Ductal Carcinoma In Situ

Introduction

Over the past several decades, the widespread adoption of mammographic screening has had a significant impact on the incidence, diagnosis, classification, and treatment of ductal carcinoma in situ (DCIS). There has been a vast increase in the publications with regard to the definition, diagnostic criteria, and both short-term and long-term risks associated with specific histologic variants or types of in situ carcinoma of the breast. DCIS is extremely rare in patients <30 years. Its incidence increases with age. One in 33 patients is expected to be diagnosed with DCIS in their lifetime. In situ carcinomas of the breast were first recognized in the early 20th century and were identified morphologically as cells cytologically similar to those of invasive carcinomas but confined to ductal structures within the breast parenchyma. Such lesions were generally found to be located adjacent to areas of invasive carcinoma. The original definitions given to in situ carcinomas of the breast were arbitrary. Opportunities to study the natural history and behavior of such in situ lesions independent of an invasive component of disease or after a surgical procedure less than that of mastectomy were previously rarely encountered. Since that time, several studies relying on the review of archival slide material have demonstrated basic differences between distinct histologic patterns of in situ carcinomas. This subsequently resulted in the distinction between those lesions representing purely markers of increased risk (e.g., lobular carcinoma in situ [LCIS] and atypical ductal hyperplasia [ADH], with increased breast cancer risk that was essentially equally distributed to either breast) and committed premalignant lesions (e.g., DCIS, with increased breast cancer risk that was more often reported to be confined to the ipsilateral breast ).

The classical studies of Wellings and Jensen focused attention on the terminal ductal-lobular unit (TDLU) as a common anatomic site for the development of hyperplastic changes of both the ductal and the lobular type as well as corresponding neoplastic lesions. The terms DCIS and LCIS were once meant to signify separate anatomic origins, with the former originating within the ductal structures and the latter originating within the lobular structures. However, this anachronous concept is now recognized to be inaccurate. DCIS is essentially a diagnosis of exclusion, including in its broad sweep any lesion deemed in situ carcinoma that does not exhibit the cytologic features of lobular neoplasia cells. In that regard, those studies that have specifically addressed the incidence rate or relative risk of developing a subsequent ipsilateral and contralateral invasive breast cancer in women with in situ carcinoma have traditionally shown a higher incidence rate or relative risk of contralateral invasive breast cancer for those women with LCIS compared with those women with DCIS.

The following sections will review some of the unique aspects of LCIS as compared to DCIS. The chapter will also provide the most up-to-date information on classification of DCIS. Some of the less common types of DCIS and their presentations will be discussed. This chapter will also discuss the ancillary tests, immunomodulation, and role of the extent of cancer on the recurrence potential.

Recent Insights into the Unique Biology of Ductal Carcinoma in Situ, Lobular Carcinoma in Situ, and Related Entities

DCIS possesses all of the molecular and biological abnormalities as its frankly invasive carcinoma component. It has the same loss of heterozygosity patterns as its invasive counterpart. Using comparative genomic hybridization studies, DCIS exhibits no gains or losses of chromosomal regions compared with its invasive counterpart. Ductal hyperplasia of the usual type, based on loss of heterozygosity and genomic hybridization studies, is no longer thought to be a precursor lesion of either ADH or DCIS. DCIS is held in check by a surrounding layer of myoepithelial cells that exert paracrine suppressive effects on invasion. Because of this, the histopathologic patterns of DCIS and grade usually strongly correlate with its invasive counterpart ( Figs. 19.1–19.3 ). As the images show, the morphology of the invasive component mimics the in situ component. Comparative genomic hybridization has also proven dedifferentiation from low- to high-grade breast cancers as unlikely.

Tall columnar cell changes with atypia (flat epithelial atypia [FEA]) is often detected due to the presence of amorphous calcifications on mammograms. It is closely related to ADH and low-grade DCIS as defined by comparative genomic hybridization. These lesions are characterized by variably enlarged and dilated acini lined by monomorphic, primarily columnar, cells with apical snouts. Intraluminal secretion and calcification are present ( Figs. 19.4 and 19.5 ). These lesions are seen in association with atypical ductal/lobular hyperplasia. They show an immune profile similar to ADH/low-grade DCIS. The upgrade of tall columnar changes with atypia is <5% based on most published data. The upgrade is usually minute and incidental (low-grade invasive carcinoma, measuring <2 mm); therefore most publications suggest radiologic-pathologic correlation followed by radiologic follow-up.

