Tumors of the Breast

Fibrocystic Change

Fibrocystic change is the preferred term used by pathologists to note combinations of breast changes, including cyst formation, apocrine metaplasia, columnar cell change, and various forms of adenosis. Minor changes, including fibrosis, microcyst formation, lobular involution, and minor degrees of sclerosing adenosis or columnar cell change, should be regarded as variants of normality and classified as minimal alteration or “aberrations of normal development and involution.” The use of an umbrella category such as fibrocystic change, which covers a variety of alterations in the breast that may have differences in etiology or clinical relevance, is acceptable when a dominant distinct histologic process is not present.

After publication of retrospective studies of breast cancer risk related to benign breast disease, the importance of the presence of epithelial proliferation in association with fibrocystic change and other benign conditions, such as papillomas, has increasingly become recognized. The presence or absence of epithelial hyperplasia and its character is now regarded as a mandatory subclassification of fibrocystic change (see later discussion of proliferative breast disease).

Cysts

Histologic Appearances.

Two main forms of cyst are recognized; those lined by a single layer of cuboidal ( Fig. 16.1 ) or flattened, attenuated epithelium and, more common, those lined by apocrine epithelium (see Fig. 16.1 ). This latter type of epithelium resembles normal apocrine sweat gland epithelium, the cells being large and columnar with abundant granular eosinophilic cytoplasm and usually basally positioned nuclei. The cytoplasm may protrude into the lumen in the form of apical snouts. The granules show Sudan black and periodic acid–Schiff (PAS) diastase-resistant positivity. The apocrine epithelial layer is usually single, but hyperplasia resulting in a papillary structure can occur. The significance of papillary apocrine change and apocrine metaplasia is dealt with later in the section on epithelial hyperplasia.

The distinction of two principal cyst types is supported by biochemical studies of cyst fluid. Cysts with a high pH and a low sodium to potassium ratio (similar to that of intracellular fluid) are more frequently recurrent and usually of apocrine type, in contrast with those with a lower pH and higher sodium to potassium ratio (corresponding to plasma), which are less likely to recur and are lined by a cuboidal or attenuated epithelial layer.

Apocrine Metaplasia

Apocrine metaplasia is a frequent finding in the breast and is associated with cyst formation (see earlier discusssion). In its usual form, regular columnar cells are arranged in a single layer above a normal myoepithelial layer. Apocrine proliferation may be more florid and take on a papillary configuration; usually these are well structured often with fine fibrovascular cores. More rarely, complex patterns of apocrine hyperplasia may be found in the form of sheets or complex architectural patterns with multiple irregular luminal spaces.

In addition to its usual association with cysts, apocrine metaplasia can extend to involve other benign processes, including sclerosing adenosis (which is sometimes described as apocrine adenosis) ( Fig. 16.2 ), papillomas, or fibroadenomas. No particular significance is attributed to this phenomenon.

Although usually regular in nuclear and cytologic morphology, DNA tetraploidy has been identified in apocrine metaplasia. This is believed to be the explanation for forms of “apocrine atypia,” in which nuclear enlargement and variation in size with prominence of nucleoli can be observed (see Fig. 16.2 ). At present, the relevance and clinical significance of apocrine atypia is poorly understood. Some authorities believe this to be a purely benign phenomenon, whereas others have suggested it may be a precursor lesion of some forms of carcinoma such as apocrine carcinoma or medullary carcinoma. Although pure apocrine change is currently regarded as carrying a low but significant increased risk of subsequent development of breast cancer, most of the increased risk is related to the presence of coexisting atypical hyperplasia.

Clinical Features

Sclerosing lymphocytic lobulitis usually presents as a mass and can mimic carcinoma. The masses can be multiple or bilateral. By recognizing this condition clinically and on core biopsy, unnecessary surgery may be avoided. It can be seen in women from about 20 to 65 years old but is most common in women in their 30s. It has also been described in men.

Macroscopic Features

Typically a poorly defined firm gray-white mass is seen, but sometimes the appearance is essentially that of normal fibrous breast tissue.

Histologic Appearances

The characteristic feature is circumscribed clusters of lymphocytes in and around lobules and ducts and around blood vessels ( Fig. 16.3 ). Typically the lymphocytes are predominantly B cells. Germinal centers may be present. Lobular atrophy is frequently evident. Fibrosis and sclerosis are usually present between, and sometimes within, lobules. Epithelioid fibroblasts may be present but are not specific for this disorder. Patients with sclerosing lymphocytic lobulitis who have had a series of biopsies often show progression from dense inflammation to increasing lobular atrophy and fibrosis with decreasing inflammation. The changes seen later in the disease (lobular atrophy and fibrosis with little inflammation) are not specific to this disorder. The pathologic features in patients with and without diabetes mellitus are similar. Thus a general pathologic term such as sclerosing lymphocytic lobulitis or lymphocytic mastopathy is preferable to diabetic mastopathy.

Differential Diagnosis

The diagnosis of sclerosing lymphocytic lobulitis is usually straightforward. It may be confused with lymphoma, but lymphoma typically has a different architecture, composed of sheets of cells, often with distinct morphology. An association between lymphoma and sclerosing lymphocytic lobulitis has been described in Japan but was not seen in European or American series. A pattern of inflammation like that of sclerosing lymphocytic lobulitis may be associated with carcinomas, particularly invasive lobular carcinoma, and, if marked, the inflammation may obscure the carcinoma.

Clinical Features

Fibroadenoma is one of the most common causes of a benign lump in the breast. They may occur at any age after puberty but are most frequent in the third decade. They may be single or multiple and unilateral or bilateral. Clinically they present as firm or rubbery, mobile, well-defined masses that are painless. Most measure between 1 and 2 cm in diameter, but fibroadenomas measuring up to 4 cm are not uncommon. After the advent of screening programs for breast cancer, impalpable fibroadenomas have increasingly been detected by mammography.

The great majority of fibroadenomas behave in a benign fashion clinically and do not recur after adequate resection. Because they may be multiple it is not unusual for new lesions to become apparent, even close to the site of a previous biopsy. The trend is increasing for clinicians to adopt a more conservative approach to the management of women with suspected fibroadenoma using the triple approach of clinical examination, imaging, and fine-needle aspiration cytology or needle core biopsy. Provided that all the criteria for benignity are satisfied, surgery may safely be avoided in most patients. However, if the lesion is large (e.g., >3 cm) or is known to be increasing in size, excision is typically recommended.

