The prognostic significance of conventional TNM staging remains the standard for determining prognosis in breast carcinoma. The presence (or absence) of axillary lymph node metastases remains the single most important parameter for predicting patient outcome. The presence of regional lymph node metastases implies that the primary tumor has the capacity for successfully completing the steps of the metastatic cascade. However, the absence of regional lymph node metastases does not ensure that distant or systemic seeding of tumor cells has not occurred, only that it is less likely. Staging data appear to be refined by addition of several standard morphological parameters. Although there is considerable overlap and interaction between these factors, as well as with staging data, there is strong evidence that grade, necrosis, inflammatory cell "immune response," and possibly pattern of invasion and intravascular tumor each independently supplement staging information. Some data appear to have independent significance only when applied to specific patient subsets, raising serious question as to their biologic importance. Nevertheless, morphological data are subjective and susceptible to observer variation and have less statistical power in predicting patient outcome than staging data. It was initially thought that DNA analysis of breast cancer by flow cytometry might supplant morphological data in assessing tumor behavior. The following conclusions can now be drawn: (1) there is no clear association between aneuploidy and SPF and stage; (2) aneuploid tumors are associated with higher SPF and shorter disease-free survival while diploid-range tumors generally have lower SPF and longer disease-free survival; (3) aneuploid DNA content is significantly associated with markers of decreased morphological (grade) and biochemical (ER status) differentiation. Determination of S-phase fraction by FCM appears to be a rapid and potentially easy method for obtaining kinetic information on individual breast tumors, although the technology for improving the accuracy of SPF measurements is still under development (e.g., tumor cell gating, debris subtraction). SPF appears to be comparable to other kinetic measurements, such as TLI, and shows many of the same associations with morphological and clinical data as ploidy. This is due to the close association of ploidy and SPF. Which of these parameters is more important for predicting patient outcome has not been clearly defined. Additional technological refinements for determining SPF may result in a more prominent prognostic role for this measurement. Three problems have limited our ability to draw specific conclusions about the biologic significance of tumor ploidy and SPF.(ABSTRACT TRUNCATED AT 400 WORDS)