Before the use of screening mammography, ductal carcinoma in situ (DCIS) was an uncommon entity. In the United States in 1975, only 5.8 women per 100,000 were diagnosed with in situ carcinoma, but by 2007, this had increased to 34.5 per 100,000. Our inability to predict which DCIS will progress to invasive carcinoma has resulted in a menu of treatment options, including excision alone, excision and radiotherapy, and mastectomy. Tamoxifen may be added to any of these treatments. The risk of breast cancer death after any surgical approach to DCIS is extremely low, ranging from 2.3% to 4.7% at 15 years in the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-18 and B-24 trials of breast-conserving surgery to less than 1% after mastectomy, and is not changed with tamoxifen. However, the risk of local recurrence is considerably higher after breast-conserving approaches than after mastectomy, and patients have indicated that concern about recurrence is an important factor in the choice of local therapy. Tamoxifen reduces the incidence of both local recurrence and contralateral breast cancer, but the absolute benefit is relatively small. The ability to individualize risks and benefits of therapy for a particular patient is an important component of the selection of therapy, and in the article that accompanies this editorial, Allred et al provide information that allows endocrine therapy to be offered to the subset of patients with DCIS most likely to obtain benefit. The benefit of tamoxifen in reducing local recurrence in women with DCIS treated with excision and radiotherapy was initially reported by investigators from the NSABP in 1999, with a 37% reduction in the relative risk of ipsilateral and contralateral breast cancer events observed with tamoxifen treatment. Although the relative risk reduction was substantial and highly statistically significant, the absolute difference in event rates between treatment arms was only 5.2% and enthusiasm for tamoxifen use in DCIS after the publication of NSABP B-24 was variable. Only 19% of patients with DCIS entered into a national audit database in Australia and New Zealand between January 1998 and December 2004 were reported to be taking tamoxifen. In the United States, single-institution reports indicated that, although 60% to 91% of patients with DCIS were offered tamoxifen, only one-half to three-quarters of those offered the drug chose to take it. An analysis of SEER data found that, in 2000, 35.3% of patients with DCIS received tamoxifen, but that use declined to 20.6% in 2005. A retrospective subset analysis of NSABP B-24 examining the use of estrogen receptor (ER) status in selecting patients with DCIS for tamoxifen therapy was reported at the San Antonio Breast Cancer Symposium in 2002 and is presented in detail by Allred et al. The absolute reduction in breast cancer events in the ER-positive subgroup increases to 11% compared with 5.2% for patients not selected on the basis of ER. The observation that women with ER-positive DCIS have a reduction in ipsilateral breast tumor recurrence (IBTR) with tamoxifen and those with ER-negative disease experience no benefit is not surprising. In invasive breast cancer, tamoxifen is only beneficial in the presence of hormone receptor expression, and in the setting of prevention, tamoxifen reduces the incidence of ER-positive cancers by 69% but has no impact on the incidence of ER-negative breast cancers. Women with DCIS are recognized to be at increased risk for the development of invasive breast cancer, so it is also not surprising that an approximately 50% reduction in contralateral breast cancers was seen in patients with both ER-positive and ER-negative disease. This study allows us to refine our definition of the subgroup of women with DCIS most likely to have a favorable risk/benefit ratio with tamoxifen treatment—namely, premenopausal women with two breasts at risk and ER-positive DCIS. The fact that tamoxifen benefit in patients with ER-positive DCIS was observed using an unplanned, retrospective subset analysis is of concern to some. However, the 41% of patients entered onto the NSABP B-24 trial who made up the study subset did not differ from the rest of the trial population in factors impacting the incidence of IBTR, including age, tumor size, and margin status. This finding, coupled with the knowledge that the tamoxifen outcomes observed in this study are consistent with those seen in large numbers of prospective randomized trials conducted in women with invasive breast cancer, increases confidence in the results, although this is not levelone evidence. The other methodologic concern regarding the study is the variation in the determination of ER status. Blocks were available for assessment of ER status in a central laboratory in 449 patients. In an additional 283 patients without material for central testing, the local determination of ER status was used to classify the patient as having ER-positive or ER-negative disease, despite the fact that in the 102 patients in whom ER was analyzed both centrally and locally, concordance in ER status was present in only 74.5% of cases. The patients in which ER was determined at the central lab did not differ in demographics, tumor characteristics, or treatment from patients in whom JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L S VOLUME 30 NUMBER 12 APRIL 2