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

1)研究背景包括潜在的影响阿尔茨海默病（AD），一种神经退行性疾病，是痴呆症的主要原因，随着人口老龄化，这种疾病还没有变得更加普遍。AD的一个潜在原因是大脑中淀粉样蛋白b（Ab）的积累导致AD。因此，许多治疗剂靶向Ab。然而，药物试验已经强调，这些药物可能导致一种潜在的严重副作用，称为“淀粉样蛋白成像相关抑制剂”（ARIA）。这些异常仅在MRI扫描中可见，但根据ARIA的类型有不同的表现。特别是ARIA-E，它表明脑沟中的血管源性水肿或液体渗漏，在液体衰减反转恢复序列中表现为信号高强度。相反，ARIA-H，代表脑微出血，在梯度回波/T2* 加权图像中引起低信号。尽管ARIA通常是短暂的和无症状的，最高风险是在治疗开始时，但已有住院和死亡病例的报道。值得注意的是，ARIA的病因尚未确定。然而，一个假设的原因是药物相关的血脑屏障（BBB）的破坏和其完整性的丧失，其破坏可导致脑组织中的液体积聚。血脑屏障对于保护大脑的微环境和防止有毒物质流入大脑至关重要，但在AD的存在下会受到损害。ARIA对AD的关键治疗的开发和接受构成了障碍-临床试验暂停严重副作用的参与者。因此，需要一种方法来促进ARIA的早期检测及其定量。2）目的和目标本项目的主要目的是创建灵敏的MRI采集和分析方法，以检测ARIA的发生和进展，即使是在轻微的情况下。为实现这一目标，将实施两项战略。首先，将开发新的磁共振指纹（MRF）方法来识别敏感的BBB破坏，可用于检测早期ARIA。具体而言，MRF能够通过同时量化多个组织特性和参数来精确跟踪病理。其次，将使用机器学习开发纵向MRI分析方法，以使用常规MRI和超快速MRI采集从现有数据集和当前正在采集的数据集识别细微的ARIA。3）研究方法的新奇目前，ARIA对药物开发构成了挑战-对副作用的了解相对缺乏。这项工作中的创新包括引入机器学习方法，以允许使用MRF快速准确地量化ARIA，以计算整个BBB的水交换率。此外，使用机器学习来分析纵向ARIA数据集还有待建立。4）与ESPRC的战略和研究领域保持一致这项工作符合ESPRC的工程和医疗技术研究主题。该项目旨在通过ARIA的检测和灵敏定量来帮助开发阿尔茨海默病药物，属于改变健康和医疗保健以及人工智能和机器学习的战略优先事项。

1) Context of the research including potential impactAlzheimer'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 objectivesThe 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 methodologyCurrently, 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 areasThis 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 involvedTo date this is not applicable.