The literature on breast cancer reports conflicting prognostic results with respect to DNA ploidy of flow cytometric DNA histograms. This might result from different DNA ploidy classification methods. Our study evaluated the prognostic power of DNA ploidy, using different classification methods, in a large prospective group (n = 1301) of breast cancer patients. Flow cytometric DNA histograms obtained from fresh frozen material were interpreted with use of a commercially available computer program. On the basis of the number of stemlines and the DNA Index, we classified the DNA ploidy by different methods. In all of the cases, the classification method "DNA diploid versus DNA nondiploid" provided the best prognostic significance for overall survival (OS) (Mantel-Cox (MC) = 5.4, P = .02; relative risk (RR) = 1.3, P = .05) and for disease-free survival (DFS) (MC = 11.8, P = .0006; RR = 1.3, P < .05). This was also true for the OS of the lymph node-positive (but not the lymph node-negative) subgroup (MC = 4.1, P = .04; RR = 1.3, P = .05). In subgroups classified on the basis of tumor size, DNA ploidy showed prognostic significance for DFS only in the subgroup of tumors smaller than 2 cm and larger than 5 cm. In multivariate analysis, DNA ploidy showed no additional prognostic power to lymph node status and tumor size. The classification "DNA diploid versus DNA nondiploid" was mostly consistent with respect to prognostic power for OS and DFS, especially in small or lymph node-positive tumors. The RR of DNA nondiploid patients was only marginally higher, however, so large study groups are required to reach statistical significance. This could partly explain the disagreements in the literature. Therefore, DNA ploidy seems to be of little clinical importance in breast cancer patients, compared with other prognostic parameters.