AIM/BACKGROUND: Conflicting results have been reported concerning the prognostic value of DNA flow cytometric variables (DNA ploidy, DNA index, %S phase fraction) in breast cancer. Selection bias and differences in treatment may have contributed to these conflicting prognostic results. Differences in tissue processing, the number of nuclei measured, DNA histogram/cell cycle analysis, and intra-tumour heterogeneity may also have played a role. The aim of the present study was to assess intra-tumour heterogeneity of DNA flow cytometric variables in breast cancer. METHODS: Fresh frozen specimens (n = 274) (0.3 x 0.3 x 0.3 cm) of 17 breast cancers and 167 slices, 50 microns thick, of 58 paraffin wax embedded blocks of 21 breast cancers were studied. All samples were prepared individually for DNA flow cytometry. DNA histograms were interpreted by semi-automated cell cycle analysis (MultiCycle) by two observers to avoid biased interpretation. An artificial averaged DNA histogram of each case was composed to simulate a sample prepared from whole tumour tissue. RESULTS: With regard to DNA ploidy, classified as diploid or aneuploid, the fresh frozen and paraffin wax embedded breast cancers showed intra-tumour heterogeneity in 53% and 38% of cases, respectively. For fresh frozen and paraffin wax embedded material, respectively, six samples had to be measured to detect the highest DNA ploidy class in 71% and 86% of cases. Averaged DNA histograms showed a loss of DNA aneuploidy in 36% and 6% of fresh frozen and paraffin wax embedded samples, respectively. High intra-tumour heterogeneity (wide ranges) was found for the %S phase fraction. Average %S phase fraction and average aneuploid %S phase fraction had the widest ranges at 9.5-31.6% and 0.0-62.7%, respectively. There was no correlation between the number of stemlines and intra-tumour %S phase variability on the one hand and tumour size and grade on the other. CONCLUSIONS: High intra-tumour heterogeneity for breast cancer was found for DNA ploidy, the DNA index and %S phase fraction as measured by flow cytometry, which may explain the conflicting prognostic results reported in the literature. To detect aneuploid cells, six samples may have to be prepared and measured separately. Measurement of these variables may be more reliable in paraffin wax sections because the thick slices provide a more representative sample. Prospective studies are required to determine whether the highest %S phase fraction value or the average value is more useful in the clinical context.