LCIS is a noninvasive neoplastic proliferation of small, dyscohesive epithelial cells within TDLU with more than 50% expansion. LCIS is multicentric in 80% of patients and bilateral in 30% to 67%. The epithelial cells of LCIS are poorly cohesive and lack E-cadherin. LCIS (classic pattern) is both a risk factor and a nonobligate precursor of invasive breast carcinoma. The relative risk for classic LCIS is 8 to 10 times the risk in the general population. The increased risk can be associated with any type of infiltrating breast carcinoma, including invasive ductal and lobular carcinoma. The risk is bilateral; however, ipsilateral cancers predominate. Other LCIS patterns include pleomorphic and florid. Pleomorphic LCIS shows large discohesive cells with marked nuclear pleomorphism >4 times lymphocytes/equivalent to the cells of high-grade DCIS, with or without apocrine features. Pleomorphic LCIS can be associated with comedo-type necrosis. Up to 87% of pleomorphic LCIS are associated with an invasive component. Most pleomorphic lobular carcinomas are not estrogen receptor (ER)-positive. The proliferation index is usually higher in the pleomorphic LCIS. In addition, some of them overexpress HER2 and show TP53 mutation. Florid LCIS should have one of two architectural features: little to no intervening stroma between markedly distended acini of involved TDLUs and a size cut-off equivalent to ~40 to 50 cells in diameter. As compared to pleomorphic LCIS, diffuse and classic LCIS are usually ER-positive and show low Ki67 expression. They almost never show HER2 overexpression or TP53 mutation ( Figs. 19.6–19.9 ). Some LCIS spreads laterally through the ductal system analogous to Paget disease ( Fig. 19.10 ). This spread of LCIS is aptly termed pagetoid spread .

Molecular studies of LCIS by loss of heterozygosity and genomic hybridization confirm the molecular heterogeneity of LCIS. CDH1 inactivation has been documented in 81% of all types of LCIS. The inactivation of CDH1 causes the loss of E-cadherin. Other mutated genes in LCIS include PIK3CA (41%) and CBF2 (12%).

Most authorities agree that diagnosis of florid LCIS and pleomorphic LCIS on core needle biopsy requires surgical excision due to the high number of concurrent higher-grade lesions, including invasive lobular carcinomas. On the other hand classic LCIS could be treated by “watchful waiting.”

Mammographic findings associated with LCIS include calcification (~53%), mass/architectural distortion (31%), and nonmass enhancement on MRI (~13%). In the TNM American Joint Committee on Cancer (AJCC) staging 8th edition LCIS (no variant specified) is classified only as a high-risk lesion. The most recent European Society for Medical Oncology (ESMO) guidelines recommended considering classic LCIS as a “nonobligate precursor to invasive cancer and … a risk factor for future development of invasive cancer in both breasts. ESMO acknowledges that pleomorphic variant of LCIS may behave similarly to DCIS and should be treated accordingly, after multidisciplinary discussion.” Even though florid LCIS is not mentioned in the ESMO report, most authorities feel that these lesions also require surgical excision.

Classification of DCIS including special patterns will be discussed in the following sections.

Pathology and Classification of Ductal Carcinoma in Situ

DCIS comprises a heterogeneous group of noninvasive neoplastic proliferations of epithelial cells. DCIS is a clinically, radiologically, and histologically heterogeneous nonobligate precursor of invasive breast carcinoma. DCIS is usually a segmental disease arising in the TDLU and extending toward the nipple. DCIS and ADH can occur within other lesions, including papillomas and fibroadenomas. The distinction of ADH versus low-grade DCIS is predominantly quantitative and not qualitative. ADH either may show partial involvement of ductal spaces or is limited in size. Most authorities consider two involved duct spaces with ≤2 mm lesion as ADH. ADH usually presents as mammographic abnormalities, predominantly in association with calcification. ADH does not form mass lesions, unless it involves preexisting masses, such as fibroadenoma or papillary lesions. All ADH cases are ER-positive. In fact, they show diffuse and strong ER positivity. Pathologists do not always agree on whether small lesions should be considered ADH or in situ carcinoma. In general, lesions that involve only a few membrane-bound spaces and show low nuclear grade should be regarded as atypical hyperplasia and not in situ carcinoma. Diagnostic agreement rates for ADH are low (40%–60%). A second opinion is often sought following the diagnosis of ADH. There is a greater degree of concordance in larger lesions. Pathologists are more likely to agree on the diagnosis of difficult, smaller, and borderline lesions if they have agreed on criteria. It may at times be difficult to distinguish between DCIS and LCIS. Especially DCIS cases characterized by small uniform cells with a solid growth pattern can simulate LCIS ( Fig. 19.11 ). E-cadherin and p120 immunostains are useful in such overlap cases.

