Artificial intelligence applied to the standard electrocardiogram could help clinicians better identify patients with repaired tetralogy of Fallot who require closer follow-up with cardiac MRI, potentially improving access, efficiency and personalization of care.
Researchers from the Mount Sinai Kravis Children's Heart Center have led a multicentre study to develop and validate an artificial intelligence (AI) model capable of detecting clinically relevant heart changes in patients with repaired tetralogy of Fallot using a routine electrocardiogram (ECG). The study, supported by the National Institutes of Health, has been published in European Heart Journal: Digital Health.
Lifelong follow-up remains a challenge
Tetralogy of Fallot is a congenital heart defect that is usually corrected surgically in early childhood. Despite successful repair, patients require lifelong surveillance to monitor changes in heart size and function, particularly ventricular remodelling that may indicate deteriorating cardiac health.
Cardiac magnetic resonance imaging (MRI) is considered the gold standard for this follow-up. However, MRI is costly, time-consuming and not always readily available. As a result, many patients do not receive imaging at the recommended intervals, increasing the risk that clinically important changes go undetected.
Learning from ECG and MRI data
In the new study, investigators trained an AI model using paired ECG and cardiac MRI data from patients with repaired tetralogy of Fallot. The model was subsequently validated across five additional hospitals in North America, allowing the researchers to assess how well it performed in different clinical environments.
The AI system learned to recognise subtle ECG patterns associated with ventricular remodelling—changes in heart size and pumping function that are normally identified through MRI. According to the researchers, the findings suggest that a simple, widely available ECG, enhanced with AI, could help estimate which patients are most likely to benefit from earlier or more frequent MRI scanning.
Key findings
The study highlights several important points for clinical practice and digital health innovation:
- AI-enhanced ECGs can support risk stratification: A routine ECG may help identify patients at higher risk of ventricular remodelling who should be prioritised for MRI.
- Potential gains in access and efficiency: By safely delaying MRI in lower-risk patients, clinicians may be able to allocate imaging resources more effectively.
- Local validation is essential: Model performance varied between hospitals, underlining the need to test AI tools in each clinical setting before deployment.
“This research shows how artificial intelligence can unlock new value from a routine ECG,” said Son Duong, MD, MS, Assistant Professor of Pediatrics and Artificial Intelligence and Human Health at the Icahn School of Medicine at Mount Sinai and lead author of the study. “Our goal is to make lifelong heart monitoring more accessible and efficient for people born with congenital heart disease.”
Complement, not replacement
The researchers stress that the AI model is not intended to replace cardiac MRI. Instead, it is designed as a decision-support tool to help clinicians determine when advanced imaging is most urgently needed.
“As AI becomes more integrated into health care, it is critical to rigorously validate these tools across diverse clinical settings,” said Girish Nadkarni, MD, MPH, co–senior author and Chair of the Windreich Department of Artificial Intelligence and Human Health at the Mount Sinai Health System. “Our findings show both the promise of AI-enabled screening and the importance of testing performance at each site before real-world implementation.”
Why it matters
Patients with congenital heart disease often face decades of specialised follow-up care. By combining AI with a simple ECG, the approach outlined in this study could:
- Reduce unnecessary testing and associated healthcare costs
- Improve access to advanced imaging for patients who need it most
- Enable more personalised follow-up strategies and better long-term outcomes
The research team plans to evaluate the AI-ECG approach in prospective clinical studies and trials, as well as refine the model for use in younger patients. Ultimately, the aim is to integrate the technology into routine clinical workflows, supporting more targeted and sustainable lifelong cardiac care.
