Ductal Carcinoma In Situ , Complexities and Challenges

Abstract

The incidence of ductal carcinoma in situ (DCIS), a noninvasive form of breast cancer, has increased markedly in recent decades, and DCIS now accounts for approximately 20% of breast cancers diagnosed by mammography. Laboratory and patient data suggest that DCIS is a precursor lesion for invasive cancer. The appropriate classification of DCIS has provoked much debate; a number of classification systems have been developed, but there is a lack of uniformity in the diagnosis and prognostication of this disease. Further investigation of molecular markers should improve the classification of DCIS and our understanding of its relationship to invasive disease. Controversy also exists with regard to the optimal management of DCIS patients. In the past, mastectomy was the primary treatment for patients with DCIS, but as with invasive cancer, breast-conserving surgery has become the standard approach. Three randomized trials have reported a statistically significant decrease in the risk of recurrence with radiation therapy in combination with lumpectomy compared with lumpectomy alone, but there was no survival advantage with the addition of radiotherapy. Two randomized trials have suggested an additional benefit, in terms of recurrence, with the addition of adjuvant tamoxifen therapy, although in one trial the benefit was not statistically significant. Current data suggest that tamoxifen use should be restricted to patients with estrogen receptor–positive DCIS. Neither trial demonstrated a survival benefit with adjuvant tamoxifen. Ongoing and recently completed studies should provide information on outcomes in patients treated with lumpectomy alone and on the effectiveness of aromatase inhibitors as an alternative to tamoxifen.

Ductal carcinoma in situ (DCIS) of the breast is a proliferation of malignant-appearing cells of the ducts and terminal lobular units of the breast that have not breached the ductal basement membrane ( 1 ) . Although DCIS was first noted in 1893, the concept of a benign lesion that could evolve into invasive cancer was not fully recognized until the 1930s, with the evolution of the term “carcinoma in situ ” ( 2 – 4 ) . Early thoughts were that DCIS was a “malignant lesion” that always required mastectomy ( 5 ) . However, an improved understanding of the biology of DCIS came with the recognition that the terminal ductal lobular unit is the origin of most pathologic breast lesions, including DCIS, and shifted the treatment paradigm toward lumpectomy ( 6 ) . Further information on the pathogenesis and natural history of DCIS has contributed to a variety of approaches in the current management of DCIS.

In the last several decades, the incidence of DCIS has increased dramatically, due largely to screening mammography ( 7 , 8 ) . The clinical relevance of a diagnosis of DCIS is uncertain. Although DCIS is a benign disease, women with DCIS have an increased propensity to develop invasive disease, and so therapy for DCIS is ultimately therapy for the prevention of invasive cancer. Unfortunately, the natural history of DCIS and the likelihood that DCIS will progress to invasive disease is unclear; consequently, the risk–benefit ratio of instituting potentially toxic therapy to treat DCIS needs to be considered carefully. Pathologic features of DCIS may predict invasive potential; when these features are combined with patient factors such as age, comorbidities, and breast size, recommendations for the optimal management of DCIS can be individualized and discussed with the patient. At present, management of DCIS remains controversial, as is reflected in the various treatment options offered to patients. Although existing data fail to provide all the answers on the clinical relevance of DCIS, assimilation of these data will improve the ability to inform patients of their management choices and provide a template on which to base future trials. In this review, we analyze the current data on DCIS, with particular emphasis on areas of controversy and exciting new research areas, such as molecular profiling of DCIS, and we describe current and future trials aiming to improve our understanding of this complex and challenging disease.

P athology

DCIS results from a disruption in the architecture of the breast glandular epithelium involving loss of the hollow lumen and epithelial cell proliferation in acinar units that occurs via an imbalance between apoptosis and proliferation ( 9 ) . DCIS is generally categorized by architectural description into five groups: comedo (layer of neoplastic cells surrounding a central area of necrosis), cribriform (radially oriented neoplastic cells forming glandular lumina), papillary (large papillations with fibrovascular stalks), solid (ductal filling with neoplastic cells), and micropapillary (fingerlike papillary projections into dilated ductal spaces). However, these groups do not take into account important prognostic features such as nuclear grade (high, intermediate, or low), necrosis (presence or absence), and polarization (architectural differentiation) ( 10 , 11 ) .

Due to the subjective interpretation of architecture and the overlap in architectural features, many experienced pathologists differ in their diagnosis of DCIS ( 12 ) . Lack of consistency in the diagnosis of DCIS can have important treatment and prognostic consequences. Widespread adoption of a single classification system (see below) could improve consistency in the interpretation and reporting of DCIS.

C lassification

A number of classification systems have been proposed to standardize the diagnosis of DCIS (Table 1 ). Many of these are based on the Bloom–Richardson nuclear grading system ( 14 ) . The Van Nuys classification stratifies patients into three groups ( 19 – 21 ) . Non–high-nuclear-grade cases are placed in group 1 if they lack comedo-type necrosis and in group 2 if they have necrosis, and cases with high nuclear grade are placed in group 3. The 8-year actuarial disease–free survival rates are 93%, 84%, and 61%, respectively ( P = .05 comparing group 1 to 2; P = .003 comparing group 2 to 3) ( 19 ) . The Van Nuys prognostic index, which also includes age, tumor size, and margin width, can be used to place patients in categories corresponding to treatment algorithms ( 20 , 21 ) . However, tissue processing by the Van Nuys protocol is complex, thus limiting its generalizability to clinical practice.

The Armed Forces Institute of Pathology (AFIP) classification system separates DCIS into three grades (grades 1–3) on the basis of the presence or absence of nuclear atypia and necrosis, with grade 3 representing the presence of nuclear atypia and necrosis and grade 1 representing the absence of both ( 25 ) . The DCIS grade is then incorporated in a broader classification called ductal intraepithelial neoplasia (DIN), which includes non-DCIS entities, intraductal hyperplasia (IDH), and atypical ductal hyperplasia (ADH). There are three DIN categories, DIN-1 to DIN-3. DIN-1 includes IDH (grade 1a), ADH (grade 1b), and grade 1 DCIS (grade 1c); DIN-2 includes grade 2 DCIS; and DIN-3 includes grade 3 DCIS.

No single classification scheme has been universally accepted, and experts disagree as to which is the most reproducible. Moreover, many physicians favor classification schemes that also have prognostic capabilities. A case-control study of 141 patients compared the abilities of architectural classification and four other classification schemes ( 18 , 19 , 23 , 24 ) to predict recurrence after local excision ( 27 ) . A statistically significant association between nuclear grade and recurrence was found with both the European Breast Screening Working Groups classification ( P = .009) and the Van Nuys classification ( P = .001) ( 19 , 23 ) . Because researchers are using multiple classification systems, it is essential that they describe the pathologic features that they observe in sufficient detail to allow comparisons to be drawn between different studies. Molecular markers should be studied to see whether they have the potential to facilitate improved categorization of DCIS and stratify patients into prognostic groups that facilitate appropriate treatment choices.

