Accumulation of fat specifically around the heart has long been linked to cardiovascular and metabolic disease but until now there hasn’t been a simple way to measure this. A new artificial intelligence tool has been developed that can quantify these fat deposits from regular MRI images.
Pericardial adipose tissue (PAT) is a particular collection of fat tissue surrounding the surface of the heart. High levels of PAT, separate to body weight or body mass index, have been linked to increased risk of diabetes and coronary heart disease but the association has remained a hypothesis due to measurement challenges.
The best way we can currently measure PAT levels is using a computed tomography (CT) scan. The problem is CT scans expose patients to small but relevant levels of ionizing radiation, which limits the ability of researchers to ethically measure PAT levels in large, diverse cohorts of healthy volunteers.
"To address this problem, we've invented an AI tool that can be applied to standard heart MRI scans to obtain a measure of the fat around the heart automatically and quickly, in under three seconds,” says Zahra Raisi-Estabragh, lead researcher on the new study.
After training a novel neural network its accuracy was assessed using several independent datasets. The researchers validated the tool could indeed accurately measure total PAT volumes from simple cardiac MRI imaging.
To demonstrate the new tool’s clinical potential the researchers looked at a large UK Biobank dataset of over 40,000 people. The tool measured PAT levels and then looked at the associations between PAT and diabetes.
The tool accurately evaluated a patient’s risk of diabetes based on the amount of fat it measured around the heart. Raisi-Estabragh notes this doesn’t immediately mean the tool is good for a diagnostic diabetes test but instead offers insights into how associations between disease and PAT levels can now be effectively investigated.
“This tool can be used by future researchers to discover more about the links between the fat around the heart and disease risk, but also potentially in the future, as part of a patient's standard care in hospital,” adds Raisi-Estabragh.
The researchers say the tool is now ready to be used in new clinical research. This will be the first time PAT levels can be easily measured in large cohorts of people.
Steffen Petersen, a supervisor on the research from Queen Mary University of London, says ultimately this tool may very well move into clinical applications if further study clearly validates PAT measures alongside disease risk.
“This novel tool has high utility for future research and, if clinical utility is demonstrated, may be applied in clinical practice to improve patient care,” says Petersen. “This work highlights the value of cross-disciplinary collaborations in medical research, particularly within cardiovascular imaging."
The new study was published in Frontiers in Cardiovascular Medicine.
Source: Queen Mary University of London