Neuroimaging and neurochemical ageing biomarkers for optimising prognosis in motor neurone disease

1) Brief description of the context of the research including potential impact

Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease with no known cure. The time between disease onset and the end stages of disease vary widely from patient to patient. Patients with ALS endure rapid widespread brain tissue loss and this process typically progresses quicker in elder people.

ALS is therefore thought to interact with the ageing process, which by itself usually causes gradual loss of brain tissue. It may be the case that individuals with larger amounts of age-related tissue loss may have a more fatal (shorter survival time) prognosis if they develop neurodegenerative diseases like ALS. To this end, machine learning algorithms in tandem with magnetic resonance imaging (MRI) data, have been used to predict disease progression in neurodegenerative diseases.

This is done through the 'brain age' index, which is a measure that is informative for brain health. Additionally, neurofilaments data taken from blood samples, are informative for brain atrophy. In health, neurofilaments typically reside in the cytoplasm of neurons, and are released into the blood after neurons die.

As individuals age and normal brain tissue loss occurs the levels of neurofilaments in blood also increases. Both MRI and neurofilaments data have been useful for predicting outcomes in ALS but have not previously been used together. The present research aims to combine data from both modalities in-order to develop techniques which aid ALS prognostication.

This will be done through the application of computational methods, with a particular focus on machine learning. These statistical/machine learning models will be sensitive to accelerated age-related brain tissue loss as well as increased neurofilaments in the blood. Ultimately, such models will be able to recognise the rate at which ALS will progress in an individual patient.

This will increase the speed and cost effectiveness of clinical trials for ALS interventions, as well as enabling targeted treatments. 2) Aims and Objectives -The specific objectives are to:

- Develop multimodal (MRI and blood) machine learning and other statistical models that are sensitive to accelerated age-related changes to the brain. - Use these models to correctly classify whether ALS will progress quickly or slowly within an individual.

- Apply these approaches to aid in the development of targeted treatments and more efficient clinical trials for ALS drug treatments. 3) Novelty of Research Methodology

Our research methodology is novel primarily due to the multi-modal nature of the machine learning/statistical models that will be developed. MRI and blood samples data have independently shown promise in predicting ALS prognosis, but their combination should provide even greater power to predictive models. Also, the use of 'brain age' as well as, machine learning techniques in aiding prognostication of neurodegenerative diseases is a recent development.

Therefore, this project will be able to unlock more of the potential these techniques have already shown. 4) Alignment to EPSRC's strategies and research areas

This project is aligned with the EPSRC's healthcare technology strategy. Within that, the project is aligned with the medical imaging, clinical technologies and analytic science research areas. 5) Any companies or collaborators involved N/A