2524 Background: To elucidate the genomics of tumor responses to different classes of chemotherapy, we analyzed breast cancer gene expression before and after in vivo treatment with adriamycin or docetaxel. Methods: Tumor biopsies were obtained before and 3 weeks after one chemotherapy cycle and tumor RNA amplified and hybridized on the Affymetrix HG-U133+2 array containing 33,000 genes. Results: Pre- and post-treatment tumors from 46 chemonaive patients with unresectable breast cancers were studied, of which 24 and 22 respectively received adriamycin and docetaxel in the first cycle, and 14 in each group had early response sensitive tumors (=25% shrinkage after 1 cycle). Comparison of our baseline gene signatures with drug-specific panels generated in vitro (Nevins, Nat Med 2006,12:1294) revealed 12 and 2 common genes (p<0.05) that predicted for adriamycin and docetaxel response respectively, with the 12 common adriamycin-response gene panel correctly predicting response in 76% of patients. Analysis of the relative change in tumor gene expression (ratio of post- and pre-treatment differential values to pre-treatment values) in our dataset revealed adriamycin to up- or down- regulate 209 transcripts (p<0.005) including genes that encode for nuclear protein, cell cycle regulation, aminopeptidases, and Ankyrin repeats, while docetaxel up- or down-regulated 469 transcripts (p<0.005) including genes that encode for extracellular matrix, transmembrane signaling, endocytosis, EGF-like calcium binding, tubulin and actin binding functions. Adriamycin and docetaxel concordantly up- or down-regulated 269 transcripts (p<0.01) that may be common response markers, including genes involved in cell cycle proliferation, mitosis, DNA damage, and carboxypeptidase activities. Adriamycin and docetaxel differentially induced 92 transcripts (p<0.01) that distinguished between the two drugs with 96% accuracy. 27 adriamycin- and 100 docetaxel-induced transcripts (p<0.005) predicted response to each drug with >90% accuracy. Conclusions: Drug-specific genomic changes can predict clinical response, and may yield insights to targets to overcome drug resistance. No significant financial relationships to disclose.
Breast cancer specific survival (BCSS) in young women <40 years with node negative luminal breast cancer (BC) treated based on tumor gene expression
