Imaging Studies
MRI or CT scanning may be used to assess tumor extent and map out local invasion into nerves, muscle, cartilage, or bone. [28, 29, 37, 39, 49] Lesions on MRI tend to have a low T2-signal intensity, given its low cellularity and fibrosing nature. [49] Imaging may also be used to help define metastatic disease, if suspected. However, imaging is not necessary in most cases of microcystic adnexal carcinoma (MAC).
Other Tests
Dermoscopy can be used but known specific features are limited and are rarely described in the literature. Some notable dermatoscopic features include dense, white, structureless areas centrally with white clods of variable diameter peripherally and a brown-pigmented outer rim; white clods may represent keratin retention cysts also common to other mimicking tumors, such as trichoepithelioma. [50, 51, 52]
Optical coherence tomography is a biomedical tissue-imaging technique used to obtain high-resolution subsurface images similar to a low-power microscope. The process is similar to confocal microscopy. In one instance, this technique was used to diagnose microcystic adnexal carcinoma (MAC) that was subsequently confirmed by histologic evaluation. [53] Similarly, reflectance confocal microscopy can be used in defining MAC in vivo. [54] The availability of these devices is limited, but their potential use as a noninvasive means of diagnosing subtle tumors and outlining ill-defined malignancies (eg, MAC) could prove to be an invaluable tool, particularly on cosmetically sensitive areas.
Procedures
A deep incisional or excisional biopsy is required. Superficial biopsies lead to misdiagnosis because deep extension and perineural invasion are key features of the tumor. Fine-needle aspiration has been attempted, [55] but its use is not standardized.
Ensuring that the biopsy specimen is adequate for diagnosis and that the histopathologist is aware of this entity is important because as many as 30-85% of cases of microcystic adnexal carcinoma (MAC) are misdiagnosed. [40, 56]
Histologic Findings
Note the images below.
A low-power view of microcystic adnexal carcinoma demonstrates superficially located keratocysts and variably sized tumor nests and ducts. Note the diminution in size of the nests and cysts with the depth of dermal invasion. Courtesy of Dirk M. Elston, MD
Small ductular structures lined by 2-3 cell layers of small eosinophilic cells showing little pleomorphism set in a dense fibrous stroma. Courtesy of Dirk M. Elston, MD.
A high-power view of small, irregularly shaped nests and strands of small tumor cells without obvious ductal formation. Courtesy of Dirk M. Elston, MD.
Hematoxylin and eosin
Hematoxylin and eosin of deep biopsy specimens (preferably to subcutaneous fat) still remain the criterion standard in diagnosing microcystic adnexal carcinoma (MAC). Overall, MAC is a poorly circumscribed, deeply infiltrative, asymmetric tumor composed of variable proportions of keratocysts, squamoid or basaloid nests, infiltrating cords, and ductular structures set in a variably hyalinized but usually paucicellular desmoplastic stroma. In a minority of cases, MAC shows attachment to the epidermis or the follicles, but generally, a striking zonal separation is noted between the tumor and the epidermis.
In the superficial aspects of the tumor, small-to-medium keratinizing cysts lined by squamous epithelium can be seen. Atypia and mitoses are sparse. Nests of small basaloid or squamoid cells, sometimes showing a whorled-appearing infiltrate between the cysts with depth, are present. These may be solid, show central keratinization, can rarely be calcified or demonstrate central lumen formation, and have a tadpole or "paisley-tie" appearance.
In deeper sections, variable numbers of small ducts lined by 1-2 layers of cuboidal cells are present throughout the tumor. Narrow cords of cuboidal cells usually characterize the deepest portion of the tumor. Some of the nests and ducts show tail-like cellular extensions reminiscent of syringoma. Glycogen-rich, clear-cell change, foci of decapitation secretion, and sebaceous cell or duct differentiation may be present. The nuclei of all the components are mainly small, slightly irregular, and hyperchromatic, but little pleomorphism or mitotic activity occurs. Perineural infiltration is common (80%). [57] The presence of eosinophils, particularly with the presence of lesional clear-cell change, can be seen. [58, 59] Lymphoid aggregates are common, particularly at the dermal-subcutaneous junction.
With regard to histologic subtype, tumors displaying the architectural features of MAC but with increased nuclear pleomorphism, hyperchromasia, vascular invasion, and necrosis may be suggestive of a more aggressive subtype. [42, 57] Clinically, these can demonstrate rapid growth, possible carcinosarcomatous metaplastic transformation with relapse, and possible metastases. [42] Careful examination of sections would also be needed to completely exclude a cutaneous squamous cell carcinoma with MAC-like differentiation in this setting, as this entity can be identical to MAC but has higher-grade features, and adequate sampling often reveals an in situ or keratinizing invasive squamous cell carcinoma component (see Histologic differential diagnosis below). [43, 60] Sebaceous [8, 59] and trichoepitheliomatous [61] differentiation has been reported, which is of no prognostic significance. Monophasic eccrine variants are also seen. Solid carcinoma, also known as syringomatous carcinoma, is considered a solid variant of MAC found mostly on the scalp of elderly men. [62] This variant tends to have large nodular cords and islands of infiltrative neoplastic cells that reveal the typical ductal differentiation and perineural invasion of MAC.
