Most women diagnosed with primary invasive breast cancer are potential candidates to receive adjuvant systemic treatment. Tumor markers that predict the likelihood of response to therapy might help select optimal treatment for individual patients. Of these, c-erbB-2 is the most promising marker. In the past 15 years, over 200 reports related to c-erbB-2 and breast cancer have been published. However, its clinical role remains unclear. c-erbB-2 might serve as either a pure prognostic factor, with no association with therapy, or as a predictive factor of benefit from specific types of systemic treatments. A recent overview suggested that c-erbB-2 is only a weak prognostic factor. A review of the literature suggests that c-erbB-2 overexpression might predict for relative, but not absolute, resistance to endocrine therapy in estrogen receptor-positive women. When adjuvant chemotherapy is indicated, c-erbB-2 positive patients may receive more benefit from anthracycline-containing regimens than alkylating agents. However, if anthracyclines are contraindicated, women with c-erbB-2 positive tumor do derive benefit from alkylating agents. Importantly, c-erbB-2 status appears critical for selecting patients who should receive trastuzumab. Well-designed prospective randomized clinical trials or formal meta-analysis will help to establish the predictive role of c-erbB-2.