Background

Desmoplastic small round cell tumor (DSRCT) is an extremely rare and aggressive neoplasm that most commonly affects adolescents and young adults with a male predominance [1,2,3,4,5,6,7]. DSRCT preferentially involves the abdominal and pelvic cavities [8,9,10,11]. DSRCT in the pleura, lung, eye, ear, and testis has been reported only in a few cases (< 5%) [3,4,5,6,7], but it is not specifically associated with any organ. Clinical symptoms are usually associated with the tumor site and lack specificity. DSRCT can metastasize in the early stage and quickly recurs, despite treatment [12,13,14,15]. Hence, its prognosis is poor. Herein, we have presented a rare case of primary DSRCT in the submandibular gland. We have described and summarized the clinicopathological and cytological features of DSRCT, in addition to discussing its diagnosis and treatment based on the literature review.

Case presentation

Clinical history

A 26-year-old Chinese man with a chief complaint of a mass in the right submandibular region for the past 1 year was admitted to Xiangya Hospital, Central South University, Hunan, China. He had no significant past medical or family history. Routine physical and laboratory examinations were performed. Ultrasonography revealed a hypoechoic mass measuring approximately 28 mm × 18 mm in the right submandibular region, with an irregular shape and clear boundary (Fig. 1). Abdominal computed tomography (CT) scan revealed no other lesion. There was no evidence of metastasis to the local or distant organs. Hence, lumpectomy was performed under general anesthesia.

Fig. 1
figure 1

Cervical US: A hypoechoic, well-defined mass measuring 28 mm × 18 mm located in the right submandibular region; no obvious blood flow signal was observed in the lesion

Full size image

Pathology

Histological examination showed sheets, cords, and nests of small round cells separated focally by desmoplastic stroma (Fig. 2a). Under higher magnification, tumor cells showed round to oval hyperchromatic nuclei with an increased nuclear/cytoplasmic ratio and inconspicuous nucleoli. The cytoplasm of the tumor cells was scanty with indistinct cytoplasmic borders (Fig. 2b). Mitotic activity and individual cell necrosis were common. Immunohistochemical analysis was performed using formalin-fixed paraffin embedded sections from representative tumor blocks and the antibodies listed in Table 1. Immunohistochemical results indicated the multi-directional differentiation of tumor cells. The immunohistochemistry results were as follows: desmin (+) (Fig. 3), FLI-1 (+), CD99 (+), E-cadherinD (+), chromogranin-A (+), neuron-specific enolase (+), vimentin (+) (Fig. 4), pan-cytokeratin (+), epithelial membrane antigen (+), CD56 (+), synaptophysin (weakly positive [+/−]), NKX2.2 (−), WT1 (−), myogenin (−), and S-100 (−). Moreover, the Ki-67 proliferation index was estimated as 50%. The tumor cells were negative for Epstein-Barr virus-encoded small RNA on fluorescence in situ hybridization (FISH). The FISH analysis with a break-apart probe proved that there was EWSR1 gene spilt in the neoplastic cells (Fig. 5). However, EWSR1-WT1 fusion detection by reverse transcription-polymerase chain reaction was not performed owing to certain limitations. Based on the above findings, primary lesions in the abdominal cavity and pelvic cavity were excluded, and a final diagnosis of primary DSRCT in the submandibular gland was made.

Fig. 2
figure 2

a Tumor cells were irregular sheet-like and nest-like distribution, surrounded by proliferative fibrous stroma (H&E, magnification × 200). b The tumor cells are small round or oval, with few cytoplasm, unclear cell boundaries, round or oval hyperchromatic nuclei and unclear nucleoli (H&E, magnification × 400)

Full size image
Table 1 List of immunohistochemical antibodies used in diagnosis
Full size table
Fig. 3
figure 3

Neoplastic cells showed a diffuse perinuclear staining pattern with desmin (a magnification× 400 and b magnification × 600)