Toluidine blue
Special staining with toluidine blue has been shown to highlight infiltrating tumor strands with a distinctive pink halo and perineural invasion with a maroon tint. [63] This stain may be of benefit in Mohs micrographic surgery (MMS).
Immunohistochemistry
Definitive immunostaining to differentiate MAC from other sclerosing neoplasms, especially in superficial samples, is lacking. However, a few immunostains can be of some assistance.
p63, a member of the p53 gene family highly expressed in the basal cells and adnexa of human epithelial tissues, was compared with MAC, sclerosing basal cell carcinoma (SBCC), and desmoplastic trichoepithelioma (DTE) in one study. All tumors were stained with p63, but the nuclear staining pattern was more complete and diffuse in SBCC (20 of 20 cases) and DTE (10 of 10 cases). Individual MAC tumor islands had selective "scattered" positivity mostly confined to the periphery (5 of 5 cases) and even extinguished with tumor depth. The disadvantage is that the scattered pattern was most obvious in deeper portions of MAC and may not be of help in superficial biopsy specimens. [64]
p75 neurotrophin receptor has been found to play a role in hair follicle biology, and one study [65] extrapolated its use in differentiating morpheaform basal cell carcinoma (14% positive with < 30% cells staining) and desmoplastic trichoepitheliomas (100% positive with >80% cells staining). MAC was not used in the study, but it was used in another study in which 4 (44%) of 9 MACs stained for p75 neurotrophin receptor. [66] DTEs and infiltrative basal cell carcinoma stained 94% and 36%, respectively, in this particular study.
CK (cytokeratin)–15 was found in a study using a battery of immunomarkers (ie, CK15, CK7, CK20, CK903, carcinoembryonic antigen [CEA], CD10, CD15, BerEP4) to stain MAC (92% of 13 cases) and DTE (100% of 8 cases), but not infiltrative basal carcinomas (10 cases) or squamous cell carcinomas (8 cases, 5 with ductal differentiation). Although the other stains were not reliably distinguishing, CK15 may be of help to exclude MAC and DTE from the latter two entities. [67]
BerEP4, a basal cell marker, was reported to differentiate basal cell carcinoma from MAC, [68] but reappraisal found that 38% of MACs revealed positivity. [67]
CD5, a lymphoid tumor marker, was found to stain MACs 71% of the time, particularly in deeper portions of the tumor, which may help differentiate them from DTEs and morpheaform basal cell carcinomas. [69] This, again, may not facilitate diagnosis in superficial samples.
CK20 staining for the presence of Merkel cells in the tumor was negative in one case report. [68] Its utility was suggested to help differentiate MAC from other benign adnexal tumors that typically carry scattered Merkel cells throughout. Some studies appear to support this concept, [67, 70] but the variable presence of Merkel cells within sections makes its use less predictable. [65]
SOX-10 may also be of future diagnostic assistance. In one study, staining was seen in the majority of sweat gland tumors with the exception of syringoma and MAC. [52]
Other stains of interest but of little utility in diagnostic distinction include pancytokeratin, AE1/AE3, and cytokeratin-1; these all positively highlight tumor cells. Reactivity to hard keratin subclasses AE13 and AE14 has been demonstrated and suggests pilar differentiation. [71] Epithelial membrane antigen and CEA highlight ductal structures or intracytoplasmic lumen formation and suggest a sudoriferous origin, but ductal differentiation can be seen in other mimicking tumors. At times, MAC can also be negative for CEA. [67] Additionally, Leu-M1 is positive and S-100 is negative. [71]
A suggested immunohistochemistry panel may include the following:
-
Superficial biopsies: p63 may be of assistance if the peripheral pattern is seen, but this is best visualized in deeper portions of the dermis. CK15 may help exclude other malignancies, such as basal cell and squamous cell carcinomas, but DTE remains in the differential. Presence of tumoral CK20 cells may support a benign process, but lack of CK20 cells may not exclude a malignant process. Multiple step sections should be considered.
-
Deep (midreticular dermis or deeper) biopsies: p63 with a scattered peripheral pattern within in each tumor island, or extinction with depth, may help confirm a diagnosis of MAC. CK15 and CK20 can be used, as noted above. CD5 could also be used to verify MAC but may be less reliable.
Histologic differential diagnosis
With hematoxylin and eosin staining, attention to the infiltrative growth pattern and, if present, perineural invasion, usually easily discriminates MAC from the other tumors in suitable biopsy specimens. Rendering a definitive diagnosis from a shave or superficial punch biopsy specimen is often impossible. In this context, duct formation, if present, would favor MAC over DTE, but perineural invasion can be seen in both. DTE tends to be more circumscribed and contain stromal calcification and rudimentary papillary mesenchymal bodies, but an eosinophilic infiltrate, particularly with lesional clear-cell change, has been postulated to favor MAC. [58, 59] In the case of syringoma, the presence of keratocysts, slight atypia, and few mitoses would favor MAC.
