Encyclopedia of Pathology

Living Edition
| Editors: J.H.J.M. van Krieken

Atypical Ductal Hyperplasia

  • Werner BoeckerEmail author
Living reference work entry
DOI: https://doi.org/10.1007/978-3-319-28845-1_4740-1



According to the current WHO Classification of Tumors of the Breast, atypical ductal hyperplasia (ADH) is defined by a local proliferation of evenly spaced monotonous cells with an atypical architecture, characterized by smooth geometrical growth patterns as micropapillae, arcades, Roman bridges, bars crossing the glandular space, and finally full-blown lesions with solid or cribriform growth (Fig. 1). As discussed below, the definition of ADH has undergone remarkable changes. It is the great achievement of Azzopardi who shifted our view from usual ductal hyperplasia to ductal in situ microcancer as the decisive step in early breast cancer development.
Fig. 1

Atypical ductal hyperplasia. (a) Scanning view showing unfolded TDLU with atypical ductal hyperplasia (center) with two foci of columnar cell change (arrows) and lobular hyperplasia (asterisks); (bd) higher magnification of the lesions in a. (b) ADH with monotonous atypical cells with atypical architecture characterized by cribriform growth. (c) Columnar cell change with psammomatous type microcalcification. (d) Atypical lobular hyperplasia

In rare cases, atypical ductal proliferation (ADH/DCIS) may evolve from primarily benign proliferative breast lesions, or they may show apocrine features (see below).

Clinical Features

  • Incidence: The frequency of ADH diagnosis has multiplied with the advent of mammographic screening with figures of 12–17% of cases performed due to the presence of microcalcifications. In biopsies from mass lesions identified in screening programs, ADH is only found in 2–4% of specimens.

  • Age: Late 40s, screening programs.

  • Sex: Female

  • Site: Either breast, any quadrant, probably preference of upper outer quadrant.

  • Clinic, Treatment, and Core Biopsy: ADH, generally, is asymptomatic. Clustered microcalcifications (Fig. 2) are thought to be the most common findings on a screening mammogram that usually require minimal invasive biopsy (MIB) techniques such as vacuum-assisted core biopsy. However, due to the limited amount of tissue in core biopsies, ADH cannot be reliably diagnosed with these techniques because the atypical epithelial foci in these biopsies may form part of an established in situ neoplastic lesion, with or without associated invasion. For this reason, the European Working Group for Breast Screening Pathology (EWGBSP) recommended to abandon the term ADH in core biopsies in favor of “atypical epithelial proliferation, ductal type” (AEPDT) (Wells et al. 2006). In the five category reporting system of core needle biopsies, which was initially proposed by the UK National Coordinating Group for Breast Screening Pathology and later adopted and recommended by the EWGBSP, “atypical epithelial proliferation, ductal type,” is listed as B3 which includes lesions with uncertain malignant potential. This explains that in the literature “upstaging” of excisions after stereotactic breast core biopsies of patients with nonpalpable suspicious microcalcifications and/or lesions of AEPDT to DCIS or invasive carcinoma are found in 12–52%; likewise, using vacuum-assisted biopsy devices, which reveal more contiguous tissue than core biopsy needles, the reported rate of “upstaging” of AEPD to breast cancers ranges from 0% to 28%.

    Recently Weigel et al. (2011) analyzed the rate and the histological spectrum for malignancy of minimally invasive biopsies with “uncertain malignant potential (B3)” in digital mammography screening. One hundred forty-eight of 979 MIBs (15.1%) were categorized as B3. Among these B3 lesions, calcifications (61.5%) were the most important screening abnormalities. In agreement with several analog studies, the authors found AEPDT to be the most frequent entity (35%), followed by radial scars (28%) and papillary lesions (20%). Furthermore, with 40.4% AEPDT was the subtype with the highest rate of finally detected malignancy (DCIS n = 16, invasive cancer n = 3) in surgical excision biopsies (19/47). The lesion-specific positive predictive value (PPV) was highest for AEPDT compared to other B3-lesions with a PPV of 0.40, similar to the respective PPVs of 0.32 to 0.59 reported from analog screening settings (see also Menes et al. 2017).

