Doctors using OpenEvidence's AI-based clinical decision support platform will soon be equipped with a new AI tool that can flag certain types of structural heart disease from electrocardiograms. The tool, called EchoNext, has received FDA approval for AI-powered detection of six types of hidden heart disease, including valve disease, congenital heart disease and other forms that ECGs alone have historically been unable to detect.
Researchers from NewYork-Presbyterian and Columbia University developed the tool, with Pierre Elias, M.D., medical director for artificial intelligence at NewYork-Presbyterian and assistant professor of medicine and biomedical informatics at Columbia University Vagelos College of Physicians and Surgeons, at the helm.
Now, Elias' company, Pathway Labs, is deploying the tool at scale, with the help of an $8.5 million seed round to support research and development and expand its deployment across health systems.
"While we have mammograms and colonoscopies for cancer, we have never had an equivalent form of early detection for the most common cause of death in the world -- heart disease," Elias, founder and CEO of Pathway Labs, said in a press release.
"Through EchoNext, we are able to detect high-risk conditions that the human eye can't and may otherwise be missed."
Pathway Labs' partnership with OpenEvidence will also expand the tool's reach. OpenEvidence is used by nearly two-thirds of U.S. physicians and touts clinical decision support, real-time documentation capabilities and the ability to synthesize medical literature.
The company has added a slew of new capabilities to its platform in recent months, from AI coding suggestions to prior authorization submission and an audio telehealth feature.
EchoNext demonstrates high accuracy in structural heart disease detection
EchoNext was born out of a desire to bridge the gap between ECGs and echocardiography, NewYork-Presbyterian explained in its announcement on EchoNext's FDA approval.
ECGs are common, inexpensive cardiac tests that measure the heart's electrical activity. However, these tests have limits. They are unable to detect structural heart disease.
Physicians would need to perform an echocardiogram, a test that uses high-frequency sound waves to create images of the heart's structure, to effectively diagnose certain structural heart problems, such as cardiomyopathy and pulmonary hypertension.
EchoNext was developed to bridge the gap by analyzing ECG data to determine when a follow-up echocardiogram is needed.
A 2025 study published in Nature found that the deep learning model, trained on more than 700,000 EKG-echocardiogram pairs, identified structural heart disease from EKG readings more often than cardiologists. EchoNext identified 77% of structural heart problems amid a sample of 3,200 ECGs, compared to 64% accuracy among cardiologists looking at the same dataset.
Additional research, this time in the form of a case study published in Nature Medicine on June 22, underscored these findings. The case involved a NewYork-Presbyterian patient who presented to the emergency room in February 2025 with four days of worsening shortness of breath and a cough after recent exposure to wildfire smoke.
The patient's lab tests appeared normal, and he was told he had an asthma flare-up and was discharged.
However, during his visit, the patient had an ECG, which was later analyzed by EchoNext. The AI model flagged the patient as high risk for undiagnosed structural heart disease, triggering an outpatient echocardiogram.
The echocardiogram revealed that the patient had a weakened heart, leading to a heart failure diagnosis. He later underwent a successful heart transplant that saved his life.
"This patient represents the first case of AI-detected advanced heart failure leading to heart transplantation. The same AI-ECG model has facilitated the first transcatheter aortic valve replacement for AI-detected severe aortic stenosis (a potentially life-threatening narrowing of the aortic valve), and the first surgical mitral valve repair for AI-detected severe mitral regurgitation," the study stated.
"These cases demonstrate the real-world applicability of AI-based diagnostics, suggesting the possibility for improving disease detection, particularly early in the disease course when treatments may be most effective."
Now, Pathway Labs is marketing EchoNext to reach more health systems and their patients.
OpenEvidence broadens access to FDA-approved AI tools
Given OpenEvidence's reach, Pathway Labs partnership with the company has the potential to vastly accelerate EchoNext's growth.
OpenEvidence is free to clinicians across the U.S., and access to reliable tools has always been at the core of the company's mission, Travis Zack, M.D., Ph.D., chief medical officer at OpenEvidence, said in a statement emailed to Healthtech Analytics.
"If it isn't implemented intentionally, the age of AI could deepen disparities in care across this country, rather than close the gap. We already see it with costly AI scribes that only large academic health systems can afford, even though it's community practices that are most overwhelmed by patient volume and documentation," Zack said.
"The same risk applies to this new generation of AI biomarkers like EchoNext. We don't want FDA-approved AI tools siloed in the ivory tower of academic medical centers, which historically have been the only places with the infrastructure to put them in front of physicians and patients."
Instead, Zack said, OpenEvidence envisions a model in which doctors have access to well-researched, trusted AI tools the moment they are approved.
"That ensures high-quality, high-impact technology doesn't die in academic procurement cycles, and that every physician, and their patients, can benefit from these breakthroughs without waiting years for them to diffuse from academia into the real world," he added.
With fresh FDA approval and seed funding, more health systems across the country could be using EchoNext in the near future.
Jill Hughes has covered health tech news since 2021. Her coverage areas include cybersecurity, HIPAA compliance, interoperability, AI and EHRs.
