Using Artificial Intelligence for Rheumatic Heart Disease Detection by Echocardiography: Focus on Mitral
Kelsey Brown1, Pooneh Roshanitabrizi2, Joselyn Rwebembera3
1Department of Pediatric Cardiology Children's National Hospital Washington DC USA.
Insights
Artificial intelligence can accurately detect rheumatic heart disease (RHD) in children using echocardiograms. This technology shows promise for widespread RHD screening and early intervention.
Area of Science:
- Cardiology
- Artificial Intelligence
- Medical Imaging
Background:
- Early identification of latent rheumatic heart disease (RHD) via echocardiography enables secondary prophylaxis.
- Limited studies explore AI/machine learning for RHD detection on portable echocardiograms.
Purpose of the Study:
- To assess the potential of AI and deep learning models to detect RHD in children using echocardiograms.
- To evaluate the accuracy of automated mitral regurgitation analysis for RHD diagnosis.
Main Methods:
- Utilized 511 pediatric echocardiograms, focusing on mitral valve color Doppler images.
- Developed automated methods using convolutional neural networks and deep learning with attention mechanisms for RHD detection.
- Compared AI-derived measurements with expert manual assessments.
Main Results:
- AI achieved high accuracy in identifying views (0.99) and systolic frames (0.93-0.94).
- Mitral regurgitation jet measurements by AI closely matched expert manual measurements (P=0.83).
- A 9-feature analysis demonstrated strong performance (AUC 0.93), and the deep learning model showed high recall (0.98).
Conclusions:
- AI demonstrates potential to detect RHD with accuracy comparable to expert cardiologists.
- AI-driven echocardiography holds promise for scalable RHD screening and early intervention.
Background:
Identification of children with latent rheumatic heart disease (RHD) by echocardiography, before onset of symptoms, provides an opportunity to initiate secondary prophylaxis and prevent disease progression. There have been limited artificial intelligence studies published assessing the potential of machine learning to detect and analyze mitral regurgitation or to detect the presence of RHD on standard portable echocardiograms.
Methods And Results:
We used 511 echocardiograms in children, focusing on color Doppler images of the mitral valve. Echocardiograms were independently reviewed by an expert adjudication panel. Among 511 cases, 229 were normal, and 282 had RHD. Our automated method included harmonization of echocardiograms to localize the left atrium during systole using convolutional neural networks and RHD detection using mitral regurgitation jet analysis and deep learning models with an attention mechanism. We identified the correct view with an average accuracy of 0.99 and the correct systolic frame with an average accuracy of 0.94 (apical) and 0.93 (parasternal long axis). It localized the left atrium with an average Dice coefficient of 0.88 (apical) and 0.9 (parasternal long axis). Maximum mitral regurgitation jet measurements were similar to expert manual measurements (P value=0.83) and a 9-feature mitral regurgitation analysis showed an area under the receiver operating characteristics curve of 0.93, precision of 0.83, recall of 0.92, and F1 score of 0.87. Our deep learning model showed an area under the receiver operating characteristics curve of 0.84, precision of 0.78, recall of 0.98, and F1 score of 0.87.
Conclusions:
Artificial intelligence has the potential to detect RHD as accurately as expert cardiologists and to improve with more data. These innovative approaches hold promise to scale echocardiography screening for RHD.