M icroinvasion

A further dilemma in the classification and histological analysis of DCIS is microinvasion. DCIS with microinvasion (DCISM) was defined in 2003 by the American Joint Committee on Cancer as “the extension of cancer cells beyond the basement membrane into the adjacent tissues, with no single focus larger than 1 mm in greatest dimension (T1mic)” ( 28 ) . The incidence rate of DCISM among all breast cancer cases is 0.68%–2.4%, and DCISM is seen in approximately 14% of DCIS cases ( 29 – 31 ) . The potential for DCISM should be suspected for DCIS tumors that are large, have comedo-type histology, and contain necrosis ( 31 – 33 ) . For example, in 115 mastectomy specimens with a biopsy diagnosis of DCIS, occult invasion was found in one of 60 lesions 25 mm or smaller but in 16 of 55 lesions larger than 26 mm ( 34 ) . Silverstein and Lagios ( 32 ) noted that DCISM lesions are usually larger and more likely to be palpable than DCIS lesions. In 24 DCISM specimens, the average lesion size was 38 mm and 29% of the lesions were clinically palpable, whereas in 909 DCIS specimens, the average lesion size was 26 mm and 14% were clinically palpable ( 32 ) . DCISM may be difficult to differentiate from noninvasive entities such as cautery artefact, sclerosing adenosis, entrapment of DCIS cells, or retrograde DCIS extension. It is imperative, therefore, that large DCIS specimens be examined carefully because they may well contain evidence of DCISM. Inadequate sampling could result in misdiagnosis and consequent undertreatment of patients with DCISM.

DCISM may also result in axillary lymph node metastases, whereas patients with DCIS should not, by definition, have axillary metastases. Nevertheless, axillary metastases (0–20%) have been reported in both DCIS and DCISM, implying either misdiagnosis or occult DCISM (Table 2 ). A higher suspicion for axillary metastases with DCISM can be obtained from the primary tumor characteristics. Statistically significant independent predictors of lymph node metastases in DCISM are comedo DCIS ( P <.03) and the number of DCIS-involved ducts ( P <.002) ( 61 ) . Another study examined outcomes in a total of 1248 patients, including 722 patients with DCIS; 72 patients with DCISM type 1, i.e., with a few single infiltrating tumor cells; 171 patients with DCISM type 2, i.e., a cluster of infiltrating cells; and 283 patients with IDC–DCIS, i.e., a gradable infiltrating focus of carcinoma classified as infiltrating ductal carcinoma with a predominant DCIS component ( 62 ) . Patients with DCIS and DCISM type 1 had better metastasis-free and overall survival than patients with DCISM type 2, and patients with DCISM type 2 had better metastasis-free and overall survival than patients with IDC–DCIS. Axillary lymph node metastases were observed in none of the patients with DCISM type 1, in 10% of the patients with DCISM type 2, and in 27.6% of patients with IDC–DCIS. Until prospective data are obtained incorporating the new definition of DCISM and comparing its outcome with those of DCIS and invasive carcinoma, it seems prudent to categorize DCISM as a small invasive tumor with a favorable outcome and to base prognostic and treatment decisions thereupon accordingly.

D iagnosis

DCIS is diagnosed primarily via mammography followed by stereotactic needle biopsy. However, new techniques such as magnetic resonance imaging (MRI) and analysis of ductal cytology aim to improve DCIS detection.

Mammography

Mammography can sometimes be used for histologic characterization of DCIS. For example, linear calcifications and multiple clusters of fine granular calcifications are usually consistent with poorly and well-differentiated DCIS, respectively ( 63 , 64 ) . However, prediction of histology and size by mammography can be unreliable ( 65 , 66 ) . In one study, DCIS lesions were found to be 2 cm larger on histologic examination than the size estimated by mammography for eight (44%) micropapillary tumors and five (12%) pure comedo tumors ( 67 ) .

MRI

MRI may be more sensitive for DCIS detection than mammography but can lack specificity. The enhancement pattern on MRI of DCIS lesions is usually focal but can be diffuse or ductal ( 68 ) . MRI can miss small lesions; in one study, enhancement was demonstrated in only 10 of 13 patients, and the mean lesion diameter in the three unidentified cases was 3.7 mm, whereas that in the identified cases was 13 mm ( 69 ) . In another study, tumor extent was determined accurately for 95% of the 22 tumors with RODEO (rotating delivery of excitation off resonance) MRI and for 74% of 19 tumors with mammography ( 70 ) . Gilles et al. demonstrated contrast enhancement with contrast material–enhanced dynamic MRI in 34 of 36 women with DCIS tumors of 4–75 mm in diameter ( 71 ) . In this study, tumors were stained with immunoperoxidase to identify factor VIII–related antigen, which is associated with neoangiogenesis. Contrast enhancement was associated with the size and density packing of DCIS-involved ducts. The two cases that did not appear on the MRI demonstrated weak tumor angiogenesis around the ducts involved by DCIS. An earlier study demonstrated that the mean microvessel size evaluated with a morphometric method is higher in DCIS tissue than in normal breast tissue ( 72 ) . This finding suggests that contrast enhancement is due to the presence of tumor angiogenesis in the stroma surrounding DCIS-involved ducts.

MRI may improve the ability to detect and determine the size and nature of DCIS. It may be particularly useful in evaluating residual disease, occult invasion, and multicentricity ( 73 ) . MRI may also have a role in the evaluation of patients at high risk of invasive disease or in patients with painful, dense, or nodular breasts for whom physical exam or mammography is difficult or unreliable. Cost–benefit analyses will be required to determine the applicability of MRI to everyday clinical practice.

Biopsy and New Diagnostic Techniques

Retrospective analysis of the diagnostic and therapeutic procedures used in the European Organization for Research and Treatment of Cancer (EORTC) 10853 trial revealed varied diagnostic practices among physicians ( 74 ) . DCIS is usually diagnosed by stereotactic core-needle biopsies (SCNB) or vacuum-assisted needle biopsies, although SCNB can underestimate the incidence of invasive disease ( 75 – 77 ) . For example, of 59 patients diagnosed with DCIS by SCNB, 29% were subsequently found to have invasive disease after surgery ( 78 ) . There was no association between age, nuclear grade, tumor size, or the presence of a mass and the likelihood of invasion.

The potential inaccuracy of current biopsy techniques has prompted investigation of other techniques that may be able to detect DCIS preclinically. These include nipple aspirate fluid (NAF), ductal lavage, and random periareolar fine-needle aspiration ( 79 – 81 ) . However, these techniques also have limitations–in particular, they have poor sensitivity for the detection of invasive cancer ( 82 ) . At present, therefore, mammography and SCNB remain the standard diagnostic approaches for DCIS.

