Ross, Dara S. MD; Wen, Yong Hannah MD, PhD; Brogi, Edi MD, PhD
Ductal carcinoma in situ (DCIS) is an established morphologic precursor of invasive breast carcinoma. Historically, DCIS constituted only 1% to 2% of breast carcinomas, but it currently represents about 20% to 25% of newly diagnosed breast carcinomas in the United States.1 The predicted number of newly diagnosed cases of breast carcinoma in situ, which consists for the most part of DCIS, was 62,280 for 2009, amounting to 32% of 192,370 estimated new cases of breast carcinoma in the same year.2 The increase was approximately 110% in the 1991 to 2001 decade, with a 20% increase between 1996 and 2001.3 The increase in incidence is for the most part due to improved detection secondary to the widespread adoption of screening mammography. In recent years, the incidence of DCIS with extensive necrosis, often referred to as “comedo” DCIS, has not increased as rapidly as the incidence of “non-comedo” DCIS, a less aggressive form. In particular, an analysis of the Surveillance, Epidemiology, and End Results (SEER) cancer registries showed that between 1991 and 2001 the age-adjusted incidence of comedo DCIS remained unchanged at approximately 7 per 100,000 women, whereas the age-adjusted incidence of non-comedo DCIS increased from 16.5 to 31 per 100,000 women.3
The risk of DCIS has been correlated with multiple parameters, including age, race, family history and genetics, breast density, body mass index, parity, and the use of hormone replacement therapy. DCIS is extremely uncommon before the age of 35 to 39 years. Starting in the fifth decade of life, the incidence of DCIS rises steadily to a peak of 96.7 per 100,000 at ages 65 to 69 and then declines slowly until age 79.4 Between 1996 and 2001 the rate of noncomedo DCIS has increased predominantly in women older than 50 years, whereas the rate of comedo DCIS has decreased predominantly in women younger than 60 years.3 Overall, the magnitude of changes in the incidence of DCIS has been greater in women older than 50 years.3
DCIS is more common in white women compared to African American, Asian, and Hispanic women. An increased risk for DCIS has been reported in women with a positive family history of breast carcinoma and mutation in the BRCA1 or BRCA2 genes, although in one study all DCIS cases were detected in women older than 50 years.5 Women with dense breast tissue also have increased risk for DCIS,6 whereas no consistent association with body mass index has been documented. Women with no children or one child and older age at first birth also show increased risk of developing DCIS. No definitive association between use of hormone replacement therapy and the incidence of DCIS has been documented in randomized trials.4
In 2012 the World Health Organization defined DCIS as “a neoplastic proliferation confined to the mammary ductal-lobular system and characterized by increased epithelial proliferation, subtle to marked cytologic atypia, and an inherent but not necessarily obligate tendency for progression to invasive breast carcinoma.”7 The extent, morphology, and final margin status of DCIS, and also the use of adjuvant radiotherapy (RT) and hormonal therapy, impact the risk for local recurrence. DCIS morphology and growth pattern have some association, and certain subtypes of DCIS occur together more often than others. In any given case of DCIS, focal intratumoral heterogeneity of nuclear grade, architecture, and immunoprofile can be observed.
The nuclear grade of DCIS is classified as low, intermediate, and high on the basis of nuclear size, pleomorphism, chromatin pattern, presence of nucleoli, and mitotic activity. DCIS with low nuclear grade consists of a monotonous population of cells with slightly enlarged nuclei, smooth nuclear membranes, finely dispersed chromatin, inconspicuous nucleoli, and rare mitoses (Fig. 1A). The cells of low-grade DCIS are typically polarized with the longest axis of the cell oriented perpendicularly to the basement membrane, and radially distributed around neoformed extracellular lumina, with resulting orderly cribriform architecture. Polarity of the cells is usually a feature of low-grade to intermediate-grade DCIS, whereas the cells of high-grade DCIS are rarely polarized. DCIS of intermediate nuclear grade consists of a population of cells with intermediate-sized nuclei with coarser chromatin and evident but still inconspicuous nucleoli (Fig. 1B). Mitotic activity may be encountered, but it is rarely high. The neoplastic cells may be polarized. DCIS of high nuclear grade consists of cells with large (>2.5 times the size of a red blood cell) and pleomorphic nuclei that show vesicular and coarse chromatin, with prominent and often multiple and irregular nucleoli (Fig. 1C). The neoplastic epithelial cells usually lack polarity, and mitotic activity is easily detected.8
The architectural patterns of DCIS include solid, cribriform, micropapillary, papillary, spindle cell, and flat (clinging) (Figs. 2A–F). Some combinations of nuclear grade and architectural pattern tend to be more frequent than others (ie, low nuclear grade and micropapillary growth pattern, or high nuclear grade and flat growth pattern), but any combination can occur. The current World Health Organization classification7 does not recognize a low-grade (monomorphic) variant of DCIS with exclusive flat type (clinging) pattern.
Solid DCIS is characterized by a monotonous proliferation of neoplastic cells that fill, expand, and distort a duct. Nuclear grades can range from high to low, although pure solid DCIS with low nuclear grade is fairly rare. In contrast to usual ductal hyperplasia (UDH), the cells of solid DCIS are often orderly distributed and do not show a streaming or swirling arrangement (Figs. 3A, D). Low-grade DCIS with solid architecture sometimes raises the differential diagnosis of lobular carcinoma in situ (LCIS), classical type. The characteristic morphologic features of classical LCIS (lobulocentric growth, pagetoid extension which undermines the normal epithelium, absence of cell dyshesion, and lack of central necrosis and coarse calcifications) help to distinguish between solid low-grade DCIS and classical LCIS in most cases.9 Immunoperoxidase staining for E-cadherin, which shows membranous immunoreactivity in ductal lesions and is absent in lobular neoplasia, is often used to resolve problematic cases or to confirm the morphologic impression.
Cribriform DCIS is characterized by well-defined lumens lined by neoplastic cells. In low-grade DCIS the neoformed lumina have a smooth outline, are fairly regularly spaced, and, most importantly, they are lined by polarized cells. Such an orderly arrangement is rare in high-grade DCIS. The neoplastic cells lining the neoformed glandular lumina are regularly distributed and polarized and lack the haphazard arrangement characteristic of UDH. The differential diagnosis of cribriform DCIS occasionally includes adenoid cystic carcinoma, invasive cribriform carcinoma, and collagenous spherulosis involved by LCIS (Figs. 4A, B). The microlumens in DCIS may contain necrosis, apoptotic cells, calcifications, and secretions.
