AI project aims to advance personalised MS care

An eight-year £2.72m AI project aims to improve personalised MS care by learning from information already collected in hospitals.

The Wellcome-funded Data-driven Representation Learning for Equitable, Personalised Care in Multiple Sclerosis project, known as DreaMS, will develop AI tools to make better use of decades of existing hospital data.

The project will be led by Dr Arman Eshaghi, who will join King’s College London from UCL to establish the DreaMS Lab within the Institute of Psychiatry, Psychology and Neuroscience’s Department of Biostatistics and Health Informatics.

The laboratory will develop open research tools for scientists and clinicians and lay the foundations for future decision-support systems that could help people with MS receive earlier, fairer and more personalised care.

Dr Eshaghi said: “For far too long we have exclusively relied on high-quality research data to understand diseases that rarely generalise to the clinic, where life-changing decisions are made with patients.

“Enormous amounts of hospital data that can revolutionise our understanding of neurological disorders go underused in the hospital.

“The key to curing many complex and poorly understood diseases, including MS, lies in this data – if we had the tools to analyse it. DreaMS will deliver those tools and tackle centuries-long questions about how MS progresses to eventually cure it one day.”

More than 3m people worldwide live with multiple sclerosis, or MS, a lifelong neurological condition.

MS can affect movement, vision, thinking, fatigue and independence.

The condition affects people differently. Some remain stable for many years, while others develop worsening disability.

Understanding these differences earlier could help doctors make more personalised treatment decisions.

Hospitals already hold large amounts of information about people with MS, including brain and spinal cord MRI scans, radiology reports, doctors’ letters, nurses’ notes and other clinical records.

However, much of this information is difficult to use because it is stored as images or free-text documents rather than in formats that can be easily analysed.

One part of the project will focus on MRI scans.

Software will be designed to analyse routine hospital scans, including older images and those produced using different machines or methods.

The aim is to measure signs of MS-related damage in the brain and spinal cord, including lesions and tissue loss, and track how they change over time.

A second part of the project will focus on clinical text.

Secure AI models will be developed to identify important information in doctors’ letters, nurses’ notes and radiology reports.

These records often describe symptoms, relapses, walking ability, fatigue, cognition, treatment changes and how MS affects daily life.

DreaMS aims to turn this information into useful measures of disease progression.

The project will also combine scans and written records.

By bringing together information from MRI images and clinical notes, the team aims to develop models that could help predict the risk of future disability, identify different patterns of MS, support treatment decisions and assist with MRI reporting.

DreaMS will involve 16 hospitals and research centres across four countries.

Thirteen are in the UK, with one each in the Netherlands, Canada and South Korea.

Through a collaboration with the NIHR Biomedical Research Centre: Maudsley, the project will also expand to several other hospitals in South East England.

Patient data will remain within each hospital.

Instead of transferring identifiable patient records between centres, the AI models will learn from anonymised data within NHS and other secure data environments. Only model updates will be shared.

This approach is designed to protect privacy while allowing the models to learn from a large and diverse group of people with MS.

Mark Richardson, Paul Getty III professor of epilepsy and head of the School of Neuroscience at the Institute of Psychiatry, Psychology and Neuroscience, said: “I am delighted to see Dr Eshaghi moving to King’s, to work with our exceptional electronic health records research infrastructure to improve outcomes for the large number of people living with MS.”

Patient and public involvement and engagement will be a central part of the project.

People living with MS and caregivers have already helped design the research through a patient and public advisory board.

Their input has influenced key parts of the project, including the decision to use nursing notes to better reflect patients’ experiences.

The board will continue to meet regularly to guide the project, provide ethical oversight and ensure transparency over how patient data are used.

Members will also help produce accessible summaries, videos and infographics and share the findings through networks including MS societies.

The partnership is intended to keep DreaMS relevant, inclusive and focused on improving care and outcomes for people with MS.

Dr James Teo, professor of neurology at King’s College Hospital, said: “With the support of the Wellcome, Dr Eshaghi’s work will take forward the frontier of brain disease through combining the latest software and AI technologies with large scale patient data to develop better precision and personalised treatments for our patients.”