Related Experiment Video
Updated: Jul 4, 2025
Noninvasive Electrocardiography in the Perinatal Mouse
Published on: June 12, 2020
Congenital heart disease detection by pediatric electrocardiogram based deep learning integrated with human concepts
Jintai Chen1, Shuai Huang2,3,4, Ying Zhang3,5
1State Key Laboratory of Transvascular Implantation Devices of the Second Affiliated Hospital, Zhejiang University School of Medicine, Zhejiang University, 310009, Hangzhou, China.
Insights
A novel deep learning model, CHDdECG, effectively detects congenital heart diseases (CHDs) in children using electrocardiograms. This non-invasive method shows high accuracy, surpassing human experts in early CHD detection.
Area of Science:
- Pediatric Cardiology
- Artificial Intelligence in Medicine
- Biomedical Signal Processing
Background:
- Early detection of congenital heart diseases (CHDs) is crucial for improving treatment outcomes in pediatric patients.
- Current detection methods may require invasive procedures or lack cost-effectiveness.
- Non-invasive, low-cost techniques for pediatric CHD detection are highly desirable.
Purpose of the Study:
- To develop and validate a deep learning approach, CHDdECG, for automated detection of CHDs using pediatric electrocardiograms (ECGs).
- To assess the performance of CHDdECG against human expert diagnosis.
- To explore the contribution of automatically extracted ECG features versus human-defined features in CHD detection.
Main Methods:
- A deep learning model, CHDdECG, was developed to automatically extract features from pediatric ECGs and wavelet transformations.
- The model integrated automatically extracted features with key human-concept features.
- The model was trained on 65,869 cases and validated on multiple real-world and external test sets (totaling over 27,000 cases).
Main Results:
- CHDdECG achieved a ROC-AUC of 0.915 and specificity of 0.881 on a 12,000-case test set.
- External validation demonstrated strong performance with overall ROC-AUCs of 0.917 and 0.907, and specificities of 0.937 and 0.907.
- CHDdECG outperformed cardiologists in CHD detection accuracy, with automated ECG features showing greater importance than human-concept features.
Conclusions:
- The proposed CHDdECG deep learning model offers a highly effective and non-invasive method for pediatric CHD detection using ECGs.
- The model's performance surpasses that of human experts, suggesting it can identify patterns beyond human cognitive capabilities.
- Pediatric ECG analysis with deep learning holds significant potential for advancing CHD diagnosis and improving patient outcomes.
Abstract:
Early detection is critical to achieving improved treatment outcomes for child patients with congenital heart diseases (CHDs). Therefore, developing effective CHD detection techniques using low-cost and non-invasive pediatric electrocardiogram are highly desirable. We propose a deep learning approach for CHD detection, CHDdECG, which automatically extracts features from pediatric electrocardiogram and wavelet transformation characteristics, and integrates them with key human-concept features. Developed on 65,869 cases, CHDdECG achieved ROC-AUC of 0.915 and specificity of 0.881 on a real-world test set covering 12,000 cases. Additionally, on two external test sets with 7137 and 8121 cases, the overall ROC-AUC were 0.917 and 0.907 while specificities were 0.937 and 0.907. Notably, CHDdECG surpassed cardiologists in CHD detection performance comparison, and feature importance scores suggested greater influence of automatically extracted electrocardiogram features on CHD detection compared with human-concept features, implying that CHDdECG may grasp some knowledge beyond human cognition. Our study directly impacts CHD detection with pediatric electrocardiogram and demonstrates the potential of pediatric electrocardiogram for broader benefits.
Related Concept Videos
01:28Electrocardiogram Fundamentals
An electrocardiogram (ECG) is a diagnostic tool for identifying cardiac conditions such as arrhythmias, conduction abnormalities, and myocardial ischemia.
Definition
An electrocardiogram (ECG) visualizes the heart's electrical activity by tracing the electrical movement associated with each heartbeat on a graph or monitor. As the heart beats, an electrical wave passes through it, correlating with the cardiac cycle events.
Parts of an ECG
An ECG utilizes electrodes on the skin...
01:30Pulse rhythm
Conversely, an irregular pulse pattern is termed dysrhythmia, stemming from disruptions in cardiac...
01:29Electrocardiogram
Three major waveforms are present in a typical ECG recording: the P wave, the QRS complex, and...