【摘要】 目的 探究脑白质变性（WML）患者认知障碍的相关因素，构建列线图模型。方法 选取2022-04—2024-04承德市中心医院收治的168例WML患者为研究对象，根据是否发生认知障碍分为认知障碍组和非认知障碍组，比较2组患者一般资料、实验室指标，采用方差膨胀因子、容忍值及Logistic回归方程筛选认知障碍发生的相关因素，R软件构建认知障碍发生列线图模型，校准曲线、受试者工作特征（ROC）曲线、决策曲线（DCA）、临床影响曲线（CIC）分析列线图模型的诊断效能。结果 所有对象的认知障碍发生率48.80%。年龄（OR=2.697）、高血压（OR=2.609）、糖尿病（OR=1.836）、血清淀粉样蛋白P(SAP)(OR=1.491)、同型半胱氨酸（Hcy,OR=1.965）、可溶性低密度脂蛋白相关受体蛋白1(sLRP-1,OR=0.780)、基质金属蛋白酶-9(MMP-9,OR=1.713)、WML严重程度（OR=2.469）是WML患者发生认知障碍的影响因素（P<0.05）。以Logistic回归方程得到的8个影响因素构建WML患者发生认知障碍的列线图模型，曲线下面积（AUC）为0.913，诊断概率接近实际概率。在0～0.8范围内，列线图模型诊断WML患者发生认知障碍的临床获益值最大。当风险阈值>0.8时，列线图模型诊断患病数接近实际患病数。结论 基于WML患者认知障碍相关因素构建列线图具有良好的诊断效能，有助于指导临床，降低认知障碍发生风险。更多还原

【Abstract】 Objective To investigate the factors related to cognitive impairment in patients with white matter lesions(WML) and construct Nomogram model.Methods A total of 168 patients with WML admitted to Chengde Central Hospital from April 2022 to April 2024 were selected as the study subjects.They were divided into cognitive impairment group and non-cognitive impairment group based on the presence or absence of cognitive impairment.The general information and laboratory indicators of the two groups were compared.Factors related to the occurrence of cognitive impairment were screened using variance inflation factor,tolerance value,and Logistic regression equation.R software was used to construct Nomogram model for the occurrence of cognitive impairment.Calibration curve,receiver operating characteristic(ROC) curve,decision curve analysis(DC A),and clinical impact curve(CIC) were used to analyze the diagnostic efficacy of the Nomogram model.Results The incidence of cognitive impairment in all subjects was 48.80%.Age(OR=2.697),hypertension(OR=2.609),diabetes(OR=1.836),serum amyloid P(SAP)(OR=1.491),homocysteine(Hcy,OR=1.965),soluble low-density lipoprotein receptor-related protein 1(sLRP-1,OR=0.780),matrix metalloproteinase-9(MMP-9,OR=1.713),and severity of WML(OR=2.469) were influencing factors for cognitive impairment in WML patients(P<0.05).Using the eight influencing factors obtained from the Logistic regression equation,Nomogram model was constructed for the occurrence of cognitive impairment in patients with WML,with the area under the curve(AUC) of 0.913,indicating that the diagnostic probability was close to the actual probability.Within the range of0 to 0.8,the Nomogram model had the greatest clinical benefit value for diagnosing cognitive impairment in WML patients.When the risk threshold was>0.8,the number of patients diagnosed by the Nomogram model approached the actual number of patients with the disease.Conclusion Constructing Nomogram based on factors related to cognitive impairment in patients with WML has good diagnostic efficacy,which is helpful in guiding clinical practice and reducing the risk of cognitive impairment.更多还原