Malignant change in fibroadenomas is uncommon. When seen it is usually in the form of carcinoma in situ; lobular carcinoma in situ (LCIS) being slightly more common than ductal carcinomas in situ (DCIS) while invasive carcinoma is rare. In a proportion of cases these lesions are also present in the adjacent tissues, and it may be difficult to establish whether involvement of the fibroadenoma is merely coincidental. Epidemiologic studies have shown a very slight increase in relative risk for subsequent invasive carcinoma of approximately twofold in patients with fibroadenoma. This risk is increased to threefold for fibroadenomas with a “complex” morphologic pattern (see later discussion) and to fourfold for patients with a family history of breast cancer. However, the absolute risk for uncomplicated fibroadenomas remains very low at 4%.

Macroscopic Appearances

The gross appearances of a fibroadenoma are distinctive ( Fig. 16.4 ). They are sharply circumscribed, spherical or ovoid nodules, gray-white in color, and clearly separate from the adjacent breast tissue. The cut surface is usually slightly lobulated and often has a glistening myxoid appearance. Most lesions are firm to the touch, but in older women they may be hard and sclerotic, even calcified.

Histologic Appearances

The dominant element is a proliferation of loose cellular stromal connective tissue, which surrounds a variable number of ductular structures ( Figs. 16.5 and 16.6 ). The stromal nuclei are spindle shaped and normally exhibit little pleomorphism with infrequent mitoses. The quality of the stromal matrix can vary markedly, some fibroadenomas having a definite myxoid background whereas others are hyalinized. The ductules also vary in configuration, and two classic patterns are described: intracanalicular when they are compressed by the stroma into clefts (see Fig. 16.5B ), and pericanalicular when the stroma appears to surround ductules in a circumferential fashion (see Fig. 16.6 ). Both patterns are often seen in the same lesion. The terms have no practical or prognostic significance and are purely descriptive. In about 20% of cases cystic changes, sclerosing adenosis, epithelial calcifications, or papillary apocrine change are seen, either alone or in combination; such lesions have been termed complex fibroadenoma by Dupont and colleagues. Epithelial hyperplasia is common, and, although tangential sectioning may produce worrying patterns, true atypical hyperplasia is rarely seen. In older patients, especially after the menopause, the stroma of fibroadenomas typically becomes less cellular and hyalinized. Calcification may also develop.

Differential Diagnosis

The most important lesion to consider in the differential diagnosis is benign phyllodes tumor (see later discussion). Although the latter is more often seen in an older age group, difficulties may be experienced with occasional fibroadenomas with a cellular stroma and prominent epithelial clefts. The distinction between fibroadenoma and benign phyllodes tumor may uncommonly be impossible to establish in excised lesions, but low stromal cellularity with relative uniformity of stromal nuclei and lack of mitoses would favor the former. It may be necessary to issue an equivocal report in rare circumstances. The WHO classification suggests that such lesions are better classified as fibroadenoma. In the past, complete excision of such cases was advised to reduce the risk of local recurrence, but current opinion is to manage as fibroadenoma. The distinction between fibroadenoma from mammary hamartoma is discussed later. Myofibroblastoma may show some features resembling fibroadenoma, but myofibromablastoma is devoid of breast parenchymal element, and the lesional cells are positive for desmin and estrogen receptor. Fibromatosis is a rare cause of a breast lump (albeit usually not well defined), but the fibroblastic proliferation around ductular structures may resemble a fibroadenoma. The infiltrative edge, cellularity of stroma, and relatively scanty epithelial component are all points in favor of a fibromatosis (see later discussion). The stroma of a fibroadenoma is typically strongly positive for CD34, whereas fibromatosis is CD34 negative and may be nuclear β-catenin positive.

Clinical Features

The clinical features of tubular adenoma are similar to those of fibroadenoma. A variety of benign noninflammatory mass lesions may present during pregnancy. In a consecutive series of 28 patients with an age range of 16 to 48 years (mean 23 years), three types of lesions were identified : Adenomas in the form of fibroadenoma (16 cases) and tubular adenoma (2 cases) were the most common lesions and usually showed lactational changes, with “lactational adenoma” (10 cases) being less frequent. The lesions presented either during pregnancy or in the postpartum period (up to a few months), usually as nontender masses.

Macroscopic Appearances

Tubular adenomas are well-circumscribed nodules measuring between 1 and 4 cm in diameter, the majority being no greater than 2 cm. The cut surface is finely nodular.

Histologic Appearances

The lesions are sharply demarcated from adjacent breast tissue but lack a true capsule. They consist of closely packed tubular structures of approximately the same size as the acini within a normal lobule ( Fig. 16.8 ). The tubules are lined by a single layer of luminal epithelial cells, but scanty flattened myoepithelial cells are also present. No nuclear atypia is seen, and mitoses are usually infrequent. A small amount of fine connective tissue stroma is seen between the tubules.

In lactational adenomas, dilated acini are present and show the typical alveolar pattern of lactating breast tissue. Luminal secretion is usually present.

Differential Diagnosis

The only lesions to be considered are fibroadenoma and tubular carcinoma. In the former the stroma is much more prominent, and the tubular structures present are arranged in well-defined islands. A clearly two-layered epithelium is seen in the tubules. Tubular adenoma should not be mistaken for tubular carcinoma. The latter has a characteristic stellate configuration with central fibrosis and elastosis with radiating tubules. The tubules are lined by a single layer of epithelial cells and infiltrate into stroma and adipose tissue at the periphery.

Clinical Features

Mammary hamartomas are relatively uncommon lesions that occur predominantly in the perimenopausal age group. They may present as a large palpable mass, but surprisingly they are often impalpable and detectable only mammographically. Clinically they may be confused with fibroadenoma or phyllodes tumor. Radiologically they are large, well-circumscribed mass lesions with characteristic central lucency. Although they may cause marked distortion of the breast, hamartomas are benign and rarely recur after adequate local excision.

Macroscopic Appearances

The lesions vary considerably in size from 1 to 20 cm, although the majority measure between 2 and 5 cm. They form well-circumscribed oval masses that have a cut section that may resemble normal breast tissue, a lipoma, or a fibroadenoma.

