基于磁共振成像特征的血管性认知障碍预测模型构建分析  

Construction and analysis of predictive model of vascular cognitive impairment based on magnetic resonance imaging features

作　　者：张舞青 全勇[2] 王正印[1] 邓会娟 李晓红[3] Zhang Wuqing;Quan Yong;Wang Zhengyin;Deng Huijuan;Li Xiaohong(Department of Neurology,Shan County Central Hospital,Heze 274300,China;Department of Imaging,Shan County Central Hospital,Heze 274300,China;Department of Neurology,Jinan Central Hospital affiliated to Shandong University,Jinan 250013,China)

机构地区：[1]单县中心医院神经内科,山东菏泽274300 [2]单县中心医院影像科,山东菏泽274300 [3]山东大学附属济南市中心医院神经内科,济南250013

出　　处：《中国临床实用医学》2021年第6期22-25,共4页China Clinical Practical Medicine

基　　金：医药卫生科技发展计划项目(S2017ASF026)。

摘　　要：目的构建基于磁共振成像(MRI)特征的血管性认知障碍(VCI)预测模型,并验证其预测效能。方法对2018年1月至2021年1月单县中心医院神经内科收治的524例缺血性脑卒中患者进行随访,最终将发生VCI的43例患者纳入发生组,男20例,女23例,年龄(63.21±8.45)岁,年龄范围为50~80岁。采用最临近匹配法纳入43例未发生VCI的患者为未发生组,男22例,女21例,年龄(62.38±8.47)岁,年龄范围为50~80岁。所有患者均接受MRI成像,采用logistic回归方程分析MRI特征与缺血性脑卒中患者VCI的相关性,并构建基于MRI特征的预测模型,分析其预测VCI的价值。结果发生组与未发生组标准化吸收值比率(SUVR)、海马体积、皮层梗死、弥漫性白质高信号(WMH)体积、血管腔隙扩大组间比较,差异有统计学意义(P<0.05);颅内体积、脑微出血比较,差异无统计学意义(P>0.05)。多因素logistic分析显示MRI特征SUVR(OR=8.254)、海马体积(OR=0.222)、WMH体积(OR=4.380)、血管腔隙扩大(OR=7.415)与缺血性脑卒中患者VCI具有相关性(P<0.05)。列线图模型显示基于MRI特征的VCI预测模型预测缺血性脑卒中患者VCI发生的C-index为0.897,校正曲线显示列线图模型预测可能性绝对误差为0.040。结论MRI特征SUVR、海马体积、WMH体积、血管腔隙扩大与缺血性脑卒中患者VCI发生具有相关性,基于此建立的预测模型能为临床医师识别缺血性脑卒中患者VCI发生的高危患者提供指导。Objective To construct a prediction model of vascular cognitive impairment(VCI)based on the characteristics of magnetic resonance imaging(MRI),and to verify its prediction performance.Methods From January 2018 to January 2018,a total of 524 patients with ischemic stroke admitted to the department of Neurology of Shan County Central Hospital were followed up,43 patients with VCI were included in the group,20 males and 23 females,aged(63.21±8.45)years old,and the age range was 50 to 80 years old.The 43 non-VCI patients included by nearest matching method were the non-VCI Group,the non-VCI Group was 22 males and 21 females,the age was(62.38±8.47)years old,and the age range was 50 to 80 years old.All patients underwent structural MRI.The logistic regression equation was used to analyze the correlation between MRI features and VCI in patients with ischemic stroke,and a prediction model based on MRI features was constructed to analyze its value in predicting VCI.Results There were statistically significant differences in sandardized uptake value ratio(SUVR),hippocampal volume,cortical infarction,diffuse white matter hyperintensity(WMH)volume,and vascular space enlargement between the occurrence group and the non-occurring group(P<0.05);There was no significant difference in internal volume and cerebral microhemorrhage(P>0.05).Multivariate logistic analysis showed that MRI features of SUVR(OR=8.254),hippocampus volume(OR=0.222),WMH volume(OR=4.380),vascular space enlargement(OR=7.415)are correlated with VCI in patients with ischemic stroke(P<0.05).The nomogram model shows that the C-index of the VCI prediction model based on MRI features to predict the occurrence of VCI in patients with ischemic stroke is 0.897,and the calibration curve shows that the absolute error of the prediction probability of the nomogram model is 0.040.Conclusion MRI features of SUVR,hippocampus volume,WMH volume,and vascular space enlargement are related to the occurrence of VCI in patients with ischemic stroke.The predictive model established

关 键 词：磁共振成像 血管性认知障碍 缺血性脑卒中 预测模型 

分 类 号：R445.2[医药卫生—影像医学与核医学] R749.13[医药卫生—诊断学]