Reducing Misdiagnosis in Alzheimer's Disease: AI-Driven Biomarker Identification

University Of California, San Diego, La Jolla CA

Investigators

Abstract

PROJECT SUMMARY /ABSTRACT Alzheimer's disease (AD) is the sixth leading cause of death in the U.S. and the fifth for adults aged 65 and older. Projections suggest the number of affected individuals could reach 13.8 million by 2060 without advancements in diagnostics, prevention, or treatment. AD poses significant challenges in elderly healthcare due to high misdiagnosis rates and limited early intervention options. Diagnostic accuracy largely depends on symptomatic assessment, leading to misdiagnosis in up to 25% of cases. Unfortunately, these misdiagnoses often go unrecognized until postmortem examinations, resulting in inappropriate treatments in 18-67% of cases. These issues create substantial challenges for AD research, including A. Inaccurate diagnostic labels that undermine research reliability, B. The depletion of patient samples over time, which restricts research, and C. Timely intervention is crucial for AD, rather than postmortem discoveries. To address these challenges, reduce misdiagnosis, and facilitate early interventions in Alzheimer's disease (AD), this project introduces innovative technologies and methodologies with three specific aims: Aim 1: Construction of a Comprehensive Multidimensional Database. Utilizing SILVER-seq, a groundbreaking small input liquid volume extracellular RNA sequencing technique, the project will repurpose over 1,500 limited-volume samples that would otherwise be wasted due to their remaining volume being too small. This will generate a multidimensional database spanning 25 years of diverse dementia samples, providing sufficient samples confirmed by the gold standard for subsequent research aims. Aim 2: Identification and Validation of Crucial exRNA Markers for Precision Diagnosis of AD. By employing artificial intelligence (AI) technologies, the project will identify crucial extracellular RNA (exRNA) markers for precise AD diagnosis, significantly reducing misdiagnosis risks. These biomarkers will be integral for developing protocol and computational pipeline. The project aims to establish an AI-driven gene detection process to extend practical applications for early intervention diagnostics based on SILVER-seq technology. The significance of this project lies in its transformative technology and precision diagnosis methods, aiming to revolutionize AD research and healthcare efficiency. The introduction of SILVER-seq and AI-driven marker identification represents unique innovations. In conclusion, this project advances AD diagnosis and early intervention, potentially transforming AD research and healthcare practices. This training will enable me to complete the proposed studies within three years. It will deepen my understanding of AD-related diseases, facilitate data generation using novel techniques, and support the identification and validation of biomarkers. The F32 grant will advance my progress toward becoming an NIH funded independent investigator, focusing on studying multiple diseases through a genetic lens, with bioinformatics as the core skill.