Machine Learning for the Prediction of Amyloid Positivity in Amnestic Mild Cognitive Impairment
- Authors
- Kang, Sung Hoon; Cheon, Bo Kyoung; Kim, Ji-Sun; Jang, Hyemin; Kim, Hee Jin; Park, Kyung Won; Noh, Young; San Lee, Jin; Ye, Byoung Seok; Na, Duk L.; Lee, Hyejoo; Seo, Sang Won
- Issue Date
- 2021
- Publisher
- IOS PRESS
- Keywords
- A beta PET; amnestic mild cognitive impairment; A beta positivity; machine learning; magnetic resonance imaging features; neuropsychological tests; prediction model
- Citation
- JOURNAL OF ALZHEIMERS DISEASE, v.80, no.1, pp.143 - 157
- Indexed
- SCIE
SCOPUS
- Journal Title
- JOURNAL OF ALZHEIMERS DISEASE
- Volume
- 80
- Number
- 1
- Start Page
- 143
- End Page
- 157
- URI
- https://scholar.korea.ac.kr/handle/2021.sw.korea/130257
- DOI
- 10.3233/JAD-201092
- ISSN
- 1387-2877
- Abstract
- Background: Amyloid-beta (A beta) evaluation in amnestic mild cognitive impairment (aMCI) patients is important for predicting conversion to Alzheimer's disease. However, A beta evaluation through A beta positron emission tomography (PET) is limited due to high cost and safety issues. Objective: We therefore aimed to develop and validate prediction models of A beta positivity foraMCIusing optimal interpretable machine learning (ML) approaches utilizing multimodal markers. Methods: We recruited 529 aMCI patients from multiple centers who underwent A beta PET. We trained ML algorithms using a training cohort (324 aMCI from Samsung medical center) with two-phase modelling: model 1 included age, gender, education, diabetes, hypertension, apolipoprotein E genotype, and neuropsychological test scores; model 2 included the same variables as model 1 with additional MRI features. We used four-fold cross-validation during the modelling and evaluated the models on an external validation cohort (187 aMCI from the other centers). Results: Model 1 showed good accuracy (area under the receiver operating characteristic curve [AUROC] 0.837) in crossvalidation, and fair accuracy (AUROC0.765) in external validation. Model 2 led to improvement in the prediction performance with good accuracy (AUROC 0.892) in cross validation compared to model 1. Apolipoprotein E genotype, delayed recall task scores, and interaction between cortical thickness in the temporal region and hippocampal volume were the most important predictors of A beta positivity. Conclusion: Our results suggest that ML models are effective in predicting A beta positivity at the individual level and could help the biomarker-guided diagnosis of prodromal AD.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - College of Medicine > Department of Medical Science > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.