Full size image
Fig. 4
figure 4

Neoplastic cells are positive for FLI-1 (a), CD99 (b), E-cadherin (c), chromogranin A (CgA) (d), neuron specific enolase (NSE) (e), and vimentin (f)

Full size image
Fig. 5
figure 5

Dual Color Break Apart specific locus FISH probe targeting EWSR1 gene; green and red signals mark the 5′ and 3′ ends of the gene respectively

Full size image

Follow-up

Comprehensive anti-tumor therapy mainly based on chemotherapy and radiotherapy was first proposed. However, synchronous chemotherapy was not performed owing to the risk of bone marrow suppression. Therefore, cyclophosphamide combined with doxorubicin and vincristine chemotherapy was used for maintenance treatment. The patient is currently alive and well with no evidence of tumor recurrence.

Discussion

DSRCT, which was first described as a specific disease by Gerald and Rosai [16, 17], is a rare and aggressive soft-tissue sarcoma. Generally, DSRCT develops on the serosal surface of the abdominal cavity [8,9,10,11], but it can also be found in the lung, eye and salivary gland [18,19,20,21,22,23,24,25,26]. DSRCT has no specific clinical symptoms. Most patients present with initial symptoms of abdominal pain, constipation, ascites, and vomiting [10,11,12,13,14, 27, 28]. These can be accompanied by the manifestations of cachexia, such as fatigue and emaciation. Patients may develop intestinal obstruction, hydronephrosis, and urinary tract irritation owing to tumor compression [29]. In our case, a right submandibular mass without apparent clinical manifestations was detected incidentally. DSRCT is often widely disseminated throughout the peritoneal cavity, and some patients may present with metastasis to the lymph nodes, liver, and occasionally the lungs [8, 12, 30]. Hence, its prognosis is exceedingly poor. The clinical features of previously published DSRCT cases in the last 5 years are summarized in Supplementary Table 2, Additional file 1.

Imaging examinations of DSRCT lack characteristic features. Ultrasound examination usually shows a lobulated soft tissue mass with an uneven internal echo [31]. CT usually reveals single or multiple lobular nodules or lumpy soft tissue masses, with an uneven density of the tumor body and multiple spotted calcifications [10]. The lesions tend to crowd out, surround, and invade the surrounding tissues [19]. DSRCT is usually accompanied by flakes of low intensity when there is a necrotic area in the tumor. Enhanced CT presents mild uneven enhancement and edge enhancement may be observed in some larger masses. Moreover, positron emission tomography (PET)-CT has the potential to monitor residual disease and detect relapse or tumor progression in the early stages [32]. Imaging findings are non-specific, but they can indicate the location, size, and the number of tumors, thereby contributing to biopsy, surgery, and radiotherapy.

The definitive diagnosis of DSRCT is based on typical morphological and immunohistochemical features, especially distinctive molecular characteristics. The pathological and molecular features of previously published cases of DSRCT in the last 5 years are summarized in Supplementary Table 3, Additional file 2. Histologically, the tumor consists of small round cells and peripheral desmoplastic stroma, with occasional cystic degeneration and hemorrhagic necrosis. The tumor has a variegated histology revealing rhabdoid, clear or pleomorphic cells, in addition to typical small round cell morphology. Moreover, the tumor can have intermittent areas of primitive tubules or rosette-like structures [12]. Immunohistochemically, tumor cells show a pattern of multiphenotypic differentiation [8, 15, 33, 34]. This multiple antigen expression profile is a characteristic of DSRCT and can be used to distinguish DSRCT from the other histologically related small round cell tumors. Further, para-nuclear dotted desmin positivity has important diagnostic significance. However, in our case, immunohistochemical staining showed a diffuse perinuclear staining pattern with desmin, while the characteristic dotted positivity was not prominent. Almost all cases of DSRCT are positive for WT1 with cytoplasmic and paranuclear staining pattern. Although the immunohistochemical analysis of classic cases of DSRCT tends to reveal WT1 positivity, N- and C-terminals may be useful as a form of “molecular immunohistochemistry” to identify the EWS–WT1 transcript, as the immunostaining pattern may be altered by variant transcripts and WT1 immunostaining may be negative (as in our case) [35,36,37]. To establish a DSRCT diagnosis, the interpretation of WT1 immunostaining requires information of antibody target epitopes and correlations with clinical, morphological, and molecular findings. DSRCT is distinguished by the t (11;22) (p13; q12) chromosomal translocation involving a fusion between the transcriptional activating domain of EWSR1 and the WT1 gene [38,39,40]. Studies have also suggested that the EWSR1-WT1 fusion protein can induce the expression of platelet-derived growth factor-A, which can induce the growth and proliferation of fibroblasts and the production of collagenous stroma; this may explain the characteristic reactive fibrosis of DSRCT [41]. Downstream activation of EWSR1-WT1 gene fusion includes signaling pathways of vascular endothelial growth factor, IL2RB, and insulin growth factor-1 [42,43,44]. A better understanding of the effects of these target genes will provide avenues for future treatment.