Micropapillary DCIS consists of ducts in which the neoplastic cells form small epithelial tufts that lack fibrovascular cores. The micropapillae of low-grade DCIS usually are club-shaped, have a narrow base and an expanded bulbous tip with smooth contour. The micropapillae of low-grade DCIS are often long and filiform, with similar height, and sometimes tend to arch to one side. The nuclei of low-grade micropapillary DCIS are fairly monotonous in size and shape from the base to the tip of a micropapilla. Adjacent micropapillae occasionally merge together, forming “Roman” arches lined by cells that are polarized toward the central point of the arch.10 In contrast, the micropapillae of UDH have a broad base, irregular outline, and the nuclei show chromatin condensation and reduced size as they “mature” from the base to the apex of the micropapillae. In high-grade micropapillary DCIS the micropapillae have a broader base, more irregular outline, and necrosis is often present within the lumen.
The neoplastic ductal proliferation composing papillary DCIS is supported by long, thin, and delicate fibrovascular fronds. Even when the DCIS has a more solid growth pattern, the presence of fibrovascular cores is still identifiable within the neoplastic proliferation in the form of minute capillary vessels. Expansion of the duct may lead to attenuation of the surrounding myoepithelial layer, although use of a panel of myoepithelial markers usually allows a definitive distinction between DCIS and invasive carcinoma. For a long time intracystic papillary carcinoma has been regarded as a form of DCIS, but recent studies have shown that this tumor lacks a myoepithelial layer, although a peripheral layer of basement membrane is present in most cases. On the basis of these findings, this rare type of low-grade carcinoma has been renamed “encapsulated papillary carcinoma,” and it is regarded as an indolent variant of invasive carcinoma. Encapsulated papillary carcinoma carries an extremely favorable prognosis and should still be managed as DCIS, despite its morphologic reclassification.11,12 On this issue, the 2012 World Health Organization classification of tumors of the breast states that “at this time there is no universal agreement on how to stage encapsulated papillary carcinomas. In the absence of conventional invasive carcinoma, … such lesions should be staged and managed as Tis disease.”7 Similarly, some cases of solid and papillary carcinoma devoid of myoepithelium likely represent indolent forms of invasive carcinoma rather than DCIS.11 Solid and papillary DCIS often shows neuroendocrine differentiation.13
Spindle cell DCIS is usually composed of neoplastic cells with low to intermediate nuclear grade. The neoplastic spindle cells can mimic the streaming arrangement typically seen in UDH,14 but they completely fill the duct lumen and show no maturation across its diameter. The nuclei are also fairly uniform and have fine chromatin. The neoplastic cells often express neuroendocrine markers including chromogranin, synaptophysin, and neuron-specific enolase.14 Rarely, spindle cell DCIS shows high nuclear grade and can have central necrosis. Spindle cell DCIS usually constitutes only a focal component in the context of a DCIS lesion that also shows other more common architectural patterns.
Flat (clinging) DCIS consists of one to few layers of cells with high nuclear grade. Ducts and acini involved by flat DCIS often appear overly dilated.
Necrosis and Calcifications
Necrosis is a frequent finding in intermediate-grade or high-grade DCIS and often harbors coarse and pleomorphic calcifications. It is extremely rare in low-grade DCIS, although scattered apoptotic cells can be present. Occasionally, necrosis can mimic inspissated and calcified secretions. In contrast to the latter, necrosis contains nuclear debris, which appears as small and irregular fragments of hematoxylin-stained material. At present, most cases of DCIS are diagnosed after identification of calcifications on a screening mammogram.7 The mammographic calcifications associated with necrotic DCIS tend to be coarse and pleomorphic and distributed with a linear or segmental pattern. In contrast, the inspissated secretions associated with low-grade DCIS, atypical ductal hyperplasia (ADH), and flat epithelial atypia and the mammographic calcifications associated with low-grade DCIS and related lesions are small, fine, and granular and often clustered together. They are usually very similar to the small calcifications associated with benign epithelium.
The size of DCIS is an important parameter to estimate the probability of residual disease in the breast after surgery. It also correlates with final margin status15 and influences the risk of local recurrence. In the rare cases when DCIS is present in only one slide, the largest microscopic span between the two furthest foci of DCIS is reported. The optimal sampling of an excision specimen containing DCIS involves serial sectioning with sequential and complete histologic evaluation of the entire specimen. In these cases, the size of DCIS is determined by multiplying the number of consecutive blocks with DCIS by the thickness of the tissue present in each block. The latter measurement is estimated dividing the longest dimension of the specimen by the corresponding number of tissue slices. Most tissue slices/blocks have an approximate thickness of 0.4 cm. If DCIS involves opposing margins of an oriented lumpectomy specimen, the distance between the two margins can be used as the size of the DCIS in the specimen. When a surgical excision specimen is large and it is only partially submitted, orientation should be kept for the specimen slices that are not evaluated histologically. These tissue slices can be submitted at a later time to complete evaluation of the specimen and rule out the possibility of invasive carcinoma. If this is not done, the span of the tissue slices intervening between the blocks with DCIS is included in the final estimation of the size of DCIS by adding them to the number of involved blocks, and the sum is then multiplied by the thickness of each tissue section (about 0.4 cm).8 In all cases the largest size estimate of DCIS should be reported.
In women treated with breast-conserving surgery (BCS) and radiotherapy (RT), margin status is an important parameter of the recurrence rate. Prospective studies of patients with DCIS treated with BCS+RT have documented a significantly reduced the rate of local recurrence in patients with negative margins.16 Even though there is complete agreement on the definition of a positive margin (tumor present at ink), the definition of a “negative” margin remains elusive, and different thresholds have been used in different studies, including at different centers involved in the same study.16 At present, a “negative” margin ranges from “no tumor at ink” to a clearance of one, two, or even three mm. Some investigators have reported a very low rate of local recurrence for DCIS excised with a 1-cm-wide margin.17 However, a prospective study of surgical excision without adjuvant treatment (no tamoxifen and no RT) for DCIS of low to intermediate nuclear grade spanning ≤2.5 cm documented a local recurrence rate of 2.4% per patient-year, amounting to a recurrence rate of 12.5% at five years.18 In addition to margin width, the volume of disease near or at a margin also influences ipsilateral breast tumor recurrence (IBTR). A study by Rudloff et al19 found that patients with higher volume of DCIS near the margin (defined as the number of ducts <1 mm from an inked margin) derive greater benefit from the addition of RT than patients with only one close focus. In particular, patients with ≥2 foci of DCIS at <1 mm from the margin were at a high risk of local recurrence (hazard ratio 3.37, P=0.002) if no adjuvant RT was administered, but the risk associated with a greater volume of disease was substantially reduced if they received adjuvant RT (hazard ratio 0.14, P=0.004).
