The role of sentinel lymph node biopsy (SLNB) in patients with a core needle-biopsy diagnosis of ductal carcinoma in situ (DCIS) has been intensely debated. Core needle-biopsy has an inherent sampling error leading to histologic underestimation of invasive disease. If SLNB is not performed at the time of the definitive operative procedure, patients found to have an invasive cancer, will require a second operative procedure. The study was designed to determine when the risk of finding invasive disease on final pathology in patients with an initial diagnosis of DCIS was sufficiently high to justify the use of SLNB. We identified 587 women with an initial core needle-biopsy diagnosis of DCIS in the prospective Breast Test Wales (BTW) database from 1995 through 2005. A variety of clinical, mammographic and histologic features were identified and correlated with the presence of invasion at excision using univariate and multivariate analyses. Median age of patients at the time of diagnosis was 58 years (range 41 to 83 years). 201 patients (36%) were treated by mastectomy and 354 (64%) by breast conservation surgery. 220 of 587 patients (38%) were found to have invasive disease on final pathology. On univariate analysis, the rate of upstaging was related to the presence of a clinically palpable mass and size of the mass (both p<0.0001, Mann-Whitney test); mammographic presence of a mass and size of the mass (both p<0.0001, Mann-Whitney test). Multivariate logistic regression analysis revealed 2 independent predictors of invasive cancer on final pathology: mass on clinical examination (odds ratio [OR], 5.09; p<0.0001) and mammographic mass (OR, 7.37; p<0.0001). Age, grade of DCIS, microinvasion and presence of comedonecrosis did not help in distinguishing between patients with DCIS and those upstaged to invasive carcinoma at definitive surgery. Axillary nodal staging (four node sampling or clearance) was done at the time of surgery in 269 patients. Axillary nodal metastases were found in 35 of 269 patients (13%). All 35 patients had invasive carcinoma on final pathology. The indiscriminate use of SLNB in patients with DCIS seems excessive. Our study suggests that patients with a mass on clinical examination or mammogram have an increased risk of invasive disease at the time of definitive operative procedure and should undergo SLNB at the initial procedure. In addition, SLNB should be performed in patients undergoing mastectomy because mastectomy precludes SLNB if invasive disease is subsequently discovered.