Histologic Appearances

Microscopically hamartomas lack a true capsule, although they separate easily from adjacent breast tissue. They are composed of a variable mixture of connective tissue stroma and breast lobules ( Fig. 16.9 ). The latter usually have a normal, albeit architecturally somewhat disorganized, configuration. Occasional ectatic ducts are present, and these may become cystic. Epithelial hyperplasia is not a feature. Islands of adipose tissue are found frequently, often in the center of the lesion ( Fig. 16.10 ); the term adenolipoma may be applied to such lesions. Pseudoangiomatous hyperplasia or smooth muscle may be present in the stroma.

Differential Diagnosis

For palpable masses the differential diagnosis includes fibroadenoma for smaller lesions and juvenile fibroadenoma, pseudoangiomatous stromal hyperplasia (PASH) and phyllodes tumor for larger lesions. At the microscopic level, the presence of more extensive stroma with relatively few, if any, lobular structures distinguishes hamartomas from fibroadenoma and the cellular stroma surrounding characteristic leaflike epithelial clefts, the phyllodes tumor. Although PASH is described as a component of some hamartomas, it is usually a relatively minor component and should not be confused with mass-forming PASH in which the stromal component predominates. Definite diagnosis of a hamartoma, particularly the distinction from normal breast tissue, may be impossible in a core biopsy.

Clinical Features

Phyllodes tumors are rare and account for less than 1% (0.2%–0.9%) of all mammary neoplasms and about 2% of fibroepithelial lesions in the breast. They occur predominantly in middle-aged or elderly women and are relatively rare below the age of 40 years. Phyllodes tumors form lobulated, firm masses that may grow rapidly and cause unilateral breast enlargement or even ulceration of overlying skin. Traditionally phyllodes tumors were said to be very large, but, perhaps because of greater breast awareness and mammographic screening, many patients now present with tumors of 2 cm diameter or less.

Macroscopic Appearances

Phyllodes tumor forms a firm lobulated mass varying in size from 2 to 10 cm in diameter, with an average size of 5 cm. It is usually well circumscribed, often with a bosselated contour. The cut surface shows a characteristic whorled pattern, resembling a compressed leaf bud, with visible clefts. Larger tumors frequently exhibit cystic spaces, and foci of hemorrhage and necrosis are not unusual.

Histologic Appearances

Microscopically, phyllodes tumors are composed of two major elements: clefts lined by epithelial cells and an associated cellular stroma ( Fig. 16.11 ). The epithelial element consists of the usual two layers of myoepithelial and luminal epithelial cells. Focal epithelial hyperplasia is not uncommon, and although this may occasionally be florid, carcinomatous change is extremely rare. Both lobular neoplasia and DCIS have been recorded; associated invasive carcinoma is an even greater rarity.

The appearances of the stromal element vary considerably from case to case. In the majority the features are those of a benign process. Although the stroma is more cellular than in a fibroadenoma, the spindle cells do not exhibit pleomorphism, and mitoses are infrequent. The presence of occasional bizarre giant cells ( Fig. 16.12A ) does not indicate malignant change (and may also be seen in fibroadenomas). At the other end of the spectrum a minority of tumors will show frankly sarcomatous change, characterized by stromal overgrowth and hypercellularity, marked nuclear atypia, and an increased mitotic count (see Fig. 16.12B ). In particular, specific patterns such as liposarcoma, rhabdomyosarcoma, chondrosarcoma, or osteosarcoma are clear indicators of malignancy. An intermediate group of tumors remains with appearances that pose problems for the pathologist and clinician in predicting the likelihood of local recurrence and metastatic malignant potential.

Although the majority of phyllodes tumors are benign (>60%), no objective system is globally recognized for grading phyllodes tumors. Numerous studies have used features of the stromal cells, including degree of cellularity, pleomorphism, mitotic count, and stromal overgrowth, as well as an infiltrative rather than a pushing margin to grade phyllodes tumors. Different studies have used different criteria and variable clinical management of patients (such as excision margin width), few have had more than 100 patients, and all are retrospective. Nevertheless, most studies find an association between grade and distant metastasis: Most distant metastases arise from malignant tumors. Distant metastasis is uncommon, occurring in fewer than 10% of malignant lesions (although published data vary); lung, pleura, and bone are the most frequent sites. Some studies find an association between higher grade and local recurrence, but the major determinant of local recurrence appears to be completeness of excision. No consensus exists on what is an adequate margin as this has not been studied sufficiently.

Differential Diagnosis

Differential diagnosis of phyllodes tumor has been considered previously in the section on fibroadenoma . In most cases the distinction from fibroadenoma is straightforward, but difficulties may sometimes be encountered. Hypercellular stroma is an important feature favoring a diagnosis of phyllodes tumor. A leaflike pattern is usually present in phyllodes tumor, particularly benign tumors, but can be seen in fibroadenoma. Another important consideration is that some phyllodes tumors show variable appearances of the stroma, including bland areas resembling fibroadenoma.

This heterogeneity may make diagnosis on a core biopsy difficult. Useful features favoring phyllodes tumor on core biopsy are increased stromal cellularity, stromal overgrowth and increased mitoses in the stroma, and tissue fragmentation. In cases showing overlapping features between fibroadenoma and benign phyllodes tumors, the WHO Working Group has proposed that the term benign fibroepithelial neoplasm be used, to avoid overtreatment. This term, however, should be limited to rare equivocal cases, as it does not represent a new diagnostic category.

Fibromatosis, primary sarcoma, and metaplastic carcinoma of the breast must be distinguished from malignant phyllodes tumors. Again, the mixtures of elements, with the leaflike epithelial clefts, are features of phyllodes tumor not seen in the other lesions. Expression of cytokeratins favors metaplastic carcinoma, but focal expression may be seen in phyllodes tumor. Most phyllodes tumors express CD34 in the stromal component, whereas spindle cell carcinoma is negative.

Clinical Features

The majority of patients who present with a clinically palpable mass are premenopausal or perimenopausal, but the age range varies from 20 to 70 years. The mass is usually ill defined, but skin or deep fascial fixation does not occur. In contrast with carcinoma of the breast, associated pain and tenderness are not uncommon. After the advent of screening programs for breast cancer, sclerosing adenosis is detected with increasing frequency by mammography because of the presence of microcalcifications (see later discussion).

