Apocrine (nodular) Hidradenoma and Adenomyoepithelioma Most Likely Do Not Form a Biological Continuum of Adnexal Neoplasia

To the Editor: I read with great interest the study by MacKinnon et al published in the July 2020 issue of your journal with the title “Apocrine Hidradenoma and Adenomyoepithelioma: Entities on a Biological Continuum of Adnexal Neoplasia.”1 In the study, the authors performed immunohistochemical studies in a cohort of 29 apocrine hidradenomas (AHs) with a panel of myoepithelial markers, including p63, S-100 protein, calponin, glial fibrillary acid protein, and alpha actin, and found myoepithelial differentiation in 5 cases of AH (17%). The myoepithelial cells were unifocal in some cases and multifocal in others, occupying a varying proportion of the tumor volume (1%, 15%, 25%, 30%, and 70%). This is an interesting histologic finding because previous studies have reported a lack of myoepithelial cells in AHs compared with other adnexal tumors, such as spiradenoma, hidradenoma papilliferum, and hidrocystoma.2,3 Based on their histologic and immunohistochemical findings, the authors seemed to suggest a chronological progression of AH to primary cutaneous adenomyoepithelioma (pcAME) with AHs with myoepithelial hyperplasia as a transitional step. Using the cutoff of 25% of the tumor volume composed of myoepithelial cells, the authors reclassified 3 of their cases as AHs associated with myoepithelial hyperplasia, whereas 2 cases harboring 30% and 70% myoepithelial cells were reclassified as an adenomyoepithelioma (AME). The authors concluded that “this finding illustrates a biological link between AH and AME” and proposed that “a myoepithelial component representing ≥25% of tumor volume serves as a minimum requirement for the diagnosis of AME”. In my opinions, this view is a little simplistic because the issues are definitely more complex as any neoplasm should be classified based primarily on morphology and supported or confirmed by immunohistochemical and/or molecular studies. Admittedly, it is difficult to characterize the morphology of pcAME because this tumor type is not a recognized category of adnexal tumors in the most recent WHO classification of skin tumors.4 Because AME and epithelial-myoepithelial carcinoma are more commonly encountered in the breast and salivary glands, their histologic description is better defined in the breast or head and neck literature. In the most recent WHO Classification of Breast Tumors, an AME is described as “a biphasic neoplasm (usually benign) characterized by small epithelium-lined spaces with inner luminal ductal cells and a proliferation of variably enlarged and clearly noticeable abluminal myoepithelial cells.”5 The counterpart of AME in the salivary glands is epithelial-myoepithelial carcinoma, which is defined as a “salivary gland tumor composed of a biphasic arrangement of inner luminal ductal cells and outer myoepithelial cells” in the last WHO Classification of Head and Neck Tumors.6 As defined by the WHO, the most characteristic histologic feature of AME is the arrangement of the cellular components, to wit, an inner layer of ductal epithelial cells surrounded by myoepithelial cells. Because the ductal epithelial cells are usually more eosinophilic/basophilic compared with the surrounding myoepithelial cells, which frequently display a clear cytoplasm, this histologic characteristic is easily recognizable under low-to-medium magnification. However, it should be kept in mind that it is at times difficult to identify the characteristic architectural arrangement of the 2 cell types in low-power magnification because both epithelial and myoepithelial cells can display cytologic variations including clear cell or oncocytic changes, and one of the 2 cellular components can represent the dominant cellular population in an AME. Nevertheless, if one carefully analyzes an AME, this typical arrangement is almost always present in the tumor, at least focally, which prompts the pathologist to consider the possibility of AME and therefore perform immunohistochemical studies to highlight the biphasic nature of the tumor. Hence, the intimate pattern of arrangement of the 2 cellular components with luminal epithelial and abluminal myoepithelial cells is almost as significant, if not more, as the presence of the 2 cell types in the diagnosis of an AME. The significance of the intimate architectural arrangement of the 2 cellular components in the diagnosis of an AME is illustrated when one compares an AME with cutaneous mixed tumors, another biphasic neoplasm of the skin. In mixed tumors of the skin, the typical adenomyoepitheliomatous arrangement of both cellular components is lacking, and therefore, the tumor is classified as a mixed tumor and not an AME, although both epithelial and myoepithelial cells are present in the tumor. Two cases of AME are presented in this article to illustrate the subtle histologic and immunohistochemical differences between AH and pcAME. Case 1 A 74-year-old male patient presented with a cystic mass on the scalp. The mass was a cystic and solid neoplasm (Fig. 1A) harboring the characteristic adenomyoepitheliomatous arrangement of eosinophilic epithelial cells forming ducts surrounded by spindle myoepithelial cells in the solid areas (Fig. 1B). The ducts were highlighted by Epithelial Membrane Antigen (EMA) (Fig. 1C), whereas the myoepithelial population was positive