SEF is a rare tumor characterized histopathologically by a predominant population of epithelioid cells arranged in strands, nests, cords, or sheets, which are embedded within a sclerotic collagenous matrix[1-4
]. Immunohistochemically, staining for vimentin is positive in all SEF whereas staining for CD34, leukocyte markers, HMB45, CD68, desmin, glial fibrillary acidic protein, and TP53 is negative. Focal and weak immunostaining for EMA, S100 and more rarely for cytokeratins may be seen in a minority of cases.
The histopathological differential diagnosis of SEF generally includes a wide variety of tumors with sclerotic or epithelioid features, and thus immunohistochemical analysis is essential for precise diagnosis of SEF. In the present case, differential diagnosis was challenging for the following reasons. Undifferentiated hepatocellular carcinoma and infiltrating adenocarcinoma should be differentiated from SEF, but immunohistochemical staining for AFP, CAM5.2, and AE1/AE3 was negative in the present case. Synovial sarcoma should also be differentiated from SEF. Cytogenetic identification of t (X: 18), which is found in synovial sarcoma, could differentiate synovial sarcoma from SEF[12,13
]. No chromosomal translocation was observed in the present case. Furthermore, smooth muscle neoplasms, such as hyalinized leiomyoma or leiomyosarcoma, frequently resemble SEF histopathologically, but they are characterized by immunohistochemical positivity for α-SMA and desmin. Clear cell sarcoma and malignant peripheral nerve sheath tumors may need to be distinguished from SEF. However, these tumors are positive for S100 immunostaining. Moreover, sclerosing lymphoma and malignant melanoma of the soft part are also on the list of differential diagnoses, but these tumors are usually positive for LCA and HMB45, respectively, and were negative in our case. Epithelioid hemangioendothelioma could be ruled out based on immunonegativity for CD34, and alveolar rhabdomyosarcoma could be excluded by the negative result of desmin and absence of rhabdomyoblasts. Moreover, absence of osteroid formation excluded traosseous osteosarcoma in the present case. Thus, taking into consideration not only the histopathological findings but also the results of cytogenetic analysis and immunohistochemical examinations, the tumor was finally diagnosed as SEF.
In reviewing previous reports, SEFs usually arise in the deep soft tissue and are frequently associated with the adjacent fascia or periosteum. Most SEFs are located in the lower extremities and limb girdles, followed by the trunk, upper extremities, and the head and neck area[1-10
In this case, although the tumor was associated with invasion to the wall of IVC and diaphragm, the most part of the tumor existed in the liver. Moreover, in this case, tumor thrombi and intrahepatic metastases were concurrently identified. These features might suggest that the tumor had originated in the liver. To our knowledge, there have been no reports of SEF originating from the liver.
The etiology of most malignant soft tissue tumors, including SEF, is generally unknown. A recent report reviewed 90 patients with SEF and found no significant gender difference with a mean patient age of 47 years (range, 14-87 years)[5
]. The average tumor size at diagnosis is about 8 cm. Several investigators reported previously the association between radiation and occurrence of fibrosarcoma, though our patient had no history of radiation[9
The established treatment for SEF is complete resection. There is no evidence to support the effect of chemotherapy and/or radiotherapy, though these therapies were used in some previous reports[3,10
]. As for prognosis, Chow et al[4
] reported 57 patients with SEF with a local recurrence rate of 48%, metastasis rate of 60%, and mortality rate of 35%. Moreover, in a study of 16 patients with SEF by Antonescu et al[3
], the local recurrence rate, metastasis rate, and mortality rate were 50%, 86% and 57%, respectively[5
]. Distant metastasis is reported to be most common in the lung, followed by bone, soft tissue, brain, and lymph nodes[1-5
]. These previous reports of poor prognosis suggest that SEF is a clinicopathologically distinct soft tissue tumor with malignant potential although it is categorized as a low-grade neoplasm in the sarcoma group.
In our case, since no distant metastasis was found preoperatively and since complete resection was thought possible, we selected surgical resection although the tumor had extended at the time of surgery into the IVC, in addition to local intrahepatic metastasis. Indeed, we could completely resect the tumor macroscopically, but the tumor recurred postoperatively. It is quite possible that malignant cells remained at the tumor bed after the resection despite our careful endeavor. Alternatively, considering the high recurrence rate of SEF reported previously, it is also possible that malignant cells might have already exfoliated from the tumor extending to IVC and presumably reached other organs before the resection. Thus, when surgical resection is selected for SEF, perioperative chemotherapy and/or radiotherapy, in addition to macroscopic complete resection, may be necessary to prevent recurrence, though there is no established effective and standardized regimen of chemotherapy and/or radiotherapy. To clarify the effect of perioperative chemoradiotherapy, in addition to surgical resection, further study with a larger number of SEF cases treated perioperatively is needed.