The use of gene expression arrays in the evaluation and classification of tumors is becoming increasingly important in a number of malignancies. This is a powerful technique able to disclose interpatient variance in gene expression. Such variation in gene expression may be the cause of different disease outcome and may reflect disease phenotypes or chemoresistance. Acute myeloid leukemia is a malignant disease of the bone marrow where overall long-term disease-free survival is less than 50%. The need for better disease classification and evaluation is consequently evident. Gene expression profiling in acute myeloid leukemia has, in recent years, proven able to distinguish acute myeloid leukemia subclasses and predict clinical outcome and is, as such, a promising technique for improved disease evaluation. The early detection of gene expression in response to chemotherapy may be a novel way of monitoring disease management. The immediate gene response may be an indication of whether the drug of choice is efficient in leukemic cell eradication and may early indicate the need for other therapeutic measures. Furthermore, these early alterations in gene expression could facilitate identification of new treatment targets, thereby enabling better patient care and follow-up in the future.