The purpose of the present study was to investigate the association of glutathione S-transferase P1 (GSTP1) expression with resistance to neoadjuvant paclitaxel followed by 5-fluorouracil/ epirubicin/cyclophosphamide (P-FEC) in human breast cancers. The relationship of GSTP1 expression and GSTP1 promoter hypermethylation with intrinsic subtypes was also investigated. In this study, primary breast cancer patients (n = 123, stage II-III) treated with neoadjuvant P-FEC were analyzed. Tumor samples were obtained by vacuum-assisted core biopsy before P-FEC. GSTP1 expression was determined using immunohistochemistry, GSTP1 promoter methylation index (MI) using bisulfite methylation assay and intrinsic subtypes using DNA microarray. The pathological complete response (pCR) rate was significantly higher in GSTP1-negative tumors (80.0%) than GSTP1-positive tumors (30.6%) (P = 0.009) among estrogen receptor (ER)-negative tumors but not among ER-positive tumors (P = 0.267). Multivariate analysis showed that GSTP1 was the only predictive factor for pCR (P = 0.013) among ER-negative tumors. Luminal A, luminal B and HER2-enriched tumors showed a significantly lower GSTP1 positivity than basal-like tumors (P = 0.002, P < 0.001 and P = 0.009, respectively), while luminal A, luminal B and HER2-enriched tumors showed a higher GSTP1 MI than basal-like tumors (P = 0.076, P < 0.001 and P < 0.001, respectively). In conclusion, these results suggest the possibility that GSTP1 expression can predict pathological response to P-FEC in ER-negative tumors but not in ER-positive tumors. Additionally, GSTP1 promoter hypermethylation might be implicated more importantly in the pathogenesis of luminal A, luminal B and HER2-enriched tumors than basal-like tumors. (Cancer Sci 2012; 103: 913-920) N eoadjuvant chemotherapy (NAC) for primary breast cancer patients is known to enhance the operability of patients with advanced tumors previously considered inoperable, as well as making breast-conserving surgery more feasible for patients for whom such surgery was previously not feasible due to large tumor size. In addition, it is well established that patients who show a pathological complete response (pCR) to NAC can have a better prognosis than those who do not, (1)(2)(3) so the response to NAC can provide valuable information regarding patient prognosis. These advantages of NAC have led to its widespread use including recently for a growing number of breast cancer patients. However, pCR rates for NAC of only 20-30% of patients are still rather low. (4) Because adverse effects of various degrees of severity are seen in virtually all patients, it seems to be very important to develop predictive factors for the response to NAC to avoid the unnecessary use of NAC for patients who are unlikely to derive benefits from such therapy.Among predictive factors, estrogen receptor (ER), progesterone receptor (PR), HER2, histological grade (HG) and Ki-67 have been most extensively studied and significant associations of ER negativity, PR negativity, HER2 amp...