Reference: JAMA Cardiol. 2025 Jan 15 early online
Practice Point: AI-guided software may allow even non-experts to capture quality lung ultrasound images and diagnose pulmonary edema without training in point-of-care ultrasonography, which could be invaluable in low-resource settings.
EBM Pearl: The term “ground truthing” refers to using a practical reference standard in place of the gold standard for diagnosis. In this situation, a full right-sided coronary catheterization wouldn’t be practical, so a panel of expert POCUS-trained clinicians served as a surrogate reference standard.
Imagine a world where even the most novice healthcare provider can wield ultrasound probes with the finesse of a Jedi knight using a lightsaber. With the emergence of advancing artificial intelligence (AI) technology, this vision is a little closer to becoming reality. A recent study published in JAMA Cardiology assessed the ability of healthcare professionals not trained in ultrasonography to obtain high-quality diagnostic images on lung ultrasound with the assistance of an AI Lung Guidance software algorithm designed to detect pulmonary edema in patients with shortness of breath. Using an 8-zone lung ultrasound protocol, the AI program ensures captured images are of diagnostic quality and determines if B-lines (which are very sensitive for acute pulmonary edema) are present.
In this diagnostic validation study, a total of 163 adults presenting with shortness of breath were enrolled and had 2 ultrasound examinations: one performed by a non-expert healthcare professional aided by Lung Guidance AI and one performed by an ultrasound fellowship–trained physician or advanced cardiac sonographer without the assistance of AI. The non-expert healthcare professionals were mostly registered nurses and medical assistants, but a group of 4 physicians who had completed an emergency medicine residency with prior lung ultrasound training were also included as a subset of the non-expert healthcare provider group. The authors reasoned that including these physicians in a planned, separate analysis would allow an assessment for inferior performance by ultrasound-trained physicians using the AI software, important for real-world use.
Diagnostic image quality was assessed by a panel of 5 expert physicians with training in point-of-care ultrasound who determined whether each image clip provided enough information to classify it as insufficient, normal, or abnormal. Images were randomly presented and masked for both operator experience and use of AI technology. This panel served as a source of “ground truth” in place of an objectively-measured reference standard which would have been invasive and costly.
The study found that over 98% of the ultrasound images obtained by non-expert healthcare professionals with AI guidance (including the ultrasound-trained ED physicians) were of sufficient quality to meet diagnostic standards and were not statistically different compared with images acquired by ultrasound experts without AI guidance. In fact, in lung ultrasound zone 6, nonexpert ultrasound operators using AI guidance significantly outperformed expert ultrasound operators not using AI. The results were found to be consistent across diverse patient demographics and BMIs as well as multiple healthcare settings and operators of variable ultrasound experience from novice to expert.
Beyond enhancing diagnostic capabilities, this AI innovation could lead to a paradigm shift in patient care. Fast, accurate diagnoses mean timely treatments and better patient outcomes. POCUS is cheaper and faster and much more portable than most standard radiology studies. POCUS is already proving invaluable in low resource and remote rural settings but sometimes training and the volume needed to become expert are barriers in these same settings. With AI-assisted lung ultrasounds, the sole reliance on expert ultrasound operators could diminish with the potential to enhance diagnostic reach in areas with limited access to specialized healthcare professionals. In essence, AI-guided lung ultrasound is not just a technological advancement, it's a catalyst for more efficient and equitable healthcare delivery.
For more information, see the topic Ultrasound in the Intensive Care Unit (ICU) in DynaMed.
DynaMed EBM Focus Editorial Team
This EBM Focus was written by Rich Lamkin, MPH, MPAS, PA-C, Clinical Writer at DynaMed. Edited by Alan Ehrlich, MD, FAAFP, Executive Editor at DynaMed and Associate Professor in Family Medicine at the University of Massachusetts Medical School; Katharine DeGeorge, MD, MS, Senior Deputy Editor at DynaMed and Associate Professor of Family Medicine at the University of Virginia; Dan Randall, MD, MPH, FACP, Senior Deputy Editor at DynaMed; McKenzie Ferguson, PharmD, BCPS, Senior Clinical Writer at DynaMed; Matthew Lavoie, Senior Medical Copyeditor, BA, at DynaMed; Hannah Ekeh, MA, Senior Associate Editor II at DynaMed; and Jennifer Wallace, BA, Senior Associate Editor at DynaMed.