瘤内及瘤周CT影像组学联合临床预测甲状腺乳头状癌颈部淋巴结转移  

Intratumoral and peritumoral CT radiomics combined with clinic for predicting cervical lymph node metastasis of papillary thyroid carcinoma

作　　者：姚琪 刘奇峰 吕磊 张成孟 丁治民 YAO Qi;LIU Qifeng;LYU Lei;ZHANG Chengmeng;DING Zhimin(Department of Radiology,Yijishan Hospital of Wannan Medical College,Wuhu 241001,China)

机构地区：[1]皖南医学院弋矶山医院放射科,安徽芜湖241001

出　　处：《中国介入影像与治疗学》2025年第2期112-117,共6页Chinese Journal of Interventional Imaging and Therapy

基　　金：安徽省教育厅2022年度新时代育人质量工程项目(研究生教育)(2022zyxwjxalk165)。

摘　　要：目的观察瘤内及瘤周CT影像组学联合临床预测甲状腺乳头状癌(PTC)颈部淋巴结转移(CLNM)的价值。方法回顾性分析338例PTC患者,按7∶3比例将其分为训练集(n=236)与验证集(n=102),并根据有无CLNM于集内划分亚组;采用多因素logistic回归分析筛选预测PTC CLNM的独立危险因素并构建临床模型;基于CT平扫+三期增强图像勾画肿瘤ROI并将其适形外扩2 mm,分别构建平扫、动脉期、静脉期、延迟期及多期瘤内、瘤周及瘤内+瘤周影像组学模型,以其中效能最高者联合临床模型构建联合模型。绘制受试者工作特征曲线,计算曲线下面积(AUC),评估各模型预测PTC CLNM的效能。结果男性及年龄<45岁均为PTC CLNM的临床独立危险因素(P均<0.05)。影像组学模型中,多期瘤内+瘤周模型最优;以之联合临床模型构建的联合模型预测训练集PTC CLNM的AUC为0.859,高于临床模型(0.684)及瘤内+瘤周多期影像组学模型(0.831)(P均<0.05);联合模型在验证集的AUC为0.832,与多期瘤内+瘤周影像组学模型相当(0.819,P=0.368)而高于临床模型(0.605,P<0.001)。结论瘤内及瘤周多期CT影像组学可有效预测PTC CLNM;联合临床可进一步提高其预测效能。Objective To observe the value of intratumoral and peritumoral CT radiomics combined with clinic for predicting cervical lymph node metastasis(CLNM)of papillary thyroid carcinoma(PTC).Methods Totally 338 patients with PTC were retrospectively enrolled and divided into training set(n=236)and validation set(n=102)at a ratio of 7∶3,also further assigned into subgroups according to CLNM or not within sets.The clinical independent risk factors for PTC CLNM were identified using multivariate logistic regression analysis,and a clinical model was then constructed.Intratumoral ROI were delineated on CT images including non-enhanced and three-phase enhancement images,which were then enlarged by 2 mm.Radiomics models of intratumoral,peritumoral and intratumoral+peritumoral ROI were established based on non-enhanced CT,as well as arterial phase,venous phase,delayed phase and multi-phase enhanced CT,respectively,the best one was selected and combined with clinical model to construct a combined model.Receiver operating characteristic curve was drawn,and the area under the curve(AUC)was calculated to evaluate the efficacy of each model for predicting PTC CLNM.Results Male patient and aged<45 years were both clinical independent risk factors for PTC CLNM(both P<0.05).Multi-phase intratumoral+peritumoral radiomics model was the optimal radiomics model,which was used to construct the combined model combining with clinical model.The AUC of combined model for predicting PTC CLNM in training set was 0.859,superior to that of clinical model and multi-phase intratumoral+peritumoral radiomics model(0.684,0.831,both P<0.05),which in validation set was 0.832,similar to that of multi-phase intratumoral+peritumoral radiomics model(0.819,P=0.368)but superior to that of clinical model(0.605,P<0.001).Conclusion Multi-phase intratumoral and peritumoral CT radiomics could be used to predict PTC CLNM.Combining with clinic could further improve its efficacy.

关 键 词：甲状腺肿瘤 淋巴转移 体层摄影术 X线计算机 影像组学 

分 类 号：R736.1[医药卫生—肿瘤] R814.42[医药卫生—临床医学]