To establish the diagnostic accuracy of the Liver Imaging Reporting and Data System (LI-RADS) for hepatocellular carcinoma (HCC) and compare its performance to that of international criteria from European Assofor the Study of the Liver (EASL), Japan Society of Hepatology (JSH), Asian Pacific Association for the Study of the Liver (APASL), and Organ Procurement and Transplantation Network (OPTN), and to the reporting radiologist's overall opinion regarding the probability of a nodule being a HCC by correlating with a histological diagnosis from whole liver explants. The present single-centre, retrospective review selected participants based on the following criteria: adults (≥18 years) listed for liver transplantation in 2014/2015, with liver cirrhosis at the time of magnetic resonance imaging (MRI) with hepatocyte specific contrast agent, and at least one liver lesion ≥10 mm on MRI with histology from subsequent liver explant for comparison. Each lesion was assessed against international criteria and given a "radiologist opinion" score of 1-5 (1=definitely benign, 5=definitely HCC). Total 268 patient records were reviewed, with 105 eligible lesions identified from 47 patients. Median lesion size was 15.5 mm (range 10-68 mm). Sensitivity (%), specificity (%), and positive predictive value (PPV; %) for LI-RADS LR5 was 45, 89, and 89, for LI-RADS LR4+5+TIV was 61, 80, and 86, for EASL was 44, 86 and 86, for JSH/APASL was 64, 81, and 87, for OPTN was 36, 90, and 88, and for "radiologist impression" of probably or definitely HCC was 79, 79, and 88 respectively. MRI has moderate sensitivity and good specificity for the diagnosis of HCC with considerable variation depending on criteria used. OPTN criteria have the best specificity, but low sensitivity. "Radiologist opinion" gives highest overall accuracy with increases in sensitivity and reduction in specificity when compared to the imaging criteria.
Read full abstract