Out of 753 patients with breast cancer treated with radical mastectomy from 1968 to 1970 at the National Cancer Institute of Milan, Italy, 308 had histologically proven positive nodes. The number of positive nodes was not dependent on the location of the primary tumour, its diameter and the patient's age. Extracapsular invasion was related to the number of positive nodes at a statistically significant level: P value 2 X 10(-9). Survival was influenced by the number of positive nodes and extension of metastases beyond their capsule and age. Each of these criteria had an independent impact on survival. Three subgroups with different prognosis were identified in patients older than 40: (a) with a single involved node and 69.9% 10 year survival rate, (b) patients with two or more nodes with metastatic deposit still confined within node capsule and 47.4% 10 year survival rate, (c) patients with two or more involved nodes and extracapsular invasion and 25.3% 10 year survival rate. In patients younger than 40 no subgroup was identified: this group had an intermediate 10 year survival rate (50.9%). The authors conclude that there is a need (a) of re-consideration of the prognosis of patients with positive nodes and (b) to agree on the definition of "high-risk patients".