AI and MRI: improving the diagnosis of people with muscle wasting conditions 

In most cases of muscle wasting conditions, diagnosis isn’t straightforward. It requires linking clinical information with various medical and genetic tests. Magnetic resonance imaging (MRI) of muscles has been used to help diagnose conditions as it can identify areas where fat replaces muscle in these conditions – often called the pattern of fat replacement.  

However, the number of conditions with a similar pattern of fat replacement is growing, making it more and more difficult for clinicians to use MRI to provide an accurate diagnosis. We previously supported Professor Diaz-Manera and his team to develop a machine learning tool called MYO-Guide which analyses muscle MRIs and predicts a diagnosis of 20 different conditions with a higher level of accuracy. The project returned positive results, showing that artificial intelligence (AI) could be applied to the understanding of MRI data, which can then be used by clinicians to speed up the process of diagnosis. 

The main aim of this two-year research project is to speed up the diagnosis of muscle wasting conditions by enhancing the current version of MYO-Guide – you can read about the previous MYO-guide project here. The new version will include conditions such as congenital myopathies, distal myopathies, myasthenic syndromes and neuropathies.  

As part of this project, Professor Diaz-Manera and colleagues aim to develop a new tool that will help clinicians understand the MRI scans. This tool will be able to identify different muscles and the level of fat replacement.  

This research will bring further improvements to a diagnosis of over 40 muscle wasting conditions, making it quicker and more accurate. Early diagnosis is important for many conditions as early approaches to treating or managing a condition – such as physiotherapy – can help slow down its progression.