The histologic differential diagnosis of DCIS includes non-neoplastic epithelial proliferations, such as UDH, as well as atypical ductal lesions, including columnar cell change with atypia (also known as flat epithelial atypia), and ADH. ADH consists of cells cytomorphologically indistinguishable from those of low-grade DCIS, but it is smaller in extent (<2 mm size or <2 ducts) and shows only partial architectural complexity that does not involve the full cross-section of the ducts. ADH is also an established morphologic precursor of invasive breast carcinoma, and carries a four- to five-fold risk of subsequent carcinoma. Distinction between DCIS and the abovementioned benign proliferative lesions of the breast is usually achieved on the basis of review of hematoxylin and eosin (H&E)-stained sections. Immunohistochemical (IHC) stains may aid in the characterization of ductal and lobular lesions that cannot readily be distinguished. Frequent diagnostic dilemmas include solid low-grade DCIS versus classical LCIS, and intermediate-grade or high-grade solid DCIS versus pleomorphic LCIS. E-cadherin and p120 are the most frequently used stains. E-cadherin is a transmembrane glycoprotein involved in cell-to-cell adhesion. It is encoded by the CDH1 gene, which is located on 16q22.1. Normal and neoplastic ductal epithelia show strong continuous linear membranous immunoreactivity for E-cadherin, whereas atypical lobular hyperplasia (ALH) and LCIS lack membranous reactivity for this antigen because of mutation or epigenetic inactivation of the CDH1 gene. Occasionally, pagetoid growth of ALH may incompletely fill the lumen of the acini of a lobule leaving some residual normal epithelium in the center, mimicking ADH or UDH. Another scenario that may be misdiagnosed as UDH or ADH can occur when lobular neoplasia infiltrates UDH (Fig. 4C). When E-cadherin is applied to evaluate these cases, it is important to recognize the presence of two cell populations, namely E-cadherin-positive normal ductal cells and E-cadherin-negative neoplastic lobular cells (Fig. 4D). These cases should not be overinterpreted as ADH or carcinoma in situ with mixed ductal and lobular features. The absence of membranous E-cadherin results in mobilization of p120, the E-cadherin intracellular ligand, from a submembranous location to a diffuse and uniform distribution throughout the cytoplasm. This change is reflected in a shift of p120 immunoreactivity from linear and membranous in ductal epithelium to diffuse and cytoplasmic in lobular neoplasia.9,20 LCIS variants, especially pleomorphic LCIS, can have overlapping histologic features with solid DCIS, and IHC staining for E-cadherin and/or p120 can be very helpful in this differential diagnosis. The use of other markers, such as K903 (negative in DCIS, positive in LCIS) and CAM 5.2 (peripheral staining in DCIS and perinuclear staining in LCIS),21 is not uniformly acknowledged.
Normal mammary ducts and lobules are composed of two cell types. The luminal epithelial cells line the glandular lumen, whereas the myoepithelial cells serve as contractile cells interposed between the luminal cells and the basement membrane. Some studies have suggested the existence of a putative population of “basal” cells, which express basal/high–molecular weight cytokeratins (CKs) (i.e. CK5, CK6, CK14, CK17). These cells are thought to be very few and scattered within the myoepithelial layer or, less commonly, admixed with the luminal layer. The putative stem cell–like properties and the role of basal cells in the pathogenesis of breast carcinoma are a subject of debate.
Normal luminal epithelial cells express low–molecular weight CKs including CK7, CK8, CK18, and CK19 and do not express high–molecular weight CKs. Scattered luminal cells also express ER and PR. The myoepithelial cells express the high–molecular weight CKs mentioned above as well as p63, smooth muscle actin, smooth muscle myosin-heavy chain, and calponin. Low-grade DCIS usually express only luminal CKs, but some high-grade DCIS will rarely express basal cell markers.22,23
IHC stains for CK5/6 and ER can be used together to differentiate ADH and low-grade DCIS from UDH. The neoplastic cells in ADH and low-grade DCIS show very focal positivity for CK5/6 (Fig. 3B), as only 10% express high–molecular weight CKs.21 In UDH, hyperplastic epithelial cells and myoepithelial cells are closely admixed, consistent with a polymorphic non-neoplastic proliferation. In this setting, positive reactivity for CK5/6 results in a heterogenous and mosaic (checkerboard staining) pattern (Fig. 3E). ADH and low-grade DCIS display diffuse and strong reactivity for ER (Fig. 3C) in contrast to a heterogenous staining pattern in UDH (Fig. 3F). It is noteworthy that columnar cell changes with or without atypia show diffuse and strong staining for ER and no staining for CK5/6. These two markers are not helpful to distinguish columnar cell change with or without atypia.
Otterbach et al24 assessed CK5/6 immunoreactivity in ductal epithelial proliferations and noted differential expression in UDH and noninvasive neoplastic proliferations (ADH, DCIS, ALH, and LCIS), whereas the hyperplastic cells of UDH expressed strong CK5/6 staining in 87.6% of cases, and no UDH case was CK5/6 negative. In contrast, 47.5% of the cases diagnosed as ADH on the basis of morphology completely lacked any epithelial reactivity for CK5/6. In the remaining 43.1% of ADH cases only a few CK5/6-positive cells were present in the center of the ducts, which were thought to represent either residual normal or hyperplastic glandular cells on the basis of morphology. Nearly all of the cases of DCIS and LCIS showed no CK5/6 staining of the neoplastic cells, irrespective of the nuclear grade.24 A subsequent study by Boecker et al25 also confirmed differential expression of CK5/6 in UDH versus neoplastic ductal proliferations using comparative genomic hybridization (CGH).
A subset (6% to 8%) of DCIS is ER, PR, and HER2 negative and shows positivity for CK5/6 and/or epidermal growth factor receptor (EGFR), an immunoprofile similar to that of invasive basal-like carcinoma.23,26 Basal-like DCIS usually has high nuclear grade and solid growth, and morphologic evaluation is usually sufficient to distinguish it from UDH22,27 (also see the Molecular advances section).
Lack of interobserver agreement for ADH bordering on low-grade DCIS remains a problematic area in breast pathology.28,29 Use of specific morphologic criteria has been shown to improve interobserver reproducibility.30 A few studies have also advocated the use of immunoperoxidase stains for CK5/6, ER, and myoepithelial markers to aid in the diagnosis of intraductal epithelial proliferations. For example, a study29 evaluated the utility of a commercially available cocktail of antibodies targeting basal (CK5 and CK14) and luminal (CK7 and CK18) CKs, as well as p63, a myoepithelial nuclear antigen, to improve interobserver reproducibility in the diagnosis of 81 ductal proliferations encompassing UDH, ADH, and DCIS. Interobserver agreement among nine pathologists was only fair (k=0.34) upon review of H&E slides, whereas a slight improvement in interobserver concordance (k=0.50) was observed upon review of immunostains obtained using the antibody cocktail.29 Another study assessed interobserver agreement among twenty pathologists in the interpretation of 105 cases of noninvasive proliferative breast lesions using H&E-stained, CK5/6-stained, and E-cadherin-stained slides. In this study, the baseline concordance rate was higher and the use of IHC stains increased it only slightly (from k=0.47 to k=0.53, not significant).31 These results demonstrate that the use of immunoperoxidase stains for selected markers can assist pathologists in the diagnosis of challenging ductal epithelial proliferations, but it cannot replace morphologic evaluation as the primary and critical assessment.
