摘要
目的 基于急性脑卒中血管内机械取栓切除术(mechanical thrombectomy,MT)前MRI影像组学特征和机器学习,分析其在预测出血转化(hemorrhagic transformation,HT)中的价值。材料与方法 回顾性分析我院神经内科行MRI检查和MT治疗的214例急性脑卒中患者临床和影像学资料。采用ITK-SNAP软件对治疗前MRI图像上的弥散加权成像(diffusion weighted imaging,DWI)高信号梗死区和灌注加权成像(perfusion weighted imaging,PWI)灌注异常区进行分割,应用AK软件进行影像组学特征提取和降维,最终使用最小绝对收缩与选择算子算法(least absolute shrinkage and selection operator,LASSO)确定HT相关的最佳影像组学特征,通过支持向量机分类器评估其在HT预测中的价值。结果 每例患者各提取792个影像组学特征,降维后筛选出10个与HT预测最相关的特征。ROC分析显示该模型预测训练集患者HT的AUC为0.984,敏感度和特异度分别为0.932、0.967;预测测试集患者HT的AUC为0.921,敏感度和特异度分别为0.826、0.852。结论 基于MRI的影像组学和机器学习分析是预测急性脑卒中HT的重要工具,对早期准确识别HT高风险患者具有较高的效能。
Objective: To investigate MRI radiomic features before mechanical thrombectomy(MT) in acute stroke and machine learning and analyze their value in the prediction of hemorrhagic transformation(HT). Materials and Methods: A total of 214 acute stroke patients receiving MRI and MT therapy in the neurology department of our hospital were retrospectively enrolled. The ITK-SNAP software was used to segment the high signal areas of diffusion weighted imaging(DWI) and the abnormal perfusion areas of perfusion weighted imaging(PWI). The AK software was used to extract the radiomic features and reduce the dimensionality. The least absolute shrinkage and selection operator(LASSO) regression analysis was used to determine the radiomic features related to HT and support vector machine classifier was used to evaluate its value in HT prediction. Results: Seven hundred and ninety-two radiomics features of each patient were extracted and 10 features highly related to HT were screened after dimension reduction. ROC analysis showed that the area under curve(AUC) of the prediction model based on the training set was 0.984, the sensitivity and specificity were 0.932 and 0.967 respectively;the AUC of the prediction model based on the test set was 0.921, the sensitivity and specificity were 0.826 and 0.852 respectively. Conclusions: The analysis based on MRI radiomics and machine learning are the important tools for predicting HT, and have high efficiency in early accurate identification of HT.
作者
缪丽琼
彭明洋
王同兴
陈国中
殷信道
吴刚
MIAO Liqiong;PENG Mingyang;WANG Tongxing;CHEN Guozhong;YIN Xindao;WU Gang(Department of Radiology,Jiangyin Hospital of Traditional Chinese Medicine,Wuxi 214400,China;Department of Radiology,Nanjing First Hospital,Nanjing Medical University,Nanjing 210006,China)
出处
《磁共振成像》
CAS
CSCD
北大核心
2022年第3期18-21,75,共5页
Chinese Journal of Magnetic Resonance Imaging
基金
国家自然科学基金(编号:82001811)
江苏省自然科学基金(编号:BK20201118)。
关键词
卒中
出血转化
磁共振成像
影像组学
机器学习
stroke
hemorrhagic transformation
magnetic resonance imaging
radiomics
machine learning
