Developing faster, cheaper, and higher quality brain scans for dementia diagnosis

Lead Investigator: Professor Geoff Parker

Institution: University College London

Grant type: Alzheimer’s Society Project Grant

Start date: January 2023

Duration: 30 months

Amount: £389,834.08

This project is supported by the Heather Corrie Impact Fund.

“The length of time for a person to have an MRI being shortened and the images being better would really make a difference to people's experience, those in the machines and those interpreting the findings”

Alzheimer's Society Research Network Volunteer

Project summary:

Currently, dementia diagnosis involves attending memory clinics with blood tests, memory and thinking assessments, and having MRI or CAT scans. Diagnosis and MRI scanning is time consuming, inefficient and expensive. 

Professor Parker’s project will test a special ultra-fast type of magnetic resonance imaging (MRI) scan originally designed for diagnosing brain tumours and apply it to dementia diagnosis. This could help speed up diagnosis of Alzheimer’s disease and could also improve access to scanning methods that are not currently available.  

Project background:

Dementia diagnosis involves brain imaging, for example using MRI or CAT scans. These brain scans allow clinicians to look for causes of symptoms and may help to determine which type of dementia people may have. Ideally, MRI scans would be a preferred method as they provide better quality brain images and do not expose people to radiation unlike CAT scans. However, in practice, 87% of people will be referred for a CAT scan due to the low availability, high cost, and time taken to have an MRI scan. 

One of the biggest challenges in using MRI scans to diagnose dementia is the slow scanning time. Developing faster ways of completing high-quality MRI scans would release scan slots, decrease the costs associated with MRI scanning, and allow clinicians to look at high-quality images to make an accurate diagnosis. 

What does this project involve?

The project aims to cut down on MRI scan time by producing lower-quality images, which will then be enhanced for quality using specially developed computer systems. 

The project will use machine learning methods to automatically learn, analyse data, and look for patterns. This method will be tested for reliability by scanning eight participants multiple times on different MRI scanners. Then to test the accuracy it will be tested in a clinical trial of 130 people. Whilst being scanned as part of routine appointments, they will be asked to stay in the scanner for an extra 7 minutes to run the new method and will have an extra blood test taken. Specialists will compare the ultra-fast and standard MRI image sets. They will review both sets of images (at least 1 month apart to prevent them remembering the other set) and will be asked to provide a diagnosis for each. This will show if the new faster technique is as accurate as the existing MRI method.

How will this project help people with dementia?

With shorter scan times, this new MRI method should be cheaper without sacrificing accuracy and therefore be accessible to more people across the UK. 

If successful, the project will allow more people with dementia to gain access to safer, faster, and more accessible diagnostic brain scans ultimately leading to a faster and more accurate diagnosis. People with dementia would benefit from shorter waiting times, which would speed up access to disease-specific support, therapies, and access to research.

This project could help improve and standardise image-based dementia diagnosis across the NHS.