Based on the morphological and genetic features, PEComa is a strong consideration for the diagnosis. PEComas can be composed of epithelioid and spindle cells arranged in a hemangiopericytoma-like vascular pattern. TFE3 is positively expressed in a number of PEComas resulting from gene fusion12. TFE3 gene amplification, a characteristic finding in the present case, has also been documented in three cases of PEComas; all of these PEComas were found in adults and showed predominant or pure epithelioid morphology. Among these, two cases displayed marked atypical large cells with abundant eosinophilic cytoplasm, nuclear anaplasia, and multipolar mitotic figures, sharing histological overlaps with the present case. However, tumor cells in PEComas are commonly positive for melanocytic and smooth muscle markers; all the three cases were positive for Melan-A and actin/SMA, two were positive for HMB458-10, whereas none of Melan-A, HMB45 and actin/SMA were expressed in the present case. Therefore, evidence for classifying the tumor as PEComa is insufficient.
Several sarcomas should also be taken into consideration owing to their morphologic resemblance, including solid variant of ASPS, epithelioid sarcoma, epithelioid angiosarcoma, and epithelioid rhabdomyosarcoma. ASPS typically is composed of large eosinophilic cells in nest or pseudoalveolar growth pattern delineated by fibrous septa. In solid variant of ASPS, tumor cells arrange in a diffuse sheet growth pattern without conspicuous nesting, fibrous septa, or dilated sinusoidal vascular channels. Increased cytologic pleomorphism and nuclear pleomorphism are uncommon but have been described13. ASPS demonstrates nuclear expression of TFE3, resulting from the oncogenic fusion or rearrangement of TFE35. However, such characteristic chromosomal translocation was not detected by FISH in the present tumor. The intracytoplasmic accumulation of rod-shaped crystals, which is another specific character of ASPS, could be found in a majority of the cases, was not identified. Both epithelioid sarcoma and epithelioid angiosarcoma are composed of diffusely arranged large cells with abundant eosinophilic cytoplasm, vesicular nuclei, and prominent nucleoli. As for epithelioid sarcoma, especially in the proximal-type, pleomorphism is prominent14. It is one of the few mesenchymal neoplasms that metastasize to lymph nodes. CD 34 is expressed in 50% of cases. However, epithelioid sarcoma is positive for CK and EMA, and usually loss of nuclear INI1 expression15. Epithelioid angiosarcoma presents irregular vascular spaces lined by protuberant epithelioid tumor cells. Endothelial differentiation should be demonstrated by the expression of endothelial markers such as CD31, CD34, ERG and FLI-113. In contrast to the obvious cytologic pleomorphism in the present case, cells and nuclei in epithelioid rhabdomyosarcoma tend be relatively uniform in size. Skeletal muscle differentiation should be proven based on immunostaining for myogenin and myoD116. Therefore, the diagnoses of ASPS, epithelioid sarcoma, epithelioid angiosarcoma, and epithelioid rhabdomyosarcoma were easily excluded.
Immunochemical panels were used to explore possible differentiation of the present tumor. Apart from vimentin, TFE3, and INI-1, only CD68 and CD34 were positive. Because of the non-specificity, whether such immunophenotype would be a diagnostic feature remained to be illuminated. CD68 expression might indicate histiocytic proliferation and be found in fibrous histiocytomas and granular cell tumor13. However, the tumor cells were negative for CD163, a biomarker that is truly histocyte specific superior to CD68, rendering the evidence for histiocytic differentiation weak. CD34 is a non-specific biomarker that should always be used in conjunction with morphology and other markers. Diffuse expression is usually significant; weak or focal expression is often nonspecific. Tumors with fibroblastic/ myofibroblastic or peripheral nerve sheath differentiation might be positive for CD34. Most of them should simultaneously express other more specific markers, such as SMA and/or desmin, and S-100 protein respectively. Other markers explored in the present case include MiTF, STAT6, CDK4, MDM2 and MPO, all of which were negative, demonstrating the undifferentiated nature of the tumor.
TFE3 immunoreactivity has been identified in several neoplasms, ranging from weak to intense expression levels (Table 1). The most common chromosomal abnormalities involving TFE3 are translocations and rearrangements, and are among the earliest reported gene fusion in tumorigenesis17,18. TFE3 gene fusion can occur with different partners, such as PRCC, ASPSCR1, SFPQ/PSF, and NONO, all of which provide active gene promoters, and therefore, lead to higher levels of TFE3 fusion proteins than wild-type TFE35,19,20. Apart from ASPS, PEComas, and renal cell carcinoma, which were most extensively studies, epithelioid hemangioendothelioma with YAP1-TFE3 gene fusion21 and malignant chondroid syringoma with PHF1-TFE3 gene fusion22 were also identified respectively. Although nuclear immunoreactivity for TFE3 protein by immunohistochemistry staining was initially considered as a highly sensitive and specific diagnostic tool for neoplasms bearing TFE3 gene fusions23, subsequent FISH analysis demonstrated that there was no relationship between immunoreactivity and gene fusion24,25. For instance, although TFE3 was positive in 53%-91% of granular cell tumors and 93.5% of desmoid-type fibromatosis, gene rearrangement was absent according to FISH analysis, whether the immunohistochemical staining was weak or strong24,26,27. Therefore, molecular and/or cytogenetic analysis should be performed to avoid false positive staining28. TFE3 gene amplification has only been identified in several cases. Apart from PEComas, it has also been confirmed in four cases of renal cell carcinomas with moderate to strong nuclear expression of TFE311. TFE3 is located on X-chromosome, aneuploidy for X-chromosome was detected and resulted in increasing TFE3 copy numbers10,11. In the present case, we also inferred that it was the X polyploidy that caused TFE3 gene amplified. An interesting finding is that such genetic alteration might be associated with aggressive biological behavior8-11. As for renal cell carcinoma, patients with TFE3 amplification exhibited a significantly poorer cancer-specific survival rate than those with translocation11. As for PEComas, distant metastasis and recurrence were documented8-10. In the present case, tumor aggressiveness has been suggested by radiological examination, and one lymph node metastasis was identified.
The latency time between the diagnosis of CLL and the present sarcoma was 8 years. There is no strong evidence of the relationship between these two tumors in this patient. Studies have shown increased risks of second malignancies after CLL. Langerhans cell sarcoma29, myeloid sarcoma30, and histiocytic/dendritic cell sarcomas31 would transdifferentiate from CLL. Skin neoplasm was also reported preceding the diagnosis CLL, including malignant melanoma32, Kaposi sarcoma33-35, Merkel cell carcinoma, malignant fibrous histiocytoma, dermatofibrosarcoma protuberans, sebaceous carcinoma36, and leiomyosarcoma37. However, the morphological and immunochemical features of the present case did not support any of these neoplasms. The most common adverse effect of chlorambucil is myelosuppression. One conjunctival Kaposi's sarcoma was developed in one patient with sympathetic ophthalmia treated with high-dose, short-term chlorambucil therapy38. No study has documented undifferentiated sarcoma resulting from chlorambucil therapy.
In conclusion, we present a case of undifferentiated sarcoma with characteristic genetic alteration. Even comprehensive immunohistochemistry and molecular examination could not help in classifying the tumor under any of the existing soft tissue tumor classification. Therefore, we designated it as epithelioid undifferentiated sarcoma with TFE3 amplification. Our report indicated TFE3 amplification as an additional chromosomal abnormality might result in increased expression of protein, and suggested the possible relationship between the genetic alteration and biological behavior of tumor.