基于冠状动脉CT血管成像影像组学的稳定型心绞痛痰瘀互结证与气虚血瘀证鉴别研究  

Research on the differential diagnosis of phlegm and blood stasis pattern and qi deficiency and blood stasis pattern in stable angina pectoris based on coronary artery CT angiography radiomics

作　　者：魏东升 齐佳杰 刘孝生 李路珍 李涵[1,2] 刘雨婷 邓成康 戴旭 赵宝英 张哲 WEI Dongsheng;QI Jiajie;LIU Xiaosheng;LI Luzhen;LI Han;LIU Yuting;DENG Chengkang;DAI Xu;ZHAO Baoying;ZHANG Zhe(First Clinical College of Liaoning University of Traditional Chinese Medicine,Shenyang 110847,China;Liaoning University of Traditional Chinese Medicine,Key Laboratory of Ministry of Education for TCM Viscera-State Theory and Applications,Shenyang 110847,China;Affiliated Hospital of Liaoning University of Traditional Chinese Medicine,Shenyang 110033,China)

机构地区：[1]辽宁中医药大学第一临床学院,沈阳110847 [2]辽宁中医药大学中医脏象理论及应用教育部重点实验室 [3]辽宁中医药大学附属医院

出　　处：《北京中医药大学学报》2024年第4期545-554,共10页Journal of Beijing University of Traditional Chinese Medicine

基　　金：辽宁省重点研发计划项目(No.2020JH2/10300070);国家中医药管理局青年岐黄学者支持项目(No.20201A2180);辽宁省科学技术计划项目(No.2019-ZD-0445)。

摘　　要：目的应用影像组学技术构建稳定型心绞痛痰瘀互结证与气虚血瘀证鉴别诊断模型。方法收集2021年1月—2022年1月辽宁中医药大学附属医院91例行冠状动脉CT血管成像的稳定型心绞痛患者,其中痰瘀互结证47例、气虚血瘀证44例,采用分层随机抽样方法将患者按照7∶3的比例分为训练集(64例)与测试集(27例),使用3D-slicer软件提取冠状动脉周围脂肪组织(PCAT)图像影像组学特征。使用主成分分析可视化痰瘀互结证与气虚血瘀证影像组学特征分布情况。采用最小绝对收缩和选择算子回归分析及支持向量机递归特征消除进行特征筛选,使用Logistics多因素回归算法构建鉴别诊断模型,在训练集与测试集中应用接受者操作特征(ROC)曲线对模型进行验证,评价影像组学特征鉴别痰瘀互结证与气虚血瘀证的效能。使用Spearman相关分析进行鉴别特征与临床理化数据的相关性分析。结果3D-slicer软件共提取出837个PCAT图像影像组学特征。主成分分析中第一主成分和第二主成分分别解释了77.9%和8.1%的总体变异,两证候组间存在较为明显的分离趋势。经过特征筛选,7个影像组学特征用于构建痰瘀互结证与气虚血瘀证的鉴别诊断模型。训练集中鉴别诊断模型的ROC曲线下面积(AUC)为0.844,测试集中AUC为0.834。Spearman相关性分析表明,鉴别特征与心肌肌钙蛋白I、中性粒细胞、甘油三酯、总胆固醇及白细胞具有明显相关性。结论基于PCAT的CT影像组学模型对稳定型心绞痛痰瘀互结证与气虚血瘀证具有较高的鉴别作用。Objective To establish a differential model of phlegm and blood stasis pattern and qi deficiency and blood stasis pattern in stable angina pectoris using radiomics.Methods A total of 91 patients with stable angina pectoris who underwent coronary artery CT angiography in Affiliated Hospital of Liaoning University of Traditional Chinese Medicine from January 2021 to January 2022 were collected,including 47 cases of phlegm and blood stasis pattern and 44 cases of qi deficiency and blood stasis pattern.The patients were divided into train set(64 cases)and test set(27 cases)according to the ratio of 7∶3 by stratified random sampling method.3D⁃slicer software was used to extract the radiomics features of pericoronary adipose tissue(PCAT)images.Principal component analysis was used to visualize the distribution of radiomics features of pattern of phlegm and blood stasis and pattern of qi deficiency and blood stasis.The least absolute shrinkage and selection operator regression analysis and support vector machine decreasing feature elimination were used for feature selection.The multinomial logistics regression was used for model construction.The receiver operating characteristic(ROC)curve was used to verify the model in the train set and the test set to evaluate the effectiveness of the radiomics features in differentiating phlegm and blood stasis pattern and qi deficiency and blood stasis pattern.Finally,Spearman coefficient was used to analyze the correlation between the differential features and clinical physicochemical data.Results A total of 837 radiomics features were extracted from PCAT images by 3D⁃slicer software.In the principal component analysis,PC1 and PC2 explained 77.9%and 8.1%of the total variance,respectively,and there was a relatively obvious separation trend between the two pattern groups.After feature screening,7 radiomics features were used to construct the differential model of phlegm and blood stasis pattern and qi deficiency and blood stasis pattern.The area under the ROC curve(AUC)of the d

关 键 词：稳定型心绞痛 影像组学 冠状动脉CT血管成像 鉴别诊断 痰瘀互结证 气虚血瘀证 

分 类 号：R259.414[医药卫生—中西医结合]