The FDA has cleared EchoNext, the first artificial intelligence tool designed to detect underlying structural heart disease from routine electrocardiograms (ECGs). Developed by researchers at NewYork-Presbyterian and Columbia University and commercialized by Pathway Labs, EchoNext scans standard ECG waveforms to flag patients at higher risk of structural problems who should be referred for an echocardiogram.
Why this matters
Cardiovascular disease is the leading cause of death worldwide, and until now there has been no simple screening test for many forms of structural heart disease. Echocardiography is the diagnostic standard for valve disease, chamber abnormalities and other mechanical heart problems, but echos are more costly and less available than ECGs. EchoNext aims to turn the inexpensive, widely used ECG into an effective early-detection screen that identifies people who need follow-up imaging.
How the tool works and evidence so far
Pathway Labs says the model was trained on more than 700,000 paired ECG and echocardiogram records from the NewYork-Presbyterian health system. In a 2025 Nature study, EchoNext correctly identified about 77% of structural heart problems from ECGs alone, compared with roughly 64% accuracy for cardiologists reading the same ECGs. The FDA clearance covers six indications related to structural heart disease; the company has publicly disclosed five of those conditions, including forms of heart failure and valve disease.
A published case in Nature Medicine described EchoNext flagging severe, previously undiagnosed heart failure in a 45-year-old man, a detection that ultimately led to a heart transplant. Larger clinical trials are underway, and Pathway Labs has raised $8.5 million to expand deployment into more health systems.
Potential benefits
EchoNext could improve early detection by identifying structural disease before symptoms appear, when treatment options may be more effective. It may also reduce unnecessary echocardiograms by helping clinicians rule out structural disease in low-risk patients, saving costs and patient time. Pathway Labs is partnering with the clinical platform OpenEvidence to make the tool available to physicians across hospitals and community practices.
Cautions and equity concerns
Cardiologists stress that AI should be an aid, not the sole arbiter, in diagnosis. Experts emphasize the continued importance of clinical history, physical exam and confirmatory testing. There are also concerns about bias: an algorithm is only as good as the data used to train it, and if training data underrepresents certain groups the tool could perform unevenly. The Nature study reported no clinically relevant differences in performance by race, ethnicity or sex in its test set, but clinicians and researchers say wider validation across diverse populations is essential.
Broader context
Adoption of an ECG-based screening tool could help address uneven referral patterns for echocardiography and subsequent treatments — a documented issue for some racial and ethnic groups. Still, prevention remains key: guideline-recommended management of modifiable risk factors and heart-healthy lifestyle measures continue to be central to reducing cardiovascular disease burden.
Bottom line
EchoNext represents a novel, FDA-cleared use of AI to extract structural heart disease signals from routine ECGs, with promising early results and real-world impact already described in a published case. It may expand access to early detection and help prioritize confirmatory imaging, but physicians caution that further validation, careful deployment, and attention to equity will be important as the tool is rolled out more broadly.