    The NHSBSP guidelines (2001) suggest that if a low- or intermediate-grade atypical epithelial proliferation is found in a vacuum-assisted biopsy that is lacking in extent or degree of ductal/lobular involvement to be classified as DCIS, the lesion should be categorized as B3, and an indication for a diagnostic biopsy to exclude a DCIS should be made. Nevertheless, the number of foci of atypical ductal proliferation and the number of cores involved may help in making a clinical decision about a surgical diagnostic excision. In the study of Sneige et al. (2003) on 61 cases of vacuum-assisted biopsy atypical ductal proliferations, the lesions were confined to an average of 1.5 large ducts or lobular units and were associated with microcalcifications in all cases. Surgical specimens showed ADH in 15 cases, no residual lesion in 24 cases, and ductal carcinoma in situ only in 3 cases. It was found that microcalcifications that contain atypical ductal proliferations in less than three lobules or ducts and/or that are removed completely do not reveal higher-risk lesions on excision; thus, it was concluded removal in such cases to be unnecessary. Similar results were published by Wagoner et al. (2009). In practice, we therefore modify the NHSBSP and EWGBSP guidelines for cases in which sufficient material is available and in which the mammogram shows only minor changes that have been removed by the procedure and finally where the atypical epithelial proliferation is confined to single terminal duct-lobular units (TDLUs) and spares the ducts. Follow-up of these patients with annual control mammography seems a reasonable management option. The same holds true for ADH in diagnostic biopsies. In contrast, if ducts or several TDLU areas are involved by atypical proliferations in core biopsies or in cases in which the mammogram shows more extended changes, which usually indicates grade 1 DCIS, we prefer further diagnostic and, if necessary, therapeutic procedures.

  • Outcome: Dupont and Page (1989) proposed the term ADH for lesions in which “some but not all the features of DCIS” are present and demonstrated that the subsequent general risk of invasive breast carcinoma for these lesions is about four to five times that of the general population.

Fig. 2

Mammographic microcalcification (a) and corresponding picture in histologic section (b)


ADH is not associated with any specific macroscopic features.


The presence of a focal proliferation (usually less than 3 mm) of atypical uniform ductal-type cells with rounded or oval slightly hyperchromatic nuclei and with an atypical architecture with even spacing of cells is now universally recognized as indicating ADH. In most cases, this atypical proliferation is confined to a TDLU. At low magnification, the lobular architecture with its moderately “unfolded” ductules or acini and the relationship to ducts is easily appreciated (Fig. 1). The atypical architecture includes micropapillae, rigid bars, Roman bridges, and cribriform patterns. The key features of ADH are shown in Fig. 3 and Table 1.
Fig. 3

Cellular algorithm of ductal-type lesions of the breast. Usual ductal hyperplasia (UDH) contains phenotypically the whole set of luminal cells contrasting with the robust K8/18+ phenotype of low-grade intraepithelial neoplasias (LG-IEN), including flat epithelial atypia, atypical ductal hyperplasia, low-grade DCIS, and lobular neoplasia. (From Boecker et al. 2017)

Table 1

Key features of atypical ductal hyperplasia


Monomorphic epithelial proliferation with regular placement of cells and atypical architecture with smooth geometrical growth patterns such as micropapillae, arcades, Roman bridges, bars crossing the glandular space, and in full-blown lesions with cribriform growth

The cells contain eosinophilic to pale cytoplasm and round to oval monotonous nuclei

Confined to individual TDLU(s), lacking the ductal segmental spread of low-grade DCIS

Often associated with lamellar microcalcifications

Demarcation of ADH at the Lower and Upper Limits

At the lower limit, “flat epithelial atypia” (FEA) (Columnar Cell Lesions) and ADH are currently recognized as members of the low-grade carcinogenesis pathway. Any form of atypical architectural growth pattern (micropapillae, bars, Roman bridges, etc.) is regarded as the essential criterion for ADH and for the distinction between FEA and ADH.

