Gene expression profiling in acute myeloid leukaemia

Acute myeloid leukaemia (AML) is a heterogeneous disease characterised by clonal malignant haematopoiesis with a differentiation arrest and excessive proliferation of leukaemic blasts. Over the past decades, the heterogeneity of AML has been illustrated by evolving classifications based on morphology (French-American-British classification (FAB classification), cytogenetic abnormalities (e.g. t(8;21), monosomies etc.), phenotype and/or molecular abnormalities (e.g. Fms-like tyrosine kinase 3 gene internal tandem duplication (FLT3-ITD), mutations in nucleophosmin I (NPM1) and the transcription factor CCAAT/enhancer binding protein alpha (CEBPA), etc.). The current World Health Organisation (WHO) 2008 classification has integrated these classification modalities. Clinically, dissection of AML into various subtypes allows better survival prediction, but has still limited impact on treatment strategies, with the exception of all-trans retinoic acid treatment for AML-M3 and no allogeneic haematopoietic cell transplantation in complete remission (CRI) for patients with normal karyotype bearing an NPM1 mutation without FLT3-ITD. However, enhanced understanding of the molecular biology of AML will likely result in more 'tailor-made' therapies, for example by adding specific tyrosine kinase inhibitors to standard chemotherapy. In this review, we summarise the variables currently used to classify AML. Specifically, the contribution of microarrays in classification, prognosis and understanding of pathobiology of AML is discussed.