To evaluate the effectiveness of artificial intelligence (AI) guidance in cardiac point-of-care ultrasound performed by advanced practice providers.
Approach:
Study Design: A prospective observational study involving 100 examinations performed by advanced practice providers, with 64 using AI guidance and 36 without.
Evaluation Method: Advanced practice provider interpretations were compared with blinded cardiologist overreads to assess agreement and clinical management decisions.
Key Findings:
Cardiologists agreed with advanced practice provider interpretations on major findings in over 96% of examinations.
AI-guided examinations had fewer corrections by cardiologists (27%) compared to non-AI-guided examinations (47%).
AI guidance was associated with improved acquisition completeness and fewer missed or underestimated left ventricular and valvular abnormalities.
Scan duration was shorter with AI guidance (13.7 minutes) compared to without (16.9 minutes).
Point-of-care ultrasound findings led to management changes in 57% of patients.
Interpretation:
AI-guided cardiac point-of-care ultrasound performed by advanced practice providers demonstrated high diagnostic agreement with cardiologist reviews and supported clinical management decisions.
Limitations:
The study did not provide disclosure information.
Conclusion:
AI-guided examinations supported clinical management and discharge decisions without the need for inpatient transthoracic echocardiography in over half of patients.
