Novel MRI Acquisition and Analysis Methods for Sensitive Detection of Dementia Treatment Side Effects

1) Context of the research including potential impact

Alzheimer's Disease (AD), a neurodegenerative disease, is the leading cause of dementia-and is yet to become more widespread with an ageing population. One potential cause of AD states that the build-up of amyloid-b (Ab) in the brain leads to AD. Hence, many therapeutics target Ab.

However, drug-trials have highlighted that these drugs can lead to a potentially serious side-effect called 'Amyloid Imaging Related Abnormalities' (ARIA). These abnormalities are exclusively visible in MRI scans, but have different appearances based upon the type of ARIA present. Specifically, ARIA-E, which indicates a vasogenic oedema or fluid leakage in the sulci, appears as signal hyperintensities in Fluid Attenuated Inversion Recovery Sequences.

Conversely, ARIA-H, which represents a cerebral microhaemorrhage, causes hypointensities in Gradient Recalled Echo/T2* weighted images.

Even though ARIA is commonly transient and asymptomatic, with the highest risk being at the start of treatment, there have been reported cases of hospitalisation and fatality. Markedly, the aetiology of ARIA is undetermined. However, a postulated cause is the drug-related disruption of blood-brain barrier (BBB) and its loss of integrity-where its disruption can lead to fluid build-up in brain tissues.

The BBB is essential for preserving the microenvironment of the brain and preventing the influx of toxic substances into the brain, but is compromised in the presence of AD.

ARIA poses an obstacle in the development and acceptance of crucial treatments for AD-with clinical trials suspending participants with severe cases of the side-effect. Therefore, there is a need for methods to facilitate early detection of ARIA, and its quantification. 2) Aims and objectives

The principal aims of this project are to create sensitive MRI acquisition and analysis methods to detect the occurrence and progression of ARIA, even in subtle cases. There are two strategies that will be implemented to achieve this. First, novel Magnetic Resonance Fingerprinting (MRF) methods to identify sensitive BBB disruption, which can be used to detect early ARIA will be developed.

Specifically, MRF enables accurate tracking of the pathology via the simultaneous quantification of multiple tissue properties and parameters. Secondly, longitudinal MRI analysis methods will be developed using machine learning to recognise subtle ARIA using conventional MRI and ultra-rapid MRI acquisitions from pre-existing datasets, and those currently being acquired.

3) Novelty of the research methodology

Currently, ARIA poses a challenge in drug development-where there is a relative lack of understanding of the side-effect. The innovations within this work include the introduction of machine learning methods to allow for quick and precise quantification of ARIA using MRF to compute the rate of water exchange across the BBB. Further, the use of machine learning to analyse longitudinal ARIA datasets is yet to be established.

4) Alignment to ESPRC's strategies and research areas

This work complies with the ESPRC's research themes of engineering and healthcare technologies. The project aims to aid in the development of drugs for Alzheimer's Disease via the detection and sensitive quantification of ARIA, falling within the strategic priority of transforming health and healthcare, as well as artificial intelligence and machine learning.

5) Companies or collaborators involved To date this is not applicable.