In our study we have compared the prognostic value of two distinct methods of immunohistochemical Ki-67 determination, tissue microarray (TMA) and classical whole section analysis. "Cut-off" values were used according to the 2009 St. Gallen Consensus. Tissue specimens were obtained from a consecutive retrospective series of 215 female patients with primary invasive tumours. Two hundred and thirteen patients were included in the study. Data on Ki-67 was collected by both tissue microarray (TMA) and whole section analysis. Follow up data on overall (OS) and disease-free survival (DFS) were collected. Median follow-up was 95 months (range from 7.8 through 107 months). Mutual correlation of two Ki-67 determination methods was non-significant (Person's r = 0.13417; p = 0.0528). There was statistically significant association of whole section Ki-67 expression with histological and nuclear grade, progesterone receptor and HER2/neu status. The expression of Ki-67 protein in TMAs correlated only with histological and nuclear grade, but not with other traditional clinicopathological factors. Statistically significant differences in DFS (p = 0.0156) and OS (p = 0.0028) were confirmed between subgroups with low and high whole section Ki-67 expression. When subgroups with high and intermediate expression were compared, significant difference was found in DFS (p = 0.0272), but not in OS (p = 0.0624). On the other hand, there was no statistically significant difference either in DFS, or in OS, according to the expression of Ki-67 in TMAs (p = 0.6529; p = 0.7883; p = 0.7966 for DFS, and p = 0.8917; p = 0.6448; p = 0.4323 for OS, respectively). In our study, classical whole section was superior to TMA analysis in terms of prognosis and clinicopathological correlation. Our results indicate that the method used may have impact on prognostic significance of Ki-67. Further studies are needed, covering a greater number of patients and including a precisely defined stage and treatment patient cohorts, in order to solve controversies in Ki-67 assessment methodology.