A number of microarray studies have reported distinct molecular profiles of breast cancers (BC), such as basal-like, ErbB2-like, and two to three luminal-like subtypes. These were associated with different clinical outcomes. However, although the basal and the ErbB2 subtypes are repeatedly recognized, identification of estrogen receptor (ER) –positive subtypes has been inconsistent. Therefore, refinement of their molecular definition is needed.

The hypothesis of this study was that genomic grade could aid in the task of refining ER-positive molecular subtypes. In particular, the aim of this study was to determine how a classification based on high or low genomic grade compares with previously described molecular classifications.

The data set consisted of 335 early-stage BC samples; 86 of these have been previously used in another study (methods described in Sotiriou et al, with the raw data available at the GEO repository database, accession code GSE2990). Only ER-positive tumors from patients who had received no systemic adjuvant treatment were selected from the GSE2990 data set. This dataset contained samples from Oxford, U.K., and Uppsala, Sweden.

The previously unpublished data set consisted of 249 samples from patients, all of whom had received adjuvant tamoxifen only. This data set contained samples from the John Radcliffe Hospital, Oxford, United Kingdom; Guys Hospital, London, United Kingdom; and Uppsala University Hospital, Uppsala, Sweden. Each hospital’s institutional ethics board approved the use of the tissue material, and written informed consent was obtained.

Eligible samples were selected from the GSE2990 dataset according to a positive ER status by ligand-binding assay. The cutoff value for classification as positive or negative for ER and progesterone receptor (PgR) was 10 fmol/mg protein. A positive ER level was confirmed by microarray expression levels using probe sets 205225_at and 208305_at, representing the ESR1 and PgR genes, respectively.

Microarray analysis was performed with Affymetrix U133B Genechips (Affymetrix, Santa Clara, CA). Gene expression values from the CEL files were normalized by use of the standard quantile normalization method in Robust Multiarray Analysis described by Bolstad et al.

Loi, Sherene, Benjamin Haibe-Kains, Christine Desmedt, Françoise Lallemand, Andrew M. Tutt, Cheryl Gillet, Paul Ellis, et al. “Definition of Clinically Distinct Molecular Subtypes in Estrogen Receptor-Positive Breast Carcinomas through Genomic Grade.” Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology 25, no. 10 (April 1, 2007): 1239–46. doi:10.1200/JCO.2006.07.1522.

Sotiriou, Christos, Pratyaksha Wirapati, Sherene Loi, Adrian Harris, Steve Fox, Johanna Smeds, Hans Nordgren, et al. “Gene Expression Profiling in Breast Cancer: Understanding the Molecular Basis of Histologic Grade to Improve Prognosis.” Journal of the National Cancer Institute 98, no. 4 (February 15, 2006): 262–72. doi:10.1093/jnci/djj052.

Bolstad, B. M., R. A. Irizarry, M. Åstrand, and T. P. Speed. “A Comparison of Normalization Methods for High Density Oligonucleotide Array Data Based on Variance and Bias.” Bioinformatics 19, no. 2 (January 22, 2003): 185–93. doi:10.1093/bioinformatics/19.2.185.