The purpose of this retrospective study was to evaluate the influence of axillary disease on patients' survival after neoadjuvant chemotherapy and to assess patient and tumor characteristics associated with post-chemotherapy axillary involvement. After six induction cycles, 277 patients with operable breast cancer (stage II-III) underwent surgery with axillary dissection, followed by radiotherapy (n = 267) or additional chemotherapy (n = 63) and adjuvant tamoxifen therapy (n = 138). At a median follow-up of 8.5 years, overall survival (OS) and disease-free survival (DFS) were analyzed as a function of node involvement. The differences in OS and DFS according to the number of positive nodes were highly statistically significant with a decreased survival associated with the increasing number of nodes (p = 5 x 10(-6) and 9 x 10(-7), respectively). Upon multivariate analysis, the node number after chemotherapy appeared as the most significant prognostic factor (p = 7 x 10(-4) for OS and p = 3 x 10(-5) for DFS). All the other classical prognostic factors were insignificant, except post-chemotherapy Scarff-Bloom-Richardson (SBR) grading for OS (p = 8 x 10(-4)) and adjuvant hormonotherapy for DFS (p = 1 x 10(-2)). Although constituting a different parameter from primary surgery data, the number of positive nodes after chemotherapy could still remain a valuable prognostic factor at secondary surgery, raising the question for high risk patients of a second non-cross-resistant adjuvant regimen, or high dose chemotherapy with peripheral blood stem cells support.