Macroscopic Appearances

In the diffuse, microscopic form, no distinctive gross appearances are seen. The nodular or tumor type forms an ill-defined, firm or hard mass varying in size from 0.5 to 3 cm, with an average of 1.5 cm. No specific coloration is seen.

Histologic Appearances

Histologically the normal configuration of a group of lobules is distorted by a disorderly proliferation of acini and intralobular stromal cells ( Fig. 16.13 ), which gives an overall whorled appearance. Compressed microtubular structures may be seen, but acinar lumina are frequently obliterated. A two-layered structure is usually visible, at least in part, but epithelial and myoepithelial cells may proliferate separately. Identification of the myoepithelial component is facilitated by the use of immunohistochemical markers such as anti–smooth muscle actin ( Fig. 16.14 ), smooth muscle myosin heavy chain, or p63. Epithelial nuclei are small and regular, usually without atypia, and mitoses are infrequent. Fine microcalcification is frequently found in the luminal spaces of the microtubular structures (see Fig. 16.13B ). In the diffuse form, several small, discrete lobular collections may be seen, often admixed with other features of fibrocystic change. In the nodular form the distorted lobules are fused together in an organoid configuration to form a mass lesion.

Differential Diagnosis

If the nodular form produces a large mass, it will become palpable and may be indistinguishable clinically from a carcinoma. Mammographically it is often impossible to distinguish the microcalcifications from those seen in low-grade cribriform DCIS. Histologically the main differential diagnoses are those of CSL and invasive carcinoma. Central elastosis with entrapped tubules, an important diagnostic feature of CSLs, is occasionally seen in sclerosing adenosis. Conversely, foci of sclerosing adenosis are found not infrequently associated with CSLs. This suggests that the two lesions may represent the ends of a spectrum of proliferative changes, and that attempting to separate them is a semantic argument. For the present it is probably sensible at a practical level to decide which pattern predominates and use that as the diagnostic category.

The distinction from invasive carcinoma, especially the tubular type, is the more important differential diagnosis. In a very small proportion of cases of sclerosing adenosis, extension into perineural spaces and blood vessels can be demonstrated. Areas of apocrine change may also be present, with cytologic atypia ; in the absence of confirmatory evidence neither feature should be taken as evidence of malignancy. The main distinguishing features from invasive tubular carcinoma are the lobulocentric configuration, the disorderly proliferation of microtubules lined by two-layered epithelium, and the lack of separate tubular structures with clear lumina, set in a desmoplastic stroma. Unlike sclerosing adenosis, tubular carcinoma may be associated with DCIS. In problematic cases, particularly in a core biopsy, myoepithelial markers unequivocally differentiate sclerosing adenosis from invasive tubular carcinoma.

Although the risk of subsequent invasive carcinoma appears to be low, both DCIS and lobular neoplasia can be present in association with sclerosing adenosis. The diagnosis of DCIS should pose little problem, because benign ductal epithelial hyperplasia is rarely seen in sclerosing adenosis. Lobular proliferations require a more cautious approach, and a diagnosis of LCIS should not be made unless all the criteria are fully satisfied.

Clinical Features

Presentation of microglandular adenosis is usually in the form of a palpable mass with relatively sharp definition, which may be up to several centimeters in diameter. Smaller foci can be identified as a chance finding in breast tissue excised for other reasons. Age at presentation varies widely, from the mid-20s to over 80 years, with the majority of patients presenting between the ages of 45 and 60 years. Mammographic features are not specific but may be suggestive of malignancy.

In the past, most authorities regarded microglandular adenosis in its pure form as an entirely benign condition with no associated risk for the development of breast cancer. However, several authors have reported the coexistence of epithelial proliferative lesions, including atypical hyperplasia and in situ and invasive carcinoma. Recent molecular evidence suggests that some cases of microglandular adenosis are clonal, and identical genetic changes have been described in cases of microglandular adenosis with adjacent carcinoma. This suggests that some examples of microglandular adenosis may be nonobligate precursors of invasive carcinoma, which is often of triple-negative phenotype. Complete excision of microglandular adenosis, particularly when showing atypical features, is therefore now advised.

The cut surface of involved breast tissue shows an ill-defined area of nondescript fibrofatty tissue. The density and appearance vary, reflecting the nature of the underlying breast stroma.

Histologic Appearances

Microglandular adenosis is defined as a proliferation of small, round acinar structures that are not lobulocentric but haphazardly arranged and appear to infiltrate fibrous stromal and adipose tissue. The glands are typically formed by a single layer of uniform, cuboidal epithelial cells, which have clear or eosinophilic cytoplasm and lack apical snouts ( Fig. 16.15 ). Mitoses are absent or very rare. A myoepithelial cell component is absent. No associated stromal reaction is seen, and occasionally hyalinization can occur. A periglandular distribution of reticulin fibers and glands may contain PAS-positive secretions. In very rare cases associated with epithelial proliferation the glands may vary in size. When microglandular adenosis shows cytologic atypia, small solid clusters, and/or fused irregular glands, the term atypical microglandular adenosis is used.

Differential Diagnosis

The principal differential diagnosis of microglandular adenosis is invasive adenocarcinoma of pure tubular type, particularly the sclerosing variant. In both conditions, abnormal infiltration of the glandular structures is seen in the mammary stroma and adipose tissue. A distinction can be made by using the following criteria: Tubular carcinomas usually have a distinctive stellate configuration with central elastosis and fibrosis. They excite a characteristic metachromatic stromal reaction that is not found in microglandular adenosis. The glandular structures in tubular carcinomas contain a high proportion of irregular shapes and elongated tubular forms, some showing branching. Cytologically, the cells in tubular carcinoma are larger with more abundant eosinophilic cytoplasm, often forming characteristic apical snouts. Secretion is rare in the luminal spaces of tubular carcinoma. Although most studies have emphasized that microglandular adenosis is characterized by a single layer of epithelial cells (see Fig. 16.15 ), some authors have reported the presence of myoepithelial cells in a minority of cases. In such cases the use of immunostaining with myoepithelial markers is of additional value in distinguishing microglandular adenosis from tubular carcinoma. Valuably, microglandular adenosis lesions express S100 protein ( Fig. 16.16 ) but lack estrogen receptor (ER) expression, whereas pure tubular carcinoma is typically ER positive and S100 negative. Sclerosing adenosis can be distinguished from microglandular adenosis by its lobulocentric configuration and the presence of a distinct layer of myoepithelial cells (see Fig. 16.14 ). The rare form of adenomyoepithelioma described by Kiaer and colleagues may have components indistinguishable from those of microglandular adenosis but typically has a dual myoepithelial and epithelial component. Its features are described in more detail in a later section.

