Nonproliferative lesions, as defined by Dupont and Page (3), include cysts, papillary apocrine change, epithelial-related calcifications, and mild hyperplasia of the usual type.

Cysts are fluid-filled, round-to-ovoid structures that vary in size from microscopic to grossly evident (Fig. 9-1). Gross cysts, as defined by Haagensen (9), are those which are sufficiently large to produce palpable masses. Cysts are derived from the terminal duct lobular unit. The epithelium usually consists of two layers: an inner (luminal) epithelial layer and an outer myoepithelial layer. In some cysts, the epithelium is markedly attenuated or absent; in others, the lining epithelium shows apocrine metaplasia, characterized by granular eosinophilic cytoplasm and apical cytoplasmic protrusions (“snouts”).

Papillary apocrine change is characterized by a proliferation of ductal epithelial cells in which all of the cells show apocrine features as described above. Epithelial-related calcifications are frequently observed in breast tissue and may be seen in normal ducts and lobules or in virtually any pathologic condition in the breast. It should be noted that calcifications may also be seen in the breast stroma as well as in blood vessel walls. Mild hyperplasia of the usual type is defined as an increase in the number of epithelial cells within a duct that is less than four epithelial cells in depth. In this type of hyperplasia, the epithelial cells do not cross the lumen of the involved space.

In the original study of Dupont and Page (3), 70% of the biopsies showed nonproliferative lesions. The risk of subsequent breast cancer among these patients was not increased, compared with that of women who have had no breast biopsy (relative risk [RR] 0.89), even in patients with a family history of breast cancer (in a mother, sister, or daughter). The only group of patients in the nonproliferative category with an increased risk of developing breast cancer was that with gross cysts plus a family history of breast cancer. The relative risk with gross cysts alone was 1.5, but was 3.0 in patients with gross cysts and a family history. It should be noted that, although Dupont and Page initially included fibroadenomas among the nonproliferative lesions, the results of a subsequent study by these investigators indicated a higher relative risk for breast cancer among patients with fibroadenoma than for patients with nonpro-liferative lesions (10). As a result, fibroadenomas are now included among the proliferative lesions without atypia (see the section on fibroadenomas).

Included within the group of proliferative lesions without atypia are usual ductal hyperplasia (11) (also known as moderate or florid hyperplasias of the usual type), intraductal papillomas, sclerosing adenosis, and radial scars (3). As noted above, fibroadenomas are now included in this category as well. Women who have had a benign breast biopsy showing proliferative lesions without atypia, as defined previously, have a mildly elevated breast cancer risk, approximately 1.5 to 2.0 times that of the reference population (intraductal papillomas, radial scars, and fibroadenomas are discussed elsewhere in this chapter).

Usual ductal hyperplasias are intraductal epithelial proliferations more than four epithelial cells in depth. They are characterized by a tendency to bridge and often distend the involved space. The proliferation may have a solid, fenestrated or papillary architecture. If spaces remain within the duct lumen, they are irregular and variable in shape. These spaces are often slit-like and arranged around the periphery of the proliferation, with their long axes parallel to the basement membrane. The cells comprising this type of proliferation are cytologically benign and variable in size, shape, and orientation, and they often are arranged in a “swirling” pattern (Fig. 9-2). It is sometimes possible to discern multiple distinct cell populations, including epithelial cells, metaplastic apocrine cells, and myoepithelial cells (11).

Sclerosing adenosis is usually an incidental finding, but may present as a mammographic abnormality (microcal-cifications, distorted architecture) or a mass lesion (also known as nodular adenosis or adenosis tumor). This lesion is composed of distorted epithelial, myoepithelial, and sclerotic stromal elements arising in association with the terminal duct lobular unit. This lobulocentric pattern is key to the correct diagnosis of sclerosing adenosis and its variants, and is best appreciated at low power microscopic examination (Fig. 9-3). The epithelium in sclerosing adenosis may undergo apocrine metaplasia, and is then referred to as apocrine adenosis. The apocrine metaplastic cells may show cytologic atypia, raising the differential diagnosis of invasive carcinoma if the lesion is examined at high microscopic power without accounting for the lobulocentric architecture appreciated at low power (12). Sclerosing adenosis may also be involved by atypical lobular hyperplasia, lobular carcinoma in situ, atypical ductal hyperplasia, or ductal carcinoma in situ (DCIS). Perineural “pseudoinvasion” may be present in approximately 2% of sclerosing adenosis cases and should not be confused with invasive carcinoma. Because of the distorted glandular pattern of sclerosing adenosis, this lesion may be confused with a lowgrade invasive carcinoma, particularly tubular carcinoma. In contrast to the lobulocentric pattern of sclerosing adenosis, tubular carcinoma is infiltrative in nature, however, and does not conform to the normal breast ductal and lobular microanatomy. Although sclerosing adenosis is composed of distorted, elongated, or obliterated glands and tubules, tubular carcinoma is composed of angulated tubules with open lumens. The stroma of sclerosing adenosis is fibrotic or sclerotic compared with the desmoplastic stroma of invasive carcinoma. Importantly, as opposed to tubular carcinoma, sclerosing adenosis contains myoepithelial cells, which may be highlighted by immunohistochemistry.

