基于肝脏影像报告和数据系统肝细胞肝癌MRI征象判读的一致性

Inter-observer Agreement on MRI Signs of Hepatocellular Carcinoma Based on Liver Imaging Reporting and Data System

目的分析MRI检查中不同阅片者对2018版肝脏影像报告和数据系统(LI-RADS v2018)中定义的肝细胞肝癌(HCC)征象识别的一致性。资料与方法收集经病理诊断为HCC的56例患者的MRI增强扫描图像,由2位经过LI-RADS v2018内容培训且考核合格的医师分别对图像进行独立分析,判断LI-RADS v2018定义征象(非环状动脉期高强化、非周边廓清、强化包膜3个主要征象及其他12个倾向于恶性的辅助征象)是否出现,并进行LI-RADS分类。按照肿瘤最大径将病例分为3组,即<1.0 cm组、1.0~1.9 cm组和≥2.0 cm组;再依据采用的对比剂类型将病例分为普美显组及非普美显组。利用加权Kappa检验分析2位阅片者间分类结果及各征象识别的一致性。结果2位阅片者间LI-RADS分类结果一致性高(Kappa=0.71),非环状动脉期高强化、非周边廓清征象识别一致性高(Kappa=0.67、0.79),强化包膜识别一致性中等(Kappa=0.56)。肿瘤最大径1.0~1.9 cm组内分类及主要征象识别一致性高于最大径≥2 cm组(Kappa=0.85~0.95、0.38~0.56)。普美显组内分类结果一致性高于非普美显组(Kappa=1.00、0.69)。恶性辅助征象的识别一致性一般至极好(Kappa=0.46~1.00)。结论基于LI-RADS v2018的阅片者间分类及征象识别一致性较好。

Purpose To analyze the agreement of hepatocellular carcinoma(HCC)signs defined in liver imaging reporting and data system version 2018(LI-RADS v2018)between two different observers in MRI examination.Materials and Methods Dynamic enhanced MRI images of 56 HCC patients diagnosed by pathology were collected.The images were independently analyzed by two qualified doctors trained in LI-RADS v2018 program to judge the occurrence of signs(nonrim arterial phase hyper enhancement,nonperipheral washout and enhancing capsule as well as twelve ancillary imaging features favoring malignancy)defined by LI-RADS v2018 and classify the lesions.The cases were divided into three groups in line with maximal diameter of tumor,namely<1.0 cm group,1.0-1.9 cm group and≥2.0 cm group.Then the cases were divided into Primovist group and non-Primovist group according to the contrast taken.Weighted Kappa testing was used to analyze the consistency of classification results and recognition of each sign between the two observers.Results LI-RADS classification results of the inter-observer were of high consistency(Kappa=0.71)and identification of nonrim arterial phase hyper enhancement and non-peripheral washout was also of high consistency(Kappa=0.67,0.79).In the meantime,identification of enhancing capsule was of moderate consistency(Kappa=0.56).Classification in 1.0-1.9 cm group with maximal diameter of tumor and consistency of major signs identification were higher than that in≥2 cm group with maximal diameter(Kappa=0.85-0.95 and 0.38-0.56).Consistency of classification results in Primovist group was higher than that in non-Primovist group(Kappa=1.00,0.69).Identification consistency of ancillary imaging features favoring malignancy was general to very good(Kappa=0.46-1.00).Conclusion Consistency of classification and sign identification of inter-readers based on LI-RADS v2018 is quite good.