Prediction of metastatic relapse in node-positive breast cancer: establishment of a clinicogenomic model after FEC100 adjuvant regimen.

Breast cancer is a very heterogeneous disease, and markers for disease subtypes and therapy response remain poorly defined. For that reason, we employed a retrospective study in node-positive breast cancer to identify molecular signatures of gene expression correlating with metastatic free survival. Patients were primarily included in FEC100 (5-fluorouracil 500 mg/m(2), epirubicin 100 mg/m(2) and cyclophosphamide 500 mg/m(2)) arms of two multicentric prospective adjuvant clinical trials (PACS01 and PEGASE01-FNCLCC cooperative group). Data from nylon microarrays containing 8,032 cDNA unique sequences, representing 5,776 distinct genes, have been used to develop a predictive model for treatment outcome. We obtained the gene expression profiles for 150 of these patients, and used stringent univariate selection techniques based on Cox regression combined with principal component analysis to identify a genomic signature of metastatic relapse after adjuvant FEC100 regimen. Most of the 14 selected genes have a clear role in breast cancer, carcinogenesis or chemotherapy resistance. Six genes have been previously described in other genomic studies (UBE2C, CENPF, C16orf61 [DC13], STMN1, CCT5 and BCL2A1). Furthermore, we showed the interest of combining transcriptomic data with clinical data into a clinicogenomic model for patients subtyping. The described model adds predictive accuracy to that provided by the well-established Nottingham prognostic index or by our genomic signature alone.