Comment
Myoepithelial cells normally surround ducts in secretory organs, such as the breasts, salivary glands, and cutaneous sweat glands. Myoepithelial neoplasms are well recognized in the salivary glands14,15; however, myoepithelial neoplasms also can arise in other sites, including the soft tissue4,5,16-18 and skin.1-3,7,11,19,20 Myoepithelioma of soft tissue was first described by Burke et al21 in 1995 and later described in the skin by Fernandez-Figueras et al22 in 1998. Since then, diagnostic criteria for cutaneous myoepithelial neoplasms have evolved, suggesting a spectrum of disease rather than a single distinct entity.11 Most often, cutaneous myoepithelial carcinomas arise as soft nodular lesions in the head and neck areas or extremities of adults. The nodules typically are nontender and range in size from 0.5 to 18.0 cm. Our review of the literature revealed 11 additional cases of cutaneous myoepithelial carcinomas have been reported, ranging in size from 0.7 to 7.0 cm (Table). In our case, the main lesion was 6 cm, mildly tender, ulcerated, and accompanied by satellite nodules.
Histologically, cutaneous myoepithelial tumors typically are well-defined, dermal-based nodules with no connection to the overlying epidermis. Similar to myoepithelial tumors of other sites, they can be diagnostically challenging due to the heterogeneity of both their architectural and cytological features. The presence of a chondromyxoid or hyalinized stroma is common but not always present. Neoplastic myoepithelial cells can exhibit spindled, epithelioid, plasmacytoid, or clear cell morphologic features and show growth patterns in clusters, cords, glands, or sheets. Focal epithelial cells can be present. Although benign myoepithelial neoplasms with overt ductal differentiation are consistent with cutaneous mixed tumors (chondroid syringomas), those without ducts are characterized as myoepitheliomas. It is uncertain if cases with only focal ductal differentiation should be classified as mixed tumors or as myoepitheliomas. Malignant myoepithelial tumors show infiltrative borders, nuclear pleomorphism, coarse nuclear chromatin, prominent nucleoli, and increased mitotic activity. A 2003 study by Hornick and Fletcher16 found that cytologic atypia was the primary predictor of malignant behavior for myoepithelial neoplasms of the soft tissue.
Despite a wide variety of expression patterns, immunohistochemistry is critical in demonstrating myoepithelial differentiation and establishing a diagnosis of a myoepithelial neoplasm. Most cases display coexpression of epithelial markers, including keratins and/or EMA as well as S-100 protein. Myogenic markers also may be variably expressed; however, the absence of myogenic markers does not exclude the diagnosis of a myoepithelial tumor. Commonly expressed epithelial markers are cytokeratin AE1/AE3, cytokeratin 8/18, and EMA, while commonly expressed myogenic markers include muscle specific actin and smooth muscle actin.5,7,11,19 Myoepithelial tumors also may express calponin, p63, and glial fibrillary acidic protein.16
Molecular studies also can aid in the diagnosis of myoepithelial tumors. A study by Antonescu et al8 demonstrated EWSR1 gene rearrangement in 45% (30/66) of extrasalivary myoepithelial tumors and the absence of EWSR1 gene rearrangement in salivary gland myoepithelial tumors. The authors also showed that EWSR1-negative tumors were more likely to be superficially located, display ductal differentiation, and possess a benign clinical course.8 In another study, Bahrami et al23 suggested that a subset of mixed tumors, specifically those with tubuloductal differentiation, are genetically linked to salivary gland pleomorphic adenomas, which was achieved by the coexpression of the PLAG1 protein and PLAG1 gene rearrangement on immunohistochemistry and fluorescence in situ hybridization (FISH), respectively. Of the 19 cases evaluated, 11 (58%) expressed nuclear staining for PLAG1 immunohistochemistry; 8 of those 11 showed positive gene rearrangement for PLAG1 using FISH. These findings raise the possibility that cutaneous mixed tumors may be more closely related to those of the salivary glands, while deep myoepithelial tumors that lack ductal differentiation may represent a distinct group. Similar to the study by Antonescu et al,8 Flucke et al10 investigated EWSR1 gene rearrangement but limited their sample to cutaneous tumors, including myoepitheliomas, mixed tumors, and myoepithelial carcinoma. The authors found that 44% of cases (7/16) expressed EWSR1; this expression suggests that cutaneous myoepithelial tumors may have a genetic relationship to their soft tissue, bone, and visceral counterparts.10