In the present case, the tumor was composed of nests of small to medium-sized cells, which might be misdiagnosed as small cell carcinoma. Small cell carcinoma can also show the immunoreactivity for epithelial and neuroendocrine markers. However, in our case, the co-expression of epithelial membrane antigen, vimentin and desmin by tumor cells strongly supports a diagnosis of DSRCT. DSRCT should also be distinguished from other carcinomas, such as malignant melanoma, malignant lymphoma, and metastatic neuroblastoma. The current case was negative for Melan-A and S100, which ruled out the likelihood of malignant melanoma. The positivity for epithelial markers helped rule out the possibility of malignant lymphoma, which often involves the lymph nodes, bone marrow and peripheral blood. In addition, stroma of massive nerve fiber network is a characteristic feature of neuroblastoma, which might be a diagnostic clue. Owing to the histological features of small round cells in the present tumor, it must be distinguished from other small round cell tumors, such as rhabdomyosarcoma, primitive neuroectodermal tumor (PNET), and Ewing sarcoma (EWS). Rhabdomyosarcoma is more common in children; the tumor cells are commonly positive for myogenic markers (such as MYOD1 and myogenin), but negative for epithelial and neuroendocrine markers. Morphologically, DSRCT and PNET reveal chrysanthemum-like structure, and both of them are positive for CD99 and neuron-specific enolase. In our case, immunohistochemical results showed positivity for desmin and epithelial markers, which favors a diagnosis of DSRCT over that of PNET. EWS shares histological and immunophenotypic similarities with DSRCT. However, the survival rate of patients with DSRCT is significantly lower than that of EWS patients, which indicates that DSRCT and EWS have different biological backgrounds. EWS mainly occurs in children and is common in bones. Soft tissue involvement is rare. EWS can also be positive for cytokeratin, desmin, CD99, FLI-1, and neuroendocrine markers, which may result in confusion. The diffuse membranous positivity for CD99 is typical of EWS, but in our case, nonspecific cytoplasmic positivity for CD99 is one of the features of DSRCT. Furthermore, negative NKX2.2 is an important clue to distinguish EWS. As both EWS and DSRCT harbor EWSR1 rearrangements, the break-apart FISH assay for EWSR1 will not be helpful in the differential diagnosis. However, characteristic translocation of EWS involves EWSR1 and the ETS family of transcription factors, not WT1. Convincingly, documentation of EWSR1-WT1 fusion is the “gold standard” for the diagnosis of DSRCT [15, 27, 33]. It was not performed in our case because the sample tissue did not meet the requirements and the patient was unwilling; this is a limitation of this case. Taken together, combined with the tumor location and morphological features, as well as a distinctive pattern of multiphenotypic differentiation on immunohistochemistry, the diagnosis of DSRCT was considered. Similar to DSRCT, both myoepithelial carcinoma and synovial sarcoma are also multiphenotypic and expresse multilineage markers. Myoepithelial carcinoma expresses cytokeratin and myogenic markers, such as myogenin, smooth muscle actin and HHF35. The present case showed positivity for desmin and negativity for all other myogenic markers, P63 and CK5/6, strongly favoring a diagnosis of DSRCT. Synovial sarcoma mainly occurs in the extremities. It expresses epithelial and mesenchymal markers, but desmin positivity is uncommon, which can help in the differentiation.

