PREDICTION OF ALZHEIMER’S DISEASE USING AN AI DRIVEN SCREENING PLATFORM

点击翻译按钮获取中文摘要

Alzheimer’s disease (AD) and related disorders leading to dementia are associated with staggering costs and suffering. Recently, there has been some progress in the search for effective therapeutic interventions and it is clear that any treatment is likely to be most effective if administered at the earliest stage of disease, but the health care system is not ready for this new scenario. There is an urgent need, therefore, to establish scalable, cost-efficient diagnostic markers, tools and procedures that can identify people at increased risk, at point of care for stratification into personalized interventions to prevent or delay dementia. PREDICTOM will develop an open-source, interoperable and customisable biomarker screening platform, utilizing an existing online resource to save time and money, to generate an evidence base for general population screening for AD and related disorders. We will bring diagnostics closer to the patient by examining the feasibility of using samples which can be obtained at home (e.g. finger-prick blood, saliva (for genetics and epigenetics) and stool for microbiom) for diagnostic biomarker analysis. We will also evaluate innovative technologies for disease risk identification, including digital technologies and novel MRI, EEG, eye tracking, and blood-based biomarkers. The platform will use artificial intelligence models to analyse data from all biomarkers to identify users at high risk of developing dementia and to direct them to personalized intervention to prevent further cognitive decline and development of dementia. We will seek to facilitate a change in current healthcare practice for early diagnosis of AD through development of new clinical practice guidelines based on evidence generated in the project. By improving the ease of identification of those with early signs of dementia we expect to have a significant impact on personal and financial burden of dementia in Europe and across the world.