Document Type


Original Publication Date


Journal/Book/Conference Title




DOI of Original Publication



Originally Published at

Date of Submission

November 2014



Early hepatocellular carcinoma (HCC) detection is difficult because low accuracy of surveillance tests. Genome-wide analyses were performed using HCV-cirrhosis with HCC to identify predictive signatures.

Methodology/Principal Findings

Cirrhotic liver tissue was collected from 107 HCV-infected patients with diagnosis of HCC at pre-transplantation and confirmed in explanted livers. Study groups included: 1) microarray hybridization set (n = 80) including patients without (woHCC = 45) and with (wHCC = 24) HCC, and with incidental HCC (iHCC = 11); 2) independent validation set (n = 27; woHCC = 16, wHCC = 11). Pairwise comparisons were performed using moderated t-test. FDR<1% was considered significant. L1-penalized logistic regression model was fit for woHCC and wHCC microarrays, and tested against iHCC. Prediction model genes were validated in independent set by qPCR. The genomic profile was associated with genetic disorders and cancer focused on gene expression, cell cycle and cell death. Molecular profile analysis revealed cell cycle progression and arrest at G2/M, but progressing to mitosis; unregulated DNA damage check-points, and apoptosis. The prediction model included 17 molecules demonstrated 98.6% of accuracy and correctly classified 6 out of 11 undiagnosed iHCC cases. The best model performed even better in the additional independent set.


The molecular analysis of HCV-cirrhotic tissue conducted to a prediction model with good performance and high potential for HCC surveillance.


Copyright: © 2012 Gehrau et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Is Part Of

VCU Biostatistics Publications (22568 kB)