Deep learning combining FDG-PET and neurocognitive data accurately predicts MCI conversion to Alzheimer's dementia 3-year post MCI diagnosis

Eric Cao; Da Ma; Siddharth Nayak; Tim Q Duong

doi:10.1016/j.nbd.2023.106310

Deep learning combining FDG-PET and neurocognitive data accurately predicts MCI conversion to Alzheimer's dementia 3-year post MCI diagnosis

Neurobiol Dis. 2023 Oct 15:187:106310. doi: 10.1016/j.nbd.2023.106310. Epub 2023 Sep 26.

Authors

Eric Cao¹, Da Ma², Siddharth Nayak³, Tim Q Duong⁴

Affiliations

¹ Department of Radiology, Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, NY 10467, United States.
² Department of Internal Medicine Section of Gerontology and Geriatric Medicine, Wake Forest, University School of Medicine, Winston-Salam, NC 27109, United States.
³ Department of Radiology, Weill Cornell Medicine, New York, 10065, United States.
⁴ Department of Radiology, Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, NY 10467, United States. Electronic address: Tim.duong@einsteinmed.edu.

PMID: 37769746
DOI: 10.1016/j.nbd.2023.106310

Abstract

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.

Keywords: Artificial intelligence; Dementia; MRI; Machine learning; Mild cognitive impairment; Positron emission tomography.