What can Artificial Intelligence do for MS Treatment?
According to the UCLH Internet site, its own Biomedical Research Centre has again aided the implementation of artificial intelligence into yet another field of medical healthcare. In this instance, we see the support being targeted towards ‘detecting the brain’s response to treatment in multiple sclerosis (MS) that is substantially better than what a human expert is able to do using conventional techniques, representing potentially ‘superhuman’ performance in the task.’
Before we delve further, multiple sclerosis (MS) is a condition which affects the brain and spinal cord, that will last throughout lifetime and will generally reduce lifespan. It is caused by the immune system attacking the brain and spinal cord, and is an example of an autoimmune condition. To be more precise, the myelin sheath, which surrounds and supports the transmission of the nerves, is attacked. Hence, this leads to damage to the myelin sheath and its underlying nerves, meaning that nerve signals would become slowed or disrupted. As for its symptoms, they are generally disabling, and the main examples include:
vision problems, such as blurred vision
problems controlling the bladder
numbness or tingling in different parts of the body
muscle stiffness and spasms
problems with balance and co-ordination
problems with thinking, learning and planning
As you can imagine, these have a profound effect on the lifestyle of MS victims, who are generally women (up to three times more likely than men) and are generally in their twenties.
Neurologist Dr Parashkev Nachev and Professor Olga Ciccarelli, whom are both UCL researchers of the UCL Institute of Neurology (UCL IoN), have entered a partnership with Kings College London researchers, in an effort to develop a new artificial intelligence-based method to predict a patient’s response to a drug before treatment is issued, which will in turn also help to decide which drug should be administered, hopefully giving the patient the best treatment with the fewest side effects and complications. The work is one of many ongoing collaborative efforts with the School of Biomedical Engineering and Imaging Sciences at King’s College London, where the former UCL team of Professor Seb Ourselin, are helping to extend the input of AI-assisted studies conducted in unison with London hospitals.
A patient’s magnetic resonance imaging (MRI) scan is a common way to assess the response of an individual to MS treatment, currently analysed by radiologists who count the abundance and volumes of lesions, directly comparing with scans that were taken before treatment had started. In this present day case, patients are at risk of severe side effects as the drug must be taken in order to assess its effects. However. the researchers hope that the new AI-based method will be able to assess scans in far more detail, well into each region of the brain, which is a far more suitable method due to the vast complexity of the brain’s structures. In my recent Ophthalmology article, which you can find here, we pondered upon the suitability of deep learning architecture for analysing OCT scans. In a similar way, if the researchers can develop a similar system, and provide it with a very large MRI scan data base, they could easily achieve an incredible effectiveness out of their system, given the potent scaling nature of ‘neural network performance’ and ‘size of data set’.
In fact, a study published by npj Digital Medicine, a renowned research journal published by Springer Nature, researchers who analysed patients at the UCLH National Hospital for Neurology and Neurosurgery with relapsing-remitting MS being treated with the humanised monoclonal antibody drug ‘natalizumab’, which is also used in the treatment of Crohn’s disease. From each scan, an ‘imaging fingerprint’ of the brain state could capture incredibly detailed changes in white and grey matter and a (machine learning optimised) rich set of data into how regions of the brain progress during treatment of time. Researchers coined the phenomenal detail achieved by the scan as ‘Machine vision’. Ultimately, AI-assisted modelling of the complex ‘imaging fingerprints’ showed much greater accuracy in assessing ‘pre-’ and ‘post'-’ treatment changes, as opposed to the traditional analysis of the total lesion and grey matter volume a radiologist can effectively extricate. In the future, it is hoped the approach could guide therapy for patients, assessing the effectiveness of treatment quicker and to be able to carry out drug trials more effectively.
The two honourable persons mentioned at the beginning of the article, namely Dr Parashkev Nachev and Professor Olga Ciccarelli, had the following to say:
Dr Parashkev Nachev; "Rather than attempting to copy what radiologists do perfectly well already, complex computational modelling in neurology is best deployed on tasks human experts cannot do at all: to synthesise a rich multiplicity of clinical and imaging features into a coherent, quantified description of the individual patient as a whole. This allows us to combine the flexibility and finesse of a clinician with the rigour and objectivity of a machine.”
Professor Olga Ciccarelli; "The method is currently focused on imaging changes only; we are extending the approach to predicting the clinical response to disease modifying treatment, in terms of cognitive and motor outcomes. I hope that this exciting field of research will lead to an individual prediction of treatment response in multiple sclerosis using AI.”
This development was funded by the NIHR UCLH Biomedical Research Centre (note that Professor Ciccarelli is in fact a NIHR Research Professor in fact) and the Wellcome Trust, and is associated with UCLH’s Research Hospital Initiative, which hopes to enable modelling methods into real clinical environments.
As always, I would like to drop a sincere sign of gratitude to Dr Parashkev Nachev and Professor Olga Ciccarelli for their efforts, as well as UCLH Internet and npj Digital Medicine. These pioneering efforts will certainly revolutionise healthcare in the coming years, thank you very much.