Grant: Project Grant

Late-onset Alzheimer's disease (late-AD) is a complex neurodegenerative condition with abnormalities in protein accumulation (amyloid, tau), vascularization, inflammation and/or cerebral atrophy. Numerous clinical trials targeting only one of these components have failed to achieve significant clinical benefits, while the treatments have not accounted for the disease's clinical and pathophysiological heterogeneity. In keeping with the tenets of Personalized Medicine, treatments need to be tailored to individual therapeutic needs as opposed to treating all patients with the same approach. In neurodegeneration, however, we currently do not have appropriate methods for pathophysiological patient staging and stratification. Furthermore, because we continue to lack robust tools to characterize each patient's distinctive disease progression, evaluating the effectiveness of different therapies remains challenging. Prompted by the imperative for customized treatment of neurodegeneration, our goal is the implementation and validation of a biologically-defined staging and stratification of the AD spectrum using an artificial intelligence (AI) framework to integrate molecular, biological and neuroimaging data. Consequently, we are integrating a large amount of biological data (including genomics, epigenomics, transcriptomics, neuroimaging) into a simplified and clinician-friendly personalized clinical profile reflecting both disease progression and variant. This framework aims to bridge the Molecular, Neuroimaging and AI fields in order to overcome the major limitations of traditional clinical diagnoses. We will validate its clinical applicability with five different populations in the AD spectrum. Critically, all resulting analytic tools will be shared with the international community via a user-friendly software and popular web-based code repositories, contributing to translate our research findings in daily clinical practice.