Generally, ER expression is inversely related to nuclear grade. As demonstrated in a comprehensive review that examined ER expression rate among 36 studies,32 approximately 70% of DCIS express ER (range, 49% to 96.6%). ER positivity shows a significant positive correlation with PR expression and is inversely correlated with HER2.32 Kerlikowske et al6 have reported that the ER-negative status in DCIS was associated with recurrence of DCIS, but found no significant correlation between ER expression in DCIS and the risk of an invasive recurrence. A study by Allred et al33 evaluated retrospectively the relationship between ER and PR status and response to tamoxifen in 732 patients with DCIS enrolled in the National Surgical Adjuvant Breast and Bowel Project B-24 (NSABP-B24) study with an overall median follow-up of 14.5 years. In this study, 76% of DCIS cases were ER positive. Patients with ER-positive DCIS treated with tamoxifen showed significant decreases in subsequent breast cancer at ten years (hazard ratio, 0.49; P<0.001), and these findings were also significant in multivariate analysis (overall hazard ratio, 0.64; P=0.003). This study also reported no significant benefit of tamoxifen treatment in patients with ER-negative DCIS, but this result is based on analysis of only 22 patients. On the basis of this study, it is standard practice to assess ER status of DCIS using immunoperoxidase stains to support the use of adjuvant hormonal therapy.
Approximately 60% of DCIS (range, 40% to 83.3%) express PR as demonstrated in a comprehensive review that examined PR expression rate in 28 studies.32 As with ER, an inverse relationship is observed between PR expression and nuclear grade: high-grade DCIS is less likely to be PR-positive than non–high-grade DCIS.
On the basis of IHC and/or fluorescence in situ hybridization (FISH) assessment, approximately 40% of DCIS express HER2.32 HER2 expression in DCIS is substantially higher than the 15% to 20% rate of HER2 positivity observed in invasive carcinoma. HER2 overexpression is more common in high-grade DCIS,32 and HER2-positive DCIS is more likely to recur (recurrence 42% HER2 positive vs. 12% HER2 negative).
Numerous studies have demonstrated that morphologically different subtypes of DCIS are characterized by distinct genetic alterations, which closely mimic those found in paired invasive carcinomas.25,34,35 Buerger et al34 first reported the existence of genetically different subtypes of DCIS and that loss of 16q occurs almost exclusively in low-grade and intermediate-grade DCIS. In the same study, intermediate-grade DCIS showed almost twice as many genetic imbalances compared with low-grade DCIS (5.5 vs. 2.5, respectively) as well as more frequent 1q gains and 11q losses.34 High-grade DCIS was much more heterogenous and showed higher frequency of amplifications at 17q12 and 11q13 and a much higher average rate of genetic imbalances (average of 7.1).
CGH analysis of paired DCIS and invasive breast carcinoma has revealed patterns of genetic alterations that are nearly identical,35 supporting the role of DCIS as a precursor of invasive carcinoma. In line with these observations, more recent studies have confirmed distinct genetic alterations in low-grade versus high-grade lesions. In particular, a “low-grade breast epithelial neoplasia family” has been recognized on the basis of close morphologic, immunophenotypic, and genetic similarities. Lesions that are part of this family include classic lobular neoplasia (ALH and classical LCIS), atypical columnar cell lesions, ADH, and low-grade DCIS, as well as tubular and tubulolobular carcinoma, grade I invasive ductal carcinoma, and invasive lobular carcinoma, classical type. These lesions are characterized by recurrent genetic alterations, specifically loss of 16q and variable gain of 1q. Lesions in this family are strongly positive for ER, Bcl-2, and cyclin D1 and negative for HER2.36,37
On the other end of the spectrum, high-grade DCIS and poorly differentiated invasive carcinoma constitute a morphologically, immunophenotypically, and genetically very heterogenous group of lesions. Deletions of 16q are rare in these tumors and alterations at 8p, 11q, 13q, 17q, and 20q are frequent; however, the patterns of genetic alterations can be quite variable. ER expression in high-grade DCIS also varies; HER2 can be positive or negative, and some high-grade DCIS cases are negative for ER, PR, and HER2 (triple negative). At present, the existence of an intermediate-grade DCIS is questionable. It is possible that some occasional low-grade DCIS may acquire additional genetic alterations and “evolve” into higher-grade DCIS, but low-grade and high-grade DCIS lesions appear for the most part separate and distinct, consistent with separate pathways of low-grade and high-grade ductal epithelial neoplasia.
Over a decade ago, two seminal studies38,39 applied gene array genomic profiling to the evolution of invasive breast carcinomas and found that they can be classified into distinct molecular groups (luminal A, luminal B, HER2, basal, and normal-like carcinomas). This investigative approach has been adopted in many subsequent studies including the recent project part of The Cancer Genome Atlas Network.40 Invasive breast carcinomas of luminal subtype express genes encoding luminal CKs and ER-related genes are subclassified into luminal type A and B according to the expression of proliferation-related genes.41 As the name implies, HER2-overexpressing carcinomas show increased levels of HER2-related genes. Most basal-like cancers lack expression of ER-related, PR-related, and HER2-related genes, but express genes found in basal/myoepithelial cells, such as CK5 and CK6, and/or the EGFR gene. The so called “normal breast-like” carcinomas, which had been described in earlier studies,38,39 have not been confirmed by other series, including in the The Cancer Genome Atlas Network project.40 The molecular subgroups of breast carcinoma correlate with different clinical outcome.39,41–43
Gene expression profiling applied to the study of DCIS consistently shows that distinct subsets of genes characterize low-grade and high-grade DCIS. In one study, Hannemann et al44 40 in situ carcinomas and 40 invasive breast carcinomas. Using hierarchical clustering, the authors identified different types of DCIS including the basal-like, ERBB2 type, and luminal type originally described for invasive breast cancer. In this study, a 43-gene classifier could separate low-grade from high-grade DCIS, and a 35-gene classifier was proposed to distinguish between DCIS and invasive carcinomas.
Using a similar approach, Muggerud et al45 separated 31 cases of pure DCIS into two groups, consisting of 24 predominantly ER-positive DCIS and seven pure ER-negative DCIS cases. The ER-positive DCIS cases were of low grade, the ER-negative DCIS cases were high grade, HER2 positive, and had a high Ki67 proliferation rate. Similarly, Allred et al46 classified 25 cases of pure DCIS into luminal (44%), basal (8%), and ERBB2 intrinsic (28%) molecular subtypes. The histologic grade of DCIS was much lower in luminal DCIS than in ERBB2-positive and basal DCIS.
