AI-Echo RCT
2026
Artificial Intelligence-Based Automated Echocardiographic Analysis and the Workflow of Sonographers: a Randomized Crossover Trial (AI-Echo RCT)
In 585 patients undergoing echocardiography, AI-based automatic measurement of echocardiographic parameters improved examination efficiency (time per patient and examinations per day) compared with manual measurement workflow.
Result
Primary endpoint: Examination efficiency: examination time per patient and number of examinations per day
Examination time 13.0±3.5 vs 14.3±4.2 min (P<0.001); examinations per day 16.7±2.5 vs 14.1±2.5 (P=0.003); parameters per examination 85±12 vs 25±1; mental fatigue score 4.1±1.1 vs 4.7±0.6 (P=0.039); AI values within clinical limits for 90% of parameters.
- Subtopic
- Echocardiography
- Design
- Single-centre randomised crossover trial (4 sonographers, AI vs non-AI days)
- Relevance
- Supporting: context rather than must-know.
- Source
- Sakamoto A, et al. J Am Heart Assoc. 2026;15(1):e045637.doi 10.1161/JAHA.125.045637
Discussion & critique
A rare randomised test of AI on the sonographer's working day rather than on diagnostic accuracy: faster examinations, more of them, far more measurements and less fatigue, with expert echocardiologists still finalising every report. Single centre with only four sonographers, a day-level crossover that cannot be blinded, and workflow rather than patient or diagnostic endpoints; whether the 3.4-fold increase in reported parameters adds clinical value is untested.