A cancer stem (cancer-initiating) cell is defined as a cell within a tumour that possesses the capacity to self-renew and to generate the heterogeneous lineages of cancer cells that comprise the tumour. This definition directly implies that an anti-cancer therapy can cure a tumour only if all cancer stem cells are killed. A recent milestone in cancer research was the introduction of flow sorting techniques to isolate cell populations based on cell surface markers that are differentially expressed in tumour cell subpopulations that are enriched for cancer stem cells. Application of this technology may allow discrimination of stem cells and non-stem cells on an individual basis, although the interpretation of this data in the context of the exact phenotype of a stem cell is currently evolving. The question of whether cancer stem cells represent a (small) subpopulation of tumour cells which may respond differently to treatment (e.g. radiotherapy or chemotherapy) compared to the bulk of non-stem tumour cells has direct implication for understanding the response of tumours to treatment. Changes in tumour volume after therapy, i.e. tumour response, are governed by the changes in the mass of tumour cells, i.e. primarily by the non-stem cells. In contrast, permanent tumour eradication is expected to be dependent on the complete inactivation of the subpopulation of cancer stem cells. This distinction is extremely important for optimization of cancer research methodology. Today the vast majority of preclinical studies in cancer research use volume dependent parameters such as tumour regression or tumour growth delay as experimental endpoints. An often unrecognised assumption in modelling such data is that cancer stem cells have the same treatment response as non-stem cells in the tumour. The validity of this assumption for different tumour types is currently unknown.