BOSTON — A gene signature that predicts survival after surgery for hepatocellular carcinoma also predicts the outcome of liver cirrhosis, a study has shown.
Compensated cirrhotic patients at risk of poor prognosis can be identified via the 186-gene signature of nontumor liver tissue, Dr. Yujin Hoshida reported at the annual meeting of the American Association for the Study of Liver Diseases.
The investigators performed whole-genome gene expression analysis of liver biopsy specimens obtained from 276 patients with compensated cirrhosis who were included in a prospective surveillance study for hepatocellular carcinoma, said Dr. Hoshida of the Broad Institute of Massachusetts Institute of Technology and Harvard University, Cambridge Boston.
“The findings suggest that we might be able to identify patients who need preventive treatment for advanced cirrhosis and possibly hepatocellular carcinoma,” Dr. Hoshida said.
He and his colleagues previously developed and reported (N. Engl. J. Med. 2008;359:1995-2004) a technique for globally profiling gene expression from formalin-fixed, paraffin-embedded (FFPE) tissues. The method enabled the researchers to profile both nontumor and cancerous tissue obtained from 307 liver cancer patients participating in prospective surveillance studies.
The technique involved a modified cDNA-mediated annealing, selection, extension, and ligation assay to interrogate about 6,000 genes expressed in the tumor and nontumor tissue. By partitioning samples into training and validation sets, the researchers were able to develop a 186-gene signature of nontumor tissues to predict hepatocellular carcinoma survival.
In the current study, the investigators used Cox regression modeling to evaluate the potential associations between the 186-gene signature and overall survival, hepatocellular carcinoma, and hepatic decompensation in patients with compensated cirrhosis.
Almost all (98%, or 270) of the 276 patients from whom biopsy specimens were obtained were classified as Child-Pugh A with respect to cirrhosis severity, Dr. Hoshida said. About 90% of the participants had hepatitis C infection, and the median baseline serum alpha-fetoprotein level was 6 mg/dL. Most patients (62%) were male.
During the prospective surveillance period (median follow-up of 9.8 years), 90 patients (33%) died, 81 (29%) developed hepatocellular carcinoma, and 88 (32%) developed hepatic decompensation.
In multivariate analyses, the 186-gene signature was associated with overall survival, with a hazard ratio (HR) of 2.2. It also was associated with hepatocellular carcinoma development (HR 1.6) and hepatic decompensation (HR, 2.1). The association between the gene signature and each of the three outcomes remained significant after adjustment for bilirubin greater than 1.0 mg/dL and platelet count, Dr. Hoshida said.
Subsequent gene set enrichment analysis revealed enrichment of metabolic-related pathways in patients with a good prognosis and enrichment of pathways associated with inflammation (including those related to interferon and tumor necrosis factor–alpha signaling) in patients with a poorer prognosis, he noted.
Although additional clinical validation is required before the genetic signature is introduced into clinical practice, the findings are promising, said Dr. Hoshida, who noted that the prognostic prediction of early-stage cirrhosis will increase the opportunity for early medical intervention, including chemoprevention of advanced cirrhosis and hepatocellular carcinoma.
Disclosures: Dr. Hoshida reported having no relevant conflicts of interest.
My Take
Clinical Implications Not Yet Clear
This study, largely involving hepatitis C patients, adds to our knowledge of factors that influence the course of hepatic cirrhosis. The ability to predict which patients with cirrhosis are likely to decompensate or develop hepatocellular carcinoma (HCC) should be helpful in clinical management.
The study shows that gene enrichment for metabolic-related pathways predicts a better outcome for patients with cirrhosis, compared with gene enrichment for inflammatory pathways—a finding that intuitively fits with existing clinical information related to prognosis. It also seems that a greater emphasis on following those patients at greatest risk for HCC could reduce the costs of screening, if we could reduce screening of patients with less risk. This genetic risk information also might permit earlier intervention for patients suspected of having an HCC, or those at risk of decompensation.
But would this information change how we deal with an individual patient? Would it benefit patients to know that they are in the group with a worse prognosis? We need further studies to see if this can truly change how we manage the individual patient with cirrhosis.
ROWEN K. ZETTERMAN, M.D., a gastroenterologist, is dean of the school of medicine at Creighton University, Omaha, Neb. He reports no relevant conflicts of interest.
