基于MRI瘤内及瘤周特征预测肝内肿块型胆管癌病理分级模型的开发和验证  

Development and validation of a model for predicting pathological grade of intrahepatic mass-forming cholangiocarcinoma based on intratumoral and peritumoral features on MRI

作　　者：李晓萌 邢立红 卓利勇 刘学彦[3] 代硕 董景辉 张岩岩[5] 李宏军[5] 殷小平[1,2] LI Xiaomeng;XING Lihong;ZHUO Liyong;LIU Xueyan;DAI Shuo;DONG Jinghui;ZHANG Yanyan;LI Hongjun;YIN Xiaoping(Department of Radiology,Affiliated Hospital of Hebei University,Baoding 071000,China;Hebei Key Laboratory of Precise Imaging of Inflammation Related Tumors,Baoding 071000,China;School of Mathematical Sciences,Liaocheng University,Liaochen 252000,China;Radiology Department,the Fifth Medical Center of Chinese PLA General Hospital,Beijing 100039,China;Department of Radiology,Beijing You'an Hospital,Capital Medical University,Beijing 100069,China)

机构地区：[1]河北大学附属医院放射科,保定071000 [2]河北省炎症相关肿瘤精准影像诊断学重点实验室,保定071000 [3]山东省聊城大学数学科学学院,聊城252000 [4]解放军总医院第五医学中心放射诊断科,北京100039 [5]首都医科大学附属北京佑安医院放射科,北京100069

出　　处：《磁共振成像》2025年第2期51-58,共8页Chinese Journal of Magnetic Resonance Imaging

基　　金：河北省高层次人才资助项目(编号:B20231008)。

摘　　要：目的探讨基于T2WI及弥散加权成像(diffusion weighted imaging,DWI)的瘤内及不同瘤周区域影像组学特征和临床影像学因素术前预测肝内肿块型胆管癌(intrahepatic mass-forming cholangiocarcinoma,IMCC)病理分级的价值。材料与方法回顾性分析197例经术后病理证实的IMCC患者的临床及术前MRI资料,于轴位T2WI及DWI图像上勾画肿瘤感兴趣区(region of interest,ROI),并分别向外扩展3、5、10、15和20 mm以获得不同范围的瘤周ROI。利用PyRadiomics提取影像组学特征,通过方差齐性检验、独立样本t检验、递归特征消除算法及最小绝对收缩和选择算子筛选特征,采用逻辑回归(logistic regression,LR)分类器及5折交叉验证法建模并验证,建立临床影像学模型、瘤内组学模型、瘤周组学模型、瘤内+瘤周组学模型、双序列融合模型及多模态联合模型,比较以上各个模型预测效能,选出最佳模型。采用受试者工作特征(receiver operating characteristic,ROC)曲线及曲线下面积(area under the curve,AUC)评价模型性能,采用DeLong检验比较AUC差异,校准曲线评价模型拟合能力,决策曲线评估模型的临床价值。结果瘤周组学模型中DWI序列3 mm模型表现最优,训练集及验证集中AUC分别为0.836和0.777。性别、年龄、病灶位置及血管受累是IMCC病理分级的独立预测因子,临床影像学模型训练集和验证集中的AUC分别为0.658和0.614。DWI序列瘤内+瘤周3 mm模型具有最佳预测效能,训练集和验证集中的AUC分别为0.892和0.814,优于双序列融合模型及多模态联合模型。结论基于DWI序列的瘤内+瘤周3 mm组学模型表现出最佳预测能力,能够在术前无创性地预测IMCC病理分级,为临床治疗决策提供理论指导。Objective:To explore the value of intratumoral and different peritumoral radiomics features based on T2-weighted imaging(T2WI)and diffusion weighted imaging(DWI),as well as clinical imaging factors,in preoperative prediction of the pathological grade of intrahepatic mass-forming cholangiocarcinoma(IMCC).Materials and Methods:A retrospective analysis was conducted on the clinical and preoperative MRI data of 197 patients with IMCC confirmed by postoperative pathology.The region of interest(ROI)of the tumor was delineated on axial T2WI and DWI images,and extended outward by 3 mm,5 mm,10 mm,15 mm,and 20 mm respectively to obtain peritumoral ROIs of different ranges.Radiomics features were extracted by PyRadiomics.Features were screened through homogeneity of variance test,independent sample t-test,recursive feature elimination algorithm and least absolute shrinkage and selection operator.Logistic regression(LR)classifier and 5-fold cross-validation method were used for modeling and verification.Clinical imaging model,intratumoral omics model,peritumoral omics model,intratumoral+peritumoral omics model,dual-sequence fusion model and multimodal combined model were established.The predictive efficacies of each of the above models were compared to select the best model.Receiver operating characteristic(ROC)curve and area under the curve(AUC)were used to evaluate the performance of the model.DeLong test was used to compare the differences in AUC.Calibration curve was used to evaluate the fitting ability of the model,and decision curve was used to assess the clinical value of the model.Results:In the peritumoral omics models,the DWI 3 mm model shows the best performance,with AUCs of 0.836 and 0.777 in the training set and validation set respectively.Gender,age,lesion location,and vascular involvement are independent predictors of the pathological grade of IMCC.The AUCs of the clinical imaging model in the training set and validation set are 0.658 and 0.614 respectively.The intratumoral+3 mm peritumoral 3 mm model has the

关 键 词：肝内肿块型胆管癌 影像组学 瘤周 病理分级 磁共振成像 

分 类 号：R445.2[医药卫生—影像医学与核医学] R735.7[医药卫生—诊断学]