Gene expression profiling (GEP) for the prediction of response to neoadjuvant paclitaxel and gemcitabine in breast cancer (BC). Preliminary results from a Phase II trial

580 Background: GEP of primary breast tumors can be useful in selecting specific chemotherapy regimens based on their probabilities of response. Methods: A neoadjuvant phase II trial for stage II-III BC patients was prospectively designed to test the ability of GEP to predict clinical and pathological responses. Treatment schedule consisted of 6 cycles of both G 2,500 mg/m2 and T 150 mg/m2 on day 1 of a 14 day-cycle before surgery. All 70 patients are already recruited. Mean age 50 yr (range, 26–71), and tumor size: 35 mm (range; 23–100). The analysis of GEP was performed with amplified-RNA from frozen tru-cut biopsies hybridized against a Universal Human Reference RNA pool. GEP analysis was performed in duplicate in 70% of the samples with a correlation coefficient of 0.95 (CNIO Oncochip is a cDNA array comprising 9,300 cancer-related genes: http://bioinfo.cnio.es/data/oncochip/). The analysis of predictor genes was performed using the Random Forests method. To validate the genes in the predictor gene se...