DSRCT was a highly aggressive and rare mesenchymal tumor first described as a separate identity by Gerald and Rosai in 1989[10]. DSRCT characteristically arose in abdominal or pelvis cavity in young men, only 6% of DSRCTs occurred in sites outside the abdomen[11]. DSRCT of submandibular gland as in the case presented here had similar morphologic and immunophenotypic features to the previously reported cases. Histologically, DSRCT was characterized by well-defined and variable-sized nests comprised of small to medium tumor cells with apparent collagen background. Immunohistochemically, DSRCT had a polyphenotypic immunoprofile with neoplastic cells expressing epithelial, mesenchymal, and neuroendocrine markers. WT-1 was considered as a useful antibody for the diagnosis of DSRCT and differentiating it from other tumors with small blue cell morphology. Frequent strong nuclear expression of WT1 had been reported in DSRCT was attributed to EWSR1-WT1 gene fusion[12]. However, WT1, Clone EP122, was negative expression in our study. Kanako C et al. demonstrated that Clone of N-WT1 and 6F-H2 antibodies resulted in negative or abnormal expression of cytoplasm. And their research recommended using C-WT1 antibody[13]. In addition, Paul et al. report showed that a small number of cases were indeed C-WT1 (Clone: C19) negative[12]. Therefore, it was recommended to perform molecular detection so as to avoid diagnostic pitfall. We then performed FISH assay in order to determine whether EWSR1 gene fracture was presented. Not surprisingly, EWSR1 split signals were detected in 86% of the tumor cells, which conclusively confirmed the diagnosis of DSRCT.
DSRCT needed to be differentiated from primary salivary gland tumors. Poorly differentiated myoepithelial carcinoma, the solid variant of basal cell adenocarcinoma (BCA) and adenoid cystic carcinoma (ACC) also had similar morphological areas which were composed of large solid tumor island. Lack of myoepithelial differentiation (actin and S-100 negative) and expression of neuroendocrine markers (CD56, NSE, Syn and CgA) would be helpful in excluding BCA and ACC. Moreover, the tumor islands of DSRCT were lack of peripheral palisade structures in BCA or the focal presence of cribriform structures and hyaline globules in ACC. CD117, frequently used for the diagnosis of ACC, had been demonstrated to be negative or focal positive expression in DSRCT[12].
Although it’s also rare, the other small round cell tumors, such as Ewing sarcoma, small cell carcinoma, rhabdomyosarcoma and malignant melanoma, were still need to be excluded. It was reported that CD99, a marker frequently used for the diagnosis of Ewing’s sarcomas/PNET, had been shown to be positive in 57% of DSRCT[12]. Fli-1 was also found to be positive expression in the current study. Given this finding, CD99 and Fli-1 had a limited role in distinguishing DSRCT from Ewing’s sarcomas and primitive neuroectodermal tumors. Primary small cell carcinoma in the salivary glands was rare, and most occur in patients older than 50 years of age[14]. Positive for CD56 and neuroendocrine marker may increase the risk of misdiagnosis as small cell carcinoma. Abundant desmoplastic stroma and collagenous background may be an important clue for DSRCT. Rhabdomyosarcoma in young patients was a differential diagnosis should also be included. Unlike rhabdomyosarcoma, DSRCT was negative for myogenin and myoD1.
So far, the origin of DSRCT remains unclear. Although the common site of DSRCT, peritoneum-lined surfaces, suggested that it may be derived from the mesothelium, immunohistochemical expression (CK5/6 and calretinin negative) and ultrastructural level did not support tumor cells with mesothelium differentiation[7]. Furthermore, there was no mesodermal distribution in the submandibular area. It seemed reasonable to explain that the tumor cells may originate from multipotential differentiated primitive mesenchymal cells or neuroectodermal tissue.
Table 2 summarizes the clinical and pathological features of submandibular gland DSRCT reported previously in the English-language literature. All 5 cases, including the present one, occurred in males, and the age ranged from 24 to 41 years (mean age, 30 years). All 5 tumors reported were less than 5cm in size. According to the previously reports, the prognosis of intra-abdominal DSRCT was particularly poor, median survival ranged from 17 to 25 months, and 5-year survival rates was around 15%[15]. Lymph node metastases had occurred in most cases at the time of diagnosis. Moreover, tumor occurring in the abdominal cavity was relatively hidden, and the tumor appeared to be progressive when it was discovered. However, the prognosis of DSRCT in this uncommon location remained unknown. Of the 5 patients presented in the Table 2, 3 were alive at the time of publication, one died with systemic metastasis at 25 months, and another one who had a history of diabetes mellitus and end-stage renal failure died of disease at 5 weeks. It was worthy of note that both of them had nodal metastasis. It was reasonable to speculate that lymph node metastasis was a risk factor for poor prognosis. On the other hand, the early occurrence of local mass in submandibular gland provided a higher chance of complete surgical resection. In our case, the patient showed recurrence after 17 months follow up.