Clinical Features

RS/CSLs are reported to occur in women from 15 to 80 years in age but are most frequent in premenopausal and perimenopausal women. They were initially thought to be uncommon, perhaps because they do not usually present as a palpable mass. However, studies of surgical mastectomy specimens and breast tissue from autopsies have shown them to be common, multiple, and frequently bilateral. They are detected with increasing frequency in breast screening programs because it is virtually impossible to distinguish them from small invasive carcinomas mammographically. For this reason, all stellate lesions detected mammographically should be thoroughly sampled (e.g., with vacuum-assisted biopsy/excision) and examined histologically. It is important to search for epithelial atypia in a core biopsy with features of an RS. If atypia is present, the chance of associated carcinoma in the subsequent wider specimen is about 40%. If no atypia exists on the core then the risk of carcinoma in the adjacent breast is about 5%, so removal with a vacuum-assisted device is a reasonable alternative.

Macroscopic Appearances

The gross appearances of small RSs are unremarkable; lesions that measure less than 3 to 4 mm may become apparent only on histologic examination. Larger RSs, together with CSLs, form firm, white stellate nodules with a dense center of variable size and fine radiating spicules of fibrous tissue. Central elastosis may be visible to the naked eye as pale yellow flecks.

Histologic Appearances

The microscopic features depend on the stage of evolution of the lesion. Small early RSs are composed of a group of radiating ductules surrounded by cellular connective tissue, with minimal elastosis ( Fig. 16.17 ). In larger lesions an accumulation of dense sclerotic fibrous tissue occurs in the center of the lesion, and elastic tissue is more evident ( Fig. 16.18 ). The central tubules become compressed, but the normal two-layered epithelium is preserved ( Fig. 16.19 ). Tubular structures are also present within the radiating arms, and more peripheral cystic change is frequently observed (see Fig. 16.17 ). A variable degree of epithelial hyperplasia may be seen; this is most frequently of usual type ( Fig. 16.20 ), but careful evaluation for the presence of atypical features, both ductal and lobular, is important ( Fig. 16.21 ).

Apart from their larger size, CSLs, as the name implies, exhibit a wider variety of morphologic patterns. All the basic histologic features of an RS are present but, in addition, foci of sclerosing adenosis may be seen, with more florid epithelial hyperplasia and occasionally with papilloma formation. Thorough sampling of any lesion is required to obtain a full appreciation of the structures present.

Differential Diagnosis

Histologically RS/CSL must be distinguished from the nodular form of sclerosing adenosis and from tubular carcinoma. The former has been discussed previously, and this section will concentrate on the distinction from tubular carcinoma. The presence of entrapped tubules within the central fibroelastotic core poses the greatest problem. In RS/CSL the tubular structures are lined by two layers of epithelial cells in contrast to the single layer of the neoplastic tubules of a tubular carcinoma (compare Fig. 16.19 with Fig. 16.59 ). Immunohistochemistry (IHC) for myoepithelial markers is useful in confirming the difference in difficult cases. The entrapped tubules of RS/CSL are flattened and elongated, whereas those of a tubular carcinoma are more rounded or oval with open luminal spaces. The nature of the associated connective tissue stroma also provides a useful contrast; this is typically dense and relatively acellular in RS/CSL, but desmoplastic in tubular carcinoma.

As noted earlier, a variable degree of epithelial proliferation often exists in RS/CSLs, ranging from usual hyperplasia to atypical hyperplasia of either (or both) ductal and lobular types. The latter needs to be distinguished from in situ carcinoma, but it is quite clear that both in situ ( Fig. 16.22 ) and invasive carcinoma may arise in association with RS/CSLs. As long as the diagnostic criteria for malignancy are met, the appropriate designation should be recorded.

Macroscopic Appearances

In the great majority of cases, duct papillomas measure less than 2 to 3 mm in diameter and are therefore difficult to visualize with the naked eye. The site of a papilloma within a dilated duct may be indicated by adjacent amorphous debris, including blood clot, but, in most cases, the pathologist receives a specimen of fibrofatty tissue in which no gross abnormality is discernible. Multiple blocks may be required to identify the site of the lesion. Larger papillomas up to 1 to 2 cm may become encysted.

Histologic Appearances

Microscopically, papillomas consist of a fronded fibrovascular core attached by a stalk to the duct wall ( Fig. 16.23 ). Multiple fronds from the same lesion may extend into several adjacent duct spaces. Occasionally the base may be sessile. The fronds are covered by two types of epithelium ( Fig. 16.24 ), an outer cuboidal or columnar luminal cell layer and an inner myoepithelial cell layer. Any doubts concerning the presence of two cell types may be resolved by use of IHC for myoepithelial markers. The myoepithelial cell layer may be incomplete, but in rare cases it forms a major part of the epithelial proliferation. In the case of single, central duct papillomas the adjacent duct structures are usually unremarkable, although they may be dilated and contain hemorrhagic debris. Apocrine change is common and is usually focal. Infarction and hemorrhage are less common.

Multiple papillomas are usually found in smaller peripheral ducts, frequently as part of a focal or more diffuse epithelial proliferative process. They may therefore be accompanied by almost any type of epithelial lesion, including usual type hyperplasia, ADH, lobular neoplasia, or DCIS. It is important to search for epithelial atypia in a core biopsy sample with features of a papilloma. If atypia is present the chance of associated carcinoma in the surgical specimen is about 40%. If no atypia is seen on the core then the risk of carcinoma in the adjacent breast tissue is approximately 5%.

