CT and MRI Liver Imaging Reporting and Data System Version 2018 for Hepatocellular Carcinoma: A Systematic Review With Meta-Analysis

Purpose The aim of this study was to determine the diagnostic performance of the LR-5 category for hepatocellular carcinoma (HCC) and the pooled proportion of HCC in each Liver Imaging Reporting and Data System (LI-RADS) category with CT and MRI, using LI-RADS version 2018. Methods The MEDLINE, Embase, and Scopus databases were searched from inception to December 7, 2019, for studies reporting the diagnostic accuracy of LI-RADS version 2018 for HCC. Risk for bias and concerns regarding applicability were assessed using the Quality Assessment of Diagnostic Accuracy Studies–2 tool. Random-effects models were used to determine the summary estimates of the diagnostic performance of LR-5 and the pooled proportion of HCC for each LI-RADS category. Results Fourteen studies were included in the final analysis, consisting of 2,708 observations with 1,841 HCCs. The pooled per-observation sensitivity and specificity of the LR-5 category for diagnosing HCC were 70% (95% confidence interval [CI], 61%-78%) and 91% (95% CI, 89%-93%), respectively. No HCCs were reported for LR-1 and LR-2. The pooled proportions of HCC were 31% (95% CI, 12%-50%) for LR-3, 64% (95% CI, 47%-80%) for LR-4, 95% (95% CI, 93%-96%) for LR-5, 54% (95% CI, 30%-77%) for LR-TIV, and 33% (95% CI, 21%-46%) for LR-M. The proportions of HCC were significantly different among the LI-RADS categories (P = .022). Conclusions The LR-5 category of LI-RADS version 2018 provided moderate sensitivity and high specificity for diagnosing HCC. Higher LI-RADS categories from LR-3 to LR-5 included greater proportions of HCC.