Vincent-Salomon et al47 also used integrated genomic and transcriptomic analysis to classify 57 cases of DCIS into luminal (32), ERBB2-positive (21), and ERBB2-negative/ER-negative (4) DCIS. Eighty-one percent of ERBB2-positive and 75% of the ERBB2-negative/ER-negative cases had high nuclear grade, respectively. Seventy-two percent of the DCIS with luminal A intrinsic subtype did not show high nuclear grade. ERBB2-amplified DCIS never showed 16q loss. In this study, unsupervised analysis of 26 DCIS (9 ERBB2-amplified and 17 luminal) showed that luminal and ERBB2-amplified/ER-negative cases clustered separately according to ERBB2 and hormone receptor status rather than according to nuclear grade.
Gene profiling techniques remain relatively expensive and are labor intensive and the data require substantial statistical evaluation. Investigators have identified surrogate molecular immunoprofiles based on ER, HER2, CK5/6, and EGFR immunoreactivity that can be used to estimate the molecular subtype of breast carcinomas. In particular, two studies26,27 have identified the immunoprofile of basal-like carcinomas as ER and HER2 negative and CK5/6 and/or EGFR positive. It is important to remember that the molecular immunoprofile of breast carcinomas provides a very good but not perfect assessment of the molecular intrinsic subtype by to gene array analysis. For example, 3/21 (14%) of carcinomas classified as basal by gene array profiling showed positive immunoreactivity for ER.26
Despite these limitations, data obtained using IHC provide insights into the distribution and rate of the different molecular subtypes of invasive breast carcinoma, and this approach has also been applied to the study of DCIS. Tamimi et al48 used a panel of five IHC markers (ER, PR, HER2, CK5/6, and EGFR) to classify 2249 cases of invasive breast cancer and 272 cases of DCIS into molecular subgroups. They found that the rate of luminal B and HER2 molecular phenotypes was significantly higher in of DCIS (13.2% and 13.6%, respectively) than in invasive carcinomas (5.2% and 5.7%, respectively) (P<0.0001). In contrast, the luminal A phenotype was significantly more frequent among invasive breast carcinomas (73.4%) than DCIS (62.5%) (P=0.0002). In the same study, high-grade DCIS and invasive carcinomas were more likely to be HER2 positive or basal-like than to have low or intermediate grade. These results highlight some disparities between the rates of molecular subtypes of DCIS and invasive carcinomas, which in turn raise biological questions.
Parallel with the identification of basal invasive breast carcinomas, few recent studies have described a basal subtype of DCIS (ER and HER2 negative and positive for basal markers).22,27 On the basis of histologic colocalization and similar immunoprofiles, basal DCIS likely represents the morphologic precursor of basal-like invasive carcinoma. In the Carolina Breast Cancer Study,27 7% of 275 cases of pure DCIS were classified as basal-like. Bryan et al found that 6% of 66 high-grade DCIS had triple negative phenotype. All triple negative cases also expressed basal cytokeratins compared with 42% of non-triple negative cases (P=0.04). Zhou et al49 also examined ER, PR, HER2, CK5/6, and EGFR profiles in DCIS from 392 women and found that 8.2% were basal-like. When correlated with long-term follow-up data (median follow-up to an invasive or general recurrence was 111.5 mo; range, 3 to 255 mo), basal DCIS showed twice the risk of local recurrence compared with non-basal DCIS and nearly twice the risk of developing any type of recurrence compared with non-basal DCIS. No differences in recurrence were observed when analysis focused on the long-term follow-up of triple-negative DCIS versus non–triple-negative cases. Nonetheless, these results for basal DCIS were not statistically significant, possibly because of the limited number of cases in this study.
Close similarities have also been observed between molecular phenotypes of DCIS and invasive breast carcinomas in BRCA1 and BRCA2 mutation carriers. These observations provide additional evidence that DCIS is a direct precursor lesion of invasive breast carcinoma in genetically predisposed individuals. Most invasive breast cancers occurring in BRCA1 and BRCA2 germline mutation carriers show basal and luminal immunophenotypes, respectively. Van der Groep et al50 studied surrogate molecular immunoprofiles of DCIS and associated invasive breast carcinomas from BRCA1 and BRCA2 germline mutation carriers. DCIS in BRCA1 mutation carriers showed high nuclear grade, high proliferation rate, and were triple negative with positive staining for CK5/6, CK14, and EGFR, consistent with a basal-like molecular immunoprofile. In contrast, DCIS lesions in BRCA2 mutation carriers were mostly of high nuclear grade but with a low proliferation rate and had a luminal molecular immunoprofile with positive reactivity for ER in 8/9 cases and PR in 4/9, and infrequent expression of the basal markers CK5/6, CK14, and EGFR. DCIS in BRCA2 germline mutation carriers was HER2 positive in 3/9 cases.
The development of breast cancer is a complex multistep process originating in terminal ductal-lobular units with progression to either low-grade or high-grade lesions through distinct molecular pathways.25,51,52 When DCIS and invasive carcinoma from the same specimen are compared with one another they show no substantial differences. Information obtained using CGH and gene array expression profiling is also consistent with these observations.
Molecular analysis44–46,53 has not only been used to differentiate between the heterogenous subgroups of DCIS, but also to better define the relationship between DCIS and invasive breast carcinoma to gain insights into the mechanisms that regulate disease progression. The similarities in molecular categories between invasive and in situ carcinoma suggest that genomic diversity is acquired in the early stages of cancer development and may correlate with progression of disease.54–59
In a study by Hernandez et al,60 genomic and mutational profiling revealed intratumoral heterogeneity in DCIS and matched adjacent invasive carcinoma. Thirteen matched pairs of DCIS and invasive breast cancer were analyzed by microarray-based CGH (findings validated by FISH) and Sequenom MassARRAY (findings validated by Sanger sequencing). The matched pairs showed similar genomic profiles; however, the copy number aberrations and somatic mutations varied, suggesting clonal selection during progression. PIK3CA mutations were more prevalent in the DCIS component.
Heselmeyer-Haddad et al61 microdissected single cells from 13 cases of synchronous DCIS and invasive breast carcinoma and analyzed them using ten FISH probes for oncogenes and tumor-suppressor genes. Intratumoral heterogeneity was observed as there was a high degree of chromosomal instability among single cells (identical signal clones were present in <20% of the cells) with six of 13 cases showed a shift to a major clone from DCIS to invasive carcinoma. The findings from these two suggest that clonal selection may play an important role in the evolution of DCIS to invasive disease, even though DCIS uncommonly shows morphologic and architectural heterogeneity (Fig. 5). Lee et al62 carried out comprehensive expression profiling of DCIS and invasive breast cancer and found 470 differentially expressed genes. Comparison with genes identified in similar studies resulted in the creation of a 74-gene profile. The investigators used a xenograft model to evaluate a subset of genes that could be important in the progression from DCIS to invasive carcinoma. They found that the suppression of four genes, including a protease inhibitor and genes involved in cell adhesion and signaling, was associated with increased progression of the xenografts to invasive carcinoma. Other studies59,63 have also provided evidence that the microenvironment plays an important role in the progression of DCIS.
