Development of a Calculator for Alzheimer's Disease Risk

 
   Lead Investigator:    Shanna Cooper
   Institution      :    UC San Diego
   E-Mail           :    stcooper@ucsd.edu
   Proposal ID      :    1132



   Proposal Description:
   Given the growing aging population and related increase in proportion of older   adults diagnosed with Alzheimer?s disease (AD), there is considerable           interest in developing preventative approaches. Accurate identification of        people at greatest risk is critical to these efforts. The scientific literature   has identified many factors conferring risk for AD at group levels however,       little is known of the application of these risk factors in determining overall   risk of developing AD at an individual level. Individualized risk calculation     of AD is possible when a large reference data set is available from which risks   can be calculated. Applying machine learning techniques to the large National     Alzheimer?s Coordinating Center (NACC) database, we propose a precision         medicine approach to calculate risk where individual variability across risk      metrics are considered to allow for a multifactorial, data-driven,                individualized approach for detection of AD in those at risk for disease          progression.                                                                      Specific aims of the proposed project                                             The NACC dataset offers an ideal breadth and depth of information on which        machine learning techniques can easily be applied to determine individualized     risk of AD. As such, we aim to 1.) Apply machine learning techniques to build a   predictive model that includes the most robust risk factors across numerous       available metrics (e.g., cognitive, medical/biological, social/lifestyle), with   both individual and combinatorial impact, that increase risk for AD. 2.)          Evaluate the performance of our risk prediction model on a separate               cross-validation sample. 3.) Translate our risk calculator to an easy-to-use      open source online platform and/or mobile application for researchers.            Innovation/novelty of the proposed p