At the upper limit, we are concerned with distinguishing ADH from low-grade DCIS (DCIS). Although some pathologists and even the current WHO classification incorporate an assessment of the size of a lesion in their analysis of ductal proliferations (two glandular spaces or aggregate diameter of 2 mm), the reliance on quantitative findings does not have a strong foundation. Rather, it should be emphasized that such size considerations do provide guidelines rather than precise biological definitions. We propose a pragmatically founded definition of ADH as a focal atypical ductal-type proliferation (usually of about 2–5 mm in diameter) involving TDLU structures, but with no, or only minor ductal involvement, in contrast with DCIS which is characterized by its spread into the ductal system in a segmental manner (Fig. 4).
Fig. 4

Growth patterns of ADH. High-power views. (a) Ductule highlighting the proliferation of the same cell type both in the periphery and within a single micropapilla. The tumor cells in the micropapilla, however, tend to be smaller with more darkly staining nuclei with slight variation in size and spacing. (b) Ductule with multiple micropapillae in a background of FEA. Same case as a. (c) Ductule with proliferation of a single cell type forming a rigid bar crossing the lumen in the background of FEA. The nuclei are round and hyperchromatic. (d) Ductule showing proliferation of cells forming irregular bars and an abortive cribriform pattern and micropapillae with psammomatous microcalcifications. (e) In this ductule, many rigid bars cross the lumen forming a cribriform growth pattern. Same case as c. (fg) Proliferation of a single cell type forming many arcades resembling the pillars supporting Roman bridges

It must be acknowledged that the extension of an otherwise classic low-grade DCIS into lobules (lobular cancerization) should be regarded as part of the DCIS and thus should be included into any size considerations of a DCIS. This suggestion is based on observations of several studies who found that DCIS with “ADH” at the periphery of the excision was associated with recurrences at the same location. In this context, the additional diagnosis of “ADH” may be misleading to the clinician and hence may be dangerous for the patient.

Rare Types of Atypical Ductal Proliferations

In rare cases, atypical ductal proliferation (ADH/DCIS) may evolve from primarily benign proliferative breast lesions, such as columnar cell change, papilloma (Papillary Lesions), mucocele-like lesions (Mucocele-like Lesions), radial scar (Radial Scar), sclerosing adenosis (Sclerosing Adenosis), and fibroadenoma (Fibroadenoma). In such instances, identifying the underlying benign lesion is important for an appropriate diagnosis. It should be emphasized that the same cytological and architectural criteria used for the identification of the benign background lesion are also applied in this context as discussed in the corresponding chapters. There are two main problems in this setting. These are related (1) to the extent of the atypical proliferation and (2) to the clinical impact of such findings. At the time of writing, there was no generally accepted solution to these problems. The author defines ADH in benign lesions as partial involvement of atypical ductal proliferation, having features of low-grade DCIS, in an otherwise typically benign lesion, most important, without DCIS outside the benign lesion.

Another rare morphological variation of this theme includes cases with apocrine type morphology. O’Malley and colleagues (2004) use cytological criteria (usual apocrine, borderline features, and “as in DCIS”) and extension criteria (<4 mm, 4–8 mm, >8 mm) to distinguish between benign, borderline, and malignant categories. According to Raju et al. (1993), the criteria of atypical apocrine hyperplasia are met whenever a lesion shows features of low-grade apocrine DCIS and is of small size. There is good reason to believe that the same fundamental extension criteria used to define classical atypical ductal proliferations (distinguishing such lesions from low-grade DCIS) also apply to lesions with low-grade atypical apocrine proliferations. Among the cytological hallmarks of low-grade apocrine lesions are a smaller cell size, cytoplasmic eosinophilia or pallor with less coarse granularity than in classic benign apocrine cells, and moderately stained regular nuclei with usually small nucleoli and a higher nucleus-to-cytoplasm ratio. In contrast, high-grade apocrine lesions are defined by pleomorphic, hyperchromatic nuclei, coarse chromatin, and prominent nucleoli.


Several groups demonstrated that the epithelial cells in the putative low-grade ductal pathway precursors, including classical ADH, were positive for glandular keratins CK7, CK19, and CK8/CK18 but negative for basal keratins CK5 (Fig. 5) and CK14 and myoepithelial markers. Strong E-cadherin membranous staining has been used to define ductal-type carcinomas, both in situ and invasive. Overexpression of HER2 is rare with only weak basal expression in ADH, in contrast to high expression and amplification rates in high-grade DCIS. The cellular constituents of ADH demonstrate increased expression rates with Ki-67 (usually <5%), ERα, the anti-apoptotic proto-oncogene bcl-2, and the cell cycle regulator cyclin D1, compared with adjacent normal breast lobular units. The myoepithelial cells stain robustly for p63, CK5/14, and myoepithelial markers such as smooth-muscle actin, calponin etc.
Fig. 5

