基于腹部CT平扫图像对急性阑尾炎的诊断:影像组学研究  

Diagnosis of acute appendicitis based on abdominal plain computed tomography scan:a radiomics study

作　　者：曲别雪蕾 王可欣 刘想 张耀峰 张晓东[1] 王霄英[1] Qubie Xuelei;Wang Kexin;Liu Xiang;Zhang Yaofeng;Zhang Xiaodong;Wang Xiaoying(Department of Medical Radiology,Peking University First Hospital;School of Basic Medical Sciences,Capital Medical University;Beijing Smart Tree Medical Technology Co.,Ltd.)

机构地区：[1]北京大学第一医院医学影像科,北京100034 [2]首都医科大学基础医学院,北京100069 [3]北京赛迈特锐医学科技有限公司,北京100011

出　　处：《重庆医科大学学报》2024年第8期1039-1044,共6页Journal of Chongqing Medical University

摘　　要：目的:探索使用影像组学模型在腹部CT平扫图像上诊断急性阑尾炎的可行性。方法:回顾性收集2015年5月至2021年8月北京大学第一医院经手术病理确诊的急性阑尾炎患者的术前腹部CT扫描的影像和临床数据210例,以及同期因其他急腹症行腹部CT平扫的患者210例用于影像组学模型训练。420例患者的CT检查采集于4台不同CT扫描设备。由2位影像科医生手工标注阑尾区域。将数据按7∶3的比例随机分为训练集和测试集。使用特征提取102种图像特征后,使用Pearson相关分析进行特征降维,以递归特征消除法选择最相关的20种特征训练支持向量机(support vector machine,SVM)进行二分类,得到影像组学模型。得到影像组学模型后,对测试集进行预测结果,以受试者工作特征曲线(receiver operating characteristic,ROC)评价影像组学模型的效能。结果:经过特征降维、特征选择后,共有3种形态特征、3种一级特征和14种纹理特征用于训练SVM模型。测试集中SVM模型正确预测114例(阑尾炎60例,非阑尾炎54例),错误预测12例(阑尾炎3例,非阑尾炎9例)敏感度为95.2%,特异度为85.7%,准确率为90.5%,ROC曲线下面积为0.931(95%CI=0.887~0.976)。结论:基于腹部CT平扫图像的影像组学模型可用于急性阑尾炎的预测,未来有望用于急腹症CT检查的流程优化。Objective:To investigate the feasibility of a radiomics model in the diagnosis of acute appendicitis based on abdominal plain computed tomography(CT)scan.Methods:A retrospective analysis was performed for the preoperative abdominal CT imaging data and clinical data of 210 patients with acute appendicitis confirmed by surgery in our hospital from May 2015 to August 2021,and 210 patients who underwent abdominal plain CT scan due to other acute abdominal diseases during the same period of time were enrolled for the training of the radiomics model.CT scan data of the 420 patients were collected from 4 different CT devices,and the region of the appendix was manually annotated by two radiologists.The data were randomly divided into training set and test set at a ratio of 7∶3.After 102 types of image features were extracted,the Pearson correlation analysis was used for feature dimension reduction,and the recursive feature elimination method was used to select 20 most relevant features for the training and binary classification of support vector machine(SVM)to obtain a radiomics model.After the radiomics model was obtained,the test set was used to predict the results,and the receiver operating characteristic(ROC)curve was used to evaluate the performance of the radiomics model.Results:After feature dimension reduction and feature selection,3 shape-based features,3 first-order features,and 14 texture features were used to train the SVM model.In the test set,the SVM model had correct prediction in 114 cases(60 appendicitis cases and 54 nonappendicitis cases)and wrong prediction in 12 cases(3 appendicitis cases and 9 non-appendicitis cases),with a sensitivity of 95.2%,a specificity of 85.7%,an accuracy of 90.5%,and an area under the ROC curve of 0.931(95%CI:0.887-0.976).Conclusion:The radiomics model based on abdominal plain CT scan images can be used for the prediction of acute appendicitis and is expected to be used to optimize the workflow of CT examination for acute abdominal disease in the future.

关 键 词：急性阑尾炎 CT 影像组学 

分 类 号：R445[医药卫生—影像医学与核医学]