Background: Neoadjuvant chemotherapy is a standard treatment in stage III breast cancer. Prognostic factors can help to identify patients (pts) with high risk of recurrence. The aim of this study was to assess several prognostic factors after a long follow-up, in stage III breast cancer pts, treated with neoadjuvant chemotherapy. Methods: We evaluated 126 pts with stage III breast cancer that participated in a phase-II randomized trial of neoadjuvant 5-fluorouracil, doxorubicin and cyclophosphamide (FAC every 21 days) compared with cyclophosphamide, methotrexate and 5-fluorouracil (CMF days 1 and 8 every 28). Chemotherapy was administered for three cycles prior to definitive surgery and radiotherapy, and then for six cycles as adjuvant. Response was assessed by WHO criteria. Results: The median age was 52 years (range 24-75). Median follow-up was 4.5 years (range 0.2-16.4), disease free survival (DFS) 4.8 years and overall survival (OS) 6.4 years. Results of the phase-II study showed no difference in efficacy between groups. Univariate analysis showed that the number of pathologically involved lymph nodes (pLN), pathologic response and estrogen and progesterone receptor status correlated with DFS and OS. Number of pLN was the only prognostic factor with statistical significance in Cox regression test for both, DFS and OS (P=0.0004 and P=0.0006, respectively). In a subgroup analysis of pts with pLN, we found no difference in survival when we compared FAC with CMF. Conclusions: The prolonged follow-up of this study provides a unique opportunity to evaluate factors that predict long-term outcomes. After 16 years of follow-up, the number of pLN remains the most powerful predictor of survival. The subset of pts with pLN had similar survival regardless of the regimen used. Clinical trial information: NCT00002696.