About the Podcast
EM:RAP (Emergency Medicine: Reviews and Perspectives) is one of the most widely followed podcasts in emergency medicine, featuring expert hosts and guest clinicians who break down the latest research and clinical tools shaping practice.
Summary of the Episode
In this episode, Dr. Anand Swaminathan and Dr. Tarlan Hedayati discuss the role of artificial intelligence in ECG interpretation, focusing on the PMcardio Queen of Hearts™ AI tool developed by Dr. Stephen Smith and Dr. Pendell Meyers.
- Queen of Hearts™ is a deep learning model designed to detect occlusive myocardial infarction (OMI) from standard 12-lead ECGs.
- The model achieved 91% accuracy (81% sensitivity, 94% specificity), performing similarly to ECG experts — and far outperforming traditional STEMI criteria (33% sensitivity).
- Unlike conventional ECG computer interpretations (historically 30–70% sensitivity for STEMI), the AI consistently identifies subtle occlusions missed by both machines and clinicians.
- The hosts emphasize that AI cannot replace the clinician at the bedside — it lacks access to serial ECGs, prior patient data, troponins, or echocardiography — but it can act as a valuable adjunct, improving diagnostic accuracy and reducing errors when used appropriately.
- Key takeaway: clinicians must understand the limitations, patient selection, and workflow integration of AI tools to maximize benefit and avoid over-reliance.
“If I could have a Steve Smith in my pocket to read every ECG, that’s what Queen of Hearts feels like. It doesn’t replace us, but it makes us better.”
— Dr. Anand Swaminathan, EM:RAP Podcast
🎧 Listen to the episode: Official page
