构建列线图模型预测急性前循环大血管闭塞患者机械取栓不良预后的研究

A nomogram prediction model for poor prognosis of mechanical thrombectomy in patients with acute anterior circulation large vessel occlusion

目的探讨联合多因素构建列线图模型预测急性前循环大血管闭塞(ACLVO)机械取栓术后发生预后不良的研究。方法选择2018年1月至2022年12月于北部战区总医院接受机械取栓治疗的304例ACLVO患者,将3个月改良Rankin量表(mRS)≤2分的患者纳入预后良好组(n=140),mRS>2分的患者则纳入预后不良组(n=164)。通过单因素分析和多元Logistic回归分析筛选与预后不良相关的独立变量,将这些变量纳入列线图模型。应用校准曲线评价模型的拟合优度,受试者操作曲线(ROC)验证模型的预测效能,决策曲线分析(DCA)评价模型的临床应用价值。结果多元Logistic回归模型分析显示,年龄、入院国立卫生研究院卒中量表评分、糖化血红蛋白和血管成功再通为ACLVO机械取栓预后不良的独立预测因子。列线图模型显示出良好的校准和区分能力,ROC分析表明联合多因素构建列线图模型预测ACLVO患者机械取栓术后预后不良的效能优于各单因素模型,DCA亦表明该模型具有一定的临床价值。结论联合多因素构建的列线图模型对ACLVO患者机械取栓后预后不良发生风险具有较好的预测效能。

Objective To investigate the prediction of poor prognosis after mechanical thrombectomy of acute anterior circulation great vessel occlusion(ACLVO)by using multi-factor nomography model.Methods A total of 304 ACLVO patients were selected,and patients with 3-month modified Rankin scale(mRS)≤2 were included in the good prognosis group(n=140),and those with mRS>2 were included in the poor prognosis group(n=164).Independent variables associated with poor prognosis were screened by univariate analysis and multiple Logistic regression analysis,and these variables were included in the nomogram model.Calibration curve was used to evaluate the goodness of fit of the model,receiver operating curve(ROC)to verify the predictive efficacy of the model,and decision curve analysis(DCA)to evaluate the clinical application value of the model.Results Multiple logistic regression analysis showed that age(OR=1.082;95%CI:1.050 to 1.112;P<0.001),NIHSS score at admission(OR=1.231;95%CI:1.056 to 1.215;P<0.001)(OR=1.131;95%CI:1.055 to 1.213;P<0.001),HbA1c(OR=1.401;95%CI:1.158 to 1.694;P<0.001),and successful vascular recanalization(OR=0.305;95%CI:0.156 to 0.595;P<0.001)were independent predictors of poor prognosis in ACLVO patients received mechanical thrombectomy.The nomogram prediction model showed good calibration and differentiation capabilities.The ROC analysis showed that combining multiple factors to construct a nomogram prediction model is more effective in predicting poor prognosis after mechanical thrombectomy in ACLVO patients than other single factor models.The DCA also showed that the model has clinical value.Conclusion The combined multi-factor nomogram model can predict the risk of poor prognosis after mechanical thrombectomy in ACLVO patients.