A major obstacle in the treatment of breast cancer is the lack of adequate methods for predicting patient response to a particular chemotherapy regime. To date, single tumour markers have provided limited success. DNA array technologies identifying thousands of genes simultaneously can help to solve this problem. We investigated cancer cell lines sensitive and resistant to the topoisomerase inhibitors doxorubicin and mitoxantrone. These drugs are used in several different breast cancer treatment protocols. We have identified the top genes best associated with resistance against each cytostatic agent. We applied our gene expression signatures to a set of pre-characterised patients receiving doxorubicin monotherapy. The patients classified as sensitive to chemotherapy exhibited longer survival than the resistant ones. In summary, in our study we have successfully transferred experimental results to a clinical problem, and managed to perform a predictive test for a selected monotherapy protocol. However, many different studies have been performed using microarrays, each producing a different gene list for the same classification problem. It is likely that future diagnostic tools will include the results of several different laboratories, focus on genes validated on different technological platforms and use large cohorts of patients.