The experience of the past 3 years, since the first case of leukemia was reported in a child cured of X-linked severe combined immunodeficiency (X-SCID) by gene therapy, indicates that the potential genotoxicity of retroviral integration in hematopoietic cells will remain a consideration in evaluating the relative risks versus benefits of gene therapy for specific blood disorders. Although many unique variables may have contributed to an increased risk in X-SCID patients, clonal dominance or frank neoplasia in animal models, clonal dominance in humans with chronic granulomatous disease, and the ability of retroviral integration to immortalize normal bone marrow cells or convert factor-dependent cells to factor independence suggest that transduction of cells with an integrating retrovirus has the potential for altering their subsequent biologic behavior. The selective pressure imposed during in vitro culture or after engraftment may uncover a growth or survival advantage for cells in which an integration event has affected gene expression. Such cells then carry the risk that subsequent mutations may lead to neoplastic evolution of individual clones. Balancing that risk is that the vast majority of integration events seem to be neutral and that optimizing vector design may diminish the probability of altering gene expression by an integrated vector genome. Several cell culture systems and animal models designed to empirically evaluate the safety of vector systems are being developed and should provide useful data for weighing the relative risks and benefits for specific diseases and patient populations. Gene therapy interventions continue to have enormous potential for the treatment of disorders of the hematopoietic system. The future of such efforts seems bright as we continue to evolve and improve various strategies to make such interventions both effective and as safe as possible.