Chronic myeloid leukemia (CML) is a clonal myeloproliferative disorder in which there is a deregulated amplification of CML progenitors at intermediate stages of their differentiation along the myeloid, erythroid and megakaryocyte pathways. Such cell populations are routinely quantified using standard in vitro colony-forming cell (CFC) assays. The excessive production of leukemic CFC that is seen in most CML patients at diagnosis may be explained at least in part by their increased proliferative activity. An anomalous cycling behavior in vivo has also been found to extend to more primitive CML progenitor populations detectable as long-term culture-initiating cells (LTC-IC). Although the molecular basis of these changes in CML progenitor regulation is not fully understood at the level of the primitive CFC compartment, a selective inability of CML progenitors to be inhibited by certain -C-C-type chemokines has been demonstrated. Failure of the CML stem cell compartment to expand in vivo at the same rate as later progenitor cell types may be explained by their unique additional possession of an intrinsically upregulated probability of differentiation. Such a mechanism would be consistent with the observed loss of LTC-IC activity by CML cells incubated in vitro under conditions that sustain or expand normal LTC-IC populations. Initial clinical studies undertaken at our center established the feasibility of exploiting the differential behavior of primitive normal and CML cells in vitro as a potential purging strategy for reducing the leukemic stem cell content of CML marrow autografts. The results of a larger, second trial now in progress on a group of unselected patients are encouraging. Future studies of nonobese diabetic/severe-combined immunodeficiency mice engrafted with CML cells should provide another useful preclinical model for evaluating treatments that may more effectively eradicate the neoplastic clone in vivo.