for p63 (Fig. 1D), S-100 protein (Fig. 1E), and SOX-10 (Fig. 1F). The case is an example of a primary cutaneous benign AME.FIGURE 1.: Example of a benign pcAME. The neoplasm is a cystic and solid mass (A, ×40). In the solid areas, the tumor displays the typical biphasic nature of a pcAME with an inner layer of ductal epithelial cells surrounded by spindle myoepithelial cells (arrow) (B, ×200). The ductal component is highlighted by EMA (C, ×200), whereas the myoepithelial cells are positive for p63 (D, ×200), S-100 protein (E, ×200), and SOX-10 (F, ×200).Case 2 A 39-year-old female patient presented with a 4.5-cm cystic and solid mass on the left flank. In one area, the tumor was composed of tubules lined by a low cuboidal eosinophilic epithelium and surrounded by a layer of myoepithelial cells with a clear cytoplasm (Fig. 2A). EMA highlighted the ductal components (Fig. 2C), whereas the myoepithelial cells were positive for p63 (Fig. 2D), S-100 (Fig. 2E), and SOX-10. This area displays the typical architecture of an AME. There was a transition into an area of myoepithelial overgrowth (Fig. 2B) with increased cytologic and mitotic atypia. EMA still highlighted the scattered ducts, and SOX-10 illustrated the myoepithelial overgrowth (Fig. 2B, inset). The immunostain p63 (not shown) displayed the same reactive pattern as SOX-10. This case has been previously published as a primary cutaneous malignant AME, epithelial-myoepithelial pattern,7 and is included to illustrate the adenomyoepitheliomatous growth and the typical immunohistochemical pattern, even in a malignant example.FIGURE 2.: Example of a malignant pcAME. The benign component of the tumor demonstrates the typical luminal eosinophilic ductal epithelial cells (vertical arrow) surrounded by myoepithelial cells with a clear cytoplasm (horizontal arrow) (A, ×200). There is an overgrowth of myoepithelial cells in the low-grade malignant component of the tumor with an epithelial-myoepithelioma carcinoma pattern, punctuated by scattered ducts (B, ×200). The tumor also harbors a high-grade component (not shown, Ref. 7). The ductal element is highlighted by EMA (C, F). The immunostains p63 (D, ×200), S-100 protein (E, ×200), and SOX-10 (B, inset, ×200) mark the myoepithelial cells.Compared with pcAME, an AH is frequently composed of sheets of squamoid or polygonal cells with a clear or eosinophilic cytoplasm (Fig. 3A). Ductal differentiation is usually not obvious in the solid areas. Nevertheless, a subset of AH displays focal ductal differentiation that could impart an adenomyoepitheliomatous appearance under low-to-medium–power magnification (Figs. 3B, C). However, under high-power view, the ducts are only lined by cuboidal or columnar secretory cells without a well-defined outer layer of myoepithelial cells (Fig. 3D). The ductal component just seems embedded in the background of polygonal cells. As in pcAME, the immunostain EMA highlights the ductal component (Figs. 4B, D) from the background of polygonal cells that are contrarily positive for p63 (Fig. 4E). In contrast to pcAME, myoepithelial markers such as SOX-10 (Fig. 4F), S-100 protein, and smooth muscle actin are negative in the polygonal cells of AH.FIGURE 3.: Solid area of an AH composed of squamoid and polygonal cells without any definitive ductal formation (A, ×200). A subset of AH demonstrates ductal differentiation imparting an “adenomyoepitheliomatous” appearance in a low-power view (B and C, ×100). In high-power magnification, the ducts are lined by a single layer of columnar cells without any definitive surrounding myoepithelial cells (arrows) (D).FIGURE 4.: Polygonal cells of an AH are heterogeneously positive for EMA (A and B, ×200). A high-power view of the adenomyoepitheliomatous area (B) shows the ductal component that seems to be embedded in the background of polygonal cells without a surrounding myoepithelial layer (C, ×400). Again, the immunostain EMA highlights the ductal component (D). The polygonal cells are positive for p63 (E, ×200) but completely negative for myoepithelial markers, including SOX-10 (F, ×200), S-100 protein, and smooth muscle actin (in Fig. 3B).Dermatopathologists are highly specialized pathologists in the field of dermatopathology and most only sign-out skin cases, at least in the United States. Therefore, most dermatopathologists are probably unfamiliar with AME/epithelial-myoepithelial carcinoma as these types of neoplasm are more commonly encountered and better defined in other fields of pathology. Based on the literature and the above illustrations, to qualify for the diagnosis of pcAME, an adnexal neoplasm must fulfill at least 2 main histologic criteria: The presence of 2 cell populations: epithelial and myoepithelial cells, which can be confirmed by immunohistochemical studies. The pathognomonic arrangement of the epithelial and myoepithelial cells: luminal epithelial cells forming ducts surrounded by abluminal myoepithelial cells. This growth pattern is usually identifiable in low-to-medium–power view and confirmed in high-power magnification. This is an important clue to recognize a potential pcAME. A minor criterion is the absence of a chondromyxoid stroma that is usually more suggestive of a mixed tumor. Applying the above histologic criteria to define a pcAME, it is difficult to conceptualize the pathway of how an AH progresses to a pcAME. After an analysis of the 5 cases of AH