INTRODUCTION: The investigation of prognostic and predictive factors for early diagnosis of tumors, their surveillance and monitoring of the impact of therapeutic modalities using hybrid laboratory models in vitro/in vivo is an experimental approach with a significant potential. It is preconditioned by the preparation of in vivo tumor models, which may face a number of potential technical difficulties. The assessment of technical success of grafting and xenotransplantation based on the type of the tumor or cell line is important for the preparation of these models and their further use for proteomic and genomic analyses. METHODS: Surgically harvested gastrointestinal tract tumor tissue was processed or stable cancer cell lines were cultivated; the viability was assessed, and subsequently the cells were inoculated subcutaneously to SCID mice with an individual duration of tumor growth, followed by its extraction. RESULTS: We analysed 140 specimens of tumor tissue including 17 specimens of esophageal cancer (viability 13/successful inoculations 0), 13 tumors of the cardia (11/0), 39 gastric tumors (24/4), 47 pancreatic tumors (34/1) and 24 specimens of colorectal cancer (22/9). 3 specimens were excluded due to histological absence of the tumor (complete remission after neoadjuvant therapy in 2 cases of esophageal carcinoma, 1 case of chronic pancreatitis). We observed successful inoculation in 17 of 28 tumor cell lines. CONCLUSION: The probability of successful grafting to the mice model in tumors of the esophagus, stomach and pancreas is significantly lower in comparison with colorectal carcinoma and cell lines generated tumors. The success rate is enhanced upon preservation of viability of the harvested tumor tissue, which depends on the sequence of clinical and laboratory algorithms with a high level of cooperation.Key words: proteomic analysis - xenotransplantation - prognostic and predictive factors - gastrointestinal tract tumors.