Heart failure & cardiomyopathiesDiagnostic or prognostic study
Digital profile of children’s hearts: automated echocardiogram strain analysis facilitates earlier detection of cardiac dysfunction
- Diagnostic accuracy study
- semi-supervised deep learning
Authors' conclusion
Leveraging a large-scale paediatric dataset, Motion-Echo provided a reliable and generalizable framework for myocardial strain analysis, demonstrating potential to facilitate earlier detection of cardiac dysfunction and generate digital cardiac function profiles of children.
Study context
- Population
- Paediatric and adult echocardiograms spanning diverse image qualities and vendors11,096 paediatric + 11,297 adult echocardiograms
Performance
As reported in the abstract
Automated global strain assessment
| Test or model | GLS mean absolute error | Global circumferential strain mean absolute error | GLS Pearson correlation | Global circumferential strain Pearson correlation |
|---|---|---|---|---|
| Motion-Echo | 2.099% (95% CI 1.803-2.401) | 2.665% (95% CI 2.339-3.026) | 0.799 (95% CI 0.715-0.871) | 0.781 (95% CI 0.687-0.844) |
Further findings
Cancer therapy-related cardiac dysfunction risk prediction (automated strain values)
- AUC
- 0.906 (95% CI 0.816-0.981)
Late gadolinium enhancement detection (automated GLS)
- AUC
- 0.782 (95% CI 0.666-0.885)
LVEF decline forecasting
- System outperformed manual strain values (DeLong)
- P < 0.001
MI detection (incorporating estimated motion flows)
- AUC improvement to
- 0.952 (95% CI 0.917-0.980)
Original publication
Digital profile of children’s hearts: automated echocardiogram strain analysis facilitates earlier detection of cardiac dysfunction. Eur Heart J. Published 5 December 2025. doi:10.1093/eurheartj/ehaf952
Summarised from the published abstract; every number is checked against it