with myoepithelial differentiation in the MacKinnon et al cohort, 4 of their cases with myoepithelial differentiation displayed only solid focus/foci of spindle myoepithelial cells within the tumor without any “adenomyoepitheliomatous” growth pattern and therefore should be classified as an AH with myoepithelial hyperplasia/differentiation. On the other hand, their fifth case seemed to represent a bona fide example of pcAME. If AH and AME truly form a biological continuum of adnexal neoplasms, it is not easy to explain how a well-organized adenomyoepitheliomatous pattern could arise from random solid foci of myoepithelial cells in an AH. Our previous analysis of pcAME has revealed that “63% of the reported pcAME harbored a coexisting second type of adnexal tumors,” including spiradenoma, hidradenoma papilliferum, and hidrocystoma.7 We postulated that these common adnexal neoplasms probably represent the precursor of pcAME because all have a myoepithelial component. Hence, it is problematic to hypothesize an AH and pcAME in the same spectrum of adnexal neoplasms merely based on the identification of myoepithelial cells in both tumor types because the same logical analogy could be postulated with other adnexal neoplasms, such as spiradenoma, which is the most common adnexal tumor associated with a pcAME.7 Conceptually, it is, however, conceivable that an AH could transform into a pcEMA through the acquisition of new mutations, particularly oncogenic drivers that commonly play a role in the histogenesis of AME. Similar phenomena can be observed in other fields of pathology such as transformation of an adenocarcinoma to a small cell carcinoma8 or trophoblastic differentiation in many types of carcinoma.9,10 In such scenarios, the different tumor types merely represent a transformational progression from one tumor type to another, but they are still distinct tumor types and therefore do not belong to a biological continuum. Therefore, I agree with Saggini A and L. Held in a Letter to the Editor published in the October 2020 issue of the American Journal of Dermatopathology11 that a skin neoplasm should only be classified as a pcAME if the tumor harbors both epithelial and myoepithelial cells that are arranged in the pathognomonic “adenomyoepitheliomatous” growth, regardless of the percentage of the myoepithelial component (such as ≥ 25% as suggested by MacKinnon et al). MacKinnon et al study, however, emphasized the importance of immunohistochemical studies in the diagnosis of this histologically alike adnexal neoplasms and lent additional evidence for our previous statement regarding pcAME “If attention is not paid to the distinct dual ductal-myoepithelial differentiation of a pcAME and a battery of immunohistochemical studies was not performed to evaluate the presence of myoepithelial cells, a pcAME might be easily misdiagnosed as a clear cell hidradenoma by pathologists, which might partially account for the paucity of pcAME cases reported in the literature.”7 Saggini A and Held L in the Letter to the Editor published in the October 2020 issue of the American Journal of Dermatopathology also raised the role of molecular studies as an integral component in the differential diagnosis of AH and AME. Their literature review revealed CRTC1/CRTC 3-MAML2 fusion genes in approximately half of cutaneous AH. The authors also postulated that “In light of the stark differences between known oncogenic alterations in breast AMEs and in cutaneous (and breast) AHs, it appears at least possible that legitimate cases of cutaneous AME may show a genetic background more in keeping with breast AMEs than with cutaneous AHs.” In accordance with their hypothesis, our literature analysis showed that PIK3CA oncogenic mutations are a common molecular event in a primary AME of the breast.5,12 Interestingly, molecular studies in our previously published report of primary cutaneous malignant AME also identified a PIK3CA hotspot mutation in both the benign and malignant components of the neoplasm.7 To the best of my knowledge, this finding is the only molecular alteration in pcAME reported in the literature. The detection of the mutational occurrence of PIK3CA in both mammary and pcAME is not unexpected as the breast is considered by some authorities as modified cutaneous appendages.13,14 The author therefore agrees with Saggini A and Held L that “molecular pathology will play an increasingly significant supporting role in the diagnosis of difficult adnexal neoplasms, as it already occurs in breast and salivary gland pathology.” In summary, AH and AME are most likely distinctly different types of adnexal neoplasms and do not form a biologic continuum of adnexal neoplasms based on current morphologic, immunohistochemical, and molecular findings. In comparison with an AH, a pcAME displays divergent histologic features characterized by the presence of both epithelial and myoepithelial cells with a pathognomonic biphasic arrangement of the 2 cellular components, irrespective of the percentage of myoepithelial constituent. In the hope of facilitating the identification of pcAME, the author has proposed stringent histologic and immunohistochemical diagnostic criteria for a pcAME based on a review of the pathology literature and personal experience. In this molecular age, future molecular studies are undoubtedly warranted to better categorize this group of confusing adnexal neoplasms.