Background: Taxane-based regimens have been widely used to treat breast cancer. Accordingly, it has become important to identify subgroups in which anthracyclines are indispensable. Thus, we initiated a randomized phase II neoadjuvant chemotherapy (NAC) study to compare taxane with and without anthracycline in hormone-negative subtypes. Methods: Eligibility criteria were hormone-negative, an age younger than 80 years and ECOG PS0-1. According to HER2 status, patients were randomly assigned to TC (75/600 mg/m²) q3wks ×6 or FEC (500/100/500 mg/m²) q3wks ×3 followed by D (100 mg/m²) q3wks ×3. The primary endpoint was the rate of pathological complete response (pCR; Grade 3 and Quasi-pCR; Grade 3+2b). Secondary endpoints were safety, breast-conserving surgery ratio, disease-free survival, overall survival, and predictive factors (HER2, Ki-67, P-53, CK5/6, EGFR, and TOP2A by IHC and TOP2A by FISH) for each regimen. Results: 97 out of 103 patients were successfully analyzed (47 for TC6 and 50 for FEC-D). Severe adverse events (Grade ≥2) were frequently observed in FEC-D-treated patients with statistical significance (poor appetite, nausea/vomiting: p<0.001; febrile neutropenia: p=0.016). The pCR rate tended to be higher in FEC-D-treated patients compared with that of TC6-treated patients (pCR: 36.0 vs. 25.5%, n.s.; Quasi-pCR: 46.0 vs. 40.4%, n.s.). There was no significant difference of pCR rates in the HER2 and triple negative (TN) subtypes between each regimen. Among predictors, only positive markers CK5/6 and EGFR predicted the superiority of the FEC-D treatment (p=0.05). Conclusions: TC6 was safe and relatively active even in HER2 subtype patients. Therefore, the concurrent use of trastuzumab with TC could be a reasonable option for NAC in HER2 subtype patients. However, anthracyclines are required to treat basal-type TN cancer. Clinical trial information: UMIN000002215.