Cancer stem cell theory gains increasingly greater significance in the world of medicine. Numerous findings of scientific research in vivo and in vitro indicate that it is the population of undifferentiated, self-renewing cells which is responsible for recurrence of cancer and metastasis. Similarly to normal stem cells, cancer stem cells (CSC) function in the environment of the other cells of the organism, called the niche, where they receive signals for differentiation and proliferation processes. Disorders in the signaling pathways between CSC and the niche that result from e.g. acquired oncogenic mutations may lead to uncontrolled proliferation of stem cells, gaining independence from the primary niche or settling a new microenvironment. CSC are identified on the basis of specific markers - membrane proteins or cell enzymes. Methods based on the measurement of dye fluorescence (obtaining side population, SP) or fluorescence of the fluorophore conjugated with a monoclonal antibody directed against the specific CSC marker are used for isolation. A different method obtains morphologically miscellaneous clones by single cell cloning: holo-, mero- and paraclones. Tumor forming assay in NOD/SCID mice is a standard in vivo test that confirms the stem character of isolated cells. However, this model may not fully reflect the complexity of cancer illnesses in human beings. Solving the mystery of oncogenesis, including the existence of cancer stem cells, is undoubtedly one of the priorities of contemporary medicine that should contribute to the improvement of cancer therapy.