分水岭脑梗死患者复发风险列线图模型的构建与验证  

Construction and validation of a nomogram model for predicting recurrence risk in patients with watershed infarction

作　　者：陈轲扬[1] 姜乐临 费凡 曹云刚[1] 陈松芳[1] CHEN Keyang;JIANG Lelin;FEI Fan;CAO Yungang;CHEN Songfang(Department of Neurology,the Second Affiliated Hospital of Wenzhou Medical University,Wenzhou 325015,China;The Second School of Medicine,Wenzhou Medical University,Wenzhou 325035,China)

机构地区：[1]温州医科大学附属第二医院神经内科,浙江温州325027 [2]温州医科大学第二临床医学院,浙江温州325035

出　　处：《温州医科大学学报》2025年第2期105-111,共7页Journal of Wenzhou Medical University

基　　金：温州市基础性科研项目(Y20210903)。

摘　　要：目的:探讨分水岭脑梗死患者复发的影响因素,并构建列线图预测模型。方法:回顾性收集2013年8月至2022年1月分水岭脑梗死住院患者的临床及影像资料。患者分为训练集和验证集。训练集进行单变量和多变量Logistic分析,建立列线图。在验证集中,使用ROC曲线和校准图来测量模型预测性能。利用决策曲线分析(DCA)评估模型能否影响临床决策。结果:年龄(OR=1.067,P=0.041)、性别(OR=0.340,P=0.020)、动脉狭窄(OR=1.015,P=0.023)、低血压(OR=2.410,P=0.005)、低心输出量(OR=2.021,P=0.058)等五个预测因素被纳入列线图模型中。ROC曲线分析结果显示,列线图在训练集中显示出良好的预测性能,AUC=0.821(95%CI=0.705~0.937),验证集中AUC=0.961(95%CI=0.906~0.999)。校准图显示了在训练集和验证集中实际和预测的复发卒中风险之间有高度一致性。结论:年龄、性别、动脉狭窄、低血压、低心输出量是分水岭脑梗死患者复发的独立影响因素。基于列线图预测分水岭脑梗死复发风险具有可行性,为临床个性化诊疗提供可视化依据。Objective:To investigate the recurrence influencing factors in patients with watershed infarction,and to construct a nomogram prediction model.Methods:The clinical and imaging data of patients with watershed infarction hospitalized from August 2013 to January 2022 were collected retrospectively.Using univariate and multivariate binary Logistic analysis on the training cohort of 149 individuals with watershed infarction,the nomogram was developed.In a validation cohort(n=68),predictive performance was measured using ROC curves and calibration plots.Using decision curve analysis(DCA),personalized treatment choices were made to balance the net benefit for each watershed infarction patient.Results:Five predictive factors including age(OR=1.067,P=0.041),sex(OR=0.340,P=0.020),arterial stenosis(OR=1.015,P=0.023),hypotension(OR=2.410,P=0.005),low cardiac output(OR=2.021,P=0.058)were incorporated into the nomogram model.The nomogram displayed good predictive performance in ROC analysis[AUC was 0.821(95%CI=0.705-0.937)in the training cohort and 0.961(95%CI=0.906-0.999)in the validation cohort].Additionally,the calibration plot demonstrated a high degree of concordance between the actual and projected risks of recurrent stroke in both the training and validation groups.DCA confirmed the clinical utility of this nomogram.Conclusion:Age,sex,arterial stenosis,hypotension,low cardiac output are independent influencing factors for patients with watershed infarction recurrence.It is feasible to predict the recurrence risks based on nomogram,which provides a visual basis for clinical personalized diagnosis and treatment.

关 键 词：分水岭脑梗死 列线图 风险因素 卒中复发 

分 类 号：R743[医药卫生—神经病学与精神病学]