| Objective The objective of this study is to retrospectively evaluate the diagnostic accuracy of the Liver Imaging Reporting and Data System(LI-RADS) by enhanced computer tomography(CT) and magnetic resonance(MR) imaging for hepatocellular carcinoma(HCC) in patients with cirrhosis or other risk factors for developing HCC.Method CT and MR imaging data of patients with cirrhosis or other risk factors for developing HCC were retrospectively analyzed. Two experienced radiologists,blinded to each other, assigned scores for CT data; one experienced radiologist assigned scores for MR data and CT&MR data, and redid it 1 month later, according to LI-RADS v2014. The Kappa value was used to evaluate the consistency between the two radiologists and the two assigns by the same radiologist. In order to avoid the cluster effect, we chose the malignant lesion and the largest one when benign and malignant lesions were present in the same patient. LI-RADS categorizes observations from LR-1 to LR-5, reflecting probability of benignity or HCC. LR-1 and LR-2categories have little risk for HCC, LR-3 category means intermediate probability for HCC, LR-4 categories suggest high probability for HCC, and LR-5 and LR-5V mean definitely HCC. LR-1 and 2 categories were regarded as negative reports, and LR-4,LR-5 and LR-5V as positive. Compared with reference results, the diagnostic accuracy of LI-RADS was assessed by sensitivity, specificity, and positive and negative predictive values, respectively.Results We enrolled 176 patients completed with enhanced CT. Compared with the reference standard, there are 140 HCCs, 4 intrahepatic cholangiocarcinomas, 1 hepatic adenocarcinoma, 4 metastasis and 27 benign lesions. The LI-RADS categories of the observations include LR-1 19(10.8%), LR-2 3(1.7 %), LR-3 20(11.4 %), LR-418(10.2%), LR-5 77(43.8%), LR-5V 34(19.3%) and LR-M 5(2.8%). 15 of 20 cases LR-3 corresponded to HCC, 3 cases remained stable during follow-up, 1 case was diagnosed as metastasis and 1 as intrahepatic cholangiocarcinoma. The Kappa value of the inter-reader was 0.863(P<0.001). The overall sensitivity, specificity, and positive and negative predictive values were 100.00%, 84.06%, 96.90%, and 100.00%,respectively.We enrolled 73 patients completed with enhanced MR. Compared with the reference standard, there are 59 HCCs, 1 intrahepatic cholangiocarcinoma, 3metastasis and 10 benign lesions. The LI-RADS categories of the observations include LR-1 6(8.2%), LR-2(0.0%), LR-3 4(5.5%), LR-4 21(28.8%), LR-5 31(37.9%), LR-5V 6(8.2%) and LR-M 5(6.8%). 1 of 4 cases LR-3 corresponded to HCC, and 3 cases remained stable during follow-up. The Kappa value of the inter-reader was 0.942(P<0.001). Except the LR-3 and LR-M, the coincidence rate of LI-RADS diagnosing HCC by MR compared with the referenced standard was 100%.We enrolled 45 patients completed with enhanced CT and MR simultaneously.Compared with the reference standard, there are 32 HCCs, 1 intrahepatic cholangiocarcinoma, 3 metastasis and 9 benign lesions. The LI-RADS categories of the observations with CT include LR-1 6(13.3%), LR-2 0(0.0%), LR-3 14(31.1%),LR-4 7(15.6%), LR-5 11(24.4%), LR-5V 3(6.7%) and LR-M 4(8.9%). 11 of 14 cases LR-3 corresponded to HCC, 3 cases remained stable during follow-up. The Kappa value of the intra-reader was 0.916(P<0.001). The LI-RADS categories of the observations with MR include LR-1 6(13.3%), LR-2 0(0.0%), LR-3 4(8.9%), LR-415(33.3%), LR-5 13(28.9%), LR-5V 3(6.7%) and LR-M 4(8.9%). 1 of 4 cases LR-3corresponded to HCC, 3 cases remained stable during follow-up. The Kappa value of the intra-reader was 1.000(P<0.001). Except the LR-3 and LR-M, the coincidence rate of LI-RADS diagnosing HCC by CT or MR compared with the referenced standard was 100%. However, there are 10 cases of HCC were classified as LR-3 by CT but as LR-4 or LR-5 by MR, which suggests the accuracy of LI-RADS with MR is higher than CT for diagnosing HCC.Conclusion The LI-RADS with enhanced CT and MR has the very good repeatability and high accuracy in the evaluation for HCC, and can be reliably adopted in clinical practice. The application of LI-RADS with MR acquired more accurate for diagnosis of HCC. The possibility for HCC of LR-3 observations on CT is quite high, thus,Gd-DTPA enhanced MRI should be recommended for such patients. When LR-3observations categorized with MR, regular follow-up should be suggested. |
PDF Full Text Request