Challenging single- and multi-probesets gene expression signatures of pathological complete response to neoadjuvant chemotherapy in breast cancer: Experience of the REMAGUS 02 phase II trial

• Published in: Breast (Edinburgh) Vol. 22; no. 6; pp. 1052 - 1059
• Main Authors: Valet, F, de Cremoux, P, Spyratos, F, Servant, N, Dujaric, M.E, Gentien, D, Lehmann-Che, J, Scott, V, Sigal-Zafrani, B, Mathieu, M.C, Bertheau, P, Guinebretière, J.M, Pierga, J.Y, Delaloge, S, Giacchetti, S, Brain, E, Tembo, O, Marty, M, Asselain, B
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier Ltd 01.12.2013
• Subjects: Adult; Aged; Antineoplastic Combined Chemotherapy Protocols - therapeutic use; Biomarkers, Tumor - genetics; Breast cancer; Breast Neoplasms - chemistry; Breast Neoplasms - drug therapy; Breast Neoplasms - genetics; Bridged-Ring Compounds - administration & dosage; Carbonic Anhydrases - genetics; Chemotherapy, Adjuvant; Estrogen Receptor alpha - genetics; Female; Hematology, Oncology and Palliative Medicine; Humans; Membrane Proteins - genetics; Middle Aged; Neoadjuvant setting; Neoadjuvant Therapy; Oligonucleotide Array Sequence Analysis; Pathological complete response; Prediction; Predictive Value of Tests; Receptors, Estrogen - analysis; RNA - analysis; tau Proteins - genetics; Taxoids - administration & dosage; Transcriptome; Treatment Outcome; Pathological complete response; Breast cancer; Neoadjuvant setting; Prediction; Transcriptome
• ISSN: 0960-9776; 1532-3080
• DOI: 10.1016/j.breast.2013.08.015

Abstract This study was designed to identify predictive signatures of pathological complete response (pCR) in breast cancer treated by taxane-based regimen, using clinicopathological variables and transcriptomic data (Affymetrix Hgu133 Plus 2.0 devices). The REMAGUS 02 trial ( n  = 153,training set) and the publicly available M.D. Anderson data set ( n  = 133, validation set) were used. A re-sampling method was applied. All predictive models were defined using logistic regression and their classification performances were tested through Area Under the Curve (AUC) estimation. A stable set of 42 probesets (31 genes) differentiate pCR or no pCR samples. Single-or 2-probesets signatures, mainly related to ER pathway, were equally predictive of pCR with AUC greater then 0.80. Models including probesets associated with ESR1, MAPT, CA12 or PIGH presented good classification performances. When clinical variables were entered into the model, only CA12 and PIGH, remained informative ( p  = 0.05 and p  = 0.005) showing that a combination of a few genes provided robust and reliable prediction of pCR.