Using AI to analyse MRI scans to better predict how muscle wasting conditions progress

Muscle wasting conditions can affect people differently, even if they have the same diagnosis. Some people lose strength quickly, while others remain stronger for much longer. This makes it difficult for doctors to predict how the condition will progress and to give patients and families clear information about what to expect. 

Muscle MRI scans are already used to look at how muscles change over time. They show which muscles are becoming weaker and being replaced by fat. However, doctors currently look at these scans one patient at a time, and there is no easy way to compare thousands of scans to identify common patterns of condition progression. 

The research team at Newcastle University has built one of the world’s largest collections of muscle MRI scans from people with muscle wasting conditions. This provides a unique opportunity to learn from thousands of people with a condition and better understand how different conditions change over time. We previously funded research into a new AI tool called Myo-Guide, which will help doctors read muscle scans more easily. It uses artificial intelligence (AI) to look at how much fat is in the muscles and compares this to patterns from known conditions. This helps doctors make faster and more accurate diagnoses, even when the scans are complex. This work will be expanded by Dr Verdu Diaz in this new study. 

This study will use muscle MRI scans from thousands of people with muscle wasting conditions to identify common patterns of condition progression. 

Using new computer-based methods, scans from many different patients will be compared to understand how muscle weakness develops over time. The hope is that the research team will be able to group together people whose muscles are changing in similar ways, even if they have different genetic diagnoses. 

The information will be used to create an online tool that allows doctors to compare a person’s muscle scan with thousands of others. This could help estimate how an individual’s condition is likely to progress and improve understanding of rare muscle wasting conditions. The tool will be freely available to doctors and researchers around the world.

People living with muscle wasting conditions often ask how their condition is likely to change in the future, but this can be difficult to predict. This project aims to make those predictions more accurate by learning from thousands of muscle scans collected from people around the world. 

Better predictions could help doctors plan care, support those with a condition and their families, and identify those who may benefit most from new treatments. It can also help families plan for their future needs. It could also improve clinical trials by bringing together those with similar patterns of the condition, making it easier to tell whether a new treatment is working.