Background: An increase in incidence of ductal carcinoma in situ (DCIS) of the breast has been documented, and concerns regarding overly aggressive treatment have been raised. This study was designed to evaluate the use of surgery and radiation therapy in treating DCIS. Methods: We used the National Cancer Institute's Surveillance, Epidemiology, and End Results database to assess treatment of patients with DCIS with no evidence of microinvasion who were diagnosed from January 1, 1992, through December 31, 1999. We assessed the rates of mastectomy, breast reconstruction, radiation therapy after lumpectomy, and axillary dissection. Associations were analyzed by logistic regression. Results: During the study period, 25 206 patients met selection criteria. The incidence of DCIS dramatically increased with time; however, the incidence of comedo lesions did not change. The rate of mastectomy decreased from 43% in 1992 to 28% in 1999, after controlling for age, race, tumor size, comedo histology, and geographic location. However, because of the increase in the diagnosis of DCIS, the age-adjusted incidence of mastectomy for DCIS in the population did not change (7.8 per 100 000 women in 1992 and 1999). Almost half the patients undergoing lumpectomy did not undergo radiation therapy (55% in 1992 and 46% in 1999); in those with comedo histology, 33% did not undergo radiation therapy after lumpectomy, even in 1999. Overall, patients were less likely to have axillary dissection over time (34% in 1992 versus 15% in 1999), yet the rate of axillary dissection was still high (30%) in patients undergoing mastectomy in 1999. Large, statistically and clinically significant variation by geographic location was found in treatment. Conclusions: Treatment of DCIS changed in a clinically significant fashion between 1992 and 1999. Throughout this study, many patients were found to undergo aggressive surgical therapy, including mastectomy and axillary dissection, whereas others appeared to be undertreated, e.g., not receiving radiation therapy after lumpectomy, even in the presence of adverse histologic features. Variation in demographic and geographic factors indicates that at least some of these treatment differences reflect individual and institutional practice patterns that may be modifiable.