临床因素联合平扫CT影像组学预测前循环缺血性脑卒中静脉溶栓后症状性出血转化的价值  

The value of a combined model of clinical factors and non-contrast CT radiomics in predicting symptomatic hemorrhagic transformation after intravenous thrombolysis in patients with anterior circulation ischemic stroke

作　　者：季丹丹 王天乐[1] 朱丽[1] 陆钰 阮席武[1] Ji Dandan;Wang Tianle;Zhu Li;Lu Yu;Ruan Xiwu(Department of Radiology,Nantong Municipal First People′s Hospital,the Second Affiliated Hospital of Nantong University,Nantong 226001,China)

机构地区：[1]南通市第一人民医院,南通大学第二附属医院影像科,南通226001

出　　处：《中华放射学杂志》2024年第10期1021-1027,共7页Chinese Journal of Radiology

基　　金：南通市科技局社会民生指令性项目(MS2023068)。

摘　　要：目的探讨基于平扫CT(NCCT)的影像组学特征联合临床危险因素构建联合模型预测前循环急性缺血性脑卒中(AIS)患者重组织纤溶酶原激活剂(rt-PA)静脉溶栓后发生颅内症状性出血转化(sICH)的效能。方法该研究为横断面研究。回顾性分析2018年10月至2022年9月在南通市第一人民医院就诊并接受rt-PA静脉溶栓治疗的前循环AIS患者316例的临床和影像资料。所有患者按照7∶3的比例采用分层随机抽样法分为训练集210例和验证集106例。采用单因素及多因素logistic回归分析筛选出预测sICH的临床独立危险因素。在NCCT图像勾画梗死区并提取影像组学特征,使用组间及组内相关系数、最小冗余最大相关以及最小绝对收缩和选择算子对提取的影像组学特征进行降维筛选,计算影像组学评分。最后将临床危险因素及影像组学评分纳入多因素logistic分析,建立临床模型、影像组学模型和影像组学-临床联合模型。采用受试者操作特征曲线及曲线下面积评估各模型的预测效能,利用决策曲线分析量化各预测模型的净收益。结果最终选取8个影像组学特征用于构建影像组学模型。多因素logistic分析显示高血压(OR=2.703,95%CI 1.153~6.334,P=0.022)、房颤(OR=3.023,95%CI 1.290~7.085,P=0.011)、入院美国国立卫生研究院卒中量表评分(OR=1.078,95%CI 1.017~1.143,P=0.012)是影响前循环AIS患者rt-PA静脉溶栓sICH的独立危险因素。验证集中,联合模型预测前循环AIS患者rt-PA静脉溶栓sICH的曲线下面积为0.763(95%CI 0.618~0.909),高于临床模型0.710(95%CI 0.552~0.868)和影像组学模型0.708(95%CI 0.568~0.848)。决策曲线分析示联合模型能使患者获得更高的净收益。结论基于NCCT的影像组学与临床危险因素构建的联合模型对预测前循环AIS患者rt-PA静脉溶栓sICH具有较高的诊断效能。Objective To investigate the efficacy of a combined model constructed by the radiomics features based on non-contrast CT(NCCT)combined with clinical risk factors in predicting the occurrence of symptomatic intracranial hemorrhagic transformation(sICH)after intravenous thrombolysis with recombinant tissue plasminogen activator(rt-PA)in patients with anterior circulation acute ischemic stroke(AIS).Methods In this cross-sectional study,clinical and imaging data of 316 patients with anterior circulation AIS who received intravenous thrombolysis with rt-PA at Nantong First People′s Hospital from October 2018 to September 2022 were retrospectively analyzed.The cases were divided into a training set of 210 cases and a validation set of 106 cases by stratified random sampling at a ratio of 7∶3.Univariate and multivariate logistic regression analyses were performed to select the independent clinical risk factors for predicting sICH.The infarct area was delineated on the NCCT images and radiomics features were extracted.The extracted radiomics features were dimensionally reduced and selected using the inter-and intra-group correlation coefficients,maximum correlation and minimum redundancy,and the least absolute shrinkage and selection operator,and then the radiomics score was calculated.Finally,multivariate logistic analysis was performed and the clinical risk factors and radiomics scores were used to establish the clinical model,the radiomics model and the radiomics-clinical combined model.The predictive efficacy of each model was evaluated by the receiver operating characteristic curve and the area under the curve,and decision curve analysis(DCA)was used to calculate and quantify the net benefits of each predictive model.Results In total eight radiomics features were selected to construct the radiomics model.Multivariate logistic analysis showed that hypertension(OR=2.703,95%CI 1.153-6.334,P=0.022),atrial fibrillation(OR=3.023,95%CI 1.290-7.085,P=0.011),and the National Institutes of Health Stroke Scale score at adm

关 键 词：卒中 症状性出血转化 体层摄影术 X线计算机 影像组学 

分 类 号：R743.3[医药卫生—神经病学与精神病学] R816.1[医药卫生—临床医学]