The grade of DCIS is far more important in most cases than the architectural pattern. Most authorities classify DCIS as a three-grade system, based on nuclear morphology. Grade I DCIS is morphologically similar to ADH. Both ADH and low-grade DCIS show the proliferation of uniform epithelial cells. By definition, there are no myoepithelial cells within the area of epithelial proliferation. The nuclei are small and do not show nucleoli. DCIS is usually a larger lesion. By definition, it measures >2 mm. Both ADH and low-grade (grade I) DCIS lack high molecular weight cytokeratin (HMWC) and show diffuse and strong ER positivity. Both show infrequent mitotic figures. Atypical mitotic figures are not present. Individual cell necrosis or apoptotic bodies may be present, but comedo-type necrosis is absent. The nuclei of grade I DCISs are 1–1.5 times the size of erythrocytes. Low-grade DCIS shows several architectural patterns. These include solid, cribriform, and micropapillary ( Figs. 19.12–19.14 ). Focal microcalcification is commonly encountered in low-grade DCIS ( Fig. 19.15 ).

In addition to ER positivity, many grade I DCISs express progesterone. They also express Bcl2. HER2 overexpression or TP53 mutation has not been reported.

Grade II or intermediate-grade DCIS is characterized by the presence of nuclei with mild to moderate pleomorphism and variability in nuclear polarization. The chromatin is coarse, and the nuclei can show prominent nucleoli. Necrosis is often seen as either individual cell necrosis, apoptotic bodies, or comedo-type necrosis (“The central portion of an involved ductal space is replaced by an area of expansive dirty necrosis that is easily detected at low magnification. Ghost cells and karyorrhectic debris are generally present.”) They usually lack the monotony of nuclei seen in grade I DCIS. “Microcalcification may be present in the tumor cells and/or in necrotic material.” They may show various architectural patterns, including solid, cribriform, and micropapillary ( Figs. 19.16–19.18 ). Grade II DCIS may present as mass lesion.

Grade III DCIS is characterized by the presence of markedly atypical nuclei, often markedly pleomorphic, and can be associated with comedo necrosis. The nuclei show an irregular contour, coarse chromatin, and frequent prominent nucleoli. The nuclei are >2.5 times the size of erythrocytes. Grade III DCIS can show solid, cribriform, or micropapillary patterns. A solid pattern is more common. In addition, they can show ducts lined by a single layer of a markedly pleomorphic cell population ( Fig. 19.19 ).

Multiple studies have shown that nuclear grade is important in predicting local recurrence (LR) and predicting progression to invasive carcinoma. Although nuclear grade and necrosis define most of the risk associated with DCIS, certain architectural patterns appear to bear clinical significance independent of the nuclear grade. For example, DCIS with almost pure micropapillary architectural features is associated with a higher recurrent index and appears to be independent of the nuclear grade. Micropapillary DCIS is characterized by the presence of numerous club-like micropapillae formed by a relatively uniform cell population. Micropapillae lack fibrovascular cores ( Fig. 19.20 ).

The classification scheme proposed by Page and Lagios is summarized in Table 19.1 . This classification most closely resembles the classification adopted by the World Health Organization (WHO) working group. The WHO classification is widely used throughout the world.

Conventional classification of DCIS covers perhaps 85% of what is recognized as noninvasive ductal carcinoma. The following sections will discuss the less common presentations and patterns of preinvasive cancer, DCIS, and lesions thought to closely behave in a similar manner, with an emphasis on the clinical implication.

