Background: Breast conservation (partial mastectomy, axillary node dissection or sampling, and radiotherapy) is the current standard of care for eligible patients with Stages I and II breast cancer. Because axillary node dissection (AND) has a low yield, some have argued for its omission. The present study was undertaken to determine factors that correlated with omission of AND, and the impact of the decision to omit AND on 10-year relative survival. Study Design: A retrospective review of National Cancer Data Base (NCDB) data for 547,847 women with Stage I and Stage II breast cancer treated in US hospitals from 1985 to 1995 was undertaken. A subset of 47,944 Stage I and 23,283 Stage II women treated with breast-conserving surgery (BCS) was identified. Cross-tab analysis was used to compare patterns of surgical care within this subset. Relative survival was calculated as the ratio of observed survival to the expected survival for women of the same age and racial/ethnic background. Results: The rate of BCS with and without AND increased steadily from 17.6% and 6.4% of patients from 1985-1989, to 36.6% and 10.6% of patients from 1993-1995 respectively. AND was more likely to be omitted in women with Stage I than women with Stage II disease (14.5% versus 5.5%). Similarly, AND was omitted more frequently in women with Grade 1 than women with higher grades (Grade 1, 14.9%; Grade 2, 10.1%; Grade 3, 7.1%; Grade 4, 7%). Although the rate of BCS with AND varied considerably according to location in the breast, the overall rate of BCS without AND appeared independent of site of lesion. Women over the age of 70 years were more than twice as likely to have AND omitted from BCS than their younger counterparts. Women with lower incomes, women treated in the Northeast, or at hospitals with annual caseloads < 150 were all less likely to undergo AND than their corresponding counterparts. Ten-year relative survival for Stage I women treated with partial mastectomy and AND was 85% (n=1242) versus 66% (n=1684) for comparable women in whom AND was omitted. BCS with AND followed by radiation therapy for Stage I disease resulted in 94% (n = 5469) 10-year relative survival, compared with 85% (n = 1284) without AND. Addition of both radiation and chemotherapy to BCS with AND for Stage I disease resulted in 86% (n = 2800) versus 58% (n = 512) without AND. In contrast, Stage II women treated with BCS with AND followed by radiation and chemotherapy experienced a 72% 10-year relative survival. Conclusions: A significant number of women with Stage I breast cancer do not undergo AND as part of BCS. The trend is most pronounced for the elderly, but significant fractions of women of all ages are also being undertreated by current standards. Ten-year survival is significantly worse when AND is omitted. This adverse survival effect is not solely from understaging.