Abstract
Hidradenocarcinomas are rare, aggressive skin adnexal tumors of sweat gland origin that demonstrate a high potential for local recurrence, metastasis, and poor outcome. These neoplasms can derive from preexisting clear cell hidradenomas but more commonly appear de novo, with the molecular events responsible for the pathogenesis currently unknown. Historically, diagnosis has been difficult because of the few cases, inconsistent nomenclature, variable morphology of cells that compose the neoplasm, and confusion with other visceral metastatic tumors. Presentation is generally benign with an indolent clinical course that typically includes local and multiple recurrences. Despite wide-excision surgery, disease at regional lymph nodes and metastatic sites is common and linked to decreased survival. Currently, molecular markers of pathogenesis as well as effective forms of adjuvant chemotherapy are lacking. Future studies are required to identify the histopathologic and immunohistochemical features, which may facilitate diagnosis and foster development of molecularly targeted forms of adjuvant therapy.
It is generally accepted that the first case of a sweat gland carcinoma was reported as early as 1865 by the renowned French pathologist Victor Andre Cornil.1 Early attempts at classification were initiated during the 1940s when Gates et al1 reviewed the world literature and accumulated descriptions of 29 different sweat gland carcinomas. Later, in 1951, Stout and Cooley2 provided a more precise definition of such tumors when they reported 11 well-documented new cases, including 6 that had metastasized and 5 that showed only extensive local growth. They classified the tumors based on particular histologic features into groups that were either adenoid cystic, mucinous, or ductal carcinomas. In 1968, Berg and McDevitt3 presented a further defined and expanded classification when they presented the largest series of sweat gland carcinoma cases, which included 101 tumors. Several aspects related to hidradenocarcinomas remain undefined and require further study. These include the classification, diagnostic criteria, prognostic features, and biologic markers responsible for dysregulated cellular growth that leads to tumorigenesis. Specifically, a detailed understanding of the steps that contribute to the transformation of a benign hidradenoma to a malignant form is lacking. Also, an organized classification scheme is lacking, which can be attributed to the rarity of cases, the diverse histology features, and the resemblance of hidradenocarcinomas to other primary and metastatic carcinomas.
CLINICAL FEATURES
Hidradenocarcinomas are very uncommon, intradermal malignant tumors of sweat glands that present as asymptomatic, nondescript cutaneous lesions. Sweat gland carcinomas are, collectively, rare with a reported incidence of around 0.05% in the United States.4 The disease is primarily diagnosed in the fifth to seventh decade of life with similar incidence in men and women and without any racial prevalance.4 The tumor accounts for approximately 6% of malignant eccrine tumors, which are seen in 1 in 13 000 dermatopathology biopsies and reported in only 2 cases in 450 000 consecutive skin biopsy specimens.5 Generally, lesions present initially on the face or extremities. However, cases have also been reported with lesions described on the abdomen, trunk, and groin and even more unusual presentations have been reported on the scalp, elbow, and digits. Typically, the tumor appears as a “benign” solitary skin lesion and either maintains a stable size of 1 to 5 cm or may slowly expand circumferentially. In general, hidradenocarcinomas exhibit a general lack of distinguishable features and following excisional biopsy or local excision, the primary lesion may remain dormant or recur slowly in situ. The lesion may demonstrate local expansion for a highly variable amount of time, which may last decades. In general, most patients remain asymptomatic without any overt effect of the primary lesion other than pain, discomfort, bleeding upon physical contact, or ulceration. At some point, through an unknown mechanism, the tumor assumes an aggressive clinical course with growth at regional or distant metastatic sites, primarily lymph nodes. Even upon metastasis, with or without local or regional lymph node involvement, the patient may remain asymptomatic. The prognosis for survival with newly diagnosed nodular hidradenocarcinoma is generally poor and, notably, the 5-year disease-free survival rate is less than 30%.5 Tumors demonstrate up to 50% local recurrence, despite aggressive surgical management.5 In addition, metastases have been reported in more than 60% of patients within the first 2 years, commonly through regional lymph nodes.6 Extensive metastasis through hematogenous routes to the periesophageal, peribronchial, periaortic, and retroperitoneal lymph nodes; bones; vertebrae; ribs; pelvis; lung space; viscera; and pleura have been reported. Distant metastases that present at a regional lymph node, in general, confer a decreased survival rate. With further disseminated disease, patients may complain of symptoms related to organ-specific metastases.