Atypical hyperplasias have been defined as proliferative lesions of the breast that possess some, but not all, of the features of carcinoma in situ and are classified as either ductal
or lobular type (3, 7). Atypical ductal hyperplasias (ADH) are lesions that have some of the architectural and cytologic features of low-grade DCIS, such as nuclear monomorphism, regular cell placement, and round regular spaces, in at least part of the involved space. The cells may form tufts, micropapillations, arcades, bridges, solid, and cribriform patterns (11). A second cell population with features similar to those seen in usual ductal hyperplasia is also typically present (Fig. 9-4).

Atypical lobular hyperplasia (ALH) is composed of cells identical to those found in lobular carcinoma in situ (LCIS). These cells are monomorphic, evenly spaced, and dyshesive, with round or oval, usually eccentric nuclei and pale cytoplasm often with intracytoplasmic vacuoles (Fig. 9-5). Although criteria for the distinction between ALH and LCIS differ among experts, we utilize the criteria proposed by Page and Anderson (13) and diagnose ALH when the characteristic cells are present but less than one-half of the acini of a lobular unit are filled, distorted, or distended. In addition to involving lobular units, the cells of atypical lobular hyperplasia may also involve ducts (14).

It is important to note that with the increasing use of mammographic screening, atypical hyperplasias are being diagnosed more frequently than in the past. For example, when a biopsy is performed because of a palpable mass, atypical hyperplasia is seen in only about 2% to 4% of cases (3). In contrast, atypical hyperplasia was identified in 12% to 17% of biopsies performed because of the presence of mam-mographic microcalcifications (15).

Women who have had a benign breast biopsy that demonstrates atypical hyperplasia are at a substantially increased risk for developing breast cancer, approximately 3.5 to 5.0 times that of the reference population. Some studies have suggested that the risk associated with ALH is greater than that associated with ADH (2, 7), but others have
not (16) (Table 9-3); at the present time this issue remains unresolved. Patients whose biopsies showed ALH involving both lobules and ducts had a higher relative risk of developing cancer (RR 6.8) than those with either ALH alone (RR 4.3) or those with only ductal involvement by atypical lobular hyperplasia (RR 2.1) (14).

Lesions of the breast characterized by enlarged terminal duct lobular units lined by columnar epithelial cells are being encountered increasingly in breast biopsies performed because of mammographic microcalcifications. Some of these lesions feature banal columnar cells in either a single layer (columnar cell change) or showing stratification and tufting, but without complex architectural patterns (columnar cell hyperplasia). In other columnar cell lesions, the lining cells exhibit cytologic atypia, most commonly of the low-grade, monomorphic type. Such lesions were included among lesions originally categorized by Azzopardi as “clinging carcinoma” (monomorphic type) (17), and were more recently included among lesions designated flat epithelial atypia (FEA) (11) (Fig. 9-6). The role of columnar cell lesions and, in particular, FEA in breast tumor progression is still emerging. FEA commonly coexists with well-developed examples of ADH, low-grade DCIS, and tubular carcinoma (11). These findings, in conjunction with the results of recent genetic studies (11), suggest that FEA is a neoplastic lesion that may represent either a precursor to, or the earliest morphologic manifestation of, DCIS. The few available clinical outcome studies suggest, however, that the risk of progression of FEA to invasive cancer is extremely low, lower even than that associated with ADH or ALH, supporting the notion that categorizing such lesions as clinging carcinoma and managing them as if they were fully developed DCIS will result in overtreatment of many patients (11). Additional studies are needed to better understand the biological nature and the level of subsequent breast cancer risk associated with these lesions.

A number of factors appear to modify the breast cancer risk associated with biopsy-proven benign breast disease, including a family history of breast cancer, time since biopsy, menopausal status, and the appearance of the background breast tissue.