AI-powered ECG could help guide lifelong heart monitoring for patients with repaired tetralogy of fallot

Researchers at the Mount Sinai Kravis Children’s Heart Center led a multicenter effort to develop and validate an artificial intelligence (AI) tool that can analyze a standard electrocardiogram (ECG) to identify patients with repaired tetralogy of Fallot who may be at risk for harmful heart changes typically detected by cardiac MRI. The study, supported by the National Institutes of Health, was published in the European Heart Journal: Digital Health.

Tetralogy of Fallot is a congenital heart defect that requires surgical repair in childhood, but patients need lifelong monitoring to detect changes in heart size and function. Cardiac MRI is the gold-standard imaging test for this follow-up care; however, MRI scans are expensive, time-consuming, and not always easily accessible. Many patients miss recommended imaging.

In this multicenter study, investigators trained an AI model using ECG and MRI data from patients with repaired tetralogy of Fallot and validated the model across five additional hospitals in North America. The AI learned patterns in ECG signals linked to ventricular remodeling—changes in heart size and pumping function that can signal worsening health.

Key findings

• AI plus ECG can estimate risk of ventricular remodeling: A quick, widely available test may help identify patients who need MRI sooner.
• There is potential to improve access and efficiency: The tool could help doctors prioritize MRI scans for higher-risk patients while safely delaying scans for lower-risk patients.
• Performance varied by hospital: Results highlight the importance of validating AI tools locally before clinical use.

“This research shows how artificial intelligence can unlock new value from a routine ECG,” said Son Duong, MD, MS, lead author and Assistant Professor of Pediatrics, and Artificial Intelligence and Human Health at Icahn School of Medicine at Mount Sinai. “Our goal is to make lifelong heart monitoring more accessible and efficient for people born with congenital heart disease.”

The researchers emphasize that the model is not intended to replace cardiac MRI. Instead, it could help clinicians decide when 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 Barbara T. Murphy Chair of the Windreich Department of Artificial Intelligence and Human Health, at 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.” Dr. Nadkarni is also the Director of the Hasso Plattner Institute for Digital Health and Chief AI Officer, Icahn School of Medicine at Mount Sinai.

Why this matters

Patients with congenital heart disease often require lifelong specialized follow-up care. By combining AI with a simple ECG, the researchers hope to:

• Reduce unnecessary testing and health care costs
• Improve access to advanced imaging for patients who need it most
• Personalize follow-up care and improve long-term outcomes

Next steps

The research team plans to test the AI-ECG approach in prospective clinical studies and trials and refine the model for younger patients. The long-term goal is to integrate the tool into routine clinical care.

Reference:

Son Q Duong, Akhil Vaid, Pengfei Jiang, Yuval Bitterman, Yamini Krishnamurthy, I Min Chiu, Joshua Finer, Brian Cleary, Benjamin S Glicksberg, Ruchira Garg, Michael DiLorenzo, Mark Friedberg, Evan Zahn, Matthew Lewis, Michael Satzer, David Ouyang, Pierre Elias, Tal Geva, Sunil Ghelani, Brett R Anderson, Ali Zaidi, Rachel M Wald, Girish N Nadkarni, Joshua Mayourian, Development and multicentre validation of an artificial intelligence electrocardiogram model for ventricular remodeling in repaired tetralogy of Fallot, European Heart Journal - Digital Health, Volume 7, Issue 2, March 2026, ztag015, https://doi.org/10.1093/ehjdh/ztag015