Premium
P1‐399: SYNERGISTIC EFFECTS OF APOE ε4 AND WORRY ON CORTICAL MORPHOLOGY IN SUBJECTIVE COGNITIVE DECLINE
Author(s) -
Sun Yu,
Zuo Xinian,
Han Ying
Publication year - 2019
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2019.06.1004
Subject(s) - worry , apolipoprotein e , psychology , population , brain morphometry , cognition , neurology , medicine , neuroscience , psychiatry , magnetic resonance imaging , anxiety , radiology , environmental health , disease
employed automated analysis techniques, FreeSurfer and SPM whole-brain voxel-based analysis, were used to process neuroimaging (MRI and [18F]FDG and [18F]Florbetapir PET) data. In our novel Multimodal-3DCNN approach (Fig. 1), we first applied 3D Convolutional Neural Network (3D-CNN) to multimodal neuroimaging (MRI and PET) and then combined the output of 3DCNN with APOE ε4 genotype and demographic information (age, sex, education, handedness etc.) using a gram matrix method (mCNN; Jo et al. AAIC2018). Finally, Deep Neural Network (DNN) was used to distinguish individuals with AD from CN. A 5-fold cross validation approach was employed to evaluate performance. Results: The proposed classification model for AD yielded 78% accuracy (area under the curve (AUC)1⁄40.87) when using only MRI as well as APOE ε4 and demographic information. The model was improved to 87% accuracy (AUC1⁄40.93) when using MRI and [18F]Florbetapir PET as well as APOE ε4 and demographic information. The best performance with 94% accuracy (AUC1⁄40.95) was achieved by combining MRI, [18F]FDG PET, and [18F]Florbetapir PET as well as APOE ε4 and demographic information (Table 1). Conclusions: Our multimodal-3DCNN deep learning approach for diagnostic classification of AD by combining multimodal neuroimaging, genetic, and demographic data yielded a high classification accuracy. Our proposed framework has potential to identify individuals with AD to support trial enrichment and therapeutic development. Future analyses will examine prediction of disease progression and the role of additional genetic markers.