ADH and low-grade DCIS. (a) ADH. Enlarged TDLU with distended ductules which contain a monomorphous cell population with a cribriform growth pattern with a normal duct on the left side. (b) Low-grade DCIS with cribriform growth pattern in ducts and distended lobules

Molecular Features and Tumorigenesis

The traditional tumorigenesis model of linear progression implies a cascade from benign ductal proliferations to invasive breast carcinoma. This concept resolves around usual ductal hyperplasia as the central and initial event in the tumorigenesis process and has at the same time its reflection in the notion of the ADH definition of Jensen and Page that “The lower boundary of ADH is defined by examples of florid hyperplasia and focal areas of cellular uniformity and even placement.” However, this concept is in contrast to the concept of J. Azzopardi, relating to the relationship between ductal microcancer and usual ductal hyperplasia, that “one of the more striking findings was the absence of any evidence of transition from areas of epitheliosis (usual ductal hyperplasia, the author) to cancer except perhaps in a single case.” As early as in 1987, Azzopardi, therefore, suggested a de novo evolution of breast cancer from glandular cells of the normal lobular breast epithelium. This view was supported by immunohistochemical studies of several groups demonstrating CK5/CK14 mosaicism in UDH lesions in contrast to robust negativity for basal keratins CK5 and CK14 and positivity for glandular keratins (CK7+ CK8/CK18+ CK19+) of ductal proliferations of the low-grade pathway. Recently, precision immunofluorescence studies showed that usual ductal hyperplasia represents a heterogeneous epithelial lesion of the luminal lineage, including CK5+/CK14+ progenitor cells. In contrast, low-grade intraepithelial neoplasia of the breast, based on their key characteristic of a robust CK18-positive phenotype, is thought to originate from the CK18+ cell population of normal breast epithelium. These findings support the view that usual ductal hyperplasia and low-grade intraepithelial neoplasia are different entities rather than a spectrum of the same disease (Table 2). As discussed above, the findings of “clinging carcinoma” of Azzopardi have introduced profound changes in the conception of early tumorigenesis of ductal breast carcinoma. Later on, several studies have demonstrated that FEA commonly coexists on the same microscopic slides as ADH, lobular neoplasia (LN), or low-grade DCIS and losses at 16q and 17p were the most frequent changes in ADH, similar to the other early lesions of the low-grade pathway (FEA and low-grade DCIS) and tubular and grade 1 invasive (no special type) breast carcinoma. In line with these observations, Aulmann et al. (2009) and colleagues examined the relationships between tubular carcinoma, low-grade DCIS, and FEA concerning the mitochondrial DNA mutation patterns and found that most cases of low-grade DCIS and FEA were directly related to tubular cancer with a possible precursor role. The current prevailing view is that FEA and ADH may represent an initial morphologic nonobligate precursor of the low-grade ER-dependent pathway to carcinoma. Concerning high-grade neoplasias, given the high number of individual genetic alterations including high-level amplifications in high-grade DCIS and their invasive counterparts, it seems unlikely that a relevant fraction of high-grade DCIS derive from these low-grade ductal lesions.

Differential Diagnosis

Concerning the decision-making of epithelial proliferations of ductal type is fraught with two main problems: (1) the lesion cannot be easily classified as either neoplastic or hyperplastic in nature, or (2) the distinction between FEA on the one hand and low-grade DCIS on the other hand may be a problem. Whenever a definitive diagnosis cannot be reached, this should be indicated, rather than prematurely diagnosing the lesion as ADH. Pathologists should refrain from using ADH as a blanket term for indecisive findings and should instead seek external consultation.

