There is no established method for assessing the prognosis of patients with breast cancer and metastasis confined initially to bone. The medical records of 82 patients with breast cancer nad metastasis confined initially to bone were reviewed. The following variables were analyzed at the time when bone metastasis was first diagnosed, to determine their relationship to length of survival: distribution of metastatic bone lesions on bone scan, presence of radiographic osteosclerosis in metastatic bone lesions, menstrual status, and disease-free interval. Univariate and multivariate analyses revealed that the distribution of metastatic bone lesions and the presence of radiographic osteosclerosis in these lesions were significant predictors of survival. Premenopausal or late postmenopausal status, and longer disease-free intervals (> or = 24 months) or no disease-free intervals (Stage IV breast cancer and metastasis confined to bone at the time of cancer diagnosis) showed a trend, although not statistically significant, toward longer survival. Distribution of metastatic bone lesions on bone scan and the presence of radiographic osteosclerosis in metastatic bone lesions should be considered prognostic variables for patients with breast cancer and metastasis confined initially to bone.