Artificial intelligence can measure heart function from MRI scans in a way that is more closely linked to patient survival than measurements made by doctors, a new study has found. The research was published on November 2 in The Lancet Digital Health.
The study was led by City St George’s, University of London and University College London. Researchers say the results suggest AI could serve as a useful first step for measuring heart function.
What the Study Measured
The research focused on ejection fraction. This is a measure of how well the heart pumps blood, and it is considered the most important measure of heart function.
Doctors use ejection fraction to guide treatment for people with heart disease. It helps them decide whether patients may need specialist heart failure treatment, implanted heart devices, or procedures to improve blood flow to the heart.
The measure is also used to judge risk before surgery. Doctors rely on it to track heart function in patients receiving chemotherapy as well.
Researchers analyzed heart MRI scans from 27,382 patients in the UK and the United States. All of the patients were being checked for known or suspected heart conditions.
Patients were followed for an average of more than five years. During that time, 3,458 of them died.
The team compared ejection fraction readings from an AI tool with readings from experienced clinicians who reviewed the same cardiac MRI scans. It is the largest study of its kind.
AI Results Faster and More Predictive
AI-based measurements were slightly better at predicting which patients faced a higher risk of death. The difference was statistically meaningful and held up after researchers accounted for factors such as age and sex.
The AI tool was also much faster. It analyzed scans in about 2 to 3 seconds per patient, while earlier work by the same team found expert analysis usually takes about 13 minutes.
Dr. George Joy, the study’s lead author, is an NIHR Clinical Lecturer at City St George’s. He said ejection fraction helps determine how patients are treated and how closely they are monitored.
Joy said the findings suggest AI can measure the marker with high consistency and precision. He described the result as “a small but meaningful improvement in identifying patients at higher risk.”
The researchers said human oversight is still essential. AI errors were uncommon, with an error rate of about 1% at the patient level, affecting just 0.08% of images.
Clinicians should still check results and interpret scans in the wider clinical context, the team said. Dr. Rhodri Davies, an associate professor at UCL, is the study’s senior author and the inventor of the AI tool.
“AI is not replacing clinicians, but it can provide faster, more consistent measurements at low-cost and support more streamlined clinical decision-making that ultimately benefits patients,” Davies said. He added that even small gains in accuracy can matter when applied across thousands of patients.
The researchers say the findings support using AI as a first-line tool for cardiac MRI analysis, with expert review when needed. The study was funded by an NIHR Clinical Lectureship, the NIHR UCLH Biomedical Research Centre, and a British Heart Foundation Accelerator Award.
The team now plans to combine AI heart scan measurements with results from other tests, especially electrocardiograms (ECGs). The goal is to spot heart muscle problems earlier so treatment can start before lasting damage occurs.