DCIS associated with grade I invasive cancers show different levels of gene expression compared with DCIS associated with grade III invasive cancer.53 Balleine et al53 studied DCIS microdissected from 46 cases with associated invasive carcinoma. Using array-based genomic hybridization they found that DCIS could be classified into two subgroups, namely low–molecular grade and high–molecular grade DCIS. The authors also found that DCIS nuclear grade and proliferation index measured by Ki67 IHC were accurate predictors of low or high molecular grades of DCIS.
Another possible approach to the subclassification of DCIS rests on the differential expression of ER and ER-related genes, as summarized by Lopez-Garcia et al.52 ER-positive and ER-negative breast cancers are fundamentally distinct diseases. According to this model, ER-positive encompasses most known precursor lesions and a range of invasive carcinomas of low to high grade. The ER-negative arm includes ER-negative DCIS, as well microglandular adenosis, a rare morphologic precursor of some triple-negative breast carcinomas,64,65 and HER2-overexpressing and triple-negative invasive breast carcinomas.
DCIS per se is not responsible for any deaths; however, invasive recurrences can be locally aggressive or result in metastatic spread. Morphologic, immunophenotypic, and molecular parameters have all been used to try and predict the risk of local recurrence of DCIS and, in particular, the risk of an invasive recurrence. The ultimate goal of large-scale genomic analysis of DCIS is to translate the information into clinically applicable predictive tools.
The risk for IBTR after BCS alone for DCIS is relatively high. RT and endocrine therapy are two effective, yet not harmless, treatments that have been shown to decrease the risk of IBTR after surgery by approximately 50% and 30%, respectively. Neither treatment option has been proven to improve survival.
Nomograms are graphical depictions of prediction models that provide overall probability of a specific outcome for an individual patient. Investigators at Memorial Sloan-Kettering Cancer Center (MSKCC) developed a nomogram that uses 10 clinical, pathologic, and treatment variables, age at diagnosis, family history, presentation, nuclear grade, necrosis, surgical margins, number of surgical excisions, year of the surgery, adjuvant RT, and adjuvant endocrine therapy, to predict the risk for local recurrence at 5 and 10 years after BCS for DCIS.66 The nomogram was based on a data set from 1868 consecutive DCIS patients treated with BCS, and was internally validated using 200 bootstrap samples. The DCIS nomogram for prediction of 5- and 10-year IBTR probabilities demonstrated good calibration and discrimination, with a concordance index of 0.704 (bootstrap corrected, 0.688) and a concordance probability estimate of 0.686. Collins et al67 applied the nomogram to a large population of patients with DCIS treated with BCS with known outcome. The risk estimates provided by the MSKCC nomogram showed approximately 90% correlation with the observed rates of local recurrence at five years.
Kerlikowske et al6,68 correlated disease recurrence with the expression of ER, PR, Ki67, p53, p16, EGFR-2, and cyclooxygenase-2 in DCIS in a nested case-control study in a population-based cohort of 1162 women with DCIS treated by lumpectomy alone from 1983 to 1994. Individual markers were not statistically significantly associated with the risk of subsequent invasive cancer; however, certain combinations of these biomarkers were statistically significantly associated with subsequent invasive cancer. The results of these study showed that women whose DCIS expressed high levels of p16, cyclooxygenase-2, and Ki67 were more likely to develop an invasive recurrence.
A quantitative multigene reverse transcription polymerase chain reaction (RT-PCR) assay has recently been developed to predict recurrence risk after surgical excision without subsequent RT for DCIS.69 This assay evaluates 12 genes that represent a subset of the 21 genes used in the commercially available RT-PCR recurrence score for invasive carcinoma. These genes include proliferation markers (Ki67, STK15, Survivin, Cyclin B1, MYBL2), hormone-related genes (PR, GSTM1), and several reference genes (β-actin, GAPDH, RPLPO, GUS, TFRC). The validity of the DCIS recurrence score was retrospectively tested in a subset of patients enrolled in ECOG E5194.70 The results of this analysis were assigned to three subgroups showing low, intermediate, and high recurrence scores. Patients with low risk DCIS still had a likelihood of any local recurrence of 10.6% at 10 years and a risk of an invasive recurrence of 3.7% at 10 years. The 10-year risk of local recurrence in the low DCIS score group is still regarded by most as too high to forgo RT.71
The use of novel molecular techniques combined with traditional morphologic and IHC evaluation has greatly broadened our understanding of DCIS. The heterogeneity of DCIS closely mirrors that of the specific subtypes of invasive carcinoma. Despite recent advances, many questions still remain unanswered and deserve future investigation. In current clinical practice, the diagnosis of DCIS rests mainly on morphologic evaluation, and only occasionally benefits from the use of IHC markers. Continued investigation of the mechanisms involved in the pathogenesis of DCIS and its progression to invasive carcinoma will improve our ability to subclassify this disease and identify patients that can be treated less aggressively, develop targeted therapies and/or preventive strategies.
The authors of the paper would like to acknowledge Meighan M. Gallagher for her editorial assistance.
1. DeSantis C, Siegel R, Bandi P, et al .Breast cancer statistics, 2011.CA Cancer J Clin. 2011; 61:409–418.
2. Jemal A, Siegel R, Ward E, et al .Cancer statistics, 2009.CA Cancer J Clin. 2009; 59:225–249.
3. Li CI, Daling JR, Malone KE .Age-specific incidence rates of in situ breast carcinomas by histologic type, 1980 to 2001.Cancer Epidemiol Biomarkers Prev. 2005; 14:1008–1011.
4. Virnig BA, Tuttle TM, Shamliyan T, et al .Ductal carcinoma in situ of the breast: a systematic review of incidence, treatment, and outcomes.J Natl Cancer Inst. 2010; 102:170–178.
5. Brekelmans CT, Seynaeve C, Bartels CC, et al .Effectiveness of breast cancer surveillance in BRCA1/2 gene mutation carriers and women with high familial risk.J Clin Oncol. 2001; 19:924–930.
6. Kerlikowske K, Molinaro AM, Gauthier ML, et al .Biomarker expression and risk of subsequent tumors after initial ductal carcinoma in situ diagnosis.J Natl Cancer Inst. 2010; 102:627–637.
