Min Jung Park, Presenter: Nothing to Disclose

Myeong-Jin Kim MD, PhD, Abstract Co-Author: Nothing to Disclose

Yong Eun Chung MD, PhD, Abstract Co-Author: Nothing to Disclose

Chansik An MD, Abstract Co-Author: Nothing to Disclose

Hyungjin Rhee MD, Abstract Co-Author: Nothing to Disclose

To retrospectively assess whether gadoxetic acid-enhanced MRI can be applicable to the non-invasive diagnosis of HCC using AASLD and LI-RADS systems

A total of 124 hepatic nodules (≤ 5cm in diameter) in 109 patients at-risk for HCCs who had no history of liver tumor treatment were analyzed. Three observers independently assigned LI-RADS categories (1-5 and M) and AASLD scores (adapted from Bruix et al; 5 as definite HCC), and recorded major (assessment of washout in portal venous phase) and ancillary features (including hypointensity in hepatobiliary phase) of HCC for each lesion. Standard reference for diagnosis was histopathology other than 18 of 26 benign lesions diagnosed based on clinical and imaging features. Scores were compared by using McNemar test. Interobserver agreement was assessed by using multirater Fleiss κ statistics.

On surgery or biopsy, 94 of 124 (75.8%) nodules were confirmed as HCC. Sensitivity and specificity for LI-RADS in three observers were 55.3-67.0% and 83.3-90.0%, respectively, and those for AASLD were 67.0-76.6% and 83.3-86.7%, respectively. AASLD showed significantly higher sensitivity than LI-RADS in two observers (P ≤ 0.003), but no significant difference for specificity. When considering LR-4 and LR-5 as HCC, sensitivity and specificity for LI-RADS were 85.1-87.2% and 70.0-86.7%, respectively, and showed significantly higher sensitivity (P ≤ 0.035) than AASLD in all observers, but no significant difference for specificity. Lesions that were false positive for both systems (when LR-4 and LR-5 as HCC in LI-RADS) in all observers were cholangiocarcinoma (n=1), combined hepatocellular and cholangiocarcinoma (n=1) and dysplastic nodule (n=2). Interobserver agreement for categories consistent with HCC was moderate or good (LR-4 [κ=0.44], LR-5 [κ=0.68], AASLD-5 [κ=0.74]).

Gadoxetic acid-enhanced MRI can be applicable to the non-invasive diagnosis of HCC using AASLD and LI-RADS systems, yielding moderate sensitivity and specificity.

Gadoxetic acid-enhanced MRI can be utilized for scoring LI-RADS and AASLD systems with acceptable diagnostic performances.

Park, M, Kim, M, Chung, Y, An, C, Rhee, H, Applicability of Gadoxetic Acid-enhanced MRI for Non-invasive Diagnosis of Hepatocellular Carcinoma (HCC) Using American Association for the Study of Liver Diseases (AASLD) and Liver Imaging Reporting and Data System (LI-RADS) Systems.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14016558.html