MacCarthy MD, Petrella JR, Sheldon FC, Shaffer JL, Doraiswamy PM, Calhoun VC.
Multi-modality fusion of neuroimaging and genetic data in predicting abnormal cognitive decline in aging.Human Amyloid Imaging 2011 Meeting Abstracts.
2011 Jan 15;
ABSTRACT
Recently, a novel data fusion method developed by Calhoun et al (2009) using parallel Independent Component
Analysis (Fusion ICA Toolbox, FIT Version 2.0b) has shown potential in discovering disease-contributing
characteristics by encompassing whole-brain image analysis and incorporating multiple data types into a single
model. To examine characteristics that may predict a decline from normal aging to MCI, we obtained data from
the Alzheimer's Disease Neuroimaging Initiative (ADNI) database and used the Fusion ICA Toolbox to analyze
the MRI, FDG-PET, and SNP genetic data of 103 normal study participants (M/F: 63/40, mean (sd), age: 75.8
(4.73), baseline ADAS-Cog: 10.4 (4.24), baseline MMSE: 29.0 (1.12)). CSF abeta and tau data are available only
in a subset, but APOE4 genotyping, available in all subjects, will be used as a surrogate for abeta levels. 27 of
these patients showed cognitive decline based on the CDR sum of boxes score at either 36- or 48-month followup
visits. Data analysis is ongoing and the multi-modality data will be fused to create a model for determining
potential clinically relevant characteristics. By comparing the imaging and genetic data of normal controls to those
who experienced cognitive decline, we anticipate developing sensitivity and specificity measures for using the
above modalities together in predicting progression from normal aging to pathologic impairment. This method
may aid in the early detection of preclinical stages of MCI and AD, and may enable the most effective use of
disease-modifying therapeutics designed to stop or slow the progression of the disease. The results of our work
in progress will be presented at the meeting.
