Keywords:
Abdomen, Liver, Oncology, CT, MR, MR-Diffusion/Perfusion, Staging, Structured reporting, Observer performance, Cancer, Neoplasia, Pathology
Authors:
A. Borgheresi, A. Agostini, D. Nicolini, R. Montalti, M. Vivarelli, A. Lorenzoni, A. Giovagnoni; Ancona/IT
DOI:
10.26044/ecr2019/C-3707
Aims and objectives
The diagnosis of Hepatocellular Carcinoma (HCC) strongly relies on cross-sectional imaging according to the 2018 European Association for the Study of the Liver (EASL) Guidelines [1].
Conversely,
the most important prognostic factors in terms of recurrence are the Edmondson-Steiner Grade (E-S) [2,3] and Microvascular Invasion (MVI) [2,4] evaluated on the pathological specimen.
Unfortunately,
the preoperative biopsy is not considered a reliable tool [5].
Several studies have correlated different radiological findings to post-surgical or post-transplant recurrence of HCC [6-10],
but at now their definition is not completely standardized and reproducible.
Therefore,
the most important prognostic information is still obtained from the surgical specimen or explanted liver.
A radiological marker predicting pathological features would allow for more accurate patient stratification and treatment selection [11,12].
Liver Imaging - Reporting and Data Systems (LI-RADS) have been developed with the main purpose of standardization of liver imaging in patients with chronic liver disease at risk of HCC.
This algorithm is still under construction; the last version has been released recently [13].
The key feature of LI-RADS is the defined lexicon and the exact definition of the lesion pattern,
aiming to an unequivocal characterization of the observations with improved reproducibility [14].
The purpose of this study was to evaluate the inter-reader agreement (IRA) between readers with different experience and to determine the correlation of CT/MRI LI-RADSv2018 standardized imaging observations with pathological features of HCC on the surgical specimen in cirrhotic patients.