基于四维CT血管成像的急性缺血性卒中患者血管内治疗后出血转化的列线图模型建立及效能预测

A nomogram for individualized prediction of the probability of hemorrhagic transformation in patients with acute ischemic stroke after endovascular treatment based on 4-dimensional CT angiography factors

目的探讨四维CT血管成像(4D CTA)对急性缺血性卒中(AIS)患者血管内治疗(EVT)后出血转化的预测因素及基于预测因素建立的列线图模型预测出血转化的效能。方法回顾性分析2016年3月至2020年11月在北京医院急诊绿色通道进行"一站式"CTA-CT灌注成像及EVT的101例颈内动脉和(或)大脑中动脉闭塞的AIS患者的影像学资料。根据是否发生出血转化分为出血转化组(45例)及未出血转化组(56例)。采用logistic回归方法筛选相关临床、影像学变量,如年龄、初诊美国国立卫生研究院卒中量表(NIHSS)评分、4D CTA侧支循环评分、Alberta卒中项目早期CT评分(ASPECTS)、血栓负荷评分等,并建立列线图预测模型。采用受试者操作特征曲线(ROC)及曲线下面积(AUC)评估模型对AIS是否发生出血转化诊断的效能。结果单因素分析发现,出血转化组与未出血转化组患者的年龄[分别为79.00(68.00,85.00)岁、73.00(62.75,80.00)岁,Z=-2.20、P=0.028]、NIHSS评分[分别为16.00(12.00,21.00)分、9.50(6.00,14.00)分,Z=-4.44、P<0.001]、ASPECTS评分[分别为5.00(3.00,8.00)分、8.00(7.00,9.00)分,Z=-4.23、P<0.001]、4D CTA侧支循环评分[分别为2.00(0,3.00)分、3.00(3.00,4.00)分,Z=-5.39、P<0.001]、血栓负荷评分[分别为4.00(1.00,7.00)分、7.50(6.00,9.00)分,Z=-3.42、P=0.001]、闭塞位置(颈内动脉/大脑中动脉分别为23/22例、11/45例,χ^(2)=9.70、P=0.002)及心房颤动(分别为27、19例,χ^(2)=5.83、P=0.016)差异均有统计学意义;纳入多因素logistics回归分析显示,ASPECTS评分(OR=0.64,95%CI 0.47~0.87)、NIHSS评分(OR=1.13,95%CI 1.01~1.26)、4D CTA侧支循环评分(OR=0.40,95%CI 0.22~0.76)是AIS患者术后发生出血转化的独立预测因素P<0.05)。以ASPECTS评分、NIHSS评分、4D CTA侧支循环评分构建列线图模型预测AIS是否发生出血转化的AUC为0.876,95%CI 0.807~0.945,灵敏度为77.8%,特异度为87.5%。结论低ASPECTS评分、高NIHSS评分及低4D CTA侧支评分的AIS患者术后更易发生出血转化,结合三者建立的列线图模型可以较好地预测AIS患者发生出血转化的概率,为临床决策提供有效的帮助。

Objective To assess the value of 4-dimensional CT angiography(4D CTA)to predict hemorrhagic transformation(HT)with a new nomogram model in acute ischemic stroke(AIS)patients after endovascular treatment(EVT).Methods Imaging and clinical data of 101 AIS patients with internal carotid artery and/or middle cerebral artery occlusion who underwent"one-stop"CTA-CT perfusion and EVT in green channel of Beijing Hospital from March 2016 to November 2020 were analyzed retrospectively.The patients were divided into HT group(45 patients)and non-HT group(56 patients).Multivariate logistic regression analysis was used to select relevant clinical and imaging variables,such as age,initial National Institute of Health stroke scale(NIHSS)score,4D CTA collateral circulation score,Alberta stroke program early CT score(ASPECTS),clot burden score,and a predictive nomogram model were developed.The receiver operating characteristic(ROC)curve and area under the curve(AUC)were used to evaluate the efficacy of predictive nomogram model for diagnosing HT.Results Univariate analysis showed that there were significant difference of age[79.00(68.00,85.00)years,73.00(62.75,80.00)years,Z=-2.20,P=0.028],NIHSS score[16.00(12.00,21.00),9.50(6.00,14.00),Z=-4.44,P<0.001],ASPECTS score[5.00(3.00,8.00),8.00(7.00,9.00),Z=-4.23,P<0.001],4D CTA collateral circulation score[2.00(0,3.00),3.00(3.00,4.00),Z=-5.39,P<0.001],clot burden score[4.00(1.00,7.00),7.50(6.00,9.00),Z=-3.42,P=0.001],location of the occlusion(internal carotid artery/middle cerebral artery occlusion was 23/22,11/45 cases,χ^(2)=9.70,P=0.002),and atrial fibrillation(27 and 19 cases respectively,χ^(2)=5.83,P=0.016)between HT group and non-HT group.Multivariate logistic regression analysis showed that ASPECTS score(OR=0.64,95%CI 0.47-0.87),NIHSS score(OR=1.13,95%CI 1.01-1.26),4D CTA collateral circulation score(OR=0.40,95%CI 0.22-0.76)were independent predictors of HT in AIS patients(P<0.05).The AUC of the nomogram based on the ASPECTS score,NIHSS score and 4D CTA collateral circulation score to predict HT of AIS patients was 0.876(95%CI 0.807-0.945),with a sensitivity of 77.8%and specificity of 87.5%.Conclusions Patients with low ASPECTS score,high NIHSS score and low 4D CTA collateral circulation score have a higher risk of HT after EVT.The nomogram model may predict the probability of HT of AIS patients and provide effective assistance for clinical decision-making.