Heart failure & cardiomyopathies
Identification of high-risk imaging features in hypertrophic cardiomyopathy using electrocardiography: A deep-learning approach
Authors' conclusion
In HCM, novel ECG-DL models reliably identified patients with high-risk imaging features while offering the potential to reduce CMR testing requirements in underresourced areas.
Study context
- Population
- Patients with hypertrophic cardiomyopathy (HCM)
- Participants
- 1,930 vs 233
- Comparison
- ECG-DL models vs cardiac magnetic resonance (CMR) for detecting high-risk imaging features
- Holdout testingn = 1,930
- External validationn = 233
- Outcome
- ECG-DL prediction of high-risk imaging features: systolic dysfunction, massive hypertrophy (≥30 mm), apical aneurysm, extensive late gadolinium enhancement
Methodological detailsdeep-learning model development, external validation
- Deep-learning model development
- External validation
Performance
As reported in the abstract
Identification of high-risk imaging features by ECG-DL models (systolic dysfunction, massive hypertrophy, apical aneurysm, extensive late gadolinium enhancement)
| Test or model | C-statistic, systolic dysfunction | C-statistic, massive hypertrophy | C-statistic, apical aneurysm | C-statistic, extensive late gadolinium enhancement |
|---|---|---|---|---|
| Holdout testing | 0.72 | 0.83 | 0.93 | 0.76 |
| External validation | 0.71 | 0.76 | 0.91 | 0.68 |
C-statistic, systolic dysfunction
Holdout testing0.72
External validation0.71
C-statistic, massive hypertrophy
Holdout testing0.83
External validation0.76
C-statistic, apical aneurysm
Holdout testing0.93
External validation0.91
C-statistic, extensive late gadolinium enhancement
Holdout testing0.76
External validation0.68
Further findings
Hypothetical screening strategy: echocardiography + ECG-DL-guided selective CMR
- Sensitivity for patients with high-risk features
- 97%
- Reduction in recommended CMRs
- 61%
- NPV for absence of high-risk features without ECG-DL recommendation for CMR
- 99.5%
Original publication
Identification of high-risk imaging features in hypertrophic cardiomyopathy using electrocardiography: A deep-learning approach. Heart Rhythm. Published 26 January 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →