Related Experiment Video
Updated: Aug 6, 2026
12:24
Noninvasive Assessment of Cardiac Abnormalities in Experimental Autoimmune Myocarditis by Magnetic Resonance Microscopy Imaging in the Mouse
Published on: June 20, 2014
Machine learning-based methods in diagnosing cardiac amyloidosis: a meta-analysis
Yuchen Song1, Qun Wang2, Lianqun Jia2,3
1College of Integrated Chinese and Western Medicine, Liaoning University of Traditional Chinese Medicine, Shenyang, Liaoning, China.
Frontiers in Cardiovascular Medicine
|July 23, 2026
Summary
Machine learning (ML) shows good accuracy in diagnosing cardiac amyloidosis (CA), a condition often misdiagnosed as hypertrophic cardiomyopathy. Further research is needed to validate these findings for developing AI-driven diagnostic tools.
Area of Science:
- Cardiology
- Medical Imaging
- Artificial Intelligence
Background:
- Cardiac amyloidosis (CA) is a restrictive cardiomyopathy caused by amyloid deposition, often mimicking hypertrophic cardiomyopathy.
- Nonspecific early symptoms lead to misdiagnosis and delayed treatment for this progressive disease.
- Machine learning (ML) shows promise for CA diagnosis, but evidence on its accuracy is limited.
Purpose of the Study:
- To systematically evaluate the diagnostic accuracy of ML algorithms for CA.
- To provide evidence for developing intelligent CA detection tools.
Main Methods:
- A meta-analysis of 30 studies was conducted, searching major databases up to September 2025.
- Study quality was assessed using QUADAS-2.
- Subgroup analyses included disease types (AL-CA, ATTR-CA) and imaging modality (echocardiography).
Main Results:
- ML for overall CA demonstrated high diagnostic accuracy (AUC 0.93).
- Specific performance metrics (sensitivity, specificity, PLR, NLR) were reported for overall CA, AL-CA, ATTR-CA, and echocardiography-based models.
- ML models showed consistent diagnostic performance across different CA subtypes and imaging modalities.
Conclusions:
- ML exhibits favorable diagnostic accuracy for CA.
- Methodological limitations necessitate cautious interpretation of current findings.
- Further validation with diverse datasets is crucial for advancing AI-based CA diagnostic tools.