Although nearly 95% of women with abnormalities on screening mammography do not have breast cancer, evaluation of their tissue has advanced scientific understanding of the spectrum of changes that occur in the breast and the risk associated with these changes (1). The relationship between breast lesions and cancer risk is complex, but certain principles have been established that can inform clinical care (2). Pathologic diagnosis from needle core biopsy samples has become the primary determinant of further intervention, and pathologists have made attempts to classify the spectrum of histologic findings into reproducible and relevant categories (3,4,5). Noninvasive changes in the glandular epithelium of the breast can be divided into four categories: nonproliferative, proliferative without atypia, proliferative with atypia, and carcinoma in situ. The validity of these categories of breast lesions is supported by the finding that there is an increased risk for developing breast cancer associated with proliferative changes. This risk is more than doubled when the proliferative changes include atypia and may persist for 10 years or more after biopsy (6,7,8,9,10).

Nonproliferative benign lesions are considered to be in the spectrum of normal and confer no increased risk of developing breast cancer. The nonproliferative category includes cysts, apocrine metaplasia, columnar cell change without atypia, fibroadenomas, nonsclerosing adenosis, and mild (usual) hyperplasia. These lesions may form palpable masses, may change in response to hormones, and can be associated with mammographic calcifications (11,12,13). Apocrine metaplasia and columnar cell change without atypia are benign alterations in the cuboidal epithelium lining the ducts and lobules. However, the designation of fibroadenomas, nonsclerosing adenosis, and mild hyperplasia as “nonproliferative” is not entirely accurate. Mild (usual) hyperplasia and nonsclerosing adenosis are physiologic proliferations of glandular epithelium and acini, respectively. The definition of hyperplasia for the glandular epithelium of the breast is the presence of more than the normal two cell layers, myoepithelial and epithelial, above the basement membrane. Hyperplasia is further characterized as “mild” or “usual” if there is no bridging or distention of the intraductal space, and these changes are considered nonproliferative (3). Fibroadenoma is a benign proliferation of the hormonally responsive intralobular stroma. However, the proliferation in these benign lesions is not due to loss of normal function or clonal proliferation of the breast epithelium. The literature is challenging to summarize, given that each investigator can include a variety of diagnoses in the overarching category of nonproliferative lesions. On the whole these lesions are associated with minimal to no increased risk of developing breast carcinoma (14,15,16,17,18,19). An early series from Dupont and Page with 17 years follow-up reports a relative risk (RR) of 0.9 (95% confidence interval [CI] 0.6 to 1.3) for developing breast cancer compared to a healthy survey cohort (17). In 2005, Hartmann et al. published long-term follow-up results from women who received a diagnosis of benign breast disease at the Mayo Clinic between 1967 and 1991 (8). This included 9,087 cases followed for a median of 15 years. Women with nonproliferative lesions had a relative risk of 1.3 (95% CI 1.2 to 1.4) for developing breast cancer compared to the population at large (Table 6.1).

Proliferative breast lesions provide the first morphologic evidence of genetic alteration, such as increased estrogen receptor expression, but they lack the full spectrum of cellular changes found in malignancy (20,21). These lesions may be detected as irregular densities or calcifications on mammography (22). The type and extent of proliferation on a biopsy specimen are major factors used in assessing risk for developing breast cancer. This category of lesions includes moderate to florid usual ductal hyperplasia, columnar cell hyperplasia, small duct papillomas, sclerosing adenosis, and complex sclerosing lesions (radial scar) (3,8). Moderate to florid hyperplasia is considered proliferative and is defined as more than three or four cell layers, with bridging, distention, or filling of the ducts and lobules (23,24). Papillomas can be multiple, and consist of a proliferation of ductal epithelium around a fibrovascular core. Sclerosing adenosis and complex sclerosing lesions consist of various combinations of increased acini entrapped in dense stroma. The term “scar” in the case of “radial scar” refers to the appearance of the complex sclerosing lesion and does not indicate prior trauma or surgery. These proliferative lesions are often found in various combinations within the same biopsy specimen, and all of the proliferative lesions without atypia are associated with a similar relative risk (RR range 1.3 to 1.9) for developing breast cancer (3,6,8,9,10,17). For example in the Nurses’ Health Study (1976 to 1996) and the Nurses’ Health Study II (1989 to 1995) Collins et al. identified more than 1,000 cases of proliferative breast disease without atypia. These included 200 cases that
later developed breast cancer. The relative risk for developing breast cancer associated with proliferative changes without atypia is 1.5 (95% CI 1.2 to 2.0) (9). If cancer occurs later, it occurs ipsilateral to the benign breast disease 50.3% of the time. This supports the current concept of these lesions as markers for increased risk, as opposed to being precursor lesions.

