影像组学诺模图在术前预测肾透明细胞癌WHO/ISUP分级的应用研究  

Application of Radiomic Nomogram in Preoperative Prediction of WHO/ISUP Grade of Renal Clear Cell Carcinoma

作　　者：顾长青 李陆 孙冬雪 王欣[1] 张玉文 沈俊杰[1] GU Changqing;LI Lu;SUN Dongxue(Department of Radiology,The First Affiliated Hospital of Benbu Medical College Bengbu,Anhui Province 233000,P.R.China)

机构地区：[1]蚌埠医学院第一附属医院放射科,233000

出　　处：《临床放射学杂志》2022年第10期1915-1920,共6页Journal of Clinical Radiology

摘　　要：目的 探讨基于增强CT的影像组学特征结合传统影像特征构建的诺模图在术前预测肾透明细胞癌(ccRCC)的WHO/ISUP分级的价值.方法 回顾性分析186例术前行增强CT检查并经病理证实为ccRCC的影像和临床病理资料.按照WHO/ISUP分级将病例分为高级别(Ⅲ+Ⅳ)和低级别(Ⅰ+Ⅱ),并按照7:3分为训练组和验证组.在训练组提取CT的影像组学特征,筛选出最佳的特征构建影像组学模型;纳入相关的传统影像学特征,经过Logistic回归,选出最有价值的特征构建传统影像模型;联合传统影像模型和影像组学模型构建联合模型,并建立诺模图.模型的诊断效能通过受试者工作特征曲线(ROC)及曲线下面积(AUC)来评价,使用决策曲线分析(DCA)和校正曲线评价诺模图的临床应用价值.结果 最终筛选出1项传统影像特征和27项影像组学特征;联合两种特征的影像组学诺模图在训练组和验证组的AUC分别为0.823、0.726,显示出良好的预测效能.结论基于增强CT的诺模图在术前预测ccRCC的WHO/ISUP分级具有可行性.Objective To investigate the value of nomogram based on enhanced CT and traditional image features in preoperative prediction of WHO/ISUP grade of renal clear cell carcinoma(ccRCC).Methods The imaging and clinicopathological data of 186 cases with ccRCC confirmed by pathology and enhanced CT before the operation were analyzed retrospectively.According to WHO/ISUP classification,patients were divided into high grade(Ⅲ+Ⅳ)and low-grade(Ⅰ+Ⅱ),and divided into training group and validation group by 7∶3.In the training group,the radiomics features of CT were extracted and the best features were selected to construct the radiomics model.Relevant traditional imaging features were included,and the most valuable features were selected to construct the traditional imaging model through Logistic regression.Combined with the traditional image model and radiomics model,the joint model was constructed and the nomogram was established.The diagnostic efficiency of the model was evaluated by receiver operating characteristic curve(ROC)and area under the curve(AUC),and the clinical application value of nomograms was evaluated by decision curve analysis(DCA)and calibration curve.Results One traditional image feature and 27 radiomics features were selected.The AUC of the combination of the two features in the training group and the verification group was 0.823 and 0.726,respectively,showing good predictive efficiency.Conclusion Nomogram based on enhanced CT is feasible for preoperative prediction of WHO/ISUP classification of ccRCC.

关 键 词：肾透明细胞癌 影像组学 诺模图 

分 类 号：R737.11[医药卫生—肿瘤] R730.44[医药卫生—临床医学]