New AI Tool Predicts Heart Failure Risk Five Years Early

Researchers at the University of Oxford have developed an artificial intelligence tool capable of predicting the risk of heart failure at least five years before the condition develops. The technology identifies signs of inflammation and unhealthy fat around the heart that are invisible to the human eye during routine medical imaging.

The findings, published on April 8, 2026, in the Journal of the American College of Cardiology, suggest a significant shift in how clinicians may identify high-risk patients. By utilizing existing diagnostic infrastructure, the tool allows doctors to determine a patient’s risk score, which can inform decisions regarding the intensity of monitoring and early intervention strategies.

How the AI Tool Functions

The AI program analyzes data from routine cardiac CT scans, which are commonly performed in hospitals to investigate chest pain or to detect fatty plaques in the coronary arteries. In the United Kingdom, approximately 350,000 patients are referred for these scans annually.

View this post on Instagram

While these scans are typically used to look for arterial blockages, the Oxford tool focuses on the textural changes in the fat surrounding the heart. These specific changes indicate that the underlying heart muscle is inflamed and unhealthy, a detail that cannot be spotted by doctors through any other routine medical imaging tests.

By detecting these markers, the AI can warn physicians if a patient is at a high risk of developing heart failure, providing a window for preventative care or earlier management of the condition.

Study Results and Accuracy

The tool was trained and validated using data from 72,000 patients across nine NHS trusts in England. These individuals were followed for a decade after their initial CT scans to determine the accuracy of the AI’s predictions.

The study reported the following outcomes:

• The AI predicted the risk of developing heart failure within the next five years with 86% accuracy.
• Patients categorized in the highest risk group were 20 times more likely to develop heart failure than those in the lowest risk group.
• Individuals in the highest risk group had approximately a one in four chance of developing the condition within five years.

Public Health Context

Heart failure is a condition where the heart cannot pump blood around the body as effectively as it should, affecting more than 60 million people worldwide. In the UK, heart disease remains the second largest cause of death; according to the Office for National Statistics, over 54,000 people died from it in 2024, representing 9.6% of all deaths in the country.

Heart failure is often preceded by heart disease, which can be caused by the buildup of fat in surrounding tissues. Experts suggest that spotting these cases before they progress to full heart failure would allow doctors to manage the condition at an earlier stage or potentially prevent it entirely.

Future Applications

The research was led by Professor Charalambos Antoniades and funded by the British Heart Foundation. While the current tool relies on cardiac CT scans, the research team is looking to expand the application of this technology.

Although this study used cardiac CT scans, we are now working towards applying this method to any CT scan of the chest, performed for any reason.

Professor Charalambos Antoniades

This expansion would potentially allow the AI to screen for heart failure risk in patients who are receiving chest scans for unrelated medical reasons, further increasing the number of high-risk individuals who could be identified and treated early.