Recent advances in cancer treatment have resulted in better initial treatment of many cancers. Despite progress, however, approximately 50% of patients with cancer will experience relapse and die from their disease. Recently it has been suggested that relapse may result from harboring of cancer stem cells, which may restimulate cancer proliferation months to years after initial treatment and after apparent elimination of mature cancer cells. Cancer stem cells may be undetectable in number, may evade chemotherapy directed at mature cancer cells, may escape from immune surveillance, and may linger and at some later point become activated and stimulate cancer regrowth. This hypothesis is attractive, because it provides a novel approach to therapy using treatment targeted against residual cancer stem cells after the initial mature tumor burden has been eradicated. Indeed, cancer stem cells have now been identified in a variety of cancers, including leukemia and cancers of the brain, breast, colon, ovary, pancreas, and prostate. Much research is ongoing to characterize these stem cells and to isolate targets against which treatment can be developed.