Nine clinical, biologic and histologic variables were evaluated for their significance in predicting the metastasis free survival (MFS) and the overall survival (OS) of 650 histologic node negative breast cancer patients. The variables studied were: menopausal status, UICC clinical stage of disease, Scarff-Bloom and Richardson (SBR) grade and its 3 components, estrogen and progesterone receptors, and anatomic tumor size. Multivariate Cox analyses revealed that histologic grade and clinical stage were the only significant prognostic factors for both MFS and OS. In the SBR grading system, grades I and III clearly have defined those patients with low and high risk for relapse, respectively. However, it is well known that more than 50% of the patients fall into the intermediate risk category, grade II, which provides essentially no useful prognostic information for those patients. To improve the assignment of patients to specific risk groups, a modified grade (MSBR), with five categories ordered according to the degree of malignancy, has been built from the nuclear pleomorphism and the mitotic index of the SBR grade. In combination with clinical stage, MSBR was found to be a prognostic indicator with high discriminatory power and caused the SBR grade to lose its significance. The first three categories of this MSBR may be gathered to designate low risk patients, whereas the last two categories, once combined, contain all the SBR grade III plus 57% of the SBR grade II tumors, and reliably identify high risk node negative patients. We suggest that a systemic adjuvant therapy should be discussed in this high risk group.