Each year, millions of Americans develop structural heart disease (SHD)—damage to valves, walls, or chambers that silently strains the heart until a crisis hits. Traditional electrocardiograms (ECGs) are inexpensive.
Turning routine ECGs into lifesaving early-warning systems with AI— bridging technology and compassionate care for every heart in the Bronx and beyond
Why early detection still slips through the cracks
Each year, millions of Americans develop structural heart disease (SHD)—damage to valves, walls, or chambers that silently strains the heart until a crisis hits. Traditional electrocardiograms (ECGs) are inexpensive and everywhere, yet they cannot “see” most structural defects. Echocardiograms (ultrasound images of the heart) can, but they’re costlier, time-intensive, and not routinely ordered for every patient. The result? Too many Bronx neighbors learn about serious heart problems only after shortness of breath, fainting, or heart failure send them to the ER. CUIMC
Enter EchoNext: AI that listens between the heartbeats
Columbia University Irving Medical Center (CUIMC) and NewYork-Presbyterian researchers have trained a deep-learning model—EchoNext—on more than 1.2 million pairs of ECG and echocardiogram studies from 230,000 patients. The algorithm now recognizes the faint electrical fingerprints of hidden valve disease, cardiomyopathy, pulmonary hypertension, and more—clues that cardiologists were taught simply “aren’t in the ECG.” CUIMC
In head-to-head testing against 13 experienced cardiologists, EchoNext detected SHD correctly 77 % of the time, eclipsing the physicians’ 64 % average and outperforming even those who had AI assistance. CUIMC
How EchoNext reshapes the screening pathway
- Patient gets a routine ECG (for chest pain, diabetes check-up, pre-op clearance—anything).
- AI analyzes the waveform instantly.
- If risk is high, the care team orders an echocardiogram—turning an inexpensive first-line test into a smart funnel for advanced imaging.
When the team ran EchoNext invisibly in the background on almost 85,000 ECGs, it flagged 9 % of patients for hidden disease. Of those who eventually received their first echocardiogram, ~75 % were confirmed to have SHD—double the usual hit rate. Projected across all flagged patients, about 2,000 additional cases might have been caught that year. CUIMC
Why this matters for the Bronx
- Cost-effective screening: Turning ECGs—already part of many primary-care visits—into dual-purpose tools keeps expenses low.
- Health-equity impact: Our community experiences higher rates of hypertension, diabetes, and limited specialty access. AI-enhanced ECGs could spot trouble earlier, before gaps in care widen.
- Faster care pathways: By identifying who truly needs an ultrasound, families avoid needless testing while high-risk patients get same-day imaging.
Starling Health: prepared for the AI wave
At Starling Health we already provide:
| Rapid Test | Definitive Follow-Up | On-Site Turnaround |
|---|---|---|
| 12-lead ECG | Echocardiogram (Echo) | Same-day |
| Stress ECG | Nuclear or Exercise Stress Echo | Same-day |
| Chest X-ray | Cardiac CT / Calcium Scoring | 24-48 h |
| Lab cardiac markers | Cardiology consult | Same-week |
As EchoNext and similar models move from trials into practice, our cardiologists and technologists are ready to integrate AI decision support into this workflow—leveraging our state-of-the-art ultrasound suites, 3 T MRI, 128-slice CT, and nuclear cardiology labs.
What you can do today
- Know your numbers. High blood pressure, cholesterol, and A1C all raise SHD risk.
- Schedule an annual cardiac check. Even without symptoms, baseline ECGs provide a fingerprint AI can monitor over time.
- Listen to your body. New fatigue, swelling, or palpitations deserve prompt attention.
- Choose a one-stop center. Coordinated imaging, labs, and specialty visits under one roof slash delays between suspicion and diagnosis.
The road ahead
EchoNext is already in an eight-hospital clinical trial, and its creators have released a de-identified dataset so other health systems can refine local models. CUIMC has filed a patent for the algorithm, but the broader goal is clear: turn the world’s 400 million annual ECGs into 400 million chances to prevent heart failure, stroke, and premature death. CUIMC
At Starling Health, we view AI not as a replacement for expert clinicians, but as a force multiplier—helping our physicians focus on complex decision-making while machines handle signal-crunching at lightning speed.
Ready for a heart-health check-up?
Early detection saves lives. Call Starling Health at (718) 319-1610 or request an appointment today. Our board-certified cardiologists, imaging specialists, and multilingual care team are here seven days a week to keep your heart in rhythm.
If you have questions or need to schedule an appointment, our team at Starling Health is here to help with timely evaluations, advanced diagnostics, and compassionate care.
Disclaimer
This article is for informational purposes only and should not be considered medical advice. Every patient's situation is unique. Always consult with your physician or qualified healthcare provider regarding your specific condition or before making any medical decisions.
