Determining the origin of liver metastases from liquid biopsy

Determining the origin of liver metastases from liquid biopsy Background Liver tumors are common and include primary and metastatic tumors. Exact differentiation of the tumor type is an essential step in choosing the optimal treatment. The most challenging is to distinguish among metastatic adenocarcinomas from various origins, and between metastatic adenocarcinomas and cholangiocarcinoma. This differentiation is sometimes difficult to make even if the most comprehensive clinical, laboratory, radiological, endoscopic and conventional pathological examinations are used, and the tumor is termed as cancer of unknown primary. Liver tumors are either primary tumors, including hepatocellular carcinoma and intrahepatic cholangiocarcinoma, or metastatic tumors, most commonly carcinomas, melanomas, lymphomas and sarcomas. It is sometimes difficult to distinguish between metastatic and primary liver carcinoma, particularly between metastatic adenocarcinoma and cholangiocarcinoma. However, given the different prognosis and treatment options, this discrimination is of vital importance. Carcinogenesis is accompanied by widespread genomic changes within the cell, including DNA alterations, protein expression and epigenetic changes (e.g. DNA methylation). These changes can be detected in circulating cancer byproducts in liquid biopsies: cell-free nucleic acids (cell-free tumor DNA, mRNA and miRNA), circulating tumor cells and extracellular vesicles. Many of these changes occur early in tumorigenesis and are highly pervasive across different tumor types. Therefore, a combination of different liquid biopsy biomarkers holds great promise for early cancer detection, primary tumor site discovery and treatment optimization. Hypotheses With bioinformatics analysis and machine learning methods we can identify genetic markers and patterns specific for each primary and metastatic liver tumor We can design custom-made genetic marker panel for discrimination between common malignant liver tumors and identify the origin of liver metastases Methods Our project proposes to use bioinformatics integration of genomics, transcriptomic and proteomics data for common primary and metastatic liver tumors, to decipher the diagnosis and primary tumor location. With bioinformatics tools, we will analyze available genomic data of approx. 2,000 samples of different primary tumor sites, which we will be used in further machine learning methods. This approach will help us uncover specific genomic patterns of each primary tumor and help us identify specific genomic biomarkers on which a custom-made marker panel will be designed. For clinical validation of panel, tissue and blood samples of patients with primary and metastatic liver tumor will be used. For detection of selected markers the next-generation sequencing, pyrosequencing and/or digital droplet PCR will be performed. Objectives To search for genomic and transcriptomic markers specific for a primary tumor with our own extensive bioinformatic analysis To identify genetic patterns for a specific primary tumor using cutting-edge machine learning methods To construct marker panels designed to discriminate among different primary and metastatic liver tumors To test the marker panels on tissue samples and liquid biopsy samples