Paraffin-embedded biopsy specimens from the primary breast carcinomas of 653 women were subjected to histopathological assessments of the potential prognostic factors. Histological type, histological grade, nuclear pleomor-phism, tubule formation, intraductal growth pattern, tumour margin circumscription, tumour necrosis and inflammatory cell reaction were semiquantita-tively analysed with special reference to disease outcome during the mean follow-up 12.8 years. Histological grade, nuclear grade and inflammatory cell reaction were related to axillary lymph node involvement at operation (p < 0.001). Intensity of the inflammatory cell reaction was directly correlated to histological grade, histological type, nuclear pleomorphism and tumour necrosis (p < 0.001). Tubule formation and tumour necrosis were significant predictors for tumour recurrence. Recurrence-free survival was related to tubule formation (p = 0.0048), histological grade (p = 0.0208) and intraductal growth pattern (p = 0.0441). Tubule formation accurately predicted the recurrence-free survival in axillary lymph node-negative tumours (p = 0.0386). In small (diameter ≤ 20 mm) axillary lymph node-negative tumours, the inflammatory cell reaction (p = 0.0377) as well as intraductal growth pattern (p == 0.0632) were related to recurrence-free survival. Cancer-related patient survival was predicted in decreasing order of significance by tubule formation (p = 0.0002), nuclear pleomorphism (p = 0.0010), intraductal growth (p = 0.0077), tumour necrosis (p = 0.0117) and histological type (p = 0.0651). In axillary lymph node-negative tumours, tubule formation (p = 0.0409), inflammatory cell infiltration (p = 0.0790) and intraductal growth (p = 0.0958) predicted the cancer-related survival. The results indicate that despite an intense search for a diversity of prognostic factors by increasingly sophisticated techniques (e.g., morphometric measurements, flow cytometry, immunohisto-chemistry, and DNA hybridization techniques), the relatively simple light microscopic assessment of the above morphological features seems to be still advocated in predicting the disease outcome in female breast cancer.