Brain cancer accounts for a substantial proportion of deaths from cancer, particularly in young people. Brain tumours are notoriously resistant to treatment, and their diversity makes them difficult to study in a scientifically meaningful way. At present, brain tumours are classified mainly on the basis of histology, and this approach provides valuable information about cellular origin and aggressiveness. However, these parameters do not always correlate with the ability of the tumour to respond to treatment. Therefore, there is a pressing need to find alternative means of categorizing brain tumours. In a review on page 782 of this issue, Mischel, Cloughesy and Nelson discuss how DNA-microarray technology might be the way forward.

Many types of information could be gleaned from DNA-microarray analysis of brain tumours. Gene-expression profiles might reveal underlying commonalities between seemingly diverse tumour types, or allow the identification of subgroups — for example, with respect to patient survival or response to treatment — within groups of histologically similar tumours. They might also provide new insights into the molecular pathways that underlie tumorigenesis, and uncover components of these pathways that could be targeted therapeutically. By studying the pathways that are malfunctioning in brain cancer, we might also learn more about the roles of these pathways in normal brain development and function.

To make the best use of DNA-microarray technology, communication and cooperation among the clinical and research communities will be vital. Protocols will need to be standardized, and tissue and data shared, particularly in the case of rare tumours. The establishment of tissue banks, databases and research networks will therefore be invaluable for ensuring that this technology achieves its full potential.