Abstract： Objective:To compare the diagnostic performance in the hepatocellular carcinoma(HCC) with cirrhosis between the 2017 version of liver imaging reporting and data system (LI-RADS v2017) and 2018 version of LI-RADS (LI-RADS v2018) based on gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA) enhanced MRI.Methods:Clinical data of 213 patients with 246 hepatic lesions with cirrhosis who underwent Gd-EOB-DTPA enhanced MRI in the Third Affiliated Nantong Hospital of Nantong University from October 2015 to July 2020 were retrospectively collected. The MRI major features and LR categories of lesions were respectively reviewed by two radiologists according to LI-RADS v2017 and LI-RADS v2018, respectively. Taking postoperative histopathological results or follow-up imaging as references, with the LR-5 and LR-4+LR-5 as the diagnosis of HCC, the sensitivity, specificity and accuracy of the LI-RADS v2017 and LI-RADS v2018 were evaluated, respectively. The McNemar test or Fisher exact test was used to compare the diagnostic performance between the two LI-RADS versions.Results:In 246 hepatic lesions, 165 were HCCs, 31 were non-HCC malignancies and 50 were benign lesions. Due to the threshold growth and more simplified definition and changes in the LR-5 classification criteria in LI-RADS v2018, the categories of 38 (15.4%, 38/246) lesions were changed. The threshold growths of 84.6% (33/39) lesions in v2017 were reclassified to subthreshold growth in v2018. Using LI-RADS v2018, 10 lesions were down-categorized compared with LI-RADS v2017, including LR-5 to LR-4 in 7 lesions and LR-4 to LR-3 in 3 lesions, and 28 lesions were up-categorized LR-4 to LR-5, in which 25 were small HCC. With LR-5 as the diagnosis criteria of HCC, the sensitivity and accuracy of LI-RADS v2018 were 66.7% (110/165) and 73.6% (181/246); and the sensitivity and accuracy of LI-RADS v2017 were 55.8% (92/165) and 67.5% (166/246), both with statistical differences (χ2=4.13, P=0.001, χ2=6.20, P<0.001). No significant difference was found in the specificity values of LI-RADS v2018 and v2017 [87.7% (71/81) vs. 91.4% (74/81)], χ2=0.59, P=0.442). Compared with v2017, LI-RADS v2018 increased the sensitivity in the diagnosis of small HCC lesions (10-19 mm) [62.9% (56/89) vs. 40.4% (36/89), χ2=9.00, P<0.001]. With LR-4+LR-5 as the diagnostic criteria of HCC, there was no significant difference in the sensitivity, specificity and accuracy of LI-RADS v2017 and v2018 in the diagnosis of HCC (all P>0.05). Conclusions:Based on Gd-EOB-DTPA enhanced MRI, LI-RADS v2018 has higher sensitivity and similar specificity in the diagnosis of HCC compared to v2017, especially in the diagnosis of small HCC (10-19 mm).