Multi-Organ Chart of Personalized Susceptibility to Alzheimer's Disease and Aging

Recent progress in clinical trials targeting Alzheimer's Disease (AD), including drugs like Aducanumab, has shown promising research outcomes. However, the effectiveness of these treatments remains to be validated, and robust biomarkers are still needed for early AD diagnosis and prognosis. Evidence suggests that AD and related dementia (ADRD) extend beyond the brain, necessitating multi-scale modeling approaches to understand the underlying etiology and biological mechanisms, provide computational models for precision diagnosis and prognosis, and eventually aid in effective therapeutic strategies. The multi-scale approaches include data, feature representation, artificial intelligence/machine learning (AI/ML), and statistical methodologies. This proposal uses large-scale multi-omics, multi-organ biomedical data, advanced AI/ML, and computational genomics statistical methods to study ADRD and aging. The long-term goal is to construct a Multi-organ Chart (Multi-organ AI-derived Endophenotype called MAE) applicable in clinical settings to personalize patient stratification and treatment strategies, predict disease risks, and select populations for future AD clinical trials. The overarching hypothesis is that the proposed MAEs holistically model cross-organ biological and pathological effects by linking the multi-scale causal pathway of AD: Genetics®Proteomics®Imaging (MAE)®Cognitive decline®Disease endpoint (diagnosis), thus emerging as robust instruments for disease diagnosis and prognosis at an individual level. The rationale for investigating this hypothesis is the multi-dimensional nature of AD, which encompasses various factors and requires comprehensive understanding across different scales of data and computational techniques to improve precision in diagnosis and prognosis. The overarching hypothesis will be tested by pursuing four specific aims: i) consolidate and harmonize a large-scale multi-omics (imaging, proteomics, and genetics), multi-organ (brain, eye, and heart imaging) consortium; ii) depict the phenotypic, proteomic, and genetic association and causality of the imaging-derived phenotypes within the brain-eye-heart connection; iii) construct an AI/ML-derived Multi-organ Chart to quantify individual-level susceptibility to AD and aging, and iv) validate the clinical utility of the proposed Multi-organ Chart to predict systemic disease endpoints, cognitive decline, as well as mortality. We will pursue these aims using an innovative combination of multi-organ, multi-omics biomedical data and multi-resolution AI/ML methodologies, which will be built and advanced from our previously established work. The primary expected outcomes of the proposal are to i) establish a reproducible AI/ML Multi-organ Chart for digitizing personalized susceptibility to AD and aging and ii) promote the adoption of multi-organ, multi-scale methodologies within the AD community through our dedicated knowledge portals, which will be publicly accessible.