Published on:
July 7, 2021
Heart MRI scan showing the fat area detected by the AI tool
With the new tool, the team was able to show that a larger amount of fat around the heart is linked to a significantly higher risk of diabetes, regardless of a person’s age, gender and body mass index.
The research is published in the journal Frontiers in Cardiovascular Medicine and is the result of funding for the CAP-AI program led by Barts Life Sciences, a research and innovation partnership between Queen Mary University of London and Barts Health NHS Trust.
The distribution of fat in the body can affect a person’s risk for various diseases. The commonly used measure of body mass index (BMI) mainly reflects the accumulation of fat under the skin and not around the internal organs. In particular, there is evidence that fat accumulation around the heart may be a predictor of heart disease and has been linked to a number of conditions including atrial fibrillation, diabetes, and coronary artery disease.
Lead researcher Dr. Zahra Raisi-Estabragh from Queen Mary’s The William Harvey Research Institute said: “Unfortunately, measuring the amount of fat around the heart manually is challenging and time-consuming. For this reason, it has not yet been possible to thoroughly investigate this in studies on large groups of people.
“To address this problem, we invented an AI tool that can be applied to standard MRI scans of the heart to automatically and quickly get a measurement of the adipose tissue around the heart in less than three seconds. This tool can be used by future researchers to learn more about the links between fat around the heart and the risk of disease, but also possibly in the future as part of standard care for a patient in the hospital. “
The research team tested the AI algorithm’s ability to interpret images from cardiac MRI scans of more than 45,000 people, including participants from the UK Biobank, a database of health information from over half a million participants across the UK. The team found that in these images, the AI tool could accurately determine the amount of fat around the heart and also calculate a patient’s risk of diabetes.
Dr. Queen Mary’s Andrew Bard, who led technical development, added: “The AI tool also has a built-in method of calculating the uncertainty of its own results so that it can be said to have an impressive ability to mark its own homework.”
Professor Steffen Petersen of Queen Mary’s William Harvey Research Institute who oversaw the project said: “This novel tool has great utility for future research and, once the clinical benefit is demonstrated, it can be used in clinical practice to improve patient care. This work underlines the value of interdisciplinary cooperation in medical research, especially in cardiovascular imaging. “
CAP-AI is led by Capital Enterprise in partnership with Barts Health NHS Trust and Digital Catapult, and the Alan Turing Institute, and is funded by the European Regional Development Fund and Barts Charity.
More information
- Research paper: ‘Automated Quality Controlled Cardiovascular Magnetic Resonance Pericardial Fat Quantification Using a Convolutional Neural Network in the UK Biobank’. Andrew Bard, Zahra Raisi-Estabragh, Maddalena Ardissino, Aaron Lee, Francesca Pugliese, Damini Dey, Sandip Sarkar, Patricia B. Munroe, Stefan Neubauer, Nicholas C. Harvey, Steffen E. Petersen. Limits in cardiovascular medicine.
