Gene expression profile predictive of response to chemotherapy in metastatic colorectal cancer.

// Purificacion Estevez-Garcia 1, 2 , Fernando Rivera 3 , Sonia Molina-Pinelo 1 , Marta Benavent 1, 2 , Javier Gomez 4 , Maria Luisa Limon 2 , Maria Dolores Pastor 1 , Julia Martinez-Perez 1, 2 , Luis Paz-Ares 1, 2 , Amancio Carnero 5 , Rocio Garcia-Carbonero 1, 2 1 Laboratorio de Oncologia Molecular y Nuevas Terapias, Instituto de Biomedicina de Sevilla (IBIS) (HUVR, CSIC, Universidad de Sevilla), Sevilla, Spain 2 Medical Oncology Department, Hospital Universitario Virgen del Rocio, Sevilla, Spain 3 Medical Oncology Department, Hospital Universitario Marques de Valdecilla, Santander, Spain 4 Pathology Department, Hospital Universitario Marques de Valdecilla, Santander, Spain 5 Laboratorio de Biologia Molecular del Cancer, Instituto de Biomedicina de Sevilla (IBIS) (HUVR, CSIC, Universidad de Sevilla), Sevilla, Spain Correspondence to: Rocio Garcia-Carbonero, e-mail: rgcarbonero@gmail.com Keywords: Colorectal cancer, chemotherapy, gene expression, predictive, microarray Received: September 26, 2014      Accepted: January 15, 2015      Published: January 30, 2015 ABSTRACT Fluoropyrimidine-based chemotherapy (CT) has been the mainstay of care of metastatic colorectal cancer (mCRC) for years. Response rates are only observed, however, in about half of treated patients, and there are no reliable tools to prospectively identify patients more likely to benefit from therapy. The purpose of our study was to identify a gene expression profile predictive of CT response in mCRC. Whole genome expression analyses (Affymetrix GeneChip ® HG-U133 Plus 2.0) were performed in fresh frozen tumor samples of 37 mCRC patients (training cohort). Differential gene expression profiles among the two study conditions (responders versus non-responders) were assessed using supervised class prediction algorithms. A set of 161 differentially expressed genes in responders (23 patients; 62%) versus non-responders (14 patients; 38%) was selected for further assessment and validation by RT-qPCR (TaqMan ® Low Density Arrays (TLDA) 7900 HT Micro Fluidic Cards) in an independent multi-institutional cohort (53 mCRC patients). Seven of these genes were confirmed as significant predictors of response. Patients with a favorable predictive signature had significantly greater response rate (58% vs 13%, p = 0.024), progression-free survival (61% vs 13% at 1 year, HR = 0.32, p = 0.009) and overall survival (32 vs 16 months, HR = 0.21, p = 0.003) than patients with an unfavorable gene signature. This is the first study to validate a gene-expression profile predictive of response to CT in mCRC patients. Larger and prospective confirmatory studies are required, however, in order to successfully provide oncologists with adequate tools to optimize treatment selection in routine clinical practice.