Differential Diagnosis

Papillary lesions can be categorized into three main groups; morphology and IHC for myoepithelial markers (smooth muscle actin, p63, and smooth muscle myosin heavy chain), basal keratins (cytokeratin [CK] 5 and CK14), and ER are useful in separating them. Intraductal papillomas have a myoepithelial layer between the fibrovascular cores and the overlying epithelium, as well as around the duct space. In addition, any associated epithelial hyperplasia of usual type shows patchy expression of basal keratins and typically ER. Papillary carcinoma in situ shows an absent myoepithelial layer at the epithelial-stroma interface of the fibrovascular cores, and the epithelium is usually negative for basal keratins. The epithelial cells usually show strong diffuse ER expression. The third and least common category of benign papilloma with involvement by DCIS or ADH shows a retained myoepithelial layer and absent expression of basal keratins in the atypical areas.

Sclerosed papillomas may contain entrapped epithelial structures. Care must be taken to interpret these correctly to avoid overdiagnosis of malignancy (see later section on encysted papillary carcinoma). Retention of a myoepithelial layer with use of IHC may be helpful. Smooth muscle myosin heavy chain and p63 are the markers of choice as they are expressed by myoepithelium, but not by myofibroblasts/fibroblasts; smooth muscle actin is expressed by both.

If a papillary lesion is completely infarcted, then definite categorization is usually not possible.

Macroscopic Appearances

The pathologist usually receives all or part of the nipple and areola, depending on the size of the lesion, which rarely measures more than 1 to 1.5 cm. The nodule is ill defined and firm but has no specific distinguishing features.

Histologic Appearances

The microscopic features are variable but consist in essence of a diffuse papillary ductular epithelial proliferation, often intermingled with adenomatous areas ( Fig. 16.25 ). The epithelium is two layered ( Fig. 16.26 ) with a variable degree of epithelial hyperplasia, typically of usual type. Normal ductular and lobular structures may be included within the lesion. A cellular stroma may accompany the proliferative process. At the periphery, a pseudoinfiltrative pattern of entrapped tubules may be seen and this, together with a tendency toward nuclear hyperchromasia, may lead the unwary into a mistaken diagnosis of malignancy. Because of its heterogeneous growth pattern, some authors have classified adenoma of the nipple into four categories or patterns according to the dominant structural feature: adenosis pattern, papillomatosis pattern, sclerosing papillomatosis pattern, and mixed proliferative pattern.

Differential Diagnosis

Adenoma of the nipple must be distinguished from the rare entity of syringomatous adenoma of nipple. In this lesion an infiltrate of tubular structures is present among lactiferous ducts, in contrast to the intraductal papillary proliferation of nipple adenoma ( Fig. 16.27 ). As indicated earlier, the pseudoinfiltrative pattern in nipple adenoma may raise the possibility of a malignant process. The regularity of the epithelial structures throughout the lesion and absence of the characteristic features of DCIS favor a benign process, as does the presence of a two-layered epithelium. Careful sampling is important as, although the risk is low, there can be associated carcinoma.

Macroscopic Appearances

Grossly most ductal adenomas appear as well defined, pale brown or gray nodules measuring up to 3 cm in diameter. In some cases a clear association with a duct lumen or a cystic structure can be identified.

Histologic Appearances

The microscopic appearances are varied, depending in large measure on the amount of associated fibrosis. Single or multiple adenomatous nodules are seen, composed of two-layered epithelium, luminal secretory cells, and basal myoepithelium ( Fig. 16.28 ). This can be confirmed by immunostaining for myoepithelial markers. In some cases the epithelial proliferation has a papillary configuration. The lesion is surrounded by dense fibrosis, which often contains fragmented elastic tissue, presumably derived from the original duct, parts of which may be discernible. A minority of duct adenomas have a central stellate sclerosing appearance. Apocrine change is frequently seen within the epithelial component, and this may exhibit nuclear atypia.

Differential Diagnosis

Ductal adenomas are benign lesions, and care must be taken not to overdiagnose malignancy. This is particularly the case if entrapped tubules are found and where apparent epithelial atypia exists. As noted earlier, considerable overlap exists between its histologic features and those of both sclerosing adenosis and CSL, and in some cases it may be impossible to make a firm distinction.

Macroscopic Appearances

There are no specific gross appearances associated with the columnar cell lesions, which are often present in conjunction with fibrocystic change. As these lesions as usually detected through mammographic breast screening programs, there is typically radiologic evidence of calcification in the specimen. Histologically this calcification is seen in secretions in the lumen of the lobular units bearing the columnar cell process.

Histologic Appearances

Columnar cell change arises in the TDLUs, and the low-power view demonstrates the lobulocentric nature of the process ( Fig. 16.29 ). Of note, cells with columnar cytomorphology line larger ducts and do not represent a columnar cell lesion. At higher power the columnar cell morphology of the epithelial cells can be identified lining the TDLU, and the structures are often mildly dilated and bear secretions that, as noted, may be calcified. The columnar cells are one or two cells in depth, are arranged perpendicular to the basement membrane, and are cytologically bland ( Fig. 16.30 ). Apical snouts are often present and can be noted as “blebs” on the luminal surface of the cells. The nuclei are uniform, bland, and oval or elongated in shape. Mitotic figures are very rare.

Columnar cell hyperplasia is formed from cells with the same appearance as in columnar cell change, but the epithelial lining is more than two cells thick, sometimes with some crowding being seen. The architecture may include tufts or hummocks ( Fig. 16.31 ) but not bridges, punched-out luminal spaces, or micropapillary structures. There may be overlapping of nuclei, but the cytonuclear features are essentially the same as the nonhyperplastic form. Apical snouts may be more obvious with a somewhat hobnail appearance and intraluminal secretions, with microcalcification sometimes of psammomatous morphology. If true micropapillae, bridges, or a cribriform architecture is seen, then a diagnosis of ADH or DCIS is appropriate (depending on the extent of the lesion [see upcoming discussion]) rather than columnar cell hyperplasia.

Columnar cell change and columnar cell hyperplasia may, rarely, show more overt, classic cytologic atypia. Such lesions are also known as flat epithelial atypia. This may be identified in either columnar cell change or in columnar cell hyperplasia.

