The role of ancillary features for diagnosing hepatocellular carcinoma on CT: based on the Liver Imaging Reporting and Data System version 2017 algorithm

• Published in: Clinical radiology Vol. 75; no. 6; pp. 478.e25 - 478.e35
• Main Authors: Ren, A.-H., Du, J.-B., Yang, D.-W., Zhao, P.-F., Wang, Z.-C., Yang, Z.-H.
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.06.2020
• Subjects: Adult; Aged; Aged, 80 and over; Algorithms; Carcinoma, Hepatocellular - diagnostic imaging; Female; Humans; Liver Neoplasms - diagnostic imaging; Male; Middle Aged; Radiology Information Systems; Retrospective Studies; Sensitivity and Specificity; Tomography, X-Ray Computed - methods
• ISSN: 0009-9260; 1365-229X
• DOI: 10.1016/j.crad.2019.08.031

To investigate the diagnostic performance of Liver Imaging Reporting and Data System (LI-RADS) version 2017 for diagnosing hepatocellular carcinoma (HCC), by using major features only and combined major and ancillary features on computed tomography (CT). A total of 147 HCC, 35 non-HCC malignancy, and 37 benign lesions in 205 patients at high risk of HCC were evaluated retrospectively, and the diagnostic performance of LI-RADS for diagnosing HCC were compared between using major features only and adopting major and ancillary features in combination. When using LR-5 as a predictor for diagnosing HCC, the diagnostic specificity (90.3% versus 91.7%), positive predictive value (92.3% versus 93.3%), and accuracy (68% versus 68.8%) were increased based on major and ancillary features in combination than just using major features on CT. When using LR-4/5 as a predictor for diagnosing HCC, the diagnostic sensitivity (78.9% versus 85.7%), negative predictive value (64.4% versus 72%), and accuracy (78.5% versus 82.2%) were increased while preserving a high specificity (77.8% versus 75%), according to major and ancillary features in combination rather than just using major features on CT. The LI-RADS categories of 8.7% (19/219) lesions were adjusted by adding the ancillary features on CT. Adding the ancillary features visible on CT can improve the diagnostic performance of the LI-RADS v2017 algorithm for diagnosing HCC, especially for LR-3 lesions. •Adding ancillary features on CT can improve the diagnostic performance of HCC.•Mosaic architecture is the most frequent ancillary feature for HCC on CT.•Nonenhancing capsule is the second frequent ancillary feature for HCC on CT.•8.7% lesions adjusted LI-RADS categories by adding the ancillary features on CT.