7. Lakhani SR, Ellis IO, Schnitt SJ, et al .WHOClassification of Tumours of the Breast. 2012; .Lyon, France:IARC Press.
8. Lester SC, Bose S, Chen YY, et al .Protocol for the examination of specimens from patients with invasive carcinoma of the breast.Arch Pathol Lab Med. 2009; 133:1515–1538.
9. Brogi E, Murray MP, Corben AD .Lobular carcinoma, not only a classic.Breast J. 2010; 16:suppl 1 S10–S14.
10. Rosen PP .Pine JW Jr .Lobular carcinoma in situ and atypical lobular hyperplasia.Rosen’s Breast Pathology. 2009; Philadelphia, PA:Lippincott Williams & Wilkins, a Wolters Kluwer business; 637–689.
11. Rakha EA, Gandhi N, Climent F, et al .Encapsulated papillary carcinoma of the breast: an invasive tumor with excellent prognosis.Am J Surg Pathol. 2011; 35:1093–1103.
12. Wynveen CA, Nehhozina T, Akram M, et al .Intracystic papillary carcinoma of the breast: An in situ or invasive tumor? Results of immunohistochemical analysis and clinical follow-up.Am J Surg Pathol. 2011; 35:1–14.
13. Tsang WY, Chan JK .Endocrine ductal carcinoma in situ (E-DCIS) of the breast: a form of low-grade DCIS with distinctive clinicopathologic and biologic characteristics.Am J Surg Pathol. 1996; 20:921–943.
14. Farshid G, Moinfar F, Meredith DJ, et al .Spindle cell ductal carcinoma in situ. An unusual variant of ductal intra-epithelial neoplasia that simulates ductal hyperplasia or a myoepithelial proliferation.Virchows Arch. 2001; 439:70–77.
15. Cheng L, Al-Kaisi NK, Gordon NH, et al .Relationship between the size and margin status of ductal carcinoma in situ of the breast and residual disease.J Natl Cancer Inst. 1997; 89:1356–1360.
16. Solin LJ, Fourquet A, Vicini FA, et al .Long-term outcome after breast-conservation treatment with radiation for mammographically detected ductal carcinoma in situ of the breast.Cancer. 2005; 103:1137–1146.
17. Silverstein MJ, Lagios MD, Groshen S, et al .The influence of margin width on local control of ductal carcinoma in situ of the breast.N Engl J Med. 1999; 340:1455–1461.
18. Wong JS, Kaelin CM, Troyan SL, et al .Prospective study of wide excision alone for ductal carcinoma in situ of the breast.J Clin Oncol. 2006; 24:1031–1036.
19. Rudloff U, Brogi E, Reiner AS, et al .The influence of margin width and volume of disease near margin on benefit of radiation therapy for women with DCIS treated with breast-conserving therapy.Ann Surg. 2010; 251:583–591.
20. Dabbs DJ, Bhargava R, Chivukula M .Lobular versus ductal breast neoplasms: the diagnostic utility of p120 catenin.Am J Surg Pathol. 2007; 31:427–437.
21. Yeh IT, Mies C .Application of immunohistochemistry to breast lesions.Arch Pathol Lab Med. 2008; 132:349–358.
22. Bryan BB, Schnitt SJ, Collins LC .Ductal carcinoma in situ with basal-like phenotype: a possible precursor to invasive basal-like breast cancer.Mod Pathol. 2006; 19:617–621.
23. Livasy CA, Karaca G, Nanda R, et al .Phenotypic evaluation of the basal-like subtype of invasive breast carcinoma.Mod Pathol. 2006; 19:264–271.
24. Otterbach F, Bankfalvi A, Bergner S, et al .Cytokeratin 5/6 immunohistochemistry assists the differential diagnosis of atypical proliferations of the breast.Histopathology. 2000; 37:232–240.
25. Boecker W, Buerger H, Schmitz K, et al .Ductal epithelial proliferations of the breast: a biological continuum? Comparative genomic hybridization and high-molecular-weight cytokeratin expression patterns.J Pathol. 2001; 195:415–421.
26. Nielsen TO, Hsu FD, Jensen K, et al .Immunohistochemical and clinical characterization of the basal-like subtype of invasive breast carcinoma.Clin Cancer Res. 2004; 10:5367–5374.
27. Livasy CA, Perou CM, Karaca G, et al .Identification of a basal-like subtype of breast ductal carcinoma in situ.Hum Pathol. 2007; 38:197–204.
28. Rosai J .Borderline epithelial lesions of the breast.Am J Surg Pathol. 1991; 15:209–221.
29. Jain RK, Mehta R, Dimitrov R, et al .Atypical ductal hyperplasia: interobserver and intraobserver variability.Mod Pathol. 2011; 24:917–923.
30. Schnitt SJ, Connolly JL, Tavassoli FA, et al .Interobserver reproducibility in the diagnosis of ductal proliferative breast lesions using standardized criteria.Am J Surg Pathol. 1992; 16:1133–1143.
31. MacGrogan G, Arnould L, de Mascarel I, et al .Impact of immunohistochemical markers, CK5/6 and E-cadherin on diagnostic agreement in non-invasive proliferative breast lesions.Histopathology. 2008; 52:689–697.
32. Lari SA, Kuerer HM .Biological markers in DCIS and risk of breast recurrence: a systematic review.J Cancer. 2011; 2:232–261.
33. Allred DC, Anderson SJ, Paik S, et al .Adjuvant tamoxifen reduces subsequent breast cancer in women with estrogen receptor-positive ductal carcinoma in situ: a study based on NSABP Protocol B-24.J Clin Oncol. 2012; 30:1268–1273.
34. Buerger H, Otterbach F, Simon R, et al .Comparative genomic hybridization of ductal carcinoma in situ of the breast-evidence of multiple genetic pathways.J Pathol. 1999; 187:396–402.
35. Buerger H, Otterbach F, Simon R, et al .Different genetic pathways in the evolution of invasive breast cancer are associated with distinct morphological subtypes.J Pathol. 1999; 189:521–526.
36. Abdel-Fatah TM, Powe DG, Hodi Z, et al .High frequency of coexistence of columnar cell lesions, lobular neoplasia, and low grade ductal carcinoma in situ with invasive tubular carcinoma and invasive lobular carcinoma.Am J Surg Pathol. 2007; 31:417–426.
37. Abdel-Fatah TM, Powe DG, Hodi Z, et al .Morphologic and molecular evolutionary pathways of low nuclear grade invasive breast cancers and their putative precursor lesions: further evidence to support the concept of low nuclear grade breast neoplasia family.Am J Surg Pathol. 2008; 32:513–524.
38. Perou CM, Sorlie T, Eisen MB, et al .Molecular portraits of human breast tumours.Nature. 2000; 406:747–752.
39. Sorlie T, Perou CM, Tibshirani R, et al .Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications.Proc Natl Acad Sci USA. 2001; 98:10869–10874.