When making a diagnosis of hidradenocarcinoma, the differential diagnosis includes several other tumors, such as lipomas, hemangiomas, lymphangiomas, squamous cell carcinomas, basal cell carcinomas, malignant melanoma, other malignant adnexal carcinomas, and metastatic tumors to the skin. In addition, many of these tumors may resemble primary carcinomas from the lung, salivary glands, and breast.
HISTOLOGIC AND PATHOLOGIC FEATURES
On gross pathology, hidradenocarcinomas usually present as well-circumscribed nodules on the superficial skin. However, definitive histologic features that distinguish hidradenocarcinomas from the benign counterpart or other related malignant tumors are lacking. Establishing histopathologic criteria of malignancy for these tumors can be difficult because nuclear anaplasia may be absent, slight, or moderate. In addition, microscopic distinction between the benign and malignant forms is usually made based on the extent of invasion, asymmetry at scanning magnification, and degree of nuclear atypia. Cases have been reported in which the malignant tumor still has portions with features of a clear cell hidradenoma. Occasionally, malignant cells may infiltrate the surrounding benign cells in an unorganized fashion and, in contrast to the benign forms, malignant clear cell hidradenocarcinomas tend to invade the surrounding tissue as asymmetric tumors that display infiltrative dermal proliferation of variable-sized lobules. At higher magnification, nests of cells may show cyst formation. Connection to the epidermis is rarely observed (Figure 1, A and B). The lesions are composed of a mixture of eosinophilic polygonal cells, squamous cells, mucinous cells, and clear cells (Figure 1, C). Pleomorphism and necrosis are also commonly seen and, generally, ducts and intracytoplasmic lumina are present, and glycogen is evident in the cytoplasm of some of the clear cells. Histologically, lesions are typically composed of atypical, uniform polygonal cells that are suspicious for malignancy of adnexal origin. Under high-power magnification, lesions demonstrate residual multinodularity with infiltrative growth, epithelial cells with clear cytoplasm, pleomorphism, atypia, and few mitoses. These features generally are consistent with the diagnosis of hidradenocarcinoma.
Other eccrine carcinoma subtypes bear a histologic resemblance to hidradenocarcinomas.7 These include adenoid cystic eccrine carcinomas, aggressive digital papillary adenocarcinomas, eccrine adenocarcinomas, and mucinous eccrine carcinomas. Importantly, primary eccrine cancers can present without the general features of benign adnexal tumors and are less-readily distinguished from hidradenocarcinomas. These related tumors may often be confused with visceral adenocarcinomas metastatic to the skin as well. Collectively, they represent important diagnostic considerations when an adenocarcinoma is encountered in the skin, in the absence of a known extracutaneous primary.
Certain characteristics may be useful in differentiating hidradenocarcinomas from related tumors.7 For example, adenoid cystic eccrine carcinomas most often are seen as tumors of major and minor salivary glands but rarely may be encountered as primary cutaneous tumors. This tumor exhibits a population of uniform basaloid cells forming cribriform and tubular structures, usually with evidence of mucin and of hyaline surround cellular masses (Figure 2, A). Aggressive digital papillary adenocarcinomas exhibit cystic zones that manifest papillary infoldings and are lined with benign cuboidal epithelium, more cellular zones of atypical adenomatous hyperplasia, and areas of overt adenocarcinoma (Figure 2, B).6,7 As shown, an aggressive digital papillary adenocarcinoma may exhibit cystic areas with true papillary formations of cuboidal epithelium. Although bland, the example displays several areas of vascular invasion. Eccrine adenocarcinomas generally resemble moderately to poorly differentiated adenocarcinomas, with regional variation ranging from true ductules to infiltrative, nonglandular anaplastic cells, to zones with glycogenated cells.6,7 Contiguity with benign eccrine structures or with overlying epidermis is not seen. Mucinous eccrine carcinoma is characterized histologically by solitary and nested anaplastic cells floating in pools of mucin within the dermis. Thin strands of fibrous tissue serve to compartmentalize these “lakes” of mucin. Perineural invasion is prevalent and should be sought specifically. Hidradenocarcinomas have been reported to be derived from preexisting clear cell hidradenomas but have also been reported to arise de novo. The pathogenetic events responsible for either the de novo generation of hidradenocarcinomas or the events responsible for the conversion of hidradenomas to the malignant hidradenocarcinoma form are presently unknown. A recent study by Takata et al8 examined the incidence of cytogenetic abnormalities in malignant eccrine tumors and demonstrated low incidences of loss of heterozygosity or TP53 alterations in a mixed group of these neoplasms, in contrast to the frequent and multiple genetic abnormalities seen in tumors arising from epidermal keratinocytes. The authors8 speculate that this difference may be partly explained by the bulk of a sweat gland lying deep in the dermis, where it is relatively protected from the sun and environmental mutagens. The precise role of ultraviolet radiation remains to be elucidated. Future studies are required to identify the precise molecular events responsible for the dysregulated cellular growth seen in hidradenocarcinomas that contributes to tumorigenesis.
