Background: Autologous dendritic cells (DC) loaded with tumor-associated antigens (TAAs) are a promising approach for anticancer immunotherapy. Polyantigen lysates appear to be an excellent source of TAAs for loading onto the patient's dendritic cells. Cancer/testis antigens (CTA) are expressed by a wide range of tumors, but are minimally expressed on normal tissues, and could serve as a universal target for immunotherapy. However, CTA expression levels can vary significantly in patients with the same tumor type. We proposed that patients who do not respond to DC-based therapy may have distinct features of the CTA expression profile on tumor cells. Patients and methods: We compared the gene expression of the principal families CTA in 22 melanoma and 27 soft tissue and bone sarcomas cell lines (STBS), received from patients and used for DC vaccine preparation. Results: The majority (47 of 49, 95.9%) cell lines showed CTA gene activity. The incidence of gene expression of GAGE, NYESO1, MAGEA1, PRAME's was significantly different (adj. p < 0.05) between melanoma and sarcoma cell lines. The expression of the SCP1 gene was detected neither in melanoma cells nor in the STBS cells. Clustering by the gene expression profile revealed four different expression patterns. We found three main patterns types: hyperexpression of multiple CTA, hyperexpression of one CTA with almost no expression of others, and no expression of CTA. All clusters types exist in melanoma and sarcoma cell lines. We observed dependence of killing efficacy from the PRAME (rho = 0.940, adj. p < 0.01) expression during real-time monitoring with the xCELLigence system of the interaction between melanoma or sarcoma cells with the T-lymphocytes activated by the lysate of selected allogenous melanoma cell lines with high expression of CTA. Conclusion: Our results demonstrate that one can use lysates from allogeneic melanoma cell lines as a source of CTA for DC load during the production of anticancer vaccines for the STBS treatment. Patterns of CTA expression should be evaluated as biomarkers of response in prospective clinical trials.