UDH is one of the differential diagnoses (Fig. 6). The key here is to recognize the heterogeneous cell proliferation of UDH contrasting with the monotonous single-type neoplastic cell proliferation in ADH. UDH shows a robust CK5/CK14 mosaicism, whereas ADH is strongly positive for CK8/CK18 and ER and negative for CK5/CK14. Importantly, Jain et al. (2011) found an 8% decrease in the number of lesions classified as atypical ductal hyperplasia in favor of usual hyperplasia using in an interobserver study a ADH-5 cocktail (cytokeratins (CK) 5, 14, 7, 18, and p63) by immunohistochemistry; in clinical practice, this could lead to a decrease in the number of surgeries carried out for intraductal proliferative lesions. One important problem remains: as intermediate-grade neoplastic intraepithelial proliferations, ductal type, may exhibit more variability in cytological and architectural characteristic than classical proliferations of the low-grade ductal pathway, the differentiation of these lesions from conventional ductal hyperplasia is one of the greatest challenges. In the authors’ experience, the typical CK5/CK14 staining pattern usually helps to solve the problem as intermediate malignant ductal-type proliferations usually fail to stain for keratins CK5 and CK14. Finally, micropapillary ductal growth patterns of epithelial proliferations may occasionally also pose problems in the differential diagnosis in which, again, CK5/CK14 immunohistochemistry may help to solve the problem.
Table 2

Comparison of histological features of UDH, ADH, and low-grade DCISa

Histological features



Low-nuclear-grade DCIS


Variable size but rarely extensive


More extensive, ductal-lobular involvement

Cellular composition

Mixed epithelial cells with variations in size and shape

Single-cell population with regular spacing

Identical to ADH

Myoepithelial cells usually around the ductal periphery

Myoepithelial cells around the ductal periphery


Fenestrating growth pattern with irregular lumina: Streaming pattern common with long axes of nuclei arranged parallel to the direction of cellular bridges, which often have a “tapering” appearance

Atypical with well-delineated geometric structures (micropapillae, bars, Roman bridges, arcades, and cribriform pattern) or solid growth

Identical to ADH


Irregular lumina, often ill-defined peripheral slit-like spaces are common and useful distinguishing features

Usually well-formed rounded spaces

Identical to ADH


Mosaicism of CK5- (CK14-) positive cells. Myoepithelial cells around the periphery

CK8-/CK18-positive cells, lacking CK5 and CK14. Residual normal CK5- and/or CK14-positive cells usually attenuated or in a luminal position

Identical to ADH

Focal ER + positivity

ER+ tumor cells

Nuclear spacing


Even, occasionally uneven

Identical to ADH



Single, small

Identical to ADH


Infrequent with no abnormal forms

Infrequent, abnormal forms rare

Identical to ADH




If present, confined to small particulate debris in luminal spaces

Major diagnostic features are shown in bold type

aModified version of the table in “European Guidelines for Quality Assurance in Mammography Screening – Fourth Edition,” Office for Official Publications of the European Communities

Fig. 6

Atypical ductal hyperplasia versus usual ductal hyperplasia. Hematoxylin-eosin stain of ADH showing monomorphous cell proliferation with cribriform growth pattern in (a) and CK5/CK14 immunostain with CK5/CK14-negative neoplastic cells contrasting with CK5/CK14-positive myoepithelial cells in the periphery in (b). Usual ductal hyperplasia showing the fenestrating growth pattern in HE routine section in (c) and strong immunostaining for CK5/CK14 in proliferating cells with the typical K5/K14+mosaicism (d)

Occasionally cells with more advanced cellular atypia may grow in just one or two layers. Currently, most experts in the field agree that flat epithelial lesions with high-nuclear-grade morphology should be termed high-grade DCIS irrespective of their size.

The most important distinguishing feature of ADH to lobular neoplasia is its cohesive growth; the latter, by its discohesive growth and typical change of cell shape from polygonal to round tumor cells, is usually easily recognizable on H&E sections. Molecularly the presence or loss of the adhesion molecule E-cadherin may be helpful in distinguishing between growth patterns of ductal and lobular type.