IMMUNOHISTOCHEMICAL AND CYTOGENETIC FEATURES
A detailed immunohistochemical study of a series of hidradenocarcinomas was performed to clarify the process of diagnosing hidradenocarcinomas.9 Six cases of hidradenocarcinomas were evaluated histologically and immunochemically. No consistent pattern was observed when neoplasms were stained with antibodies to carcinoembryonic antigen, S100 protein, gross cystic disease fluid protein 15, epithelial membrane antigen, BCL1, or BCL2. All tumors stained positively for keratin AE1/3 and cytokeratin 5/6, whereas Ki-67 and p53 staining was strongly positive in 5 of the 6 cases (83%). The authors9 concluded that Ki-67 and p53 staining may be useful histologic parameters. In addition, it may not be necessary to segregate hidradenocarcinomas into eccrine and apocrine categories.
In a separate case report, a nodule with features of a clear cell hidradenocarcinoma was biopsied from the chest wall of a 44-year-old man, and the immunohistochemical features examined.10 Immunohistochemical analysis revealed reactivity for high-molecular weight keratins, CK5 and CK7; p53; p63; carcinoembryonic antigen; androgen receptor; estrogen receptor; epidermal growth factor receptor; MUC5AC; and strong, diffuse membranous staining for HER2/neu ERBB2. The sample was negative for staining with villin, TTF-1, CDX2, S100 protein, vimentin, gross cystic disease fluid protein 15, mammoglobulin, and MUC2. Interphase fluorescence in situ hybridization demonstrated chromosomal amplification of the HER2/neu locus within the tumor and nodal metastasis as well. This was the first demonstration of HER2/neu amplification in a malignant skin adnexal tumor. Clearly, further immunohistochemical analysis is necessary to better define the criteria for diagnosis as well as for use as prognostic factors.
The distinction between subtypes and even the designation of eccrine tumor may be difficult, if not impossible, in select cases based on light microscopy alone. In these instances, a stain of eccrine-type enzymes (eg, succinic dehydrogenase or amylophosphorylase) may be obtained. The presence of ferritin is also helpful in determining the eccrine origin of a tumor, and immunostains, such as carcinoembryonic antigen, epithelial membrane antigen, EKH5, and EKH6, may also be used.
Although mucoepidermoid carcinomas rarely show distant metastasis, and even less frequently involve the skin, this entity should be included in the differential diagnosis of mucin-producing neoplasms in the skin. The histologic differential diagnosis includes eccrine carcinoma, mucin-producing adenocarcinoma with squamous differentiation, primary cutaneous adenosquamous (mucoepidermoid) carcinoma, and malignant mixed tumors of the skin. Okabe et al11 analyzed 71 cases of mucoepidermoid cancer and 51 cases of nonmucoepidermoid cancer for the presence of MECT1-MAML2 fusion oncogene. They confirmed that MECT1-MAML2 expression is specific for mucoepidermoid cancer and that the fusion-positive tumors have histologic and prognostic features. The first clue toward understanding the etiologic basis for mucoepidermoid cancer was the identification of recurrent t(11;19) chromosomal translocation in a subset of tumors.12,13 The translocation in mucoepidermoid tumors created a novel fusion product that disrupts the Notch signaling pathway. As reviewed recently,14 these and related results have helped to initiate a preliminary understanding of mucoepidermoid cancer tumorigenesis and to raise important points for both the diagnosis and management of patients with mucoepidermoid tumors.
CURRENT AND FUTURE TREATMENT MODALITIES
Most clinicians have concluded that the treatment of choice for hidradenocarcinomas and apocrine-gland carcinomas in general involves wide, local excision with regional lymph node dissection and consideration of postoperative radiotherapy in patients with moderately or poorly differentiated tumors.15,16 Early diagnosis is, therefore, critical to patient quality-of-life and outcome. Radiation therapy has been used in select cases, and chemotherapy has not been used extensively nor has its effectiveness yet been demonstrated. Because these tumors can show a high rate of local recurrence and may metastasize to lymph nodes, bone, or visceral organs, wide surgical excision should be performed rapidly after initial diagnosis. A recent study17 examined the benefits of sentinel lymph node biopsy and reported that the biopsy may detect subclinical metastases from sweat gland carcinomas to regional lymph nodes. Sentinel lymph node mapping and biopsy at the time of resection may, therefore, provide useful information with which to guide early treatment, such as wide surgical excision and lymph node dissection. Discussion regarding lymph node dissection has remained controversial because the long-term benefit has not been clearly demonstrated. However, it has been reported that surgical removal of regional lymph nodes is recommended therapeutically based upon the high incidence of metastases.17 Further studies are necessary to determine whether this procedure results in a survival benefit or local control in patients with sweat gland carcinomas.
Case reports have described transient benefit to various chemotherapy agents and, similarly, the efficiency of adjuvant chemotherapy alone or in combination with radiation treatment has not been demonstrated.18–20 Because of the reported benefit of capecitabine in breast and colon cancer and their possible similarities in cellular targets, this agent was used in a 60-year-old woman with metastatic hidradenocarcinoma.18 Treatment resulted in greater than 50% clinical remission with acceptable clinical and biologic tolerance. Such a limited objective response in a single patient requires further study. Only 2 other cases have been reported in patients with measurable disease. An elderly patient presented with a hidradenocarcinoma as well as local, regional, and distant metastases to the lungs, pleura, and myocardium.19 Despite treatment with bleomycin and vincristine, the patient died within a few months. In a second study, 2 children with metastatic hidradenocarcinoma with involvement of the lungs, pleura, and liver were studied. Disease progression was observed despite treatment of the first child with vincristine and cyclophosphamide and treatment of the second child with vincristine, Actinomyces D, VM-26, doxorubicin, and dacarabazine.20 Further studies are required to determine the benefit of chemotherapy in the prevention of metastasis and disease recurrence or in patient survival. Because hidradenocarcinomas are infrequently occurring tumors, the lack of cases remains a major impediment to large studies or clinical trials to address chemotherapeutic agents.
The use of targeted therapies may eventually emerge as a viable treatment modality for even rare tumors, such as hidradenocarcinomas. The histomorphologic overlap between carcinomas of sweat gland origin and adenocarcinomas of breast and salivary gland origin has led to investigations that used immunohistochemistry to compare HER2/neu, estrogen receptor, and gross cystic disease fluid protein 15 levels. Investigators have evaluated the utility of tamoxifen in metastatic sweat gland carcinomas that demonstrated estrogen receptor reactivity with anecdotal success.21 In a comparable study,10 with tissue from a single patient, HER2/neu protein and gene amplification by fluorescence in situ hybridization was reported. HER2/neu protein is of clinical importance in breast cancer, and protein overexpression has been shown to be an independent predictor of poor outcome. Trastuzumab, a US Federal Drug Administration–approved therapeutic drug, is a humanized immunoglobulin G1 murine antibody that binds and antagonizes the extracellular portion of the HER2/neu receptor, which leads to the inhibition of breast tumor cell proliferation. Trastuzumab has proven efficacy in metastatic breast cancer as monotherapy and as an adjunct in multiagent regimens with more traditional chemotherapeutic agents. In a recent case report,10 a patient was treated with wide, local excision and regional lymph node dissection followed by radiotherapy to the axilla and tumor bed. Adjuvant chemotherapy included a regimen of doxorubicin and cyclophosphamide. Based upon the use of trastuzumab in HER2/neu–positive breast cancers, trastuzumab was added to the patient's regimen. Reportedly, the patient tolerated the therapy well and to date has showed no signs of local recurrence. In the patient reported in our study, immunohistochemical staining for HER2/neu was negative in the primary tumor as was the metastatic regional lymph node (unpublished observations). Therefore, the relevance of HER2/neu positivity to hidradenocarcinomas and the use of trastuzumab in the treatment of hidradenocarcinomas remains to be further addressed. Exploitation of the molecular genetic profile of the tumor may provide insight into the identification of targeted modalities and remains of paramount importance.
References
The authors have no relevant financial interest in the products or companies described in this article.
Author notes
From the Department of Anatomic Pathology, National Naval Medical Center, Bethesda, Maryland (Dr Gauerke); the Department of Pathology, Naval Medical Center, Portsmouth, Virginia (Dr Gauerke); the Department of Pathology, F. Edward Hébert School of Medicine, Uniformed Services University of the Health Sciences, Bethesda (Dr Gauerke); and the Department of Medical Oncology, National Cancer Institute, Bethesda (Dr Driscoll).
The views expressed here represent the personal opinions of the authors and not those of the US Department of Defense, the Department of the Navy, the Department of Health and Human Services, nor the US Government. The authors are federal employees, and there are no copyright or restrictions on this work.