E pidemiology

Increased participation in screening programs has resulted in an increase in the diagnosis of DCIS ( 8 , 83 ) . Surveillance, Epidemiology, and End Results (SEER) ¹ data show a 557% increase in DCIS over 20 years, from 2.8% of newly diagnosed breast cancers in 1973 to 12.5% in 1992 ( 84 ) . The incidence of carcinoma in situ (includes DCIS and lobular carcinoma in situ ) in screened women is approximately 20% (Table 3 ). Disproportionate increases in DCIS among older women and higher rates of DCIS in the United States than in Europe may be due to a lead time bias and increased biopsy rates, respectively ( 99 – 101 ) . DCIS also occurs in males, usually at an older age than in females, and is more commonly low grade in males than in females ( 102 ) . The actual prevalence of DCIS in the population is difficult to estimate because most patients are asymptomatic, and rates from autopsy specimens are variable (between 0.2% and 14.7%) ( 103 – 110 ) .

R isk F actors

Risk factors for the development of invasive breast cancer and DCIS are similar ( 111 – 122 ) (Table 4 ). For example, Claus et al. compared 875 patients with DCIS with 999 age-matched control subjects ( 111 ) and found that DCIS patients were statistically significantly more likely than control subjects to have had a family history of breast cancer, especially if they were diagnosed before age 50, and to have had a previous breast biopsy. The role of hormone replacement therapy in the etiology of DCIS is inconclusive ( 116 – 120 ) . The recent Women's Health Initiative randomized controlled trial of estrogen plus progestin versus placebo in more than 16 000 healthy postmenopausal women found an increased risk of invasive breast cancer (HR = 1.24; nominal 95% confidence interval [CI] = 1.01 to 1.54) with hormonal therapy but found no statistically significant difference in rates of in situ cancer (HR = 1.18; 95% CI = 0.77 to 1.82; P = .09) ( 123 ) .

P rognostic F actors

Characteristic features of DCIS provide prognostic information. Patients with large, high-grade DCIS with necrosis are at highest risk of developing recurrence ( 18 , 124 – 130 ) . The choice of therapy may also affect recurrence, with margin width being an important determinant of outcome. Silverstein et al. ( 129 ) found no difference in local recurrence rates for patients with margins widths of greater than 1 mm who did or did not receive postoperative radiotherapy; patients with margins that were less than 1 mm had a statistically significantly higher likelihood of recurrence if they did not receive postoperative radiotherapy (relative risk = 2.54; P = .01). Other prognostic factors include age, symptomatology, and tumor markers ( 131 – 133 ) . However, Warnberg et al. found no pattern of tumor markers that could identify the progression of DCIS to invasive carcinoma ( 134 ) . The two prospective randomized trials of lumpectomy with or without radiation (see below) also allowed assessment of prognostic factors. In the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-17 trial, evaluation of nine pathologic features by univariate analysis identified comedo necrosis ( P = .002), histologic type (solid versus cribriform) ( P = .006), lymphoid infiltrate ( P = .02), and focality ( P = .02) as being independently related to ipsilateral breast tumor recurrence ( 135 ) . In a multivariable analysis, however, only comedo necrosis remained a statistically significant predictor for recurrence. A multivariable analysis in the EORTC 10853 trial identified young age, symptomatic detection of DCIS, growth pattern, involved margins, and treatment by local excision alone as being the most important factors associated with an increased risk of local recurrence ( 136 ) . Poorly differentiated histologic types were associated with higher rates of distant metastases ( P = .0133) and death ( P = .009).

Prognostic factors such as histologic grade, necrosis, and molecular markers are being incorporated in classification schemes to predict the risk of recurrence ( 26 ) . Such schemes have not been tested in prospective randomized studies, but they are already being used by some physicians in decision-making regarding the need for adjuvant therapy.

P athogenesis and N atural H istory

A spectrum of disease ranging from benign entities to invasive carcinoma has been established for a number of cancers ( 137 ) . This model may also apply to breast cancer, with hyperplasia and atypical hyperplasia preceding DCIS, which may in turn be the final step in the pathway prior to the development of invasive disease ( 138 , 139 ) . Evidence that DCIS can progress to invasive cancer is suggested by similarities in their biology. For example, in xenograft mice studies ( 140 ) , estrogen receptor (ER)–negative, comedo DCIS proliferated at a faster rate than did ER-positive DCIS, and the proliferation of ER-positive but not ER-negative DCIS was stimulated by estrogen. Injection of the antiestrogen fulvestrant led to an increase in apoptosis but no change in proliferation in ER-positive DCIS samples ( 141 ) . Further evidence of the association between DCIS and invasive disease can be seen from histologic examination of DCIS and invasive cancer lesions in humans. For example, Mommers et al. ( 142 ) used image cytometry of hyperplastic lesions, DCIS, and invasive cancer and detected progressive changes in nuclear features in the spectrum from intraductal proliferation to invasive cancer. Similarities in the molecular marker profile of DCIS and invasive cancer (see below) further substantiate the concept of DCIS progression to invasive disease.

Laboratory and clinical data therefore indicate that DCIS can progress to invasive disease. Why and how often DCIS progresses to invasive disease and whether any types of DCIS are more likely to progress than others are less well understood. An estimate of the risk of progression of DCIS may be obtained from patients previously misdiagnosed with benign breast disease who received no treatment and for whom subsequent evaluation of biopsy specimens revealed DCIS (Table 5 ). In the largest series, by Eusebi et al. ( 152 ) , only 14% of such women developed invasive cancer, although the average progression rate from all studies combined was 43% (Table 5 ). Further investigation by means of case–control or cohort studies will be required to clarify our understanding of the natural history of DCIS and will provide better information on its role in breast carcinogenesis ( 153 ) .

M olecular M arkers

An insight into the association of DCIS and invasive cancer can be obtained from their respective molecular markers. Similarities in their molecular profiles, particularly with respect to their genotypic appearance, suggest a common origin. In addition to providing information on the pathogenesis of DCIS, these markers also provide valuable information for the diagnosis, prognosis, and treatment of DCIS.

Common Markers

Many studies have examined molecular markers in DCIS that are used in clinical practice for invasive breast cancer–especially growth factor and hormone receptors. Although most reports have outlined the presence or absence of these markers, conclusions as to their relevance to diagnosis, prognosis, or treatment recommendations are lacking. There appear to be many parallels between DCIS and invasive cancer with regard to the expression of these markers and their prognostic implications. One such marker is ER status. ER is expressed in 50%–60% of patients with DCIS or invasive cancer (Table 6 ). However, levels of ER expression are higher in DCIS lesions with less aggressive histologic features, such as increasing differentiation and the absence of necrosis, than in those with more aggressive features. For example, Zafrani et al. found ER expression in 83% of well-differentiated DCIS, in 74% of poorly differentiated DCIS, in 91% of DCIS lesions with no necrosis, and in 37% of DCIS lesions with massive necrosis ( 163 ) . As will be discussed later, ER expression is the first molecular marker to have proven relevance to treatment decision-making in DCIS. Progesterone receptor (PR) expression in DCIS appears to be similar to ER expression (Table 6 ). There is less information on the expression of androgen receptors (AR) in DCIS. In one study, AR expression was not associated with ER or PR status ( 172 ) .

