The role of autologous peripheral blood progenitor cell (PBPC) transplantation for high-risk stage II/III breast cancer remains controversial. New prognostic indicators defining subsets of patients who may benefit from autologous PBPC transplantation would be clinically useful. The axillary lymph node ratio, defined by the total number of axillary nodes involved with cancer divided by the number of axillary nodes surgically sampled, has been reported to be of potential prognostic importance in transplantation for high-risk, stage II/III breast cancer. We therefore retrospectively reviewed 111 women with high-risk, stage II/III breast cancer with at least four positive axillary lymph nodes undergoing autologous PBPC transplantation from 1991 to June 1999. None of the patients had received prior radiotherapy and all had completed one, and only one, course of at least three cycles of adjuvant chemotherapy. The median number of axillary nodes sampled was 20 (range 6-40) and the median number of positive axillary nodes was 12 (range 4-35). The median node ratio, dividing the number of positive nodes by the number of sampled nodes, was 0.68. Event-free survival was strongly influenced by node ratio. Patients having a node ratio of < 0.7 had a 5-year event-free survival of 68%, vs those with a node ratio of > or = 0.7 with a 5-year event-free survival of 46% (P = 0.03). Forty percent of patients with a high node ratio have relapsed vs 20% with a low node ratio (P = 0.02). Multivariate analysis revealed that positive estrogen receptor status and a node ratio of < 0.7 were independent factors related to better event-free survival (P = 0.0001 and P = 0.004, respectively). We conclude that patients having a node ratio of < 0.7 have a significantly better prognosis following autologous PBPC transplantation than do patients with a ratio > or = 0.7.