Key result
AI-assisted pediatric echocardiography shows >90% feasibility and strong manual agreement, but falters on wall thickness.
Why the study?
The accuracy of a novel AI-assisted echocardiographic platform for automated measurements in children required evaluation.
Does an AI-assisted echocardiographic platform provide accurate and feasible automated measurements compared to manual evaluation by a pediatric cardiologist in children?
Observational (n=30)
Single-blind
No
Does an AI-assisted echocardiographic platform provide accurate and feasible automated measurements compared to manual evaluation by a pediatric cardiologist in children?
Answers are grounded in this paper and cite their sources.
A commercially available AI-assisted echocardiographic platform demonstrates high feasibility and strong agreement with manual measurements for most 2D and 4D ventricular parameters in pediatric patients, though accuracy for wall thickness and feasibility for 4D atrial volumetry remain limited.
PURPOSE: This study assessed the accuracy of a novel artificial intelligence (AI)-assisted echocardiographic platform for automated measurements in children. METHODS: In this prospective, blinded, comparative, single-center pilot study, transthoracic echocardiography was performed in pediatric subjects aged 5 to 18 years using the Siemens Acuson Origin cardiovascular ultrasound system. For spectral Doppler assessments, two echocardiographic acquisitions were required: one using the AI-assisted protocol with automatic spectral Doppler marker placement, and one with manual marker placement. For all other analyses, a single manual acquisition was required for subsequent AI and manual assessments. Images were initially analyzed with the AI-assisted protocol, followed immediately by manual evaluation by a pediatric cardiologist blinded to the AI-derived results. Indices included B-mode, M-mode, Doppler, speckle tracking, and four-dimensional (4D) parameters. Agreement between AI-assisted and manual measurements was assessed using intraclass correlation coefficient (ICC) and Bland-Altman analysis. RESULTS: Thirty subjects (23 males; age 12.2±3.8 years) were studied, including 23 with structurally normal heart, six with congenital heart diseases, and one with dilated cardiomyopathy. Feasibility exceeded 90% for 2D AI-assisted assessment and 80% for 4D ventricular volumetry, but was below 35% for 4D atrial volumetry. Agreement between AI-assisted and manual measurements was strong for all parameters except septal (ICC 0.2) and left ventricular posterior wall thickness (ICC 0.5). Bland-Altman analysis showed no significant systematic differences. CONCLUSION: Our pilot findings suggest that a commercially available AI-assisted echocardiographic system may be feasible and reasonably accurate for conventional and advanced quantitative echocardiographic assessment for children in a real-world setting.
No takes yet. Share an insight, caveat, or question.
Chan et al. (2026) conducted an observational in Pediatric cardiac assessment (23 structurally normal heart, 6 congenital heart diseases, 1 dilated cardiomyopathy) (n=30). AI-assisted echocardiographic platform (Siemens Acuson Origin) vs. Manual echocardiographic evaluation by a pediatric cardiologist was evaluated on Agreement and feasibility between AI-assisted and manual echocardiographic measurements. AI-assisted pediatric echocardiography achieved >90% 2D feasibility and strong agreement with manual measurements, except for septal and LV posterior wall thickness.
Synapse has enriched 10 related papers. Consider them for comparative context: