BACKGROUND This study considers breast cancer survival patterns by age group among women diagnosed in the Breast Cancer Detection Demonstration Project (BCDDP) to provide insight into the nature of breast cancer and suggest a possible explanation as to why the results of the randomized clinical trials differ for women younger than 50 and those 50 or older. Based on the findings of several randomized clinical trials, there is a general consensus that breast cancer screening provides a significant benefit for women aged 50-69. From these trials, there is little evidence of a short term benefit and uncertainty regarding any long term benefit of mammographic screening for women aged 40-49. METHODS The BCDDP was not a randomized trial, but a nationwide breast cancer screening program conducted between 1973-1980, in which all participants received annual physical and mammographic examinations. Using the BCDDP data, this study calculated 14-year breast cancer survival among 4051 women diagnosed with breast cancer between 1973 and 1980 and followed through 1988 and 1989. RESULTS In all, 598 women died of breast cancer, resulting in an overall 14-year breast cancer survival of 83.4% (standard error = 0.007). Breast cancer survival varied by tumor type, lymph node status, tumor size, and stage at diagnosis. For most of the cases, however, both the distribution and breast cancer survival rates were similar for women aged 40-49, 50-59, and 60-69 across all prognostic indicators. Only breast cancer survival among women with invasive breast cancer who had a primary tumor 5 cm or larger or with positive lymph nodes differed by age, with women aged 60-69 at diagnosis having poorer survival. Analyses by modality of detection conducted in a subset of 2739 cases indicated that the 34.6% of the cases detected by mammography alone had the highest overall breast cancer survival rates (90.9%), while the 32.2% of the cases detectable by both physical examination and mammography had the lowest breast cancer survival (79.0%). Overall, women diagnosed with breast cancer in the BCDDP had a survival advantage if diagnosed with either a smaller tumor or no positive lymph nodes, or if their breast cancer was detected by mammography alone. For women with large tumors (> or = 5 cm), the survival was 80.8% for ages 40-49, 72.1% for ages 50-59, and 61.7% for ages 60-69. CONCLUSIONS Due to the higher breast cancer survival rates among women aged 40-49 with poorer prognostic characteristics, the breast cancer survival advantage for having a smaller tumor, no positive lymph nodes, or breast cancer detected by mammography alone was lower for women aged 40-49 than women aged 50 or older at diagnosis. These differences in survival advantage may help to account for the differences in mortality by age in the randomized clinical trials.