In flat epithelial atypia, the TDLUs are, on low power, often darker than typically seen in columnar cell change. The lobules are often mildly to moderately dilated, as in the less cytologically worrisome processes. However, the internal luminal shape of the acini in flat epithelial atypia is typically smooth and rounded, rather than irregular in outline as often seen in nonatypical forms (columnar cell change and columnar cell hyperplasia). The cells lining the TDLUs are formed from one or more layers of small, monotonous cuboidal cells akin to those seen in ADH/low-grade DCIS ( Fig. 16.32 ). The nuclei are regular, round, sometimes with one or more nucleoli and clumped chromatin. Mitotic figures are very uncommon. Much less commonly, a more columnar shape is maintained, and the cells remain perpendicularly arranged to the basement membrane but with oval nuclei with more classic cytologic atypia, such as more overt pleomorphism and an increase in nuclear size with occasional mitoses ( Fig. 16.33 ), but not amounting to high grade; this intermediate grade process is also classified as flat epithelial atypia.

If marked pleomorphism is present or there are features of high-grade atypia, then the lesion should not be classified as flat epithelial atypia but as flat high-grade DCIS (just as there is no entity of high-grade ADH). Similarly, if there is true architectural atypia in the form of more solid bridges or well-defined, true micropapillary structures, then the diagnosis of ADH/low-grade DCIS should be considered and not flat epithelial atypia ( Fig. 16.34 ).

Differential Diagnosis

Columnar cell lesions arise within TDLUs, and recognition of the topography of the process and the columnar cell morphology of the cells enables the correct diagnosis to be made in the majority of cases. Apocrine change may mimic columnar cell change and also arises in the TDLU. In problematic cases estrogen receptor assessment can be helpful; columnar cell lesions express luminal cytokeratins (CK18 and CK19), estrogen receptor, and progesterone receptor, while apocrine cells are typically estrogen receptor negative and express androgen receptor.

Usual Epithelial Hyperplasia, Proliferative Disease Without Atypia

Histologic Appearances.

Any increase in the number of epithelial cells and in number of epithelial layers above the myoepithelial cell is regarded as epithelial hyperplasia, but this is recognized as significant if the epithelium is four cells or more in depth (i.e., of moderate degree). The detailed histologic features are described later in differential diagnosis. Further classification is based on the degree and nature of the hyperplastic process.

The term mild usual epithelial hyperplasia (which has no significant attributable increased risk of breast cancer) is used to note an increase of no more than a four-cell depth extending up from the basement membrane. Moderate hyperplasia describes an increase of more than four cells thick in which there may be bridging of the luminal space ( Fig. 16.35 ). In florid hyperplasia the space is distended and may even be obliterated ( Fig. 16.36 ). The distinction between moderate and florid epithelial hyperplasia of usual type is relatively subjective; clearly this is a continuous spectrum of changes. However, pragmatically, there is little necessity to sharpen the distinction between the categories of moderate and florid usual epithelial hyperplasia as they are grouped together to form a single category for the purpose of risk assessment.

Epithelial hyperplasia can occur in the terminal ducts, peripheral small ducts, or the large ducts and may involve areas of columnar cell change, papillomas, or other benign conditions. The site and association with other processes have no known additional significance with regard to the risk of subsequent carcinoma.

Ductal hyperplasia of usual type can be distinguished from ADH and DCIS by its architectural low-power and microscopic high-power cytologic features. The population of cells is mixed rather than monomorphic and uniform. Luminal epithelial cells predominate in most cases, but forms expressing basal cytokeratins (e.g., CK 5/6 and CK14) are also present, and occasional lymphocytes may also be seen ( Fig. 16.37 ). The epithelial cell nuclei are in general ovoid and small but show variation is size and morphology (see Fig. 16.37 ); this, in itself, is helpful in distinguishing the process from a clonal proliferation such as ADH or low-grade DCIS. Similarly, as a hyperplastic process, the cells have a haphazard and irregular arrangement (see Fig. 16.36 ) and can be seen in swirling sheets, with irregular bridges that are often somewhat ragged in outline. A streaming character may be seen in sheets or when cells are arranged parallel to the line of papillary or bridge formations. The luminal spaces formed, or remaining peripherally, are irregular, not sharply defined and often slitlike (see Figs. 16.35 and 16.36 ). A particularly helpful diagnostic feature is the preservation of a peripheral (or partial) ring of narrow slitlike luminal spaces. It is especially helpful to examine the overall arrangement of cells within usual epithelial hyperplasia; this is typically uneven and varying in density from duct space to duct space and, indeed, within each duct space. Intraluminal necrosis and hemorrhage and surrounding stromal changes may occur but are rare and usually only confined to the most florid examples, particularly when the epithelial hyperplasia is seen within another underlying process such as a papilloma. Microcalcification can uncommonly occur in usual epithelial hyperplasia and does not help in distinguishing between this condition and atypical hyperplasia or DCIS.

One form, gynecomastoid type of usual epithelial hyperplasia, may cause diagnostic difficulties. This has a rudimentary micropapillary growth pattern and may mimic low-grade micropapillary DCIS, but as the name suggests, it resembles the changes seen in the male breast in gynecomastia. Small rudimentary micropapillary clusters of epithelial cells protrude into the lumen of the duct space, but they are not well-formed structures with thin stalk and more bulbous tip composed of evenly spaced, uniform, and regular epithelial cells, as in low-grade DCIS. The nuclei in gynecomastoid type hyperplasia are typically somewhat pyknotic and are often distributed around the periphery of the small micropapillary structures.

Atypical Ductal Hyperplasia—Proliferative Disease With Atypia

Histologic Appearances.

Page's broad criteria for ADH were modeled on low-grade DCIS of cribriform and micropapillary subtypes. Lesions with some, but not all, of the features of low-grade DCIS are classified as ADH. Microscopic low-power (architectural) and high-power (cytologic) features are both required to make the diagnosis ( Figs. 16.38 through 16.40 ).

The criteria Page used to identify DCIS, and on which ADH was modeled, include:

• 1.

  A uniform population of cells

• 2.

  Smooth geometric spaces between cells or micropapillary formations with even cellular placement

• 3.

  Hyperchromatic nuclei

Separation from DCIS requires strict adherence to criteria. Page and colleagues described how the full features of DCIS should be present in at least two duct spaces for diagnosis of low-grade DCIS. Any lesion falling short of this is classified as ADH. This definition provides some objectivity to distinguish these two clinically significant entities. Other groups have used a measure of size; Tavassoli and colleagues require definite features of DCIS extending over an area at least 2 mm in maximum dimension for diagnosis of DCIS, while smaller lesions are classified as ADH. In practical terms, these two descriptions represent essentially similar lesions.