Myoepithelial tumors display a broad spectrum of morphologic features; however, one of the most common growth patterns is that of oval to round cells forming cords and chains in a chondromyxoid stroma. As such, the histopathologic differential diagnosis for myoepithelial tumors includes other epithelioid or round-cell neoplasms with similar growth patterns including extraskeletal myxoid chondrosarcoma (EMC), ossifying fibromyxoid tumor of soft parts, and extra-axial soft tissue chordoma. Extraskeletal myxoid chondrosarcoma bears the closest similarity to myoepithelial tumors both histologically and by ancillary studies. It typically possesses cords or chains of small round tumor cells set in a chondromyxoid or myxoid background. In contrast to myoepithelial tumors, which typically have more abundant cytoplasm and can show at least focal areas of spindle cell growth, the cells of EMC are more uniform, small, round cells with relatively scant cytoplasm. Extraskeletal myxoid chondrosarcomas lack the typical myoepithelial coexpression of cytokeratin and S-100 protein, with a minority of EMCs expressing S-100 protein but rarely cytokeratin. Most cases of EMC possess a balanced t(9;22) translocation involving the EWSR1 gene,24 a finding that could lead to confusion with soft tissue myoepithelial tumors, which also may show EWSR1 rearrangement on FISH. Ossifying fibromyxoid tumor of soft parts is also composed of round cells arranged in cords in a myxoid or fibrous stroma; the majority of cases also display a peripheral rim of mature bone, a feature that is not typically seen in myoepithelial tumors. Similar to myoepithelial tumors, ossifying fibromyxoid tumor of soft parts often is positive for S-100 protein; however, it rarely is positive for cytokeratins. Ossifying fibromyxoid tumor of soft parts has been shown to have a rearrangement of the PHD finger protein 1 (PHF1) gene in approximately half of cases, a molecular finding that has not been reported for myoepithelial tumors.25 Finally, extra-axial soft tissue chordomas, though quite rare, may possess striking similarities to myoepithelial tumors both histopathologically and immunohistochemically. Chordomas are composed of epithelioid cells arranged in nests, nodules, and chains with a variably myxoid background. A variable amount of cells with bubbly cytoplasm (known as physaliphorous cells) can be seen. High mitotic activity is not a characteristic feature in chordomas. They classically coexpress cytokeratins and S-100 protein, similar to myoepithelial tumors. A subset of myoepitheliomas with similar histologic features to chordoma was historically referred to as parachordoma.26,27 The distinction between these 2 entities was challenging until the relatively recent advent of brachyury, a sensitive and specific nuclear marker of chordoma; extra-axial soft tissue chordomas and their central counterparts both express nuclear brachyury, while myoepitheliomas (including those with a parachordoma histologic pattern) do not.28 Our case did not display physaliphorous cells but did demonstrate abundant nuclear pleomorphism and high mitotic activity. In addition, immunohistochemical staining was negative for brachyury.
Because cutaneous myoepithelial tumors are relatively rare, a well-defined standard of care for treatment is lacking. Surgical excision is the primary treatment method in most reported cases in the literature.17,19 Miller et al29 reported the successful treatment of recurrent cutaneous myoepitheliomas with Mohs micrographic surgery. Chemotherapy may be useful in the setting of metastatic myoepithelial carcinomas in adults, but reported results are inconsistent.30,31 Radiation treatment of recurrent or metastatic disease has not been shown to be effective. A study of children treated with surgical resection and chemotherapy using ifosfamide, cisplatin, and etoposide followed by radiation therapy showed positive results.32
Our case highlights several challenges that may arise in establishing a diagnosis of cutaneous myoepithelial carcinoma with disseminated metastases. The diagnostic difficulty in our case was compounded by the advanced nature of the lesion at the time of presentation. Given the rarity of metastatic cutaneous myoepithelial carcinomas and the lack of a prior primary diagnosis of a malignant myoepithelioma, the index of suspicion for this entity was not high. A report of myoepithelial carcinoma of the parotid gland metastatic to the skin has been reported,33 but in the absence of salivary gland involvement or other visceral lesions, metastasis from any source other than our patient’s cutaneous scalp lesion is unlikely. The histopathologic features in combination with the characteristic immunophenotype, unique clinical setting, and radiographic findings were essential to arriving at the correct diagnosis. Unlike previously reported metastatic lesions, our case is unique in that metastatic lesions were identified at the time of initial clinical presentation.
Conclusion
Cutaneous myoepithelial carcinomas are exceedingly rare tumors with a wide range of histopathologic and immunohistochemical findings. In challenging cases, studies for EWSR1 or PLAG1 gene rearrangement can be helpful. Furthermore, this case illustrates the potential for widespread dissemination of myoepithelial carcinomas requiring clinical evaluation and imaging studies to exclude metastatic lesions.