肝脏影像报告和数据系统中恶性辅助征象对小肝细胞癌的分类及诊断效能的影响

Effect of malignant ancillary features in Liver Imaging Reporting and Data System on the classification and diagnostic performance of small hepatocellular carcinoma

目的探究基于钆塞酸二钠增强MRI(Gd-EOB-MRI)2018版肝脏影像报告和数据系统(LI-RADS v2018)恶性辅助征象对<20 mm的小肝细胞癌(HCC)分类及诊断效能的影响。方法回顾性收集未经治疗且经病理证实为HCC、再生结节(RN)、高级别异型增生结节(HGDN)、低级别DN(LGDN)109例病人,共纳入<20 mm病灶130个。采用卡方检验或Fisher确切概率检验比较各组病灶的主要和辅助征象并筛选HCC特异性辅助征象。按以下方法将病灶分类:(1)主要征象分类法;(2)主要+辅助征象分类法,即先依据主要征象分类,再由辅助征象调整;(3)添加主要征象分类法,即将HCC特异性辅助征象添加为主要征象,再按方法2分类。分别计算3种分类方法中LR-5和LR-(4+5)对HCC的诊断效能,并采用McNemar’s检验比较其差异。结果移行期(TP)及肝胆期(HBP)低信号为HCC特异性辅助征象(均P<0.05),被添加为主要征象。主要征象分类和主要+辅助征象分类下LR-5对HCC的诊断敏感度、特异度、准确度相同(67.0%、75.8%、74.6%),与添加主要征象分类法(74.2%、78.8%、70.0%)比较,仅诊断敏感度降低(P<0.05),而特异度及准确度差异无统计学意义(P>0.05)。主要+辅助征象分类和添加主要征象分类下LR-(4+5)对HCC的诊断敏感度、特异度、准确度相同(88.7%、75.4%、36.4%),与仅依据主要征象分类法(83.5%、78.5%、63.6%)比较,其敏感度及准确度差异均无统计学意义(P>0.05),但特异度降低(P<0.05)。结论经辅助征象调整后的分类不能改变LR-5对小HCC的诊断效能,却降低了LR-(4+5)诊断HCC的特异度。将TP/HBP低信号作为新的HCC主要征象重新分类,可以提高LR-5诊断HCC的敏感度,但特异度和准确度不降低。

Objective To investigate the effects of malignant ancillary features in gadoxetate disodium-enhanced MRI within the Liver Imaging Reporting and Data System version 2018(LI-RADS v2018)on the classification and diagnostic performance of small hepatocellular carcinoma(HCC)<20 mm.Methods A total of 109 untreated patients with pathologically confirmed HCC,regenerative nodules(RN),high-grade dysplastic nodules(HGDN),and low-grade DN(LGDN)were retrospectively collected,and a total of 130 lesions<20 mm were included.The chi-squared test or Fisher’s exact probability test was used to compare the major and ancillary features of lesions among the groups and to screen for HCCspecific ancillary features.Lesions were classified according to the following methods:(1)major features only classification,(2)major+ancillary features classification(classification based on major features,then adjusted by ancillary features),and(3)adding major features classification,i.e.,adding HCC-specific ancillary features to major features and then classifying the lesions according to method 2.The diagnostic efficacy of LR-5 and LR-(4+5)in the three classifications for HCC was calculated,and the differences were compared using McNemar’s test.Results Transitional phase(TP)and hepatobiliary phase(HBP)hypointensity were selected as HCC-specific ancillary features(both P<0.05)and added to the major features of HCC.The diagnostic sensitivity,specificity,and accuracy of LR-5 for HCC were consistent with the classification methods of using only major features and major+ancillary features(67.0%,75.8%,and 74.6%).Compared with adding major features(74.2%,78.8%,70.0%),its sensitivity was decreased(P<0.05),while specificity and accuracy were not significantly different(P>0.05).LR-(4+5)had same diagnostic sensitivity,specificity,and accuracy for HCC in the method of major+ancillary features and adding major features(88.7%,75.4%,36.4%).Compared with the method of using only major features(83.5%,78.5%,63.6%),there were no significant differences in sensitivity and accuracy(P>0.05),but the specificity was significantly lower(P<0.05).Conclusions The classifications adjusted by ancillary features did not change the diagnostic performance of LR-5 for small HCC,but reduced the specificity of LR-(4+5)for HCC diagnosis.When TP and HBP hypointensity were added as the major features of HCC,the sensitivity of LR-5 in diagnosing HCC could be improved without reducing specificity and accuracy.