Longitudinal real world correlation study of blood pressure and novel features of cerebral magnetic resonance angiography by artificial intelligence analysis on elderly cognitive impairment

• DOI: 10.3389/fnagi.2023.1121152
• Publication Date: 2023-02-03T04:42:22Z
• Authors: Shasha Sun; Dongyue Liu; Yanfeng Zhou; Ge Yang; Long-Biao Cui; Xian Xu; Yuanhao Guo; Ting Sun; Jiacheng Jiang; Na Li; Yabin Wang; Sulei Li; Xinjiang Wang; Li Fan; Feng Cao
• Keywords: Magnetic resonance angiography; Posterior cerebral artery

This study aims to investigate novel clinical risk factors for cognitive impairment (CI) in elderly.A total of 3221 patients (259 with CI and 2,962 subjects without CI) were recruited into this nested case-control who underwent cerebral magnetic resonance angiography (MRA) from 2007 2021. All the data MRA imaging recorded followed by standardization processing blindly. The maximum stenosis score posterior circulatory artery, including basilar bilateral artery (PCA), was calculated automatic quantitative analysis method. Logistic regression (LR) used evaluate relationship between CI. Four machine learning approaches, LR, decision tree (DT), random forest (RF), support vector (SVM), employing 5-fold cross-validation establish predictive models.After matching age gender, 208 control finalized follow-up (3.46 ± 3.19 years) mean at 84.47 6.50 years old. Pulse pressure (PP) first tertile (<58 mmHg) (OR 0.588, 95% confidence interval (CI): 0.362-0.955) associated a decreased CI, ≥50% left PCA 2.854, CI: 1.387-5.872) an increased after adjusting body mass index, myocardial infarction, stroke history. Based on means various blood (BP) parameters, performance DT, RF SVM models accurately predicted (AUC 0.740, 0.786, 0.762, 0.753, respectively) adding artery.Elderly low pulse differential may have lower impairment. hybrid model combined indicators, BP parameters can effectively predict elderly individuals.