Abstract P5-10-09: Prediction of Trastuzumab treatment response for HER2-positive breast cancer by microRNA profiling

Abstract

Abstract [Background and Aim] Although Trastuzumab has been used for HER2(+) breast cancer, the treatment response of Trastuzumab therapy depends on unknown mechanisms among individual cases. In order to avoid unnecessary adverse events and to lighten financial burden for patients, pre-treatment prediction of trastuzumab treatment response would be beneficial for patients. Thus, in this study, we develop a prediction algorithm using microRNA expression profile of breast cancer tissues. [Materials and Methods] Eighty-three breast cancer patients who underwent trastuzumab-chemo combined therapy before operations were enrolled with written informed consent. FFPE specimens of pre-treatment core needle biopsy samples were collected, and regions containing cancer and adjacent stromal cells were laser-microdissected. Total RNA samples extracted from the microdissected specimens were subjected for microRNA microarray (3D-Gene®, Toray, Japan) analysis. Among these 83 patients, 39 cases had pCR (complete response in IDC regions regardless presense of DCIS without lymphnode metastasis), and the other 44 cases did not. According to the pCR/non-pCR information, we develop a prediction model using 35 signature microRNAs by a SVM technique. Prediction accuracy assessed by Leave-one-out validation was AUROC=0.889. The 35 signature microRNAs for trastuzumab treatment response included 7 out of 8 let-7 family members and miR-125a-5p/b-5p. [Conclusion] microRNA profile could predict treatment response of trastuzumab-chemo combined therapy for HER2(+) breast cancer, and the developed prediction algorithm might be a useful tool for clinical decision making. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P5-10-09.