Current cancer therapies are based on the ability to inhibit the growth of rapidly dividing cells, the majority of which constitute the tumor. Although for decades, sporadic literature has posited the existence of cancer stem cells (CSCs), only recently has this type of cell been isolated and characterized from solid tumors. Like stem cells from their normal counterpart, CSCs are a rare population that can reconstitute a new tumor with similar composition and phenotype to the tumor of origin. These CSCs represent a small subset of the original tumor, grow indefinitely in vitro, and can form tumors in animals from a very few cells. The cells are slow cycling, capable of self-renewal and give rise to daughter cells that are either self-renewing and pluripotent or transit amplifying, and terminally differentiated. Thus far, CSCs have been isolated from only a small number of tumor types. In most instances, the cells are obtained using selection of, and enrichment for, cells with prospectively identified cell surface markers (Al-Hajj M, et al., 2003). This yields a very limited number of cells, and in many cases these cells cannot be cultured. There is a need for a method for isolation, purification, and expansion of stem cells from a greater spectrum of tumors. There is also evidence for "...a link between normal stem cell regulation and the control of cancer stem cells" (NCI Think Tanks in Cancer Biology, Executive Summary of the Tumor Stem Cell and Self-renewal Genes Think Tank1). We present here a strategy for the isolation and establishment of tumor cell lines that represent a minority of cells in the original tumor. They have the ability to grow indefinitely in vitro, form tumors in mice from less than 100 cells, and share many of the growth requirements and cell surface antigens of normal tissue stem cells from which they may arise.