Related Experiment Video
Updated: Jun 22, 2025
Noninvasive Electrocardiography in the Perinatal Mouse
Published on: June 12, 2020
Pediatric Electrocardiogram-Based Deep Learning to Predict Secundum Atrial Septal Defects
Joshua Mayourian1,2, Robert Geggel1,2, William G La Cava1,2
1Department of Cardiology, Boston Children's Hospital, 300 Longwood Avenue, Boston, MA, 02115, USA.
Insights
Artificial intelligence-enhanced electrocardiogram (AI-pECG) shows promise for detecting secundum atrial septal defects (ASD2) in children. This AI tool can help in early screening and diagnosis of ASD2 in pediatric patients.
Area of Science:
- Cardiology
- Artificial Intelligence
- Pediatric Health
Background:
- Secundum atrial septal defect (ASD2) detection in children is often delayed, leading to potential complications.
- Artificial intelligence (AI) has shown promise in adult ECG analysis for ASD2 detection, but pediatric applications are underexplored.
Purpose of the Study:
- To develop and evaluate an AI-pECG model for detecting ASD2 in pediatric patients (≤18 years old).
- To assess the model's performance in internal testing and emergency department cohorts.
Main Methods:
- A convolutional neural network (AI-pECG) was trained on paired ECG-echocardiograms (≤2 days apart) from pediatric patients.
- Model performance was evaluated using AUROC and AUPRC on internal testing and emergency department cohorts.
Main Results:
- The AI-pECG model demonstrated strong performance in both cohorts (Internal Test: AUROC 0.84, AUPRC 0.46; ED Cohort: AUROC 0.80, AUPRC 0.30).
- AI-pECG outperformed traditional ECG findings of incomplete right bundle branch block.
- Explainability analysis identified key ECG features associated with ASD2 risk.
Conclusions:
- AI-pECG shows significant promise for inexpensive screening and detection of ASD2 in pediatric patients.
- Further multicenter validation and prospective trials are needed to integrate AI-pECG into clinical decision-making.
Abstract:
Secundum atrial septal defect (ASD2) detection is often delayed, with the potential for late diagnosis complications. Recent work demonstrated artificial intelligence-enhanced ECG analysis shows promise to detect ASD2 in adults. However, its application to pediatric populations remains underexplored. In this study, we trained a convolutional neural network (AI-pECG) on paired ECG-echocardiograms (≤ 2 days apart) to detect ASD2 from patients ≤ 18 years old without major congenital heart disease. Model performance was evaluated on the first ECG-echocardiogram pair per patient for Boston Children's Hospital internal testing and emergency department cohorts using area under the receiver operating (AUROC) and precision-recall (AUPRC) curves. The training cohort comprised of 92,377 ECG-echocardiogram pairs (46,261 patients; median age 8.2 years) with an ASD2 prevalence of 6.7%. Test groups included internal testing (12,631 patients; median age 7.4 years; 6.9% prevalence) and emergency department (2,830 patients; median age 7.5 years; 4.9% prevalence) cohorts. Model performance was higher in the internal test (AUROC 0.84, AUPRC 0.46) cohort than the emergency department cohort (AUROC 0.80, AUPRC 0.30). In both cohorts, AI-pECG outperformed ECG findings of incomplete right bundle branch block. Model explainability analyses suggest high-risk limb lead features include greater amplitude P waves (suggestive of right atrial enlargement) and V1 RSR' (suggestive of RBBB). Our findings demonstrate the promise of AI-pECG to inexpensively screen and/or detect ASD2 in pediatric patients. Future multicenter validation and prospective trials to inform clinical decision making are warranted.
