Gd-EOB-DTPA增强MRI影像组学预测肝细胞癌Glypican-3阳性表达的价值  

The value of Gd-EOB-DTPA-enhanced MRI radiomics in predicting Glypican-3 positive expression in hepatocellular carcinoma

作  者:李尧森 戴慧[1] 冯蒙蒙 刘原庆 苗焕民[1] LI Yaosen;DAI Hui;FENG Mengmeng;LIU Yuanqing;MIAO Huanmin(Department of Radiology,the First Affiliated Hospital of Soochow University,Suzhou 215006,China;Department of Radiology,Wuxi Huishan Traditional Chinese Medicine Hospital,Wuxi 214100,China)

机构地区:[1]苏州大学附属第一医院放射科,苏州215006 [2]无锡市惠山区中医医院放射科,无锡214100

出  处:《磁共振成像》2025年第2期44-50,58,共8页Chinese Journal of Magnetic Resonance Imaging

基  金:江苏省科教能力提升工程项目(编号:JSDW202242)。

摘  要:目的研究基于钆塞酸二钠(gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid,Gd-EOB-DTPA)MRI影像组学预测肝细胞癌(hepatocellular carcinoma,HCC)的磷脂酰肌醇蛋白聚糖3(Glypican-3,GPC3)阳性表达。材料与方法回顾性收集分析2016年1月至2023年6月苏州大学附属第一医院126名HCC患者[GPC3阳性(77例)及阴性(49例)]的临床指标、MRI平扫及增强检查的影像资料,临床指标包括年龄、性别、乙肝感染情况,乙肝核心抗体、甲胎蛋白(alpha-fetoprotein,AFP)、糖类抗原199(carbohydrate antigen 199,CA199)、糖类抗原125(carbohydrate antigen 125,CA125),患者接受了肝切除术或穿刺活检术并且术前1个月内进行了Gd-EOB-DTPA MRI检查。在MRI影像横轴位五个序列[动脉期(arterial phase,AP)、门脉期(portal venous phase,PVP)、过渡期(transitional phase,TP)、肝胆特异期(hepatobiliary phase,HBP)和T2加权成像(T2-weighted imaging,T2WI)]上手动勾画出病灶的三维体积感兴趣区,提取病灶影像组学特征。经过Pearson相关性分析进行特征筛选,最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)回归进行特征降维,构建临床指标、单序列、多序列的逻辑回归模型,将临床指标联合多序列影像组学的特征子集构建综合模型(诺图),对HCC的GPC3阳性表达进行预测。校准曲线对综合模型进行验证,决策曲线分析(decision curve analysis,DCA)评估临床实用性。结果临床指标中AFP和乙肝感染情况与GPC3阳性表达独立相关,临床指标模型在训练集的受试者工作特征(receiver operating characteristic,ROC)曲线下面积(area under the curve,AUC)值为0.827(95%CI:0.742~0.913),在测试集的AUC值为0.779(95%CI:0.632~0.925)。磁共振单序列AP、PVP、TP、HBP及T2WI序列模型具有中等的预测性能,各序列在训练集的AUC值分别为0.804(95%CI:0.713~0.894)、0.801(95%CI:0.711~0.892)、0.796(95%CI:0.706~0.887)、0.761(95%CI:0.660~0.863)及0.733(95%CI:0.62Objective:To investigate the radiomics prediction of Glypican-3(GPC3)positive expression in hepatocellular carcinoma(HCC)based on gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid(Gd-EOB-DTPA)MRI.Materials and Methods:The clinical indicators,MRI plain scan and enhanced imaging data of 126 HCC patients with GPC3 positive(77 cases)and negative(49 cases)in the First Affiliated Hospital of Suzhou University from January 2016 to June 2023 were retrospectively collected and analyzed.The clinical indicators included age,gender,hepatitis B infection,hepatitis B core antibody,alpha-fetoprotein(AFP),carbohydrate antigen 199(CA199),carbohydrate antigen 125(CA125).The patients received hepatectomy or needle biopsy,and Gd-EOB-DTPA MRI was performed within one month before the operation.Manually delineate the three-dimensional volume of interest of the lesion on five sequences of arterial phase AP,portal venous phase PVP,transitional phase TP,hepatobiliary phase HBP,and T2 weighted imaging T2WI in the transverse axis of MRI images,and extract the radiomics features of the lesion.After Pearson correlation analysis feature screening,the least absolute shrinkage and selection operator(LASSO)regression feature dimensionality reduction was performed to construct logistic regression models for clinical indicators,single sequences,and multiple sequences.Clinical indicators were combined with feature subsets from multiple sequence omics to construct a comprehensive nomogram for predicting GPC3 positive expression in HCC.Perform calibration curves to validate the comprehensive model of clinical indicators combined with multi-sequence omics,and use decision curve analysis to evaluate clinical utility.Results:The infection of AFP and hepatitis B in the clinical indicators was independently related to the positive expression of GPC3.The area under the receiver operating characteristic curve(AUC)for the clinical indicator model training set is 0.827(95%CI:0.742 to 0.913),while the AUC for the test set is 0.779(95%CI:0.632 to 0.92

关 键 词:肝细胞癌 磷脂酰肌醇蛋白聚糖3 钆塞酸二钠 磁共振成像 影像组学 

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]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象