A new AI tool called MindGlide has been created by researchers at UCL to help doctors see how well treatments are working for people with multiple sclerosis (MS).
MS is a disease where the immune system attacks the brain and spinal cord, causing problems with movement, feeling, and thinking. Around 130,000 people in the UK live with MS, and it costs the NHS over £2.9 billion each year.
MindGlide uses MRI brain scans to look for damage, brain shrinkage, and plaques (also called lesions). These signs help doctors track how MS is progressing. Before this tool, experts had to spend weeks manually checking brain scans to get this information.
Now, MindGlide can do the same job in just 5 to 10 seconds per image, even using routine hospital scans that were not detailed enough for this work in the past.
In a study, MindGlide was tested on over 14,000 brain images from more than 1,000 MS patients. It was better at spotting changes than two other well-known AI tools (SAMSEG and WMH-SynthSeg). It found 60% more plaques than SAMSEG and was 20% better than WMH-SynthSeg.
Dr. Philipp Goebl from UCL said MindGlide could help unlock valuable data from millions of brain scans already stored in hospitals. These images can give new insights into how MS affects the brain and how treatments are working.
The tool also worked well with lower-quality scans, including those missing key scan types. It found changes in both the outer and deeper parts of the brain and gave accurate results over time.
MindGlide also supported findings from earlier MS studies, showing which treatments work best.
Dr. Arman Eshaghi, the lead researcher, said this AI tool could help doctors understand MS in a wider range of patients—not just those in clinical trials.
What’s next?
Right now, MindGlide only works on brain scans. The next step is to include spinal cord images to give a full picture of MS damage.
MindGlide was trained using thousands of MRI scans from MS patients. It learned to recognize signs of MS using these images and is now ready to help doctors and researchers in real-world settings.
Source: medicalxpress
