SummaryBecause of the high probability that breast cancer patients have experienced systemic dissemination prior to their initial treatment, the determination of tumor stage and grade should not be considered an estimate of the likelihood of eventual metastasis. An examination of staging and grading parameters reveals strong associations among measures of tumor stage, and a relative independence of those measures from tumor grade. Similarly, measures of tumor grade tend to be expressed at approximately the same level in the individual patient, but are relatively independent of tumor stage, in clinical human breast cancer, tumor 'stage' would appear to be an estimate of body burden of tumor, and tumor 'grade' an estimate of growth rate. Together they provide, with some degree of probability, an estimate of the amount of time between initial presentation and recurrence. In order to increase the probability with which this estimate can be made, greater precision is needed in our determinations of prognostic indices, particularly those of tumor grade. Biochemical measurement of estrogen receptors and other biological products, and tumor morphometry, may help to alleviate this problem. Patient prognosis can best be evaluated by combining estimates of tumor stage and grade, rather than relying upon only one.