Estrogen receptor-related genes as an important panel of predictors for breast cancer response to neoadjuvant chemotherapy

Purpose: Clinical data suggest that the estrogen receptor (ER) contributes to chemotherapeutic responsiveness. However, ER status alone is not consistently predictive. In this study, we used a microarray approach to find novel ER-related genes that predicted chemotherapy responses, with the hope of providing a robust multi-variable prediction method. Methods: One hundred and ten patients with stages II and III breast cancer were included. They received four preoperative cycles of a weekly PCb (paclitaxel plus carboplatin) regimen. A total of 55 training cases were used for marker discovery and for identification of any ER-related genes that may have been associated with a chemotherapeutic response ("training cases"). The other 55 patients were available as an independent validation set ("validation cases") to test, using immunohistochemistry (IHC). Results: In the training set, 20 significantly differentially expressed genes were identified. Among these 20 genes, TFF1, ESR1, GATA3 and TFF3 were found to be ER-related. Among 55 independent validation cases, univariate analysis indicated that clinical variables and ER-related genes were all significantly associated with pCR. It was shown that the pCR rate was as high as 80% when these five factors were all negative. In contrast, these five factors were all positive in seven of nine chemo-resistant patients. Conclusion: In conjunction with levels of ER-related genes, expression of ER protein may provide important predictive outcomes for responses to neoadjuvant chemotherapy and may allow for the identification of a subgroup of patients who could significantly benefit from chemotherapy (or who may be resistant to it). (C) 2010 Elsevier Ireland Ltd. All rights reserved.