Another well-studied marker is HER2, an epidermal growth factor receptor family member. Small studies of HER2 in DCIS have typically demonstrated expression in more than 40% of DCIS patients overall, with a much higher percentage of patients with comedo/high-grade DCIS expressing HER2 than of patients with noncomedo/low-grade DCIS ( 174 ) . HER2 expression is found in more cases of DCIS than in invasive cancer ( 175 ) . Women with invasive cancer whose tumors express HER2 have a poorer prognosis than those whose tumors do not express HER2, and the higher incidence of expression found in a noninvasive condition such as DCIS may therefore seem paradoxical. The incidence of HER2 expression in invasive disease may reflect reduced expression when DCIS progresses to invasive cancer, or it may reflect the fact that most invasive cancers develop from DCIS tumors that have low expression of HER2 but high proliferative rates. Because such tumors progress rapidly, they are in the DCIS stage for only a short period and would therefore be underrepresented in population samples ( 176 ) . HER2 receptor activation via autophosphorylation is also higher in DCIS than in invasive cancer, possibly indicating a role for HER-2 in tumorigenesis ( 177 , 178 ) .

Another receptor involved in growth modulation is the epidermal growth factor receptor (EGFR). In one study EGFR expression was observed in 19 of 40 cases of DCIS (48%), but expression was not associated with histologic grade ( 171 ) . In xenograft models of DCIS, the HER2 inhibitor trastuzumab did not affect apoptosis or epithelial cell proliferation, probably due to poor antibody penetration; however, treatment with the EGFR inhibitor gefitinib, a smaller molecule, resulted in an increase in apoptosis and a decrease in proliferation of the DCIS cells ( 179 ) .

Investigational Markers

Other markers studied in DCIS are less well established but now have an emerging role in clinical practice in invasive breast cancer. Once again, these markers demonstrate the presence of key components for tumor growth and invasion, in both DCIS and invasive cancer, compared with normal tissue. Nuclear positivity for p53 was seen in 34 of 57 poorly differentiated DCIS tumors and in 0 of 22 well-differentiated DCIS tumors, for an overall incidence of 15%–two or three times less than the percentage of invasive tumors that express p53 ( 164 , 180 ) . P-cadherin expression has been strongly associated with high-grade, poorly differentiated DCIS but negatively associated with low-grade, well-differentiated DCIS ( 181 ) . Stepwise progression in enzymatic activity of matrix metalloproteinases (MMP-2) from samples of normal breast tissue to benign breast disease to DCIS and to invasive cancer has been demonstrated ( 182 ) . An increase in microvessel density is also evident in DCIS relative to that of normal tissue ( 183 ) . Guidi et al. noted an increase in stromal microvessels in comedo-type DCIS lesions with marked stromal desmoplasia, HER2/neu expression, and a high Ki-S1 proliferation index. Vascular endothelial growth factor (VEGF) expression in DCIS was reported to be greater than that in adjacent benign cells in 96% of cases ( 184 ) . Thymidine phosphorylase, also known as platelet-derived endothelial cell growth factor, was expressed at high levels in 36% of DCIS samples ( 185 ) . The expression levels of cyclin D1 also increase in a stepwise manner from ADH to DCIS and invasion ( 186 , 187 ) , possibly representing the transformation of these lesions along the neoplastic pathway. Warnberg et al. studied a number of these tumor markers expressed in breast lesions containing both DCIS and invasive cancer and found the marker profile in each histologic type to be almost identical ( 134 ) . The similarity in expression of these markers in DCIS and in invasive cancer suggests a common evolutionary path for these diseases.

Genetic Markers

Studies of genetic markers in DCIS have provided conflicting information. However, most studies have demonstrated genetic aberrations in DCIS similar to those in invasive cancer. For example, in DCIS, chromosomal gains were seen on chromosomes 1q, 6q, 8q, and Xq, and losses were seen on 17p and 22–aberrations similar to those that have been seen in invasive cancer with the exception of the gain of Xq ( 188 ) . O'Connell et al. examined 399 cases of noninvasive breast disease for loss of heterozygosity (LOH) at 15 polymorphic genetic loci known to be involved in invasive breast cancer ( 189 ) . Substantial increases in rates of LOH at certain loci were found in both comedo and noncomedo DCIS in breasts with synchronous invasive cancer relative to breasts without invasive cancer. The hypothesis was that these aberrant loci could be involved in the development of invasive cancer from DCIS. Support for this hypothesis comes from the findings of Stratton et al. ( 190 ) , who tested LOH at loci known to be aberrant in invasive cancer and found LOH at all loci in at least 11% of DCIS samples, a proportion similar to that previously documented for invasive cancer. In another study, in which 38 DCIS specimens, including six cases with synchronous invasive cancers, were examined by genomic hybridization, five of the six invasive cancers had changes identical to those in the DCIS component ( 191 ) . The average number of chromosomal aberrations per case was 2.5 in well-differentiated DCIS, 5.5 in intermediately differentiated DCIS, and 7.1 in poorly differentiated DCIS. However, in contrast to the O'Connell study, Farabegoli et al. compared samples with DCIS alone with those with DCIS and coexistent invasive cancer and found a greater rate of LOH in the 11 out of 15 pure DCIS samples ( 192 ) .

Although the chromosomal aberrations seen in invasive cancer and in DCIS appear to be similar, there is no conclusive proof that the aberrations are involved in or predict the development of invasive cancer. Increased chromosomal aberrations may reflect poorer levels of differentiation within DCIS but do not necessarily predict invasion ( 193 , 194 ) . Some genetic studies support the theory of a stepwise evolution from benign proliferative disease to DCIS to invasive cancer. For example, O'Connell et al. showed that LOH of at least one genetic locus was shared with the synchronous invasive cancer in 37% of 19 patients with usual ductal hyperplasia, 45% of 11 patients with ADH, 77% of 13 patients with noncomedo DCIS, and 80% of 11 patients with comedo DCIS ( 189 ) . Other studies, however, have not demonstrated sharing of genetic characteristics between the different stages of progression. For example, chromosomal alterations (e.g., loss of 16q, gain of 20q) were recorded in all but one of 52 cases of DCIS but were undetectable in 36 cases of benign proliferative breast disease ( 195 ) .

