Study of Hepatocellular Carcinoma in Cirrhotic Patients

Overview

The project is based on a case-control study including cirrhotic patients with (200 cases) and without (400 controls) Hepatocellular carcinoma. The determination of sample sizes in proteomic or spectroscopic studies has to be adapted to the high dimensional setting. The proteomic analysis will be conducted by Clinical Innovation Proteomic Platform of Dijon. Two approaches will be used in the proteomic study: A global approach based on the comparison of proteomic spectra profiles obtained after mass spectrometry analysis (MALDI-TOF). The particularity of this study with regard to previous studies is : the procedures used to purify the sub proteomes (Five automated methods including depletion fractionation and purification will be applied), the qualification of the generated data with the introduction of quality controls, the high number of samples included in the study. The second approach BIA-MS (Biomolecular interaction analysis mass spectrometry) is targeted approach allows the capture, quantification and characterization of proteins. The quality controls allow to quantify the various variability sources and to validate that biological variability is higher than technical variability. All the samples will be treated and analyzed with the same protocols, 100 samples will be used to validate the marker and statistical models developed after analysis of the first 500 samples. The infra-red spectroscopy analysis will be conducted by MéDIAN team, CNRS UMR 6237 of Reims university. The first 300 samples after feature selection reference spectra, are classified into different classes by means of mathematical classification methods such as multivariate statistical processes of pattern recognition, neuronal networks, support vector machines and methods of case-based classification or machine learning, genetic algorithms or methods of evolutionary programming. The analysis of a second set of samples (300) will validate the different mathematical classification methods developed. In the global study the investigators will unravel the relationship between proteomic, spectroscopic and metabolic/nutritional data. The description of these relationships will use canonical analysis and multi-block analysis in a more general extent. The goal of these methods is to explore relationships that may exist between several groups of quantitative variables observed on the same set of individuals.