With the emergence of the concept of the leukemia stem cell, assays to study them remain pivotal in understanding (leukemic) stem cell biology. Although the in vivo NOD-SCID xenotransplantation model is still the favored model of choice in most cases, this system has some limitations as well, such as its cost-effectiveness, duration, and the lack of engraftability of cells from subsets of acute myeloid leukemia (AML) patients. Here, we have described an ex vivo bone marrow stromal coculture system in which CD34+ cells, but not CD34− cells, from the bone marrow or peripheral blood of AML patients can give rise to long-term cultures (LTC) that can be maintained for over 20 weeks. Long-term expansion is associated with the formation of leukemic cobblestone area (L-CA) formation underneath the stroma. Self-renewal within these L-CAs can be determined by sequential passaging of these L-CAs onto new MS5 stromal layers, which results in the generation of second, third, and fourth L-CAs that are able to sustain long-term expansion and generate high numbers of immature undifferentiated suspension cells. Furthermore, we have optimized lentiviral transduction procedures in order to stably express genes of interest or stably downmodulate genes using RNAi in AML CD34+ cells, and this method has also been described here. Together, these tools should allow a further molecular elucidation of derailed signal transduction in AML stem cells.