Abstract 14438: Longitudinal Real World Study of Blood Pressure and Cerebral Arterial Stenosis on Elderly Cognitive Impairment Based on the Artificial Intelligence Analysis of Serial Brain Magnetic Resonance Angiography

Abstract

Background: Abnormal blood pressure (BP) is one of the most common modified risk factors for the development of cognitive impairment (CI) with increasing aging population. Objective: This study aims to investigate the association between risk factors and CI in elderly patients based on artificial intelligence (AI) analysis of BP measurements and brain magnetic resonance angiography (MRA). Methods: A total of 3221 elderly individuals undergone brain MRA were recruited into the nested case-control study from 2007 to 2021. Patients were divided into two groups according to diagnosis with or without CI. Data were collected before diagnosing CI. The stenosis rate of cerebral arteries was calculated by AI analysis of brain MRA. Four machine learning modalities were applied to establish predictive models. Results: A total of 416 cases (mean age 84.47 ± 6.50 years old) with or without CI were finally recruited into the study after matching with age and gender. Mean follow-up time was 3.46 ± 3.19 years. The number of BP measurements was 56.98 ± 77.94 times. In the univariate logistic regression model, high systolic blood pressure (SBP) (OR 1.029, 95% CIs 1.010-1.049), pulse pressure (PP) (OR 1.043, 95% CIs: 1.021-1.065) and pulse pressure index (PPI) (OR 3588.920, 95% CIs: 44.392-290148.336) increased the risk of CI. Based on various mean BP parameters, the predictive ability of the support vector machine (SVM) was improved after adding cerebral arterial stenosis rate (AUC 0.781, 95% CIs 0.639~0.923). Conclusion: Abnormal SBP, PP and PPI were risk factors for CI in the elderly Chinese population. Adding the cerebral arterial stenosis rate of internal carotid artery and posterior cerebral artery the predictive ability of SVM model of various mean BP parameters for the risk of CI was improved.