Primary Desmoplastic Small Round Cell Tumor of the Femur: Case Report and Review of a Rare Intraosseous Malignancy

Introduction

The understanding of desmoplastic small round cell tumor has evolved significantly since its establishment in a 1991 series documenting its unique multiphenotypic differentiation.^1,2 They have a predilection for young adult male patients, and are relentlessly aggressive with poor overall survival even in the face of multimodal therapy.^3–5 Cytogenetically, they are defined by a recurrent and pathognomonic t(11;22)(p13;q12) translocation. This manifests in a chimeric RNA transcript between coding regions for EWSR1 on 22q11 and WT1 on 11q13.

The description of DSRCT as a primary malignancy of intraosseous tissues is remarkably uncommon. This has only been recorded definitively, to our knowledge, on four separate occasions (calvarium, ilium, tibia, and femur; possibly a fifth which appeared more likely centered in the soft tissues and extending into the bones of the hypothenar eminence).^6–10 This lethal tumor should be included in the ever-broadening differential of primitive round blue cell tumors of bone - especially considering their multifactorial overlap with the more common osteogenic sarcomas.

As a subset of DSRCT, the overall clinical trajectory of intraosseous tumors appears comparable to those of their intraabdominal and soft tissue counterparts – distant metastases and high mortality. Herein, we describe the clinicopathologic, histologic, and molecular features of the fifth known primary intraosseous DSRCT, with widespread metastatic disease at presentation. The diagnostic challenges and differential are highlighted, along with a review of the available literature on primary DSRCT of bone.

Materials and Methods

Case Report

A healthy 18 year-old Hispanic male with a prior history of bipolar depression (on lithium, escitalopram, and aripiprazole) presented to the emergency room with a one-year history of progressively worsening left knee and thigh pain, relieved by analgesics and intermittent icing. There were no systemic symptoms or prior trauma/injury to the area. An x-ray at the time revealed no apparent abnormalities and he was discharged with a diagnosis of “runner’s knee”. Six months later, he was referred to our orthopedic clinic without improvement. Physical exam demonstrated effusion and tenderness to palpation over the medial femoral condyle, but otherwise no deficiencies in motor ability or neurovascular function. The skin overlying the lesion was intact without erythema or ecchymosis.

X-ray at that time revealed a poorly marginated, predominantly radiodense lesion involving the distal femur, extending from the diaphysis into the metaphysis. Additionally, a mineralized soft tissue mass extended medially from the femoral cortex by less than one centimeters. There was no evidence of a fracture. MRI showed the lesion to involve most of the mid- to distal femur over a span of 17 cm. It was composed of apparently separate discontinuous foci over this expanse, although these were favored to be microscopically confluent. All were dark on T1 and bright on T2-weighted images. The minor soft tissue component on x-ray was also evident on MRI, and located on the medial supracondylar flare (Figures 1 and 2). The aggressive radiologic features as well as clinical scenario in an adolescent male were concerning for osteosarcoma, Ewing sarcoma, osteoblastic lymphoma, or a benign process such as an atypical infection.

OPEN IN VIEWER

Figure 1.

Left Femur Coronal MRI. A. T1-weighted image showing multifocal lesions hypo-intense to muscle with a discontinuous lesion in the proximal femur and distal hyper-intense soft tissue component. B. Proton density-weighted image showing heterogeneous signal with hyper-intense distal soft tissue component.

OPEN IN VIEWER

Figure 2.

Distal Left Femur Axial MRI. A. T1-weighted image with hypo-intense marrow signal and hypo-intense periosteal soft tissue mass medially. B. T2-weighted image with edema and hyperintense periosteal soft tissue mass.

Histomorphology

The patient subsequently underwent an open biopsy of the lesion. Standard histologic sections (post-EDTA decalcification) showed a very crushed but cellular neoplasm infiltrating through both the intramedullary bone and adjacent soft tissues. It was composed of sharply defined clusters with irregular contours, to trabeculae of small cells embedded in a dense desmoplastic stroma (Figure 3). The cells had monomorphic, slightly enlarged round to ovoid nuclei. The chromatin was condensed to finely granular and open, with occasional small nucleoli. Cytoplasm was scant and amphophilic with indistinct borders. Neither rhabdoid cells, rosettes, nor glands were identified. The surrounding stroma consisted of bland fibroblasts and myofibroblasts elaborating a dense collagenous to focally myxoid matrix, with sparse chronic inflammation. There was some bony remodeling with osteoblastic rimming, and neoplastic osteoid was not identified (Figure 4).

OPEN IN VIEWER

Figure 3.

Morphologic features. A-B. Nests of tumor cells infiltrating through the medullary canal with accompanying fibroblastic stroma. The pre-existing trabecular bone shows some evidence of remodeling. C. Sharply defined, angulated nests with surrounding desmoplasia. D. ‘Small round blue cells’ with homogenous chromatin and scant cytoplasm.

OPEN IN VIEWER

Figure 4.

Immunohistochemical features. A. Strong, diffuse EMA expression. B. Weak, patchy membranous CD99. C. Desmin showing and with frequent ‘paranuclear dot-like’ and cytoplasmic pattern of expression. D. WT-1 (N-terminus) showing punctate cytoplasmic and paranuclear reactivity.

Accordingly, these biopsy fragments raised a rather