Chemotherapy following surgery for breast cancer is a key factor in reducing death rates, but as yet there is no way to predict how patients will respond. A recent paper [Chang, J.C. et al. (2003) Lancet 362, 362-369 (Abstract)] reports the use of microarray technology to analyse gene expression in samples from breast tumours treated with the chemotherapeutic agent docetaxel. Tissue samples from 24 patients were removed before treatment and cDNA from the tumours examined for patterns of gene expression. A set of 92 genes were identified whose expression correlated closely with the outcome of chemotherapy, as determined by changes in tumour size following therapy. Tumours were classified as sensitive or resistant to docetaxel based on their decrease in size following treatment; the profile of 92 genes correctly classified sensitivity or resistance of the tumour to docetaxel with an accuracy of 92% and 83% respectively. The authors propose that genetic profiling of tumours could identify those patients likely to respond to treatment with docetaxel before chemotherapy is started; patients whose tumours are unlikely to respond well should receive alternative treatment.
Comment: This study provides the first evidence that genetic profiling of tumours can predict the outcome of therapy with a specific drug; however, the sample size of patients involved in the study was very small and considerably more evidence would be required to validate the results. An article by Brenton and Caldas commenting on the report observes that the authors' measure of tumour response was an arbitrary one, and may not have any clinical value, unlike absolute measures such as survival outcomes, but the value of exploratory clinical studies such as this one to guide and contribute to more rigorous studies of chemosensitivity is noted.