Dynamic Predictions in  Functional Joint Models for Longitudinal and

 
   Lead Investigator:    Kan Li
   Institution      :    Univ. of Texas Health Science Center at Houston
   E-Mail           :    kan.li@uth.tmc.edu
   Proposal ID      :    887



   Proposal Description:
   The primary objective of our study is to develop a series of novel statistical    models to account for serial MRI (image/functional predictor), CSF markers, and   neuropsychological assessments to measure the progression of AD using the joint   modeling of longitudinal-survival data. In this framework, we can make            individualized dynamic prediction for patient's future health outcome and risk    of AD. Thus, earlier diagnoses can be made to subjects with high predicted risk   of deterioration, and intervention can be planned in a timely manner to           delaying the manifestation of AD in prodromal AD patients. Our preliminary        analyses were applied to ADNI data set and found that including image data,       using surface based morphology information instead of ROI volumes, boosts the     power of the model to predict AD conversion. Further model development and        external validation with an independent data set is needed to verify the          preliminary results.