DNA microarrays have also been used to analyze the association between DCIS and invasive cancer. The first study that used microarrays to characterize human breast tumors via their gene expression patterns in breast cancer, by Perou et al., included just one case of DCIS ( 196 ) . Subsequently, a small comparative study of cDNA expression in DCIS and invasive breast cancer used unsupervised hierarchical clustering to identify 978 genes whose expression was different between the two diseases. By supervised comparison, 148 of these genes were shown to have functions in progression ( 197 ) . Whether these genes are involved in progression to invasive disease or merely progression to states of poorer differentiation within DCIS cells was addressed by another study, which used microarrays to generate profiles of premalignant, preinvasive, and invasive stages of breast cancer. There were no consistent differences in the profiles between the different stages of breast cancer, although distinct patterns were seen for different histologic grades ( 198 ) . Warnberg et al. reported a similar finding in a study of tumor markers in DCIS: Expression of the seven tumor markers was associated with tumor grade but not with invasiveness ( 134 ) . Also, Mommers et al. noted extensive differences in cytometric features between well- and poorly differentiated tumors within DCIS or within invasive cancer ( 142 ) . However, well-differentiated DCIS and well-differentiated invasive cancer had similar characteristics, as did their poorly differentiated counterparts. Another study, using serial analysis of gene expression, detected no clear molecular signature differences in separate analysis of DCIS, invasive, and metastatic breast cancer, although some genes speculated to be associated with cell growth, differentiation, and survival were expressed at higher levels in DCIS than in normal or invasive/metastatic cancerous breast tissue ( 199 ) .

From the existing molecular evidence it therefore seems unlikely that there is a natural stepwise progression from low-grade DCIS to high-grade DCIS to low-grade invasive cancer and then to high-grade invasive cancer. It may be that low-grade DCIS progresses to low-grade invasive cancer and high-grade DCIS progresses to high-grade invasive cancer. So far, genetic studies in DCIS have proved provocative rather than conclusive. On the basis of LOH and microarray studies, there appear to be similarities between the DCIS genotype and invasive cancer. However, proving or identifying features consistent with the sequential progression from noninvasive to invasive disease has proved elusive. The ability to use genetic markers to predict which DCIS patients are likely to progress and what their outcome is likely to be has yet to be addressed, although preliminary evidence from microarray analyses in invasive cancer suggests that genetic markers may be used in this way in the future ( 200 ) .

Further advances in characterization of DCIS have been made with the development of proteomic analysis, which identifies protein expression at the functional level. The only reported study of matched normal breast cells and DCIS evaluated just six patients ( 201 ) . From 315 protein spots excised after analyses of proteins from 10 sets of two-dimensional gels, 57 proteins were found to be differentially expressed. Many of these proteins had not previously been associated with breast cancer. Indeed, for most of the differentially expressed proteins, the corresponding genes had not been previously identified by nucleic acid–based techniques as being differentially expressed. Therefore, the authors suggested caution in the assumption that identification of mRNA levels reflects those of functional proteins.

T reatment

There are many parallels in the treatment paradigms of DCIS and invasive cancer, as has been outlined in detail in recent reviews ( 202 , 203 ) . Consensus guidelines ( 1 , 11 , 75 , 101 , 204 , 205 ) agree that the goal of treatment for DCIS is breast conservation, with optimal cosmesis and with a minimum risk of subsequent invasive or in situ recurrence (Table 7 ).

Mastectomy

No prospective randomized trial has compared outcomes after mastectomy with those after breast-conserving surgery. A meta-analysis of studies published up to 1998 reported local recurrence rates of 22.5% (95% CI = 16.9% to 28.2%), 8.9% (95% CI = 6.8% to 11%), and 1.4% (95% CI = 0.7% to 2.1%) for lumpectomy alone, lumpectomy with radiation, and mastectomy, respectively ( 206 ) . Although other studies ( 207 , 208 ) also suggest a reduction in recurrence rates with mastectomy versus breast-conserving surgery, the lack of a difference in survival between the two approaches has led to a decline in the use of mastectomy ( 84 , 209 ) . Mastectomy should be reserved for patients with multicentric disease, large lesions, other contraindications to breast conservation, or a personal preference for mastectomy.

Lumpectomy Alone

The option of performing lumpectomy alone in good-risk patients is attractive but has not yet been substantiated by prospective data. Schwartz et al. performed lumpectomy alone on 224 patients with good-risk DCIS (size of less than 2–3 cm, absence of invasion, margins greater than 10 mm, localized disease, low nuclear grade, probable good cosmetic result); after a median follow-up of 52 months, the recurrence rate was 19.7% ( 210 ) .

An ongoing Radiation Therapy Oncology Group study (RTOG 9804) aims to randomly assign 1790 patients either to radiation therapy or observation, with the option of tamoxifen in either group. The patients must have lesions that are 2.5 cm or less in diameter, low- or intermediate-grade nuclei, and inked margins that are at least 3 mm in diameter. The primary outcomes will be the difference in local recurrence and distant disease–free survival rates. The recently closed European Cooperative Oncology Group (ECOG E-5194) trial was similar. It accrued approximately 1000 patients with DCIS lesions that were 2.5 cm or less and low- or intermediate-grade nuclei or 1 cm or less and high-grade nuclei (both groups had to have negative margins that were at least 3 mm). Patients were treated by lumpectomy alone, and the primary outcome will be local recurrence rates at 5 and 10 years. The results of these studies should provide information on the efficacy of lumpectomy alone with good-risk DCIS and may allow the development of criteria to identify subgroups of patients who do not require adjuvant radiation therapy. However, at present, lumpectomy alone should not be recommended as a standard option for DCIS patients because there are insufficient prospective data to support the claim that lumpectomy alone in patients with good-risk DCIS is equivalent to the validated approach of lumpectomy and radiotherapy.

Axillary Lymph Node Dissection

The underestimation of invasion in biopsies of DCIS and the improvement in sentinel node techniques has prompted some authors to advocate lymphatic mapping at the time of surgery ( 211 , 212 ) . In patients treated with mastectomy, it is impossible to perform subsequent lymphatic mapping if invasive cancer is subsequently found. If invasive cancer is diagnosed after lumpectomy, then a repeat surgical procedure involving axillary node dissection may be required. However, because the approximate incidence of axillary lymph node metastases with DCIS is only 1.4% and the clinical significance of such metastases is unknown, sentinel node sampling may represent overtreatment of these patients and could subject them to unnecessary risk ( 213 ) . Axillary lymph node involvement is higher with DCISM (5.1%) than with DCIS (1.4%) (Table 2 ); therefore, large DCIS lesions that have comedo-type histology with necrosis may contain an invasive component and must be evaluated carefully. Possible indications for axillary lymph node dissection are the suspicion of microinvasion, such as the presence of a clinically palpable node or large comedo-type DCIS. If a patient or physician decides that a mastectomy is necessary, then an axillary lymph node dissection at that time may also be appropriate ( 204 ) . However, at the present time sentinel lymph node testing in DCIS is investigational and should be performed as part of a clinical trial only.

