To assess anti-tumor activity of sequential epirubicin/cyclophosphamide followed by docetaxel with the randomized addition of celecoxib in HER2 negative patients or trastuzumab in HER2 positive patients. From May 2004 till October 2007, 340 patients with stage II and III breast adenocarcinoma, ineligible for breast conserving surgery, received eight sequential three weekly cycles of EC-D [epirubicin (75 mg/m(2))-cyclophosphamide (750 mg/m(2)) for four cycles followed by docetaxel (100 mg/m(2)) for four cycles]. HER2-negative patients (N = 220) were randomized to receive concomitantly with docetaxel celecoxib 800 mg/day during cycles 5-8 or no additional treatment, while HER2-positive patients confirmed by FISH (N = 120) were randomized to trastuzumab concomitant to docetaxel (8 mg/kg then 6 mg/kg IV every 3 weeks) or no additional preoperative treatment. In the HER2 negative group, pCR (grade 1 and 2 of Chevallier's classification) was observed in 11.5 and 13% of patients treated without and with neoadjuvant Celecoxib, respectively. In the HER2 positive group, pCR rate reached 26% in those who received neoadjuvant trastuzumab versus 19% in the others. There was no unexpected toxicity, no cardiac toxicity, and no toxic death. Triple negative breast cancers experience the highest pCR rate of 30%. Celecoxib is not likely to improve pCR rates in addition to EC-D in patients with HER2-negative tumor. In HER2-positive tumor patients, trastuzumab added to ECD leads to increased pCR rates. It was the only combination to deserve further study according to the two-stage Fleming's design used in this trial.