Clinical Application Of CT/MRI LI-RADS? V2017 For Diagnosis And Evaluation Of Hepatocellular Carcinoma

Objective: To provide a theoretical basis for LI-RADS update,improvement,and application in China.This study investigate the clinical application of LI-RADS v2017 in t he diagnosis and evaluation of hepatocellular carcinoma(HCC)by dynamic enhanced com puter tomography(DECT)and gadolinium ethoxybenzy diethylenetriamine pentaacetic acid magnetic resonance imaging(Gd-EOB-DTPA MRI).Material and Method: DECT and Gd-EOB-DTPA MRI data of patients with the risk factors for developing HCC(CT data: 68 patients incorporate 81 lesions;MRI data: 100 patients incorporate 128 lesions),or treated of HCC(CT data: 39 patients incorporate 48 lesions;MRI data: 40 patients incorporate 48 lesions)were retrospectively analyzed.According to LI-RADS v2017,two experienced radiologists,blinded to each other,assigned scores for CT and MRI data.When the radiologists were disagree on the classification of lesions,the final grades will be assessed by a higher-level radiologist.The Kappa value was used to evaluate the consistency between the different radiologists.Pathology or follow-up as the gold standard,LR-1、LR-2 was regarded as benign lesions,LR-4、LR-5 was considered as HCC;The LR-TR Nonviable was regarded as nonviable lesion,and LR-TR Viable was considered as viable lesion.The accuracy of LI-RADS v2017 was assessed by sensitivity(SEN),specificity(SPE),positive predictive value(PPV),negative predictive value(NPV).Then the chi-square is used test to compare CT and MRI in diagnosing and evaluating the accuracy of HCC.Paired chi-square test was used to compare the accuracy of the LI-RADS v2017 and free classification criteria on imaging data of HCC(68 lesions on CT,98 lesions on MRI)(P<0.05 for statistical difference).Result: 1)The risk factors for developing HCC in CT data.HCC 68;Hemangioma 4;Hepatocellular adenomas 3;Cholangiocellular carcinoma 2;combined hepatocellular and cholangiocarcinoma,Adenocarcinoma,DN,Liver abscess 1 each.According to LI-RADS v2017.LR-1,LR-2,LR-3,LR-4,LR-5,LR-M,LR-TIV were 1,5,9,32,28,3 and 3,respectively.The Kappa value of the intra-reader was 0.728(p<0.001).The SEN、SPE、PPV、NPV of DECT in the diagnosis of HCC using LI-RADS v2017 grading criteria were 100%,75%,96.67%,100%,respectively.LR-3 9(HCC 7,benign lesion 2).2)The risk factors for developing HCC in MRI data.HCC 98;DN 7;Cholangiocellular carcinoma,combined hepatocellular and cholangiocarcinoma,Hemangioma 5 each;RN,Hepatocellular adenomas,Inflammatory myofibroblastic tumor 2 each;Adenocarcinoma,Metastatic cancer 1 each;According to LI-RADS v2017.LR-1,LR-2,LR-3,LR-4,LR-5,LR-M,LR-TIV were 3,8,6,43,46,10 and 12,respectively.The Kappa value of the intra-reader was0.872 Gd-EOB-DTPA MRI(p<0.001).The SEN、SPE、PPV、NPV of in the diagnosis of HCC using LI-RADS v2017 grading criteria were 100%,73.33%,95.51%,100%,respectively.LR-3 6(HCC 1,benign lesion 5).3)The post-topical treatment of HCC in CTdata.Nonviable lesion 28;Viable lession 20.According to LI-RADS v2017,LR-TR Nonviable 27;LR-TR Equivocal 11;LR-TR Viable 10.The Kappa value of the intra-reader was 0.714(p<0.001).The SEN、SPE、PPV、NPV of DECT in the evaluation of HCC using LI-RADS v2017 grading criteria were 90.91%,100%,100%,96.30%,respectively.LR-TR Equivocal 11(viable lesion: 9,nonviable lesion: 2).4)The post-topical treatment of HCC in MRI data.Nonviable lesion 30;Viable lession 18.According to LI-RADS v2017,LR-TR Nonviable 23;LR-TR Equivocal 6;LR-TR Viable 19.The Kappa value of the intra-reader was 0.895(p<0.001).The SEN、SPE、PPV、NPV of evaluation in the diagnosis of HCC using LI-RADS v2017 grading criteria were 100%,92%,89.47%,100%,respectively.LR-TR Equivocal 11(viable lesion: 9,nonviable lesion: 2).5)There was no statistical difference between CT and MRI in the accuracy of the diagnosis and evaluation of HCC(P>0.05)when using the LI-RADS v2017.6).The accuracy rate of using LI-RADS classification criteria was higher that of using free classification criteria(CT: 85.29%,vs 66.18%,χ2 = 8.9,p<0.05;MRI: 86.74%,vs 74.49%,χ2=24.1,p<0.001).Conclusion:1)LI-RADS v2017 with CT and MRI has the strong validity accuracy and repeatability in diagnosis and evaluation of HCC.The possibility of HCC of LR-3 on CT data is very high,and the probability of viable tumor of HCC of LR-TR Equivocal on CT data is high,requiring further examination and treatment.For MRI of LR-3 and LR-TR Equivocal lesions,there is a high probability that they are benign lesions and have noviable tumor,regular follow-up is recommended.2)the accuracy rate of using LI-RADS v2017 classification criteria was higher that of using free classification criteria.