Gene expression profiling in breast cancer - Design of a pooled database to address open questions

Background: The Netherlands Cancer Institute used DNA microarray analyses to identify a 70-gene expression profile strongly predictive of a short interval to distant metastases in breast cancer. For patients with small tumors, the signature is not yet adequately validated. Furthermore, 95% of estrogen-receptor or triple-negative tumors are assigned to poor prognosis by the profile. Methods: A pooled database was constructed containing clinical, pathological and microarray data of 1696 patients. The database will be used to study the performance of the 70-gene profile in patients with small-sized T1 tumors. In addition, patients with triple-negative tumors will be identified and whole genome microarray analysis will be performed of these tumors to develop a new prognostic gene expression profile for this subgroup. Results and conclusions: If the 70-gene profile is accurate for small tumors, patients at risk may be assigned to adjuvant treatment. A new prognostic classifier for triple-negative tumors may help to identify women, in whom adjuvant treatment may safely be omitted.