Large-scale microarray analysis is almost ubiquitous in biological research. Recent papers in Nature Genetics illustrate the power of this technique to enhance our general understanding of cancer, and, specifically, to aid in diagnosis and in the identification of targets for the development of new therapies.

In order to shed light on the molecular basis of metastasis, Ramaswamy and colleagues used microarrays to compare the gene-expression profiles of metastatic and primary tumours from a range of human tissues, and found that the expression of 17 genes differed between the two classes of tumour and, therefore, provided a molecular signature for tumour-type.

Interestingly, some primary tumours shared the same gene-expression profile as metastatic nodules. Primary lung tumours that had the metastatic molecular signature were associated with worse prognoses compared with tumours that lacked such a signature. Furthermore, primary tumours from breast, prostate and brain with this signature were more likely to develop distant metastases, indicating that the molecular signature is biologically significant and that the basis of metastasis could be shared by different tumour types.

The authors argue that the propensity for a primary tumour to metastasise is linked to its genetic state as indicated by the gene-expression profile. This is in contrast to the current model, according to which the probability of a primary tumour becoming invasive depends mainly on its size — the bigger the tumour the more likely it is that a somatic mutation will occur that confers the metastatic phenotype.

With further refinement and testing, this work could be used to develop diagnostic tools that predict the clinical outcome of primary tumours on the basis of their gene-expression profiles at diagnosis. Microarray-based gene-expression profiling could also be used to provide detailed cancer diagnosis, as illustrated by Dyrskjøt and colleagues in the same issue of Nature Genetics. Their results showed that each class of bladder cancer has a distinct gene-expression profile that could be used in diagnosis or to generate new therapies targetted at specific stages and types of the disease. It is clear that microarray analysis will be invaluable in many aspects of cancer research. We will have to wait and see how long it takes for the practical application of these results to reach the clinic.