation and improve novice user performance for LVEF.
• AI is also explored for pericardial effusion detection, image acquisition guidance, and education.
• Complex applications like diastolic function and hemodynamic measurements are less validated.
• AI performance is highly dependent on image quality and often tested in controlled settings.

Conclusions:

• AI-assisted cardiac POCUS shows strong potential as a decision-support tool, particularly for LVEF assessment.
• Current evidence supports AI for automated LVEF calculation.
• Other AI applications in cardiac POCUS are still under investigation.