LI-RADS M (LR-M) criteria and reporting algorithm of v2018: diagnostic values in the assessment of primary liver cancers on gadoxetic acid-enhanced MRI.

Abstract

To evaluate diagnostic values of the liver imaging reporting and data system (LI-RADS) M (LR-M) category based on novel explicit criteria that accept both targetoid and nontargetoid LR-M features and the suggested reporting algorithm of LI-RADS v2018 to assess primary liver cancers (PLCs) on gadoxetic acid-enhanced MRI (Gd-EOB-MRI). This retrospective study included 165 patients at high risk for hepatocellular carcinoma (HCC) with pathologically confirmed PLCs (HCC, n = 113; intrahepatic cholangiocarcinoma [iCCA], n = 23; and combined hepatocellular cholangiocarcinoma [cHCC-CCA], n = 29). Two radiologists independently analyzed Gd-EOB-MRI features and determined LI-RADS category for each tumor and categorized the likely etiology either as HCC or non-HCC malignancy if LR-M was assigned. Diagnostic performances for HCC or those for malignancy were compared according to imaging criteria. LR-M was assigned in 95.7%/91.3% of iCCAs; 55.2%/58.6% of cHCC-CCAs; and 21.2%/17.7% of HCCs in reviewers 1/2. Combination of LR-5 plus LR-M resulted in sensitivity of 95.2%/97.6% to diagnose PLCs as malignant, which were significantly higher than that of LR-5 plus "LR-M with ≥ 1 targetoid appearances" (84.8%/91.5%, Ps < 0.01). In comparison to LR-5, LR-5 plus "LR-M of HCC as likely etiology" resulted in significant increase in sensitivity (73.5%/79.6% versus 87.6%/92.9%, Ps < 0.001) but significant decrease in specificity (76.9%/75.0% versus 57.7%/50.0%, P = 0.002 and < 0.001) in the diagnosis of HCC. The LR-M criteria v2018 are useful to differentiate non-HCC malignancies from HCCs and to accurately diagnose PLCs as a malignancy. Reporting the likely etiology in LR-M may facilitate a more sensitive detection of HCC, but along with a considerable decrease in specificity.