Meet the PhD startup using AI to prevent worsening symptoms in multiple sclerosis
When PhD student Joe Sims witnessed his aunt being left without immediate access to treatment options that might slow the progression of her dementia symptoms following a misdiagnosis, he realised AI could help reduce similar errors in the future by supporting clinicians’ decisions.
“Current AI solutions for the diagnosis of a range of neurological disorders rely on MRI scans alone, which don’t often give you a full understanding of the patient’s state — some patients suffering from multiple sclerosis, for instance, could show a lot of lesions and few symptoms, while others might have no lesions and many symptoms,” Sims says.
“But what if you could create an AI that looks at more than just images, the way neurologists use multiple data sources to diagnose their patients?”
After teaming up with University of Leeds colleague Sonia Dembowska, the two launched Theia AI to help reduce human error in diagnosis, with a focus on slowing the progression of symptoms in multiple sclerosis patients.
The startup aims to build a multimodal AI-powered tool that combines historical patient data with scans to identify subtle patterns, enabling clinicians to predict disease progression and provide personalised treatment recommendations.
“In the UK, multiple sclerosis costs the NHS almost £3 billion a year. From the point of diagnosis, patients are monitored for years with MRI scans every 12–32 months to watch out for symptoms worsening,” Dembowska says.
“We want to take this data over time to understand how the brain changes over the course of the disease and predict progression, so we can intervene earlier to avoid complications.”
The team’s AI would behave like a virtual clinician that monitors patient data to check for signs of disease development, such as changes in lesions or worsening symptoms. These insights could then be used by neurologists to assess the risk of relapse for patients and adjust treatment options.
The two co-founders joined Conception X Cohort 5 to explore alternative career options and learn how to turn their PhD research in medical imaging into a medtech startup that could enhance clinicians’ understanding of neurological conditions.
“We didn’t know what the process of going from idea to a fully fledged company looked like, and Conception X made it feel very realistic in terms of all the steps we need to take to get there,” Dembowska says.
“We wanted to understand how you could have more independence in setting your priorities while still retaining scientific integrity, and we met a lot of people with similar ambitions — realising we do fit into a category.”
By the end of the programme, Theia AI received £100k in funding through Conception X’s partnership with XTX Ventures, after being selected as one of the top performing early-stage companies in Cohort 5 building breakthrough technologies.
The startup is now focusing on expanding its network to develop its product around neurologists’ existing workflows and needs, and is looking to bring a clinician on board to support them in building an MVP.