Related Experiment Video
Updated: Jun 26, 2025
Murine Fetal Echocardiography
Published on: February 15, 2013
Deep learning-based differentiation of ventricular septal defect from tetralogy of Fallot in fetal echocardiography
Xia Yu1,2, Liyong Ma2,3, Hongjie Wang1,2
1Weihai Maternal and Children Health Hospital, Weihai, Shandong, China.
Insights
Accurate differentiation of Tetralogy of Fallot (TOF) and Ventricular Septal Defect (VSD) in fetal echocardiography is crucial. The weakly supervised data augmentation network (WSDAN) model demonstrated superior performance in distinguishing these congenital heart diseases.
Area of Science:
- Medical Imaging
- Artificial Intelligence in Healthcare
- Pediatric Cardiology
Background:
- Congenital heart disease (CHD) significantly impacts children's health and quality of life.
- Early detection of CHD is vital for reducing its long-term effects.
- Tetralogy of Fallot (TOF) and Ventricular Septal Defect (VSD) are common CHDs with similar echocardiographic features but differing prognoses.
Purpose of the Study:
- To differentiate between TOF and VSD using fetal echocardiography images.
- To evaluate the performance of various Convolutional Neural Network (CNN) models for this diagnostic task.
- To identify the most effective AI model for accurate TOF/VSD classification.
Main Methods:
- Utilized a dataset of 105 fetal echocardiography images of TOF and 96 images of VSD.
- Employed four CNN models: VGG19, ResNet50, NTS-Net, and WSDAN.
- Compared model performance using metrics including sensitivity, accuracy, specificity, and Area Under the Curve (AUC).
Main Results:
- VGG19 and ResNet50 showed comparable performance with AUCs of 0.799 and 0.802.
- NTS-Net and WSDAN, designed for fine-grained categorization, achieved higher AUCs of 0.823 and 0.873.
- WSDAN outperformed all other tested models in differentiating TOF from VSD.
Conclusions:
- The WSDAN model demonstrated the highest efficacy in distinguishing between TOF and VSD.
- WSDAN's superior performance suggests its potential for clinical application and popularization.
- AI-driven image analysis holds promise for improving the diagnosis of congenital heart diseases.
Background:
Congenital heart disease (CHD) seriously affects children's health and quality of life, and early detection of CHD can reduce its impact on children's health. Tetralogy of Fallot (TOF) and ventricular septal defect (VSD) are two types of CHD that have similarities in echocardiography. However, TOF has worse diagnosis and higher morality than VSD. Accurate differentiation between VSD and TOF is highly important for administrative property treatment and improving affected factors' diagnoses.
Objective:
TOF and VSD were differentiated using convolutional neural network (CNN) models that classified fetal echocardiography images.
Methods:
We collected 105 fetal echocardiography images of TOF and 96 images of VSD. Four CNN models, namely, VGG19, ResNet50, NTS-Net, and the weakly supervised data augmentation network (WSDAN), were used to differentiate the two congenital heart diseases. The performance of these four models was compared based on sensitivity, accuracy, specificity, and AUC.
Results:
VGG19 and ResNet50 performed similarly, with AUCs of 0.799 and 0.802, respectively. A superior performance was observed with NTS-Net and WSDAN specific for fine-grained image categorization tasks, with AUCs of 0.823 and 0.873, respectively. WSDAN had the best performance among all models tested.
Conclusions:
WSDAN exhibited the best performance in differentiating between TOF and VSD and is worthy of further clinical popularization.
