MRI Screening with AI; A Prospective study
Research type
Research Study
Full title
Feasibility of fast imaging methods for whole body AI-assisted screening in MRI
IRAS ID
300823
Contact name
Marc Rea
Contact email
Sponsor organisation
Clatterbridge Cancer Centre
Duration of Study in the UK
2 years, 0 months, 0 days
Research summary
Technical advances in MRI and computing are leading towards the capability for fast whole-body screening of all patients who attend for an MRI exam. This study will investigate different methods of MR imaging, reconstruction, and automated image analysis that may lead to a future screening program.
Patients are currently scanned in MRI over a limited region of interest to diagnose suspected disease. The major restriction for MRI examinations is the length of time examinations take, which includes the setting up of patients on the scanner and the lengthy (average time is 20-30 mins) imaging examination. There is also a restriction imposed by the availability of radiologists to provide a report of the images. Recent improvements in medical imaging and computing power now allow us to work towards taking whole-body pictures of a patient’s insides, and have computers automatically and quickly detect abnormalities.
With the development of machine learning methods for image diagnosis, an opportunity is presented to investigate whether it is feasible for a quick additional body scan to be incorporated into standard protocols and automatically processed to provide a whole body screening of all MRI patients. To make this work practically, there are various problems that need to be solved, firstly that pictures should be acquired quickly so the total scan time does not increase significantly, and that there is no extra burden on the patient or staff.
Ideally we would also have the images reviewed before the end of the exam.
It is expected that AI computers can be programmed to do this during an MRI examination, meaning all patients can be scanned and potentially benefit from the early detection of disease. This project will begin the process by finding the best type of images to acquire and optimise how they can be read by the computer program.REC name
London - Queen Square Research Ethics Committee
REC reference
22/PR/0038
Date of REC Opinion
2 Aug 2022
REC opinion
Further Information Favourable Opinion