Lumpectomy and Radiation

Nonrandomized studies of lumpectomy and radiation therapy have reported recurrence rates of 4%–25% ( 214 ) . Three prospective randomized trials have been performed to define the role of radiation therapy in the management of DCIS with breast conservation (Table 7 ). These are NSABP B-17, EORTC 10853, and the UK Coordinating Committee on Cancer Research (UKCCCR) trial.

In NSABP B-17, 818 women with DCIS were assigned to receive lumpectomy or lumpectomy combined with radiation therapy ( 215 – 217 ) . After 12 years of follow-up, the cumulative incidence of invasive and noninvasive ipsilateral breast tumors combined was 31.7% in the lumpectomy-alone arm and 15.7% in the lumpectomy-plus-radiation arm ( P <.001). The 12-year overall survival was 86% for patients in the lumpectomy group and 87% for patients in the lumpectomy and radiation therapy group ( P = .08). These results were the basis for the recommendation that lumpectomy and radiation be adopted as the standard of care for treating DCIS in patients without contraindications to this approach.

EORTC 10853 was almost identical to NSABP B-17 with regard to its design and results ( 218 ) . This trial randomly assigned 1010 patients (1002 evaluable) to lumpectomy alone or to lumpectomy plus radiation. After a median follow-up of 4.25 years, local recurrence was documented in 17% of patients in the lumpectomy-alone arm and 11% in the lumpectomy-plus-radiation arm. The shorter follow-up period may have resulted in the lower recurrence rates than those in NSABP B-17. Patients with free margins had little difference in local recurrence rates with the addition of radiation therapy (12% versus 14%). In patients with close or involved margins, the addition of radiation to lumpectomy reduced the local recurrence rate from 32% to 16%. This study further substantiates the role of radiation in the local treatment of DCIS.

The third trial, by the UKCCCR DCIS working party, was a joint study in the United Kingdom, New Zealand, and Australia that used a 2-by-2 factorial design to randomly assign 1701 patients with DCIS after lumpectomy to radiotherapy or not and to tamoxifen or not ( 219 , 220 ) . The primary outcome was the incidence of subsequent ipsilateral invasive breast cancer. The complex study design used in this trial makes interpretation of the data somewhat difficult. Among the 1030 patients who received radiotherapy or control treatment, there were new breast events in 7% of patients in the radiotherapy group and 16% of those in the no-radiotherapy group ( P <.001).

The conclusion from these three prospective randomized trials is that the addition of radiation therapy to lumpectomy results in an approximately 50% reduction in breast cancer recurrence. Patients with high-grade DCIS lesions and positive margins benefited most from the addition of radiation therapy. It is not yet clear which patients can be successfully treated with lumpectomy alone; until further prospective studies answer this question, radiation should be recommended after lumpectomy for all patients without contraindications.

Lumpectomy With Radiation and Tamoxifen

Adjuvant systemic therapy in DCIS is based on the success of adjuvant tamoxifen in reducing local breast tumor recurrences of node-negative invasive cancer from 2.7% to 0.53% ( P <.001) and contralateral breast cancers from 2.2% to 1% after 4 years of follow-up ( P = .009) ( 221 ) . Two prospective randomized trials (NSABP B-24 and UKCCCR) have examined the effect of adjuvant tamoxifen on outcomes in DCIS patients.

NSABP B-24 was a double-blind prospective randomized trial of 1804 patients followed for a median of 83 months ( 217 , 222 ) . Sixteen percent of the women on NSABP B-24 had positive resected sample margins, and 33% were less than 50 years of age. Patients were treated with lumpectomy and radiation therapy followed by either tamoxifen or placebo. After 7 years of follow-up, the cumulative incidence of all ipsilateral breast recurrences was 11.1% in the placebo arm and 7.7% in the tamoxifen arm. Patients with unknown or positive margins had a 45.1% reduction in the risk of recurrence with tamoxifen relative to placebo, compared with a 21.1% reduction in patients with negative margins. The 7-year cumulative incidence of contralateral breast tumors was 4.9% in the placebo group and 2.3% in the tamoxifen group ( P = .01). Overall survival after 7 years was 95% in both arms, and most of the deaths (75% in the placebo arm and 76% in the tamoxifen arm) occurred before the development of any breast cancer events.

A recent report has evaluated the benefit of tamoxifen according to ER status for 628 of the patients in NSABP B-24 ( 173 ) . Among the 77% of DCIS patients who were ER positive, those receiving tamoxifen had a relative risk of breast cancer events of 0.41 (95% CI = 0.25 to 0.65; P <.001) compared with placebo. In ER-negative patients, the relative risk was 0.8 (95% CI = 0.41 to 1.56; P =.51). Thus, the benefit from tamoxifen therapy was restricted to ER-positive DCIS. However, tamoxifen can result in potentially fatal toxicities. In NSABP B-24, there were nine cases of deep vein thrombosis and two cases of pulmonary embolism in the tamoxifen arm, compared with two cases and one case, respectively, in the placebo arm. Subsequently, it was reported that there were also five strokes in the tamoxifen arm and one in the placebo arm, none of which was fatal ( 223 ) . Among the stroke patients, three had a history of hypertension, one had had a previous transient ischemic attack, and four were over the age of 65.

The UKCCCR trial, involving 1701 patients, has reported results on 1576 patients regarding the use of tamoxifen in the adjuvant setting for DCIS ( 219 , 220 ) . Patients with positive margins were excluded, and 10% of patients were less than 50 years of age. After a median follow-up of 53 months, ipsilateral events were reduced from 15% in the no-tamoxifen arm to 13% in the tamoxifen arm ( P = .42), and contralateral events were reduced from 3% to 1% ( P = .07).

In summary, NSABP B-24 demonstrated a statistically significant reduction in breast cancer recurrence with tamoxifen use but no change in overall survival, whereas the UKCCCR trial did not find a reduction in either recurrence or survival. An explanation for the difference in outcomes between the studies may be the exclusion of patients with positive margins or the older age of patients in the UKCCCR trial. The design of UKCCCR may also have biased the results. On the basis of the results of NSABP B-24, tamoxifen therapy may not be appropriate for DCIS patients with cardiovascular comorbidities, those over 65 years old, or those on coumadin. Careful selection of patients who are most likely to benefit from tamoxifen and who have the least risk of serious side effects is especially important because there is no survival advantage to tamoxifen. Analysis of ER status also appears to be essential in guiding treatment decisions.

