TY - JOUR
T1 - Deep learning combining FDG-PET and neurocognitive data accurately predicts MCI conversion to Alzheimer's dementia 3-year post MCI diagnosis
AU - Cao, Eric
AU - Ma, Da
AU - Nayak, Siddharth
AU - Duong, Tim Q.
N1 - Publisher Copyright:
© 2023 The Authors
PY - 2023/10/15
Y1 - 2023/10/15
N2 - Introduction: This study reports a novel deep learning approach to predict mild cognitive impairment (MCI) conversion to Alzheimer's dementia (AD) within three years using whole-brain fluorodeoxyglucose (FDG) positron emission tomography (PET) and cognitive scores (CS). Methods: This analysis consisted of 150 normal controls (CN), 257 MCI, and 205 AD subjects from ADNI. FDG-PET and CS were obtained at MCI diagnosis to predict AD conversion within three years of MCI diagnosis using convolutional neural networks. Results: Neurocognitive scores predicted better than FDG-PET per se, but the best model was a combination of FDG-PET, age, and neurocognitive data, yielding an AUC of 0.785 ± 0.096 and a balanced accuracy of 0.733 ± 0.098. Saliency maps highlighted putamen, thalamus, inferior frontal gyrus, parietal operculum, precuneus cortices, calcarine cortices, temporal gyrus, and planum temporale to be important for prediction. Discussion: Deep learning accurately predicts MCI conversion to AD and provides neural correlates of brain regions associated with AD conversion.
AB - Introduction: This study reports a novel deep learning approach to predict mild cognitive impairment (MCI) conversion to Alzheimer's dementia (AD) within three years using whole-brain fluorodeoxyglucose (FDG) positron emission tomography (PET) and cognitive scores (CS). Methods: This analysis consisted of 150 normal controls (CN), 257 MCI, and 205 AD subjects from ADNI. FDG-PET and CS were obtained at MCI diagnosis to predict AD conversion within three years of MCI diagnosis using convolutional neural networks. Results: Neurocognitive scores predicted better than FDG-PET per se, but the best model was a combination of FDG-PET, age, and neurocognitive data, yielding an AUC of 0.785 ± 0.096 and a balanced accuracy of 0.733 ± 0.098. Saliency maps highlighted putamen, thalamus, inferior frontal gyrus, parietal operculum, precuneus cortices, calcarine cortices, temporal gyrus, and planum temporale to be important for prediction. Discussion: Deep learning accurately predicts MCI conversion to AD and provides neural correlates of brain regions associated with AD conversion.
KW - Artificial intelligence
KW - Dementia
KW - MRI
KW - Machine learning
KW - Mild cognitive impairment
KW - Positron emission tomography
UR - http://www.scopus.com/inward/record.url?scp=85172919899&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85172919899&partnerID=8YFLogxK
U2 - 10.1016/j.nbd.2023.106310
DO - 10.1016/j.nbd.2023.106310
M3 - Article
C2 - 37769746
AN - SCOPUS:85172919899
SN - 0969-9961
VL - 187
JO - Neurobiology of Disease
JF - Neurobiology of Disease
M1 - 106310
ER -