References and Further Reading

  1. Aulmann, S., Elsawaf, Z., Penzel, R., Schirmacher, P., & Sinn, H. P. (2009). Invasive tubular carcinoma of the breast frequently is clonally related to flat epithelial atypia and low-grade ductal carcinoma in situ. The American Journal of Surgical Pathology, 33, 1646–1653.CrossRefGoogle Scholar
  2. Azzopardi, J. (1979). Problems in breast pathology. London: W.B. Saunders.Google Scholar
  3. Boecker, W., Stenman, G., Schroeder, T., Schumacher, U., Loening, T., Stahnke, L., Löhnert, C., Siering, R. M., Kuper, A., Samoilova, V., Tiemann, M., Korsching, E., & Buchwalow, I. (2017). Multicolor immunofluorescence reveals that p63- and/or K5-positive progenitor cells contribute to normal breast epithelium and usual ductal hyperplasia but not to low grade intraepithelial neoplasia of the breast: New concepts on the cellular hierarchy using in-situ multicolour experiments. Virchows Archiv, 470, 493–504.CrossRefGoogle Scholar
  4. Dupont, W. D., & Page, D. L. (1989). Relative risk of breast cancer varies with time since diagnosis of atypical hyperplasia. Human Pathology, 20, 723–725.CrossRefGoogle Scholar
  5. Jain, R. K., Mehta, R., Dimitrov, R., Larsson, L. G., Musto, P. M., Hodges, K. B., Ulbright, T. M., Hattab, E. M., Agaram, N., Idrees, M. T., & Badve, S. (2011). Atypical ductal hyperplasia: Interobserver and intraobserver variability. Modern Pathology, 24(7), 917–923.CrossRefGoogle Scholar
  6. Lakhani, S. R., Ellis, I. O., Schnitt, S. J., Tan, P. H., & van de Vijver, M. J. (2012). WHO-classification of tumours of the breast. Lyon: IARC.Google Scholar
  7. Menes, T. S., Kerlikowske, K., Lange, J., Jaffer, S., Rosenberg, R., & Miglioretti, D. L. (2017). Subsequent breast cancer risk following diagnosis of atypical ductal hyperplasia on needle biopsy. JAMA Oncology, 3, 36–41.CrossRefGoogle Scholar
  8. Moinfar, F., Man, Y. G., Bratthauer, G. L., Ratschek, M., & Tavassoli, F. A. (2000). Genetic abnormalities in mammary ductal intraepithelial neoplasia-flat type (“clinging ductal carcinoma in situ”): A simulator of normal mammary epithelium. Cancer, 88, 2072–2081.CrossRefGoogle Scholar
  9. O’Malley, F. P., & Bane, A. L. (2004). The spectrum of apocrine lesions of the breast. Advances in Anatomic Pathology, 11, 1–9.CrossRefGoogle Scholar
  10. Page, D. L., & Rogers, L. W. (1992). Combined histologic and cytologic criteria for the diagnosis of mammary atypical ductal hyperplasia. Human Pathology, 23, 1095–1097.CrossRefGoogle Scholar
  11. Raju, U., Zarbo, R. J., Kubus, J., & Schultz, D. S. (1993). The histologic Spectrum of apocrine breast proliferations: A comparative study of morphology and DNA content by image analysis. Human Pathology, 24, 1973–1981.CrossRefGoogle Scholar
  12. Sneige, N., Lim, S. C., Whitman, G., Krishnamurthy, S., Sahin, A. A., Smith, T. L., & Stelling, C. B. (2003). Atypical ductal hyperplasia diagnosis by directional vacuum-assisted stereotactic biopsy of breast microcalcifications. Considerations for surgical excision. American Journal of Clinical Pathology, 119, 218–243.CrossRefGoogle Scholar
  13. Wagoner, M. J., Laronga, C., & Acs, G. (2009). Extent and histologic pattern of atypical ductal hyperplasia present on core needle biopsy specimens of the breast can predict ductal carcinoma in situ in subsequent excision. American Journal of Clinical Pathology, 131, 112–121.CrossRefGoogle Scholar
  14. Weigel, S., Decker, T., Korsching, E., Biesheuvel, C., Wöstmann, A., Böcker, W., Hungermann, D., Roterberg, K., Tio, J., & Heindel, W. (2011). Minimal invasive biopsy results of “uncertain malignant potential” in digital mammography screening: High prevalence but also high predictive value for malignancy. Fortschritte auf dem Gebiet der Röntgenstrahlen und der bildgebenden Verfahren, 183, 743–748.CrossRefGoogle Scholar
  15. Wells, C. A., Amendoeira, I., Apostolikas, N., et al. (2006). Quality assurance guidelines for pathology. In N. M. Perry, M. Broeders, & C. de Wolf (Eds.), European guidelines for quality assurance in breast cancer screening and diagnosis (p. 219). Luxembourg: Office for Official Publication of the European Communities.Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.Gerhard Domagk-Institute of PathologyUniversity of MünsterMünsterGermany