2005 Scientific Session Abstracts
Genotyping Significantly Improves Prognostic Modeling for Hepatocellular Carcinoma (HCC)
Igor Dvorchik, PhD ( dvorchikig@upmc.edu ) 1,3; Sydney D.Finkelstein,MD 2; J.Wallis Marsh, MD 1; 1Department of Surgery, Thomas E. Starzl Transplantation Institute, University of Pittsburgh School of Medicine, Pittsburgh, PA; 2RedPath Integrated Pathology, University of Pittsburgh, Pittsburgh, PA; 3Department of Biostatistics Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA
Context: Unlike expert knowledge-based pTNM staging system, currently utilized predictive models for HCC are based on analysis of large datasets and, therefore, type of information containing in these datasets becomes crucial for successful model development. This information does not include molecular markers of tumor progression. However, molecular markers of tumor progression are the strongest predictors of tumor recurrence currently available.
Technology: Molecular analysis was carried out blindly on the native liver specimens with respect to clinical outcomes. The fractional allelic loss (FAL), defined as the fraction of mutated markers among 13 markers considered, divided by the total number of informative markers, was determined in each case.
Design: The paraffin blocks from 183 patients that underwent liver transplantation in the presence of HCC between 1981 and 1997 at the University of Pittsburgh Medical Center (N=104) and Mt. Sinai School of Medicine (N=79) were collected and analyzed for FAL.
Results: FAL ranged from 0 to 89% with the mean of 30.8%. Cox regression and logistic regression models using 3- and 5-year recurrence rates as outcome were employed to identify independent predictors of tumor recurrence. Results served as the basis of development of the staging system. All three analyses found FAL, vascular invasion and largest tumor size to be independent predictors of tumor-free survival with FAL having the greatest influence on risk of recurrence.
Staging System:
I: FAL ≤ 20% and no macrovascular invasion
II: 20% < FAL ≤ 40% and no macrovascular invasion OR FAL ≤ 20% and Macrovascular invasion
III: 20% < FAL ≤ 40% with macrovascular invasion OR FAL > 40% and no Macrovascular invasion
IV: FAL > 40% with macrovascular invasion.
Three and five-year tumor-free survival rates were 94.3% and 92.8% for stage I; 79.5% and 74.8% for stage II; 46.9% and 23.4% for stage III and 10.5% and 0% for stage IV, respectively.
Conclusions: Staging system for HCC must incorporate genomic information.