• Anna Darnell, Alejandro Forner, Jordi Rimola, María Reig, Ángeles García-Criado, Carmen Ayuso, and Jordi Bruix.
• From the Department of Radiology, Barcelona Clinic Liver Cancer group, Hospital Clinic Barcelona, University of Barcelona, Spain (A.D., J.R., A.G.C., C.A.); Liver Unit, Barcelona Clinic Liver Cancer Group, Hospital Clinic Barcelona, Institut d'Investigacions Biomèdiques August Pi i Sunyer, University of Barcelona, c/Villarroel 170, Escala 7, Planta 3, 08036 Barcelona, Spain (A.F., M.R., J.B.); and Networked Biomedical Research Center in Hepatic and Liver Diseases, Barcelona, Spain (A.F., J.R., M.R., C.A., J.B.).
• Radiology. 2015 Jun 1; 275 (3): 698-707.

PurposeTo evaluate the diagnostic accuracy of the Liver Imaging Reporting and Data System (LI-RADS) with magnetic resonance (MR) imaging for hepatic nodules 20 mm or smaller detected during ultrasonographic (US) surveillance in patients with cirrhosis.Materials And MethodsBetween November 2003 and January 2010, patients with cirrhosis with a newly US-detected solitary hepatic nodule 20 mm or smaller were included in this institutional ethics committee-approved study. All patients provided written informed consent before the study; the need to obtain consent for reanalysis of the data was waived. Patients underwent MR imaging and fine-needle biopsy (the reference standard). Nodules without pathologic confirmation were followed up with MR imaging every 6 months. A LI-RADS category was retrospectively assigned to nodules seen at MR imaging. The diagnostic accuracy for each LI-RADS category was described by sensitivity, specificity, and positive and negative predictive values with 95% confidence intervals.ResultsFinal diagnoses of 133 nodules in 159 patients were as follows: 102 hepatocellular carcinomas (HCCs), three intrahepatic cholangiocarcinomas (ICCs), one neuroendocrine metastasis, and 27 benign lesions (median MR imaging follow-up, 95 months). None (0%) of five LI-RADS category 1 lesions, three (25%) of 12 category 2 lesions, 29 (69%) of 42 category 3 lesions, 24 (96%) of 25 category 4 lesions, and 44 (98%) of 45 category 5 lesions were HCCs. One category 3 lesion was ICC, one category 5 lesion was a neuroendocrine metastasis, and two (50%) of four lesions categorized as other malignancies were HCCs. In patients with nodules detected at surveillance US, LI-RADS category 4 criteria were as effective as category 5 criteria for HCC diagnosis. Combining both categories would improve sensitivity without impairing specificity or positive or negative predictive value for HCC diagnosis (42.3%, 98.2%, 97.8%, and 47.4% vs 65.4%, 96.4%, 97.1%, and 59.6%, respectively).ConclusionIn patients with cirrhosis with US-detected nodules 20 mm or smaller, both LI-RADS category 4 and category 5 have high specificity for HCC. In addition, a relevant proportion of lesions categorized as LI-RADS category 2 or 3 or as other malignancies were HCCs. Thus, active diagnostic work-up, including biopsy to allow prompt treatment, is recommended in such patients. Online supplemental material is available for this article.RSNA, 2015

24 keywords...

hide…