The current treatment of DCIS is based on a presumption that DCIS is a non-obligate precursor of invasive breast cancer. There is a paucity of natural history studies since DCIS is generally surgically resected upon diagnosis. The few retrospective reports of women who had biopsies that were assumed to be benign but on later review were found to have DCIS report a 20–50% risk of invasive cancer in the twenty years after biopsy [7-12]. Without validated measures with which to stratify future risk for invasive cancer, the current goal of all DCIS treatment is prevention of cancer progression through surgery, radiation, hormonal therapy, or a combination of these modalities[7]. In fact, despite the 99% survival rate from DCIS, these aggressive treatments are not much different than those recommended for the invasive cancers these interventions are aimed to prevent. Thus, current therapy for DCIS may represent overtreatment for many women who may never progress to invasive cancer. Nevertheless, expectant management alone is not currently considered an acceptable alternative for most women, due to fear of invasive progression.

In prospective randomized trials, adjuvant tamoxifen and aromatase inhibitor (AI) have both been associated with significantly reduced risk of contralateral invasive breast cancer [13-16]. One possible explanation for this observation is that endocrine therapy may prevent progression of in situ to invasive disease. If indeed the risk of DCIS progression could be reduced with primary medical therapy alone, some women with DCIS could potentially avoid the morbidity of surgery and radiation as well as derive benefit from contralateral risk reduction. Tamoxifen is already offered as adjuvant treatment for DCIS, based on data from a prospective placebo-controlled trial in women undergoing lumpectomy and radiation showing a benefit in DFS which favored tamoxifen[14]. Ongoing clinical trials are studying the role of AI in both the prevention and adjuvant settings for DCIS. However, the biologic impact of endocrine therapy directly on DCIS lesions themselves remains largely unexplored.

Clinical trials of neaodjuvant endocrine therapy in estrogen-positive invasive breast cancer have demonstrated a clear clinical objective response of at least 35% for both tamoxifen and aromatase inhibitors at 3 or 4 months [21-23]. If hormone therapy also induces regression in DCIS, alterations in morphology of DCIS would have been expected. In the present study, most cases treated with neoadjuvant therapy showed morphologic changes in the DCIS. These included a decrease in duct distention, an increase in periductal sclerosis and scattered degenerated cells, as well as an increase in macrophage infiltrate. These morphologic changes are compatible with treatment effect and may represent regression of DCIS. Furthermore, two of the 23 cases treated with endocrine therapy demonstrated only ADH without residual DCIS in the post-treatment excision specimens. We cannot entirely exclude the possibility that all DCIS was removed during the initial core biopsy for these two patients. However, the ADH seen in the post-treatment specimens was present immediately adjacent to biopsy site changes, demonstrated similar microcalcifications as those observed in the original DCIS, and showed similar cytologic features but less developed architecture as in the corresponding pre-treatment DCIS. Furthermore, degenerating epithelial cells and multinucleated giant cells were noted within the ADH ducts. Therefore, the morphologic features of these post-treatment ADH suggest that they may represent altered DCIS associated with treatment.

The anti-tumor effect of targeted endocrine therapy could be attributed to decreased proliferation and/or increased apoptosis at the cellular level. In this study, neoadjuvant endocrine therapy reduced cellular proliferation in DCIS cells, as measured by reduction in Ki-67. Our findings are in agreement with data reported by Boland et al. which showed a significant reduction in the percentage of proliferating cells in ER-positive DCIS following estrogen withdrawal[24]. Previous studies have also demonstrated an antiproliferative effect of tamoxifen and aromatase inhibitors in invasive breast cancer [25-31]. Most recently, in a study evaluating the DCIS component in prospective randomized window trial of letrozole versus anastrozole for ER-positive postmenopausal invasive cancer, a significant drop in Ki67 was seen for both groups in the DCIS component of the tumor[32]. In aggregate, these studies confirm that endocrine therapy effectively reduces proliferation in DCIS.

However, endocrine therapy did not have a discernible effect on apoptosis as measured by the cleaved caspase 3 assay. The lack of an observed apoptotic effect could be due to the limited sample size, or other apoptotic pathways not detected by the caspase 3 assay. Alternatively the drug concentrations at the DCIS site after oral tamoxifen or letrozole may have been below the levels required for activation of such programmed cell death pathways. Previous studies have shown conflicting results on the association between endocrine treatment and apoptosis. While in vitro[33,34], animal model[35] and one clinical[36] study demonstrated activation of apoptosis by endocrine therapy, other clinical trials on invasive breast cancer failed to show increased apoptosis by tamoxifen, raloxifene, or letrozole[31,37,38]. Further studies are required to elucidate the effect on apoptotic activity by endocrine therapy in DCIS.

Short-term treatment with AI resulted in downregulation of ER and PR expression in the DCIS cells, an effect not seen with tamoxifen treatment. This is consistent with studies in invasive breast cancer which also demonstrated changes in ER and PR expression after endocrine therapy[26,29-31,37,39]. The mechanism for ER downregulation by endocrine therapy is unclear. A previous study suggested that the effect was likely a post-transcriptional modification as there was no change of ER mRNA level[40]. On the other hand, the PR gene is an estrogen-regulated gene, so drugs with endocrine activity would be expected to reduce PR expression. Indeed, post-treatment PR H-score of less than 10 was noted in four cases, all of whom were treated with letrozole.

In this cohort, five cases overexpressed Her2. There was no difference in Ki-67 reduction between Her2-positive and Her2-negative cases in this small study. The IMPACT trial which compared neoadjuvant tamoxifen to anastrozole in ER-positive invasive breast cancer has suggested that when compared to HER2-negative tumors, HER2-positive lesions tended to show less antiproliferative effect following endocrine treatment[27]. This effect may not have been detectable in this small cohort. However, the neoadjuvant therapy approach demonstrated in our study may provide an ideal model for addressing the impact of HER2 overexpression on the response to hormone therapy in DCIS in larger clinical trials.

Importantly, two of the 23 patients had invasive ductal carcinoma in addition to DCIS in post-treatment excision specimens. One of the invasive carcinomas was HER2-positive, as was its associated DCIS. In both cases, the invasive tumors were hormone receptor-positive. The invasive carcinoma was likely present at initial diagnosis, but the focus missed due to sampling error. The rate of invasive cancer seen in our study is somewhat lower than reported rates of upstaging to invasive cancer (20–25%) in patients with DCIS only on core biopsy[41,42]. This may reflect a higher degree scrutiny for invasive cancer for those patients enrolling on this study, as assessed on both clinical examination and radiographic work-up. Both patients with invasive cancer underwent clinical examination, mammography, and MRI prior to study entry, none of which were suspicious for invasive cancer. In the current study, this would not have been expected to impact patient outcome as all patients had ER-positive disease and definitive surgery was performed in all but two patients who declined surgery. Thus essentially, these postmenopausal patients were treated with a standard course of neoadjuvant aromatase inhibitor therapy. However, the risk of invasive disease not detected at stereotactic core biopsy is an important consideration for future trials of non-operative treatment for preinvasive disease, and underscores the importance of patient selection in the design of preoperative clinical trials in DCIS. Current efforts, including some in our own group, are focusing on efforts to improve discrimination of invasive cancer and DCIS on MRI by optimizing both image acquisition and software analysis technology to address this important need.