%0 Journal Article
%T The role of gene expression profiling by microarray analysis for prognostic classification of breast cancer
%A MJ van de Vijver
%J Breast Cancer Research
%D 2005
%I BioMed Central
%R 10.1186/bcr1205
%X We used gene expression profiling using two different microarray platforms: one containing 25,000 oligonucleotide probes and one containing 18,000 cDNA probes. To obtain prognostic gene expression profiles, we isolated RNA from tumours from a series of 295 patients younger than 53 years presenting with stage I and II breast cancer treated at our institute between 1984 and 1993. The expression of 25,000 genes was assessed, and using various statistical approaches correlation of gene expression with distant metastasis-free probability and overall survival was assessed [1-3]. In addition, we started studies to obtain gene expression profiles predicting response to specific chemotherapy regimens. Within a single-institution, randomized phase II trial, patients with locally advanced breast cancer received six courses of either AC (n = 24) or AD (n = 24) containing neoadjuvant chemotherapy. Gene expression profiles for 18,000 genes were generated from core needle biopsies obtained before treatment and correlated with the response of the primary tumour to the chemotherapy administered [4]. Additionally, pretreatment gene expression profiles were compared with those in tumours remaining after chemotherapy.We previously identified a 70-gene expression profile associated with increased risk for developing distant metastases within 5 years [1,2]. More recently, we studied a Wound Signature in these same tumors [3]. By combining the 70-gene expression profile to subdivide the tumours into 'good prognosis' and 'poor prognosis' tumours, and the Wound signature to subdivide tumours into 'activated' and 'quiescent' tumours, subgroups of patients with markedly different prognosis can be identified. Additional gene expression signatures are being tested in this series of tumours to arrive at an optimal prognostic classifier and to obtain improved insight into breast cancer biology.In the study to identify predictive profiles, 10 (20%) of the 48 patients showed (near) pathological com
%U http://breast-cancer-research.com/content/7/S1/S1