Utilization of gene expression profiling to identify markers of disease recurrence in node-negative breast cancer

A major problem associated with care of breast cancer patients is the inability to make accurate estimates for risk of recurrence and to develop management plans for individual patients. Estimation of recurrence risk is becoming more important as stage I, node-negative tumors account for a greater proportion of breast cancer. The difficulty in determining this risk is due to lack of molecular prognostic markers that can be used to influence management decisions regarding adjuvant therapy. We have therefore used cDNA microarrays and subtractive suppressive hybridization techniques to identify alterations in gene expression patterns associated with disease recurrence in node negative breast cancer. Arrays representing up to 40,000 different transcripts were profiled against a panel of 25 good outcome (>5 yrs DFS) and 18 bad outcome (<3 yrs DFS) T1-2N0 breast cancers, and 5 normal breast tissues. To identify up-regulated genes not present on the 30K arrays, we also generated subtraction suppression hybridization (SSH) cDNA libraries. Analysis of the array and library data has led to the identification of a number of genes up and down regulated in breast cancer. A subset of candidate prognostic markers was subsequently validated by quantitative RT-PCR and by in situ hybridization on tissue microarrays. A group of 20 genes has been identified that can discriminate between good and bad outcome T1-2N0 cancers. Estrogen receptor and GATA-3 are 2 genes with increased expression in good outcome tumors whereas Her2 was differentially expressed in bad outcome cancers. Our results demonstrate that molecular profiling can be used to identify node-negative breast cancer with increased risk of recurrence. A description of our approaches and findings will be presented.