Related Experiment Video
Updated: May 21, 2025
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
Published on: December 15, 2023
Expert-Level Automated Diagnosis of the Pediatric ECG Using a Deep Neural Network
Joshua Mayourian1, William G La Cava2, Sarah D de Ferranti1
1Department of Cardiology, Boston Children's Hospital, Department of Pediatrics, Harvard Medical School, Boston, Massachusetts, USA.
Insights
An artificial intelligence-enhanced electrocardiogram (AI-ECG) model accurately diagnoses pediatric ECGs, outperforming commercial software and potentially improving access to expert cardiac care globally.
Area of Science:
- Cardiology
- Artificial Intelligence
- Medical Diagnostics
Background:
- Global disparities in pediatric cardiologist access and ECG interpretation persist.
- Artificial intelligence-enhanced ECG (AI-ECG) shows promise in adult ECG diagnosis but is unexplored in pediatrics.
Purpose of the Study:
- To evaluate the accuracy of an AI-ECG model for automated diagnosis of pediatric ECGs.
- To assess the model's performance in detecting various ECG abnormalities.
Main Methods:
- Retrospective cohort study of 201,620 pediatric ECGs from Boston Children's Hospital (2000-2022).
- A convolutional neural network was trained and tested on ECGs for diagnostic prediction.
- Performance evaluated using AUROC and AUPRC curves for detecting any abnormality, WPW, and prolonged QTc.
Main Results:
- The AI-ECG model demonstrated high accuracy in detecting any abnormality (AUROC 0.94), WPW (AUROC 0.99), and prolonged QTc (AUROC 0.96).
- AI-ECG outperformed commercial software interpretations across all assessed conditions.
- Expert readers favored AI-ECG classifications over original cardiologist reads in discordant cases.
Conclusions:
- The developed AI-ECG model achieves expert-level automated diagnosis for pediatric 12-lead ECGs.
- This technology has the potential to significantly enhance access to specialized pediatric cardiac care worldwide.
Background:
Disparate access to expert pediatric cardiologist care and interpretation of electrocardiograms (ECGs) persists worldwide. Artificial intelligence-enhanced ECG (AI-ECG) has shown promise for automated diagnosis of ECGs in adults but has yet to be explored in the pediatric setting.
Objectives:
This study sought to determine whether an AI-ECG model can accurately perform automated diagnosis of pediatric ECGs.
Methods:
This retrospective single-center cohort study included all patients with an ECG at Boston Children's Hospital read by an experienced pediatric cardiologist (≥5,000 reads) between 2000 and 2022. A convolutional neural network was trained (75% of patients) and internally tested (25% of patients) on ECGs to predict ECG diagnoses. The primary outcome was a composite of any ECG abnormality (ie, detecting normal vs abnormal ECG). Secondary outcomes include Wolff-Parkinson-White syndrome (WPW) and prolonged QTc. Model performance was assessed with area under the receiver-operating (AUROC) and precision recall (AUPRC) curves.
Results:
The main cohort consisted of 201,620 patients (49% male; 11% with known congenital heart disease) and 583,134 ECGs (median age 11.7 years [Q1-Q3: 3.1-16.9 years]; 56% any ECG abnormality, 1.0% WPW, and 5.3% with prolonged QTc). The AI-ECG model outperformed the commercial software interpretations for detecting any abnormality (AUROC 0.94; AUPRC 0.96), WPW (AUROC 0.99; AUPRC 0.88), and prolonged QTc (AUROC 0.96; AUPRC 0.63). During readjudication of ECGs with AI-ECG/original cardiologist read discordance, blinded expert readers were more likely to agree with AI-ECG classification than the original reader to detect any abnormality (P = 0.001), WPW (P = 0.01), and prolonged QTc (P = 0.07).
Conclusions:
Our model provides expert-level automated diagnosis of the pediatric 12-lead ECG, which may improve access to care.
