Construction of a risk prediction model for Alzheimer's disease based on multi-delay arterial spin labeling
LI Zihan
XU Xian
LI Jinfeng
WANG Yiqing
SUN Xuan
CHEN Siyu
TIAN Peng
JIA Jianjun
Abstract：Objective To develop a risk prediction model for Alzheimer's disease(AD)based on multi-delay arterial spin labeling(ASL)technology.Methods A total of 125 patients who underwent cerebral magnetic resonance imaging,magnetic resonance angiography,and multi-delay ASL in Chinese PLA General Hospital between November 2024 and July 2025 were enrolled in this study.According to scores of neuropsychological scales and results of 11C-Pittsburgh compound B positron emission tomography-computed tomography,the participants were categorized into an AD group(65 cases)and a healthy control group(60 cases).Then,they were randomly divided into a training set(n=87)and a testing set(n=38)at a ratio of 7:3.Demographic data,laboratory test indicators,routine MR indicators,and cerebral blood volume(CBV),cerebral blood flow(CBF),and arterial transit time(ATT)from 268 brain regions were collected,and then multivariate logistic regression and least absolute shrinkage and selection operator(LASSO)regression analyses were performed to screen prediction factors of AD.The prediction models were established in the training set and verified in the testing set.Results Multivariate logistic regression analysis showed that SBP and stenosis of the posterior cerebral artery≥50%were independent risk factors for AD(OR=1.05,95%CI:1.01-1.10,P<0.05;OR=3.77,95%CI:1.02-13.95,P<0.05).LASSO regression analysis indicated that the CBF values in the left superior frontal gyrus and right angular gyrus,and the right superior occipital gyrus,as well as ATT values in the right temporal lobe and left inferior temporal gyrus were identified as effective predictive indicators of AD.ROC curve analysis revealed that the clinical-routine MR-cerebral perfusion model had significantly higher efficacy in predicting AD than the clinical model and the clinical-routine MR model alone,with an AUC value of 0.924(95%CI:0.847-0.969)in the training set and 0.950(95%CI:0.828-0.989)in the testing set.Conclusion Multi-delay ASL provides significant clues for early screening of AD and shows excellent efficacy in AD risk prediction.
Keywords：Alzheimer diseasecerebrovascular circulationarterial spin labelingforecasting
Publication Date：2026-02-15
Online Publishing Date：2026-03-11(First online date of this platform, not the publication date of the document)
Pages：6( 209-214 )