An increasing body of evidence has shown that hematologic malignancies, alike normal hematopoiesis, has a hierarchical structure including a stem cell compartment with self renewal capability, endowed in a neoplastic niche bearing resemblance to its normal hematopoietic counterpart. According to experimental data on NOD-SCID mice, leukemic stem cells are characterized by a CD34+/CD38- surface profile and account for 1 in 10(3) to 1 in 10(6) of the total amount of leukemic cells. The available knowledge about leukemic stem cells (LSC) has arisen the question as to whether some targeting of LSC is achieved by current treatments; the answer is dubitative at best, with the possible exception of arsenic trioxide in promyelocytic leukemia. On the other side, the unsatisfactory results in the treatment of many hematological neoplasms has prompted many research groups to find out whether direct targeting of LSC, possibly in its niche, would result in an improvement in cure rates. This approach implies the identification of LSC specific markers, clearly distinct from their normal counterpart in order to spare normal hematopoietic stem cells. Adhesion/surface antigens, metabolic pathways involved in LSC survival and renewal, telomerase, commonly mutated genes and epigenetic phenomena have been investigated as candidate targets for newer therapeutic strategies. So far, most of the possibly effective agents have been studied in experimental models only. FLT-3 inhibitors account for a notable exception since they have resulted effective in vivo in AML with mutated, but not over expressed, FLT-3. A main task for the future is to find out whether some common LSC specific markers would be identifiable in a substantial proportion of AML cases, or whether each AML case shows a unique fingerprint of markers. In the latter event, targeting of LSC could result in an arduous task.