Newer hormonal agents, especially those with toxicity profiles that differ from that of tamoxifen, could extend the use adjuvant hormonal therapy for DCIS to patients for whom tamoxifen is unsuitable. NSABP B-35 opened in January 2003 and aims to randomly assign 3000 postmenopausal patients with ER-positive DCIS to adjuvant anastrozole or tamoxifen after local therapy (http://www.nsabp.pitt.edu/B35_Information.htm [last accessed: April 29, 2004]). The primary endpoint will be the time to any breast cancer event. An International Breast Cancer Intervention Study (IBIS-II) will be a two-arm double-blinded trial of 4000 ER-positive DCIS postmenopausal patients and will compare tamoxifen with anastrozole (http://www.IBIS-trials.org [last accessed: April 29, 2004]). The primary endpoint is local control and prevention of contralateral disease after local excision.

C onclusion

An improved understanding of the pathogenesis and natural history of DCIS will help to clarify the optimal management of this increasingly frequent diagnosis. Current data support the use of breast-conserving surgery and radiotherapy as the most appropriate therapy for DCIS. Tamoxifen appears to reduce the risk of recurrence, but no survival benefit has been documented, and the presence of comorbid illness or old age may increase toxicity. Patients should informed of the risk–benefit ratio of therapy prior to definitive treatment decisions.

Although the RTOG and ECOG studies may identify patients that can be treated by lumpectomy alone, a prospective randomized trial comparing all treatment options in patients with good-risk DCIS will be required to definitively determine the respective roles of adjuvant therapy in this select population. Perhaps most important in treatment decision making is the development of molecular profiling of DCIS. Trials in invasive breast cancer have assessed responsiveness and outcomes with chemotherapy relative to their molecular signature ( 224 , 225 ) . Outlining the molecular signature of DCIS will also help direct appropriate therapy and provide prognostic information. The American Association for Cancer Research Intraepithelial Neoplasia (AACR IEN) Task Force was devised to highlight and direct areas of research in the prevention of cancer. This group has emphasized the importance of precancerous lesions and has suggested clinical trial designs based on the prevention of DCIS ( 226 ) . An improved understanding of DCIS will help determine its clinical relevance to patients and will help in the development of effective treatment strategies. Appropriate management of DCIS may prevent the consequent morbidity and mortality associated with invasive breast cancer and so continued research in this area must be encouraged.

Table 1.

Classification and grading systems for DCIS by year *

Author and reference	Predominant histologic variable(s)	DCIS categories
Lagios ( 13 )	Nuclear grade, architecture, necrosis	High grade, intermediate grade, low grade
Ottensen ( 22 ) (Danish)	Histological growth pattern, lesion size, nuclear size, comedonecrosis, subhistologic type	Microfocal, diffuse, tumor forming
Bellamy ( 17 ) (Edinburgh)	Histologic pattern, nuclear grade, necrosis, involved duct counts	Comedo, solid, cribriform, micropapillary, all with nuclear grade
Poller ( 18 ) (Nottingham)	Architecture	Comedo, DCIS with necrosis (nonpure comedo), DCIS without necrosis
Holland ( 23 )	Cytonuclear differentiation, architectural, differentiation (cell polarization)	Poorly differentiated, intermediately differentiated, well-differentiated
European Breast Screening Groups ( 24 )	Cytonuclear differentiation	Poorly differentiated, intermediately differentiated, well-differentiated
Page ( 15 , 16 )	Nuclear grade, architecture, necrosis	High grade, intermediate grade, low grade
Silverstein ( 19 ) (Van Nuys)	Nuclear grade, comedo-type necrosis	Group 1 (non–high grade without necrosis), group 2 (non–high grade with necrosis), group 3 (high grade)
Consensus Committee ( 11 )	Nuclear grade, necrosis, polarization, architectural pattern	High grade, intermediate grade, low grade
Tavassoli ( 25 ) (AFIP)	IDH/ADH, grade (atypia, necrosis)	DIN 1a, -b, -c; DIN2; DIN3
Warnberg ( 26 )	Histologic grade, necrosis, lymphoid infiltration, mitosis, c-erb-2, p53, progesterone receptor, Bcl-2	Phenotype A, phenotype B

*

DCIS = ductal carcinoma in situ; DIN = ductal intraepithelial neoplasia; IDH = intraductal hyperplasia; ADH = atypical hyperplasia.

Table 2.

Axillary lymph node metastases in selected studies involving DCIS and DCISM *

Author and reference	No. of ALND	No. of positive ALND
DCIS
Amichetti et al. ( 35 )	97	0
Ashikari et al. ( 36 )	113	1
Brown et al. ( 37 )	21	1
Carter et al. ( 38 )	26	1
Dimpfl et al. ( 39 )	99	1
Fisher et al. ( 40 )	78	0
Gump et al. ( 41 )	64	1
Lagios et al. ( 34 )	87	2
Parker et al. ( 42 )	104	1
Patchefsky et al. ( 43 )	51 †	3
Recht et al. ( 44 )	13	0
Rosner et al. ( 45 )	210	8
Schuh et al. ( 46 )	52	1
Silverstein et al. ( 21 )	427	2
Solin et al. ( 47 )	86	0
Sunshine et al. ( 48 )	61	0
von Rueden et al. ( 49 )	32	0
Total	1621	22 (1.4%)
DCISM
Akhtar et al. ( 50 )	25	0
Kinne et al. ( 51 )	42	4
Jimenez et al. ( 52 )	69	5
Le Bouedec et al. ( 53 )	60	3
Padmore et al. ( 29 )	59	0
Patchefsky et al. ( 43 )	16	2
Prasad et al. ( 54 )	15	2
Rosner et al. ( 55 )	34	1
Schnitt et al. ( 33 )	38	0
Schuh et al. ( 46 )	30	6
Schwartz et al. ( 56 )	25 ‡	0
Silver et al. ( 30 )	38	0
Silverstein et al. ( 97 )	20	0
Simpson et al. ( 57 )	5	1
Solin et al. ( 58 )	39	2
Wong et al. ( 59 )	33	0
Zavotsky et al. ( 60 )	14	2 §
Total	547	28 (5.1%)

*

DCIS = ductal carcinoma in situ; DCISM = DCIS with microinvasion; ALND = axillary lymph node dissection.

†

Includes 16 cases of DCISM. Two of the three positive ALNDs were from DCISM primary tumors.

‡

Not known whether all 25 patients had ALND.

§

Two patients had tumor-involved sentinel lymph nodes but no lymph nodes at ALND.

Table 3.

