Pancreatic cancer is a malignant disease with an unfavorable prognosis. Currently, the diagnosis of pancreatic cancer is based mainly on CT, MR and EUS imaging methods, because a reliable biomarker of pancreatic cancer is not available. It is considered a success in the current clinical practice if patients suitable for surgical resection can be selected, because in this group the overall survival is slightly better than in the group where surgery is not technically possible. In the first part, the presented dissertation focuses on testing of already established imaging diagnostic methods used in common practice, i.e. EUS and EUS FNA, as well as modern examination possibilities, such as contrast endosonography, and the comparison of the accuracy of the methods used. The dominant role of the EUS FNA was confirmed. In the next phase of the work, the methodological procedure of EUS FNA collection and the preprocessing of pancreatic tissue samples was tested with regard to the quantity and quality for epigenetic examination and further testing using the prognostic role of KRAS and miR-21. It has been suggested that cytological smears are the most suitable source of biological material for DNA and miRNA analysis, where non-tumor tissue contamination is low. Although the prognostic role of KRAS is...