Special Types of Ductal Carcinoma in Situ With Special Implications

Paget Disease

Mammary Paget disease (MPD) is believed by most authorities to be a manifestation of grade III DCIS, which, when involving subareolar ducts, can extend within the confines of the duct and epidermal basement membrane into the epidermis. MPD characteristically presents as an eczema- or contact dermatitis–like lesion of the nipple ( Fig. 19.21 ). Other presenting symptoms include erythema, ulceration, hyperkeratosis, hyperpigmentation, or bloody nipple discharge. The pigmented variant of MPD is difficult to distinguish from malignant melanoma. The median age of the patients is 64 (20–90) years. Around 2% are men, and around 85% are White females. MPD can extend beyond the nipple to the areolar region and surrounding epidermis. On rare occasions, MPD has been reported to involve the entire chest wall and beyond.

Histologic sections show small groups or individual large pleomorphic cells with abundant clear to eosinophilic cytoplasm infiltrating the dermal/epidermal junction with individual neoplastic cells extending and involving epidermal layers. This infiltrate can be accompanied by small, mature lymphoid cells. Epidermis may show prominent hyper- and parakeratosis. Unlike epidermal cells, which show CK5/6 positivity, Paget cells show positivity for CK7 ( Figs. 19.22–19.24 ). In addition, many of them show HER2 overexpression ( Fig. 19.25 ). On rare occasions MPD is CK7-negative, but many of these cases show GATA3 positivity. MPD can also be associated with underlying invasive cancer. Almost all of these invasive cancers have an associated in situ component. Rarely MPD has been reported in pregnant women. Increased awareness of MPD in pregnant women might prevent delays in the diagnosis. MPD has also been reported in men. It comprises 1.45% of all male breast cancers, compared to 0.68% of all female breast cancers. The most frequent clinical features seen in men are ulceration and eczema-like dermatitis, followed by the presence of a mass and nipple discharge. The presence of a palpable mass increases the likelihood of invasive carcinoma. In both males and females, MPD without an underlying carcinoma is usually confined to the areola.

Ductal Carcinoma in Situ, Apocrine Type

Apocrine epithelium is estrogen/progesterone receptor–negative and does not express Bcl2; however, apocrine cells express androgen. Apocrine cells appear most frequently in benign breast lesions including fibrocystic changes, adenosis, fibroadenoma, and papilloma. Rarely ADH, DCIS, and invasive ductal carcinomas are lined by apocrine cells. In order for a lesion to be classified as apocrine-type DCIS, the lesion should show proliferation of apocrine cells (cells with abundant, finely granular eosinophilic cytoplasm and large/moderately vesicular nuclei with occasional prominent nucleoli). These cells usually show significant pleomorphism with enlarged nuclei and one or multiple prominent nucleoli. The recognition of apocrine DCIS, as a special type, continues to be debated. Focal apocrine morphology can be encountered in up to 60% of carcinomas with no specific type. It is important to distinguish apocrine DCIS from other benign apocrine proliferation and emphasize its unique immune profile. For a lesion to be classified as apocrine, more than 90% of the lesion should show apocrine features.

The criteria used to recognize nonapocrine ADH and DCIS cannot be applied to cases with apocrine differentiation, primarily because most apocrine proliferations lack any myoepithelial cells, and immunostaining for CK5/6 is not helpful. In addition, apocrine cells usually have large vesicular nuclei. O’Malley and colleagues used cytologic characteristics (irregular nuclear membranes, multiple, small nucleoli, and variety in nuclear size) and size <8 mm to classify lesions as ADH, and lesions greater than 8 mm were classified as DCIS. Tavassoli and Norris used cytologic atypia (at least threefold nuclear enlargement) and nuclear stratification or papillary tufting. They usually require that the lesion involves more than two adjacent lobular units and the intervening ducts. Most apocrine DCIS cases show high nuclear grade and necrosis, and intraluminal necrosis is not a prerequisite for the diagnosis of apocrine DCIS ( Fig. 19.26 ). LCIS with apocrine changes is considered to be a variant of pleomorphic LCIS ( Fig. 19.27 ).

In summary, apocrine ductal carcinoma is rare. The diagnosis should be made with consideration of other entities mimicking DCIS, including atypical apocrine adenosis. Molecular studies have shown that these tumors are ER/PR-negative; however, they express androgen receptors. HER2 may or may not be overexpressed.