We studied the risk of invasive breast cancer and breast cancer death after a breast carcinoma in situ during a period when mammography screening became a nationwide practice and when breast conservation was introduced. In a study base including all 4661 women registered to the Swedish Cancer Registry for a first carcinoma in situ between 1960 and 1992, we selected a cohort of 3398 women diagnosed between 1980 and 1992. The recruitment period was chosen according to the reporting routines for the registry. The corrected survival was 97.4% after 10 years. The risk of invasive cancer was similar in the ipsilateral and contralateral breast. Women diagnosed between 1989 and 1992 ran a relative risk of 0.1 (95% confidence interval, 0.0-0.9) of dying of breast cancer, as compared with women diagnosed from 1980-1982. Residence in counties where mammography screening was available was associated with a relative risk of 0.2 (95% confidence interval, 0.0-2.1) for breast cancer death in the age groups screened. Screening mammography may have contributed to the improvement of prognosis over this time period. This study cannot distinguish between lead time effects and a "true" improvement in prognosis. The increased use of breast conservation was not associated with a worse prognosis in the group as a whole. The study indicates that at least 50% of invasive cancers occurring after treatment for in situ lesions may be new cancers.