Related Experiment Video
Updated: Sep 26, 2025
Murine Fetal Echocardiography
Published on: February 15, 2013
Automatic Detection of Secundum Atrial Septal Defect in Children Based on Color Doppler Echocardiographic Images
Wenjing Hong1, Qiuyang Sheng2, Bin Dong3,4
1Department of Pediatric Cardiology, Shanghai Children's Medical Center, School of Medicine, Shanghai Jiao Tong University, Shanghai, China.
Insights
This study introduces an AI system for automatically detecting secundum atrial septal defects (ASD), a common congenital heart defect, using echocardiographic images with high accuracy. The AI system shows promise for improving the diagnosis of congenital heart diseases (CHDs) in children.
Area of Science:
- Cardiology
- Medical Imaging
- Artificial Intelligence
Background:
- Secundum atrial septal defect (ASD) is a prevalent congenital heart disease (CHD).
- Accurate and efficient diagnosis of ASD in children is crucial.
- Current diagnostic methods rely on expert interpretation of echocardiographic images.
Purpose of the Study:
- To develop and evaluate a fully automatic system for ASD detection in children using convolutional neural networks.
- To assess the feasibility and accuracy of AI-driven ASD detection from color Doppler echocardiographic images.
- To identify key echocardiographic views for robust ASD diagnosis.
Main Methods:
- A three-stage automated system was developed: echocardiographic view identification, cardiac structure segmentation and ASD candidate detection, and final detection inference.
- The system was trained on 370,057 images from 4,031 cases and validated on 203,619 images from 229 independent cases.
- Four clinically relevant echocardiographic views were utilized: subcostal, apical four-chamber, low parasternal four-chamber, and parasternal short-axis.
Main Results:
- The automated ASD detection system achieved high performance metrics on an independent test set.
- Image-level average performance included accuracy (0.8833), recall (0.8545), precision (0.8577), specificity (0.9136), and F1 score (0.8546).
- The system demonstrated robust performance across the four selected echocardiographic views.
Conclusions:
- The proposed AI system can automatically and accurately detect secundum atrial septal defects in children from echocardiographic images.
- This automated approach provides a strong foundation for AI-assisted diagnosis of congenital heart diseases.
- The findings suggest potential for improved efficiency and accuracy in pediatric cardiology diagnostics.
Abstract:
Secundum atrial septal defect (ASD) is one of the most common congenital heart diseases (CHDs). This study aims to evaluate the feasibility and accuracy of automatic detection of ASD in children based on color Doppler echocardiographic images using convolutional neural networks. In this study, we propose a fully automatic detection system for ASD, which includes three stages. The first stage is used to identify four target echocardiographic views (that is, the subcostal view focusing on the atrium septum, the apical four-chamber view, the low parasternal four-chamber view, and the parasternal short-axis view). These four echocardiographic views are most useful for the diagnosis of ASD clinically. The second stage aims to segment the target cardiac structure and detect candidates for ASD. The third stage is to infer the final detection by utilizing the segmentation and detection results of the second stage. The proposed ASD detection system was developed and validated using a training set of 4,031 cases containing 370,057 echocardiographic images and an independent test set of 229 cases containing 203,619 images, of which 105 cases with ASD and 124 cases with intact atrial septum. Experimental results showed that the proposed ASD detection system achieved accuracy, recall, precision, specificity, and F1 score of 0.8833, 0.8545, 0.8577, 0.9136, and 0.8546, respectively on the image-level averages of the four most clinically useful echocardiographic views. The proposed system can automatically and accurately identify ASD, laying a good foundation for the subsequent artificial intelligence diagnosis of CHDs.