Distinction from a morpheaform basal cell carcinoma can be made by the demonstration of duct and intracytoplasmic lumen formation and with zonation of the tumor from the epidermis, which is typically seen with MAC. Ductal formation is rare in basal cell carcinoma. If seen, it is usually in nodular subtypes. Subtle features, if seen, consistent with basal cell carcinoma (retraction artifact, fibromyxoid stroma, necrosis en mass) in basaloid islands may be supportive. CD5 positivity, CK15 positivity, or distinct p63 staining in deeper sections would also favor MAC over morpheaform basal cell carcinoma, whereas Ber-EP4 positivity favors basal cell carcinoma.
Desmoplastic squamous cell carcinoma or adenosquamous carcinoma generally lacks zonation from the epidermis, ductal structures, or intracytoplasmic lumen formation. CK15 positivity would favor MAC.
Cutaneous squamous cell carcinoma with MAC-like differentiation, also known as MAC-like squamous cell carcinoma, has identical features to MAC, including ductal differentiation with a similar immunostaining pattern. [43, 60] Distinction is made by the presence of high-grade atypia with frequent mitoses and identifying an in situ or invasive squamous cell carcinoma component. Adequate sampling and multiple step-sectioning are critical at finding these features for an accurate diagnosis.
Desmoplastic melanoma also typically lacks zonation, ductal structures, and keratocyst formation, and it is reliably S-100 positive. Lymphoid aggregates are usually situated throughout the tumor, as opposed to the dermal-subcutaneous junction, as is seen in MAC.
Trichoadenomas tend to be more well-defined tumors with larger keratin cysts surrounded by a fibrovascular stroma. Typically, they lack deep invasion and do not show perineural invasion.
Metastatic adenocarcinomas (most commonly breast and colon) demonstrate more pleomorphism and cytologic atypia, and they lack keratocyst formation. p63 positivity may help in differentiating a primary adnexal tumor over metastatic adenocarcinoma. [72] Furthermore, an immunoprofile of CK5, C14, CK17, and p63 positivity and Mammaglobin negativity may favor a primary adnexal carcinoma, such as MAC, over metastatic breast carcinoma to the skin. [73] Moreover, newly described microcystic squamous cell carcinoma of the lung has distinct features similar to MAC. Although none has metastasized to the skin, the potential is there and could be discerned from MAC via thyroid transcription factor 1 expression.
Unusual presentation of childhood onset diffuse atrophoderma vermiculata of the head and neck with histologic features of small keratocysts in the superficial dermis and sclerotic syringomatous proliferations in the mid dermis to subcutaneous fascia may mimic MAC. [48] The proliferations differed from MAC in that they lacked obvious perineural involvement, were diffuse across body surface areas, and did not progress over time. Ductal epithelial strands were also embedded in concentric fibrotic rings, and distinct ball-like papillary dermal elastin aggregates were present. These patients were thought to have possible syndromic relationship to Nicolau-Balus syndrome, Rombo syndrome, and the so-called multiple eccrine-pilar hamartoma syndrome.
Most authorities distinguish MAC from similar low-grade sweat gland carcinomas that have previously been termed syringomatous carcinoma (solid carcinoma), eccrine epithelioma, or group 2 sclerosing sweat duct carcinoma, largely based on the presence of biphasic differentiation evidenced by keratocysts or other signs of follicular differentiation typically seen in MAC. [3, 57] They may also lack the sclerosing stroma of MAC. However, some have suggested that these tumors could be variations of the same entity. Of note, some of these subtypes tend to occur in younger patients. [3, 71]
Likewise, squamoid eccrine ductal carcinoma, a sweat gland carcinoma consisting of stranded epithelium with ductal differentiation and prominent squamoid features, is considered by some authorities to be a monophasic eccrine version of MAC (lacking follicular differentiation), similar to eccrine epithelioma. Although one case report described follicular features in a squamoid eccrine ductal carcinoma, this tends to be minimal, if present, and differs from eccrine carcinoma (and MAC) owing to its conspicuous squamoid or syringosquamous metaplasialike qualities. [74]
-
Clinical photo of a microcystic adnexal carcinoma on the left upper lip of an elderly woman. Note the close clinical resemblance to basal cell carcinoma. Courtesy of Dirk M. Elston, MD.
-
A low-power view of microcystic adnexal carcinoma demonstrates superficially located keratocysts and variably sized tumor nests and ducts. Note the diminution in size of the nests and cysts with the depth of dermal invasion. Courtesy of Dirk M. Elston, MD
-
Small ductular structures lined by 2-3 cell layers of small eosinophilic cells showing little pleomorphism set in a dense fibrous stroma. Courtesy of Dirk M. Elston, MD.
-
A high-power view of small, irregularly shaped nests and strands of small tumor cells without obvious ductal formation. Courtesy of Dirk M. Elston, MD.
-
Note the subtle pink plaque on the left mid temple.