Percentage of cases of carcinoma in situ (CIS) (Includes DCIS and LCIS) from cancers diagnosed in the study arm of selected screening studies *

Author and reference	Screening project	No. of patients	No. of cancers (Invasive + CIS)	CIS (%) †
Alexander ( 85 )	Edinburgh	22 944	489	8.5
Thomas ( 86 )	Guildford	16 586	91	15
TEDBC ( 87 )	TEDBC	39 773	210	22
UKCCCR ( 88 )	UKCCCR	37 530	287	18
Miller ( 89 )	Canadian	25 214	663	11
Miller ( 90 )	Canadian	19 711	693	10
Olivotto ( 91 )	Canadian	203 141	991	17
Tabar ( 92 )	Swedish 2 county	77 080	1426	8.6
Andersson ( 93 )	Malmo	21 088	588	16
Frisell ( 94 )	Stockholm	40 318	428	10
Bjustam ( 95 )	Gothenburg	11 724	144	14
Chu ( 96 )	HIP	30 131	371	9.4
Bobo ( 97 )	NBCCDEDP	564 708	3780	24
May ( 98 )	NBCCDEDP	230 143	1250	26
Ernster ( 7 )	BCSC	540 738	3266	18

*

CIS = carcinoma in situ; DCIS = ductal carcinoma in situ; LCIS = lobular carcinoma in situ; TEDBC = UK Trial of Early Detection of Breast Cancer; UKCCCR = The United Kingdom Co-Ordinating Committee on Cancer Research; HIP = Health Insurance Plan; NBCCDEDP = National Breast and Cervical Cancer Early Detection Program; BCSC = Breast Cancer Surveillance Consortium.

†

Percentage of cancers diagnosed by screening that were carcinoma in situ .

Table 4.

Studies of risk factors for the development of DCIS *

Author and reference	Study design	No. of patients with DCIS	Main risk factors identified
Claus ( 111 )	Case–control	875	Family history, previous breast biopsy, fewer full-term pregnancies, older at first full-term pregnancy, older menopause
Kerilowske ( 112 )	Cross-sectional	102	Family history, palpable mass, increasing age
Brinton ( 113 )	Case–control	199	Family history, age at first live birth
Weiss ( 114 )	Case–control	228	Family history, previous breast biopsy, nulliparity, African American
Longnecker ( 115 )	Case–control	233	Family history, unopposed estrogen replacement therapy
Gapstur ( 116 )	Prospective cohort	131	Family history, older at first live birth
Henrich ( 117 )	Case–control	32	No risk with HRT
Stanford ( 118 )	Case–control	87	No risk with HRT
Schairer ( 119 )	Prospective cohort	154	Estrogen use
Dubin ( 120 )	Case–control	113 †	African American, first live birth before age 20 years
Lambe ( 121 )	Case–control	1368	Nulliparity, increasing age first birth
Trentham-Dietz ( 122 )	Case–control	301	Family history, previous breast biopsy, vitamin A, alcohol

*

DCIS = ductal carcinoma in situ; HRT = hormone replacement therapy.

†

Includes in situ and small invasive carcinomas.

Table 5.

Results of follow-up of untreated DCIS *

Author and reference	No. of cases	Follow-up, y	Histology	No. (%) of cases that developed invasive carcinoma
Dean ( 143 )	8	<4	Comedo	6 (75%)
Kraus ( 144 )	4	10–12	Papillary	2 (50%)
Farrow ( 145 )	25	1–6	NR	5 (20%)
Haagensen ( 146 )	11	≤10	Papillary	8 (73%)
Millis ( 147 )	8	0–19	NR	2 (25%)
Rosen ( 149 )	15	1–24	Papillary	8 (53%)
Page ( 150 , 151 )	28	3–30	Noncomedo	9 (32%)
Eusebi ( 152 )	80	5–28	Comedo + noncomedo	11 (14%)

*

DCIS = ductal carcinoma in situ; NR = not reported.

Table 6.

Selected studies of estrogen receptor expression in DCIS *

Author and reference	No. of patients with DCIS	ER expression, % †	PR expression, % †
Helin ( 154 )	5	60	80
Giri ( 155 )	48	42–45	NR
Barnes ( 156 )	32 ‡	75	NR
Malafa ( 157 )	14	57	NR
Bur ( 158 )	38 ‡	65	NR
Pallis ( 159 )	43 §	49	24
Chaudhuri ( 160 )	32	60	NR
Wilbur ( 161 )	23	65	57
Poller ( 162 )	151	32	NR
Zafrani ( 163 )	97	81	73
Bobrow ( 164 )	105	NR	45
Leal ( 165 )	36 ‖	64	66
Albonico ( 166 )	50	66	58
Bose ( 167 )	42	67	38
Shoker ( 168 )	26	90	NR
Ottensen ( 169 )	40	69	62
Warnberg ( 134 )	194	68	43
Claus ( 170 )	219	60	62
Suo ( 171 )	40	88	80
Selim ( 172 )	57	54	49
Allred ( 173 )	628	77	NR
	Total 1920	Average 64%	Average 57%

*

DCIS = ductal carcinoma in situ; ER = estrogen receptor; PR = progesterone receptor; NR = not reported.

†

Percentage of patients with ER or PR expression, as appropriate.

‡

Includes patients with all forms of carcinoma in situ (DCIS and lobular carcinoma in situ ).

§

PR expression was measured in 37 of the 43 patients.

‖

PR expression was measured in 35 of the 36 patients.

Table 7.

Comparison of breast cancer events in prospective randomized trials of DCIS treatment *

Cancer type	EORTC 10853 ( 180 ) 4 years follow-up			NSABP B17 ( 217 ) 12 years follow-up			NSABP B24 ( 217 ) 7 years follow-up			UK/ANZ DCIS † ( 220 ) 52.6 months median follow-up
		L	L+XRT	L	L+XRT	L+XRT+P	L+R+T	No XRT	XRT	No T	T
		(n = 503)	(n = 507)	(n = 403)	(n = 410)	(n = 899)	(n = 899)	(n = 508)	(n = 522)	(n = 782)	(n = 794)
Ipsilateral breast cancer
Total	83	53	124	61	100	72	69	29	114	102
Invasive	40 ‡	24	66	29	49	27	30	15	35	45
Noninvasive	44	29	57	32	51	45	38	14	77	57
Contralateral breast cancer
Total	8	21	18	30	45	25	11	9	21	11
Invasive	5	16	15	18	30	20	6	9	15	10
Noninvasive	3	5	3	12	15	5	5	0	6	1

*

DCIS = carcinoma in situ; L = lumpectomy; XRT = radiotherapy; P = placebo; T = tamoxifen; NR = not reported.

†

For four patients, the laterality and/or the invasive status of the tumor were unknown.

‡

One patient with recurrence of DCIS had a second local recurrence that was invasive.

1

Editor’s note: SEER is a set of geographically defined, population-based, central cancer registries in the United States, operated by local nonprofit organizations under contract to the National Cancer Institute (NCI). Registry data are submitted electronically without personal identifiers to the NCI on a biannual basis, and the NCI makes the data available to the public for scientific research.