We would advise caution when making the diagnosis of ADH; this is an uncommon entity and specifically should not be used for cases of usual epithelial hyperplasia showing occasional cytologic atypia. Additional useful rules of thumb are as follows:

• These lesions (ADH) are small and microfocal (at least <5 mm) and are localized.

• ADH should not be diagnosed unless the diagnosis of low-grade DCIS has been seriously considered, but the features are not sufficiently extensive for this diagnosis.

It must be remembered that ADH is not a cytologically high-grade lesion. The diagnostic term should be confined to the specific entity described earlier with known long-term risk—not simply focal cytologic atypia within another epithelial proliferative lesion. Very minor changes amounting to mild atypical hyperplasia appear to be irrelevant with regard to risk and are best disregarded when seen in a background of florid or moderate usual epithelial hyperplasia.

Differential Diagnosis.

ADH falls in the borderland or grey area of epithelial proliferative lesions of the breast. Some studies using standardized criteria have described a satisfactory interobserver concordance among groups of expert pathologists, but reproducibility in separating atypical hyperplasia from low-grade DCIS in other studies appears poor.

Another useful rule for separating ADH from usual epithelial hyperplasia is to recognize that ADH is a clonal, uniform population of cells (see Fig. 16.39 ) in contrast to usual epithelial hyperplasia where the cells are polyclonal and vary in size and shape. Aside from the regularity in cytologic shape in ADH, the architecture is also structured rather than the haphazard arrangement seen in usual epithelial hyperplasia, with evenly spaced cells forming distinctive cribriform spaces or micropapillary structures (see Fig. 16.40 ).

The clonality of ADH (and low-grade DCIS) can be demonstrated by the uniform, homogeneous lack of expression of basal cytokeratin markers such as CK5, CK5/6, and CK14 in the epithelial cells of the atypical proliferation with uniform positivity with luminal cytokeratin markers such as CK19. Conversely, in usual epithelial hyperplasia a mixed, heterogeneous expression of luminal and basal markers is seen, confirming the lack of clonality of the proliferation. Similarly, estrogen receptor can sometimes be helpful diagnostically in difficult intraductal epithelial proliferations; ADH and low-grade DCIS are invariably uniformly strongly ER positive ( Fig. 16.41 ), while in usual epithelial hyperplasia the staining pattern is often variable in extent but particularly is typically heterogeneous in intensity. In only occasional lesions of usual epithelial hyperplasia are there contiguous groups of positive cells. It should be noted that immunohistochemical examination of intraductal epithelial proliferations with basal and luminal markers is not necessarily of assistance in distinguishing hyperplasia from higher grade clonal proliferations, which may express basal markers focally rather than uniformly, although low-grade lesions express luminal cytokeratins only.

The presence of mitoses, necrosis, or hemorrhage is not a useful discriminatory feature in distinguishing ADH from usual epithelial hyperplasia, or indeed from low-grade DCIS; these features are uncommonly seen in any of these proliferations.

The distinction of ADH from low-grade DCIS may be difficult. The diagnosis of ADH is based on its similarity to low-grade DCIS on hematoxylin and eosin (H&E) stains, and both lesions are uniformly and strongly ER positive, express luminal cytokeratins, and are HER2, p53, and basal cytokeratin negative. The cytogenetic features of low-grade DCIS and ADH are also essentially similar. In essence, distinguishing the two processes is based on assessment of the extent of the process, both within the individual duct and the number of ducts that bear the proliferation in a pure form.

It can certainly be difficult to reach a definitive diagnosis when a more extensive uniform, monomorphic clonal epithelial proliferation partly involves numerous spaces with occasional micropapillae and bridges, but there is not complete involvement of any two spaces. In this setting (particularly if the process is more extensive [e.g., >10 mm]), a diagnosis of low-grade DCIS should be considered rather than classifying this as ADH.

In essence, we emphasize the need to apply Page's and/or others’ criteria very strictly and reiterate that ADH is a microfocal lesion formed from clonal, low-grade, luminal epithelial cells and should not be reported when there is focal cytologic atypia in a process which otherwise, on architectural and cytologic grounds, amounts to usual epithelial hyperplasia.

The cytologic, biologic, and genomic features of ADH and low-grade DCIS are essentially the same. If more established forms of the process are seen centrally in a surgical specimen but immediately adjacent ducts are only partially involved, this does not represent the microfocal lesion described earlier. In this setting we would not diagnose ADH but would interpret the entire lesion as one process and as low-grade DCIS, even if the peripheral spaces are not completely cytologically and architecturally abnormal.

Finally, if a monomorphic evenly spaced, low-grade epithelial proliferation involving less than two complete spaces (i.e., equivalent to ADH) is identified in a needle core biopsy, the adjacent breast tissue that has not been sampled may harbor in situ (or less commonly invasive) carcinoma. It is impossible to assess whether this represents part of a larger lesion because of the nature of the sampling undertaken. In this setting, we would argue that the term atypical intraductal epithelial proliferation (AIDEP) is more appropriate than classifying the process as ADH, to maintain the specific implications of subsequent cancer risk associated with the diagnosis of ADH. Indeed, approximately 40% to 50% of such diagnoses will reveal DCIS or invasive carcinoma in the subsequent diagnostic surgical specimen following core biopsy if smaller width core biopsy samples (e.g., 14 or 16 gauge) are examined initially. Such “upgrade” is less common with more extensive sampling, particularly with larger vacuum-assisted biopsy specimens (~10%–20% in most series), as those cases with low-grade DCIS can more often be identified preoperatively in the larger amounts of tissue received. However, in most studies, no group has been identified that does not have a significant risk of upgrade, and evaluation of further tissue, ideally by vacuum-assisted excision, or alternatively by diagnostic surgical excision is recommended.

A further, diagnostically difficult group of atypical hyperplastic lesions is recognized with apocrine morphology. These are discussed under the differential diagnosis of apocrine DCIS later. However, we currently restrict the use of the specific term ADH to nonapocrine lesions, as it is these that have been demonstrated to carry the increased risk of development of breast carcinoma in long-term follow-up series, requiring the presence of both architectural and cytologic characteristics of low-grade ductal carcinoma in situ.

Ductal Carcinoma In Situ