From Missed Signals to Actionable Insight: The Next Phase of AI in Heart Failure Care

Over the past few decades, clinical breakthroughs have helped support patients living with heart failure to achieve successful outcomes.  

However, according to research from the American College of Cardiology, the incidence of heart failure continues to rise and has led to a call for a focus on prevention and on earlier diagnosis. This presents the opportunity for digital health technology to help close the gap for patients suffering from heart failure.  

Artificial intelligence is emerging as the leading digital health technology projected to revolutionize many aspects of health care. Already applied as a critical tool in areas including imaging, cancer detection, clinical documentation, and drug discovery, AI has the potential to greatly improve the accuracy and speed of diagnosis for a difficult-to-detect form of heart failure. 

One company already translating this opportunity into clinical practice is Ultromics, a UPMC Enterprises portfolio company. Using a standard echocardiogram, Ultromics applies AI to accelerate diagnosis of heart failure and treatment through a scientifically validated solution.

Diagnosing Cardiac Amyloidosis and HFpEF

One of the most common occurrences that leads to heart failure is cardiac amyloidosis — a condition where abnormal proteins (amyloid) build up in the heart, leading to different types of heart failure, especially heart failure with preserved ejection fraction (HFpEF).   

“One of the big challenges in making this diagnosis is that the symptoms can be very nonspecific or be attributed to many different disease processes,” Michael Curren, MD, cardiologist and clinical informatics specialist at UPMC, explained in a recent webinar. 

With HFpEF, the heart’s pumping strength appears normal, but stiffened, thickened walls prevent it from properly relaxing and filling with blood. In contrast, heart failure with reduced ejection fraction (HFrEF), results from a weakened heart muscle that cannot contract effectively, reducing ejection fraction. Although both conditions cause similar symptoms, HFpEF is especially challenging to diagnose because of its seemingly normal test results.

Even with years of clinical expertise, symptoms including fatigue, shortness of breath, and swelling can be easy to misinterpret. As a result, clinicians must rely on a combination of subtle findings across multiple tests rather than a single clear indicator, making delayed or missed diagnosis more common.  

These delays not only impact patient outcomes, but also ripple across the entire health care continuum. When heart failure goes unrecognized until symptoms escalate, patients often require more intensive interventions and are more likely to return to the hospital with worsening disease. This pattern challenges care teams, increases clinical complexity, and contributes to the growing pressure on cardiovascular services nationwide. 

“Heart failure is the most common cause of readmission to the hospital amongst Medicare beneficiaries,” said Jordan Strom, MD, director of echocardiographic research at Beth Israel Deaconess Medical Center. “Even if you make a dent in a small proportion of those who are presenting for care, that has massive ramifications for both those patients and the health care system as a whole.” 

Now, for the first time, technology-enabled diagnostics exist that can make early detection possible.

Approaching Heart Failure from All Angles

Early deployments of AI-enabled echocardiography have the potential to reduce diagnostic delays and improve time to treatment; particularly when embedded into established clinical pathways.  

“There’s a lot of upstream opportunity with data models that are available to start to hit that diagnosis at risk stage prior to any development of symptomatology,” said Dr. Curren. 

The Ultromics model identifies subtle signals like minute texture changes, the “sparkly myocardium” pattern associated with amyloid, and abnormalities that are too faint for the human eye to detect. 

For Ross Upton, CEO of Ultromics, this clinical challenge provides the perfect opportunity for AI-powered technology to address a disease that’s often missed, extremely costly, and finally actionable thanks to AI. 

“I’m obviously a developer, but my company also clinically implements AI tools,” Upton said. “And from my perspective, heart failure serves as an extremely good use case for developing an AI tool for early detection.” 

Dr. Curren supports the positive impact that comes from a strong clinical foundation and proven AI technology for the benefit of his patients, and there is clear urgency for faster and more accurate diagnoses. As new therapies come to market that can slow or halt the progression of heart failure, it becomes imperative to diagnose the disease early. That is where AI can drive significant benefits for patients. 

“I’ve been in practice for about a decade and seeing the therapies for amyloidosis come to the market and watching the halting of disease progression is remarkable,” he said.

The Future of AI for Heart Failure Diagnosis 

AI’s role in heart failure diagnosis is poised to grow significantly in the next decade. As AI becomes more multimodal, it will help expand access to advanced diagnostics and enhance the utility of tools already embedded in cardiology. 

Ultromics has been part of the UPMC Enterprises portfolio since 2024, and recently completed a successful $55 million Series C funding, with support from UPMC Enterprises and others. With this funding, Ultromics seeks to expand access to hospitals and echo labs across the U.S. treating the most at-risk patients, aiming to make AI diagnostics a routine part of heart exams. The company is also growing its offerings to cover more heart conditions, reach new markets, and strengthen ties with health systems. 

UPMC Enterprises continues to partner with Ultromics as it brings clinically validated AI tools into cardiology. As AI adoption grows, success will depend on solutions that align with clinical workflows, support care teams, and deliver measurable improvements in patient outcomes. 

Key aspects of these insights were discussed during the webinar, “Earlier, Smarter, Better: Transforming Heart Failure Diagnosis with AI,” hosted by the Digital Medicine Society (DiMe) in October 2025. Watch the full conversation here. 

Next Steps

• Dive into UPMC Enterprises’ investment in Ultromics and the opportunities for growth in the cardiology space. 

• Our Technology Services team is committed to advancing health care through the application of AI tools. Learn more about Ahavi. 

Note: UPMC has a financial interest in Ultromics.