Heart failure & cardiomyopathies
Machine Learning to Automatically Differentiate Hypertrophic Cardiomyopathy, Cardiac Light Chain, and Cardiac Transthyretin Amyloidosis: A Multicenter Cardiovascular Magnetic Resonance (CMR) Study
Authors' conclusion
A trained ML model using semiautomated CMR imaging data and patient demographics can accurately identify cardiac amyloidosis and differentiate subtypes.
Study context
- Population
- Patients diagnosed with hypertrophic cardiomyopathy or AL/ATTR amyloidosis, and healthy volunteersfrom 56 institutions, 269 men aged 58.5 [48.4–69.4] years · healthy 95 · hypertrophic cardiomyopathy 94 · AL 95 · ATTR 116
- Participants
- 400 patients
- Index test or model
- A 3-stage machine learning (ML) screening cascade sequentially differentiated healthy volunteers from patients, then hypertrophic cardiomyopathy from amyloidosis, and then AL from ATTR
- Target condition
- cardiac amyloidosis and its subtypes (AL, ATTR), distinguished from hypertrophic cardiomyopathy
Methodological detailsconsecutive patients
- Consecutive patients
Performance
As reported in the abstract
Area under the curve, healthy volunteers vs patients
1.0
Area under the curve, hypertrophic cardiomyopathy vs amyloidosis
0.99
Area under the curve, AL vs ATTR
0.92
Further findings
Differentiation with data reduced to demographics and imaging data alone
- Area under the curve, healthy volunteers vs patients
- 0.99
- Area under the curve, hypertrophic cardiomyopathy vs amyloidosis
- 0.98
- Area under the curve, AL vs ATTR
- 0.88
Differentiation after removing late gadolinium enhancement imaging data from the model
- Area under the curve, healthy volunteers vs patients
- 1.0
- Area under the curve, hypertrophic cardiomyopathy vs amyloidosis
- 0.95
- Area under the curve, AL vs ATTR
- 0.86
Original publication
Machine Learning to Automatically Differentiate Hypertrophic Cardiomyopathy, Cardiac Light Chain, and Cardiac Transthyretin Amyloidosis: A Multicenter CMR Study. Circ Cardiovasc Imaging. Published 4 June 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →