por
John R. Fischer, Senior Reporter | March 28, 2022
A new AI software can diagnose heart disease in 20 seconds. (Photo source: University College London)
The National Health Services (NHS) in the U.K. has a new AI tool that can diagnose heart disease from MR scans in just 20 seconds.
The solution is designed to mimic the ability of humans to do so, but at a faster speed and with greater precision. It also diagnoses the patient while they are still on the scanner, which is faster than the 13 minutes or more that doctors need to manually examine images after the scan is complete, according to the British Heart Foundation (BHF).
When paired against three specialists, the technique was more precise at analyzing MR scans. Additionally, it detected changes to the heart structure and function with 40% greater accuracy and extracted more information than a human could. “The beauty of the technology is that it replaces the need for a doctor to spend countless hours analyzing the scans by hand. We are continually pushing the technology to ensure it's the best it can be, so that it can work for any patient with any heart disease,” said Dr. Rhodri Davies, of the University College London Institute of Cardiovascular Science and Barts Heart Centre at St Bartholomew’s Hospital and lead researcher of the project, in a statement.
Ad Statistics
Times Displayed: 80682
Times Visited: 5346 MIT labs, experts in Multi-Vendor component level repair of: MRI Coils, RF amplifiers, Gradient Amplifiers Contrast Media Injectors. System repairs, sub-assembly repairs, component level repairs, refurbish/calibrate. info@mitlabsusa.com/+1 (305) 470-8013
The tool was trained on MR scans for nine different conditions, including heart attack; high blood pressure, aortic stenosis (narrowing of the heart valve); hypertrophic cardiomyopathy (thickening of the left ventricle); dilated cardiomyopathy (weakening of the heart muscle); cardiac amyloidosis (deposits of an abnormal protein — amyloid — in the heart tissue); and Fabry disease (a rare inherited fat metabolism disorder that affects the heart).
A study on the technique utilized data from heart MR scans for 1,923 people and was published in the
Journal of Cardiovascular Magnetic Resonance.
The solution is being rolled out at University College London Hospital, the Barts Heart Centre at Barts Health NHS Trust and Royal Free Hospital, where it is being used on more than 140 patients a week. Following this initial rollout, Davies and his colleagues will use the data they acquire to further train and refine the AI and make it accessible for a wider range of heart patients in the U.K. and globally. They will also train it to quantify heart valve disease and congenital heart defects that develop in the womb.
Around 120,000 heart MR scans are performed annually in the U.K. The researchers plan to implement it at 40 other locations in the U.K. and then worldwide later this year and expect the new tool to provide medics with time to see more patients.