40. Atlas N. Cancer Genome .Comprehensive molecular portraits of human breast tumours.Nature. 2012; 490:61–70.
41. Cheang MC, Voduc D, Bajdik C, et al .Basal-like breast cancer defined by five biomarkers has superior prognostic value than triple-negative phenotype.Clin Cancer Res. 2008; 14:1368–1376.
42. Brenton JD, Carey LA, Ahmed AA, et al .Molecular classification and molecular forecasting of breast cancer: ready for clinical application? J Clin Oncol. 2005; 23:7350–7360.
43. Cheang MC, Chia SK, Voduc D, et al .Ki67 index, HER2 status, and prognosis of patients with luminal B breast cancer.J Natl Cancer Inst. 2009; 101:736–750.
44. Hannemann J, Velds A, Halfwerk JB, et al .Classification of ductal carcinoma in situ by gene expression profiling.Breast Cancer Res. 2006; 8:R61
45. Muggerud AA, Hallett M, Johnsen H, et al .Molecular diversity in ductal carcinoma in situ (DCIS) and early invasive breast cancer.Mol Oncol. 2010; 4:357–368.
46. Allred DC, Wu Y, Mao S, et al .Ductal carcinoma in situ and the emergence of diversity during breast cancer evolution.Clin Cancer Res. 2008; 14:370–378.
47. Vincent-Salomon A, Lucchesi C, Gruel N, et al .Integrated genomic and transcriptomic analysis of ductal carcinoma in situ of the breast.Clin Cancer Res. 2008; 14:1956–1965.
48. Tamimi RM, Baer HJ, Marotti J, et al .Comparison of molecular phenotypes of ductal carcinoma in situ and invasive breast cancer.Breast Cancer Res. 2008; 10:R67
49. Zhou W, Jirstrom K, Johansson C, et al .Long-term survival of women with basal-like ductal carcinoma in situ of the breast: a population-based cohort study.BMC Cancer. 2010; 10:653
50. van der Groep P, van Diest PJ, Menko FH, et al .Molecular profile of ductal carcinoma in situ of the breast in BRCA1 and BRCA2 germline mutation carriers.J Clin Pathol. 2009; 62:926–930.
51. Simpson PT, Reis-Filho JS, Gale T, et al .Molecular evolution of breast cancer.J Pathol. 2005; 205:248–254.
52. Lopez-Garcia MA, Geyer FC, Lacroix-Triki M, et al .Breast cancer precursors revisited: molecular features and progression pathways.Histopathology. 2010; 57:171–192.
53. Balleine RL, Webster LR, Davis S, et al .Molecular grading of ductal carcinoma in situ of the breast.Clin Cancer Res. 2008; 14:8244–8252.
54. Lee RJ, Vallow LA, McLaughlin SA, et al .Ductal carcinoma in situ of the breast.Int J Surg Oncol. 2012; 2012:123549
55. Porter D, Lahti-Domenici J, Keshaviah A, et al .Molecular markers in ductal carcinoma in situ of the breast.Mol Cancer Res. 2003; 1:362–375.
56. Abba MC, Drake JA, Hawkins KA, et al .Transcriptomic changes in human breast cancer progression as determined by serial analysis of gene expression.Breast Cancer Res. 2004; 6:R499–R513.
57. Schuetz CS, Bonin M, Clare SE, et al .Progression-specific genes identified by expression profiling of matched ductal carcinomas in situ and invasive breast tumors, combining laser capture microdissection and oligonucleotide microarray analysis.Cancer Res. 2006; 66:5278–5286.
58. Ma XJ, Salunga R, Tuggle JT, et al .Gene expression profiles of human breast cancer progression.Proc Natl Acad Sci USA. 2003; 100:5974–5979.
59. Knudsen ES, Ertel A, Davicioni E, et al .Progression of ductal carcinoma in situ to invasive breast cancer is associated with gene expression programs of EMT and myoepithelia.Breast Cancer Res Treat. 2012; 133:1009–1024.
60. Hernandez L, Wilkerson PM, Lambros MB, et al .Genomic and mutational profiling of ductal carcinomas in situ and matched adjacent invasive breast cancers reveals intra-tumour genetic heterogeneity and clonal selection.J Pathol. 2012; 227:42–52.
61. Heselmeyer-Haddad K, Berroa Garcia LY, Bradley A, et al .Single-cell genetic analysis of ductal carcinoma in situ and invasive breast cancer reveals enormous tumor heterogeneity yet conserved genomic imbalances and gain of MYC during progression.Am J Pathol. 2012; 181:1807–1822.
62. Lee S, Stewart S, Nagtegaal I, et al .Differentially expressed genes regulating the progression of ductal carcinoma in situ to invasive breast cancer.Cancer Res. 2012; 72:4574–4586.
63. Hu M, Yao J, Cai L, et al .Distinct epigenetic changes in the stromal cells of breast cancers.Nat Genet. 2005; 37:899–905.
64. Shin SJ, Simpson PT, Da Silva L, et al .Molecular evidence for progression of microglandular adenosis (MGA) to invasive carcinoma.Am J Surg Pathol. 2009; 33:496–504.
65. Geyer FC, Lacroix-Triki M, Colombo PE, et al .Molecular evidence in support of the neoplastic and precursor nature of microglandular adenosis.Histopathology. 2012; 60:6B E115–E130.
66. Rudloff U, Jacks LM, Goldberg JI, et al .Nomogram for predicting the risk of local recurrence after breast-conserving surgery for ductal carcinoma in situ.J Clin Oncol. 2010; 28:3762–3769.
67. Collins LC, Achacoso N, Sharafali Z, et al .Predictors of local recurrence in patients with ductal carcinoma in situ treated by breast conserving therapy: value of the Memorial-Sloan-Kettering nomogram [abstract].Mod Pathol. 2012; 25:118
68. Kerlikowske K, Molinaro A, Cha I, et al .Characteristics associated with recurrence among women with ductal carcinoma in situ treated by lumpectomy.J Natl Cancer Inst. 2003; 95:1692–1702.
69. Solin LJ, Baehner FL, Butler S, et al .A quantitative multigene RT-PCR assay for predicting recurrence risk after surgical excision without irradiation for ductal carcinoma in situ: a prospective validation study of the DCIS score from ECOG E5194 [abstract]. San Antonio Breast Cancer Symposium: S4-6.Cancer Res. 2011; 71:suppl 3.
70. Hughes LL, Wang M, Page DL, et al .Local excision alone without irradiation for ductal carcinoma in situ of the breast: a trial of the Eastern Cooperative Oncology Group.J Clin Oncol. 2009; 27:5319–5324.
71. Solin LJ, Gray R, Baehner FL, et al .A multigene expression assay to predict local recurrence risk for ductal carcinoma of the breast.J Natl Cancer Inst. 2013; 105:701–710.