Atypical lesions have proliferating, monomorphic cells and exhibit additional loss of normal cell function (25). These lesions are detected in approximately 10% of biopsies for microcalcifications or a palpable mass, and include atypical lobular hyperplasia (ALH), lobular carcinoma in situ (LCIS), atypical ductal hyperplasia (ADH), atypical columnar cell change, and flat epithelial atypia (22,26). Of note, the term “lobular neoplasia” (LN) encompasses the spectrum of atypical proliferations originating in the lobule, from ALH to LCIS, which can be difficult to distinguish (27). The term “flat epithelial atypia” was introduced by the WHO Working Group on the Pathology and Genetics of Tumours of the Breast in 2003 in an attempt to consolidate diagnostic terms for epithelial atypia that lacks the architectural complexity of ADH or the features of columnar cell change (25,28). Atypical hyperplasias (AHs) are lesions that fall short of the quantitative and/or qualitative histologic criteria for designation as carcinoma in situ and often present a diagnostic dilemma, with interobserver agreement reaching only 45% to 53% (4,24,29,30,31). Atypical lesions in general are associated with a relative risk for developing breast cancer of 4.2 (range reported 3.3 to 5.4) (7,8,9). While ALH and ADH are generally considered equivalent in regard to risk of future breast cancer, recent evidence suggests that the risk associated with ALH is higher than the risk with ADH. In one report the relative risk for women with ALH is 5.8 versus 3.1 for ADH, and if the woman is premenopausal at the time of biopsy, the risk associated with ALH is 7.3, while with ADH it is 2.7 (9). ALH and ADH are associated with future ipsilateral invasive breast cancer slightly more than half of the time (61.3% and 55.9%, respectively) and still should be considered background markers of general increased risk (9,25).

Concerns over diagnostic accuracy have been addressed with dissemination of clearer guidelines; however, discordance in diagnosing these lesions persists (4,24,31,32). In a recent assessment of accuracy, pathologists from 22 institutions reviewed 2,004 needle core breast biopsies locally from samples obtained as part of a multicenter trial. The biopsies then underwent central review, and the local and central diagnoses were compared. The investigators found high interobserver agreement (99%) for benign lesions and invasive cancers (97%). The diagnosis of ductal carcinoma in situ (DCIS) was concordant in 83% of cases. However, lower levels of agreement were seen for atypical ductal hyperplasia (ADH) (63%) and lobular neoplasia (inclusive of atypical lobular hyperplasia and lobular carcinoma in situ) (53%) (31). The lack of uniformity in the pathologic diagnosis of atypia reflects the complexity of these lesions and our imperfect understanding of them. The limitations of morphology inspire attempts to identify molecular markers and profiles to distinguish these lesions for diagnostic purposes as well as ascertain whether AH are obligate precursor lesions or reflective of a background environment. Estrogen is a well-recognized breast cancer growth factor. Higher-than-expected estrogen-receptor (ER) levels are seen in AH (>60%) compared with normal breast epithelium (7%), and this finding is in keeping with the observation that the majority of DCIS and invasive carcinomas are ER-positive, suggesting a precursor relationship (33). A number of investigators have evaluated premalignant lesions for allelic imbalance or loss of heterozygosity (LOH). In one report on 17 cancer-containing specimens where ADH and cancer coexist, LOH was evaluated by polymerase chain reaction on microdissected samples. The ADH and cancer exhibited concordance, that is, LOH in the same allele, in 82% (9 of 11) of evaluable cases (34). However in another series of seven patients who had atypical hyperplastic lesions from 2 to 16 years preceding the diagnosis of invasive cancer, no cases showed concordance, suggesting that the path between AH and cancer is not a direct route (35). Clearly, further study is needed in this area. As scientific knowledge expands, the picture of atypia will become more clearly focused.

The effect of family history and menopausal status on benign breast disease and risk of breast cancer has been examined in a number of reports. In the retrospective cohort analyses by Page and Dupont, if a woman had a first-degree relative with breast cancer and proliferative disease with or without atypia, the risk of future cancer at least doubled, for example, from 3.2 to 9.7 for women with atypia (3,17). In the nested case control trial from the Breast Cancer Detection and Demonstration Project (BCDDP) a family history of breast cancer doubled the risk associated with proliferative disease without atypia (1.3 to 2.6) and with atypia (4.3 to 22) (10). It is important to note that the number of participants on which these risk estimates are based is small and the methodology for obtaining family history is not reported. In the largest and most recent series, Hartmann more carefully defined family history of breast cancer and found no interaction between benign breast disease and family history (8). In summary, there are conflicting data on the effect of a family history of breast cancer in women with proliferative breast disease, although a more recent analysis does not support an interaction.

In regard to age and risk, both the initial and subsequent analyses of the Nurses’ Health Study found that premenopausal status increases risk of breast cancer for women with atypical lobular hyperplasia but not ductal hyperplasia or proliferative lesions without atypia (6,9). In the Mayo cohort, women who were less than 45 years old at the time atypia was diagnosed had more than twice the risk of developing breast cancer compared to women who were over 55 years of age (8). In the BCDDP analysis Dupont et al. report similar observations by menopause status (10). ALH and ADH were not reviewed separately in the Mayo or BCDDP analysis. Younger age at diagnosis of atypical hyperplasia more consistently shows higher risk for developing future breast cancer (Table 6.2). The specific risk associated with ALH versus ADH needs clarification.