Most reported cases of DSRCT are diagnosed by the histological analysis of formalin-fixed paraffin embedded tissues, and so far, very few cases have been diagnosed by fine needle aspiration cytology (FNAC) [24, 45,46,47,48,49]. FNAC is a cheap, minimally-invasive, and well-tolerated diagnostic technique [50]. Many soft tissue tumors in any location, especially deep-seated tumors, could be sampled by FNAC [51]. Increasing evidence shows that FNAC can provide rapid and valuable diagnosis in identifying soft tissue malignancy. The cytological smear of DSRCT is characterized by stromal fragments and small, round blue cells with hyperchromatic nuclei [24, 46, 47, 49, 51,52,53]. However, sometimes no fibrosis fragments can be detected in the smear [48, 54]. Occasionally, apoptotic cells and high mitotic activity are observed [24]. Moreover, multiple ancillary techniques, such as immunohistochemical or molecular procedures, can also be performed, which enormously expand the application field of FNAC in the diagnostic approach to soft tissue tumors [24, 50]. However, accurate diagnosis depends on the comprehensive evaluation of histological characteristics, including the growth pattern of neoplastic cells and stromal characteristics; these features may not be well displayed in cytological biopsies because of the cell dispersion, loss of tissue pattern, and paucity of the specimen [52, 55]. Although FNAC can distinguish benign from malignant tumors, it is difficult to provide accurate subtypes. Indeed, some tissue types may lack characteristic morphological or molecular markers, and an accurate diagnosis can only be obtained through surgical resection and histological examination. Therefore, incisional biopsy is still considered the most accurate procedure for diagnosis. In our case, the mass was superficially located and mobile without adhesion. FNAC may not be beneficial in case of such an easily resectable tumor, as complete resection is not only a diagnostic method but also a therapeutic approach. As our patient was reluctant to undergo preoperative FNAC, we routinely performed surgical resection of the mass. We believe that FNAC may emerge as an effective diagnostic tool in the future.

Despite multimodal treatment, DSRCT is highly aggressive and has a poor prognosis. The overall survival in patients is < 3–5 years after diagnosis, and the 5-year survival rate is < 20% [12, 15, 37, 56]. There is no standardized approach for the treatment of this malignant disease. Effective cytoreduction combined with comprehensive therapies, as the best treatment strategy presented in most studies, may improve patient survival rate [14, 15, 28, 33, 34, 57]. With the in-depth analysis of molecular genetics of DSRCT, targeted therapy, immunotherapy, and other methods have been gradually applied to the treatment of DSRCT in recent years [58, 59].

Conclusions

In summary, DSRCT is a poorly understood malignant tumor with characteristic morphology, immunophenotype, and cytological features. The disease does not present with specific clinical signs or symptoms. PET-CT may help diagnose recurrent disease at an early stage. The EWSR1-WT1 fusion detection by reverse transcription-polymerase chain reaction is the gold standard for the diagnosis of DSRCT. When it is not feasible, as in our case, definitive diagnosis mainly depends on a comprehensive analysis of histological and immunohistochemical studies. The submandibular gland is an unusual site for DSRCT, suggesting that the tumor may not have a site-specific predilection. Therefore, pathologists and clinicians have to be aware of its possible occurrence in extra-peritoneal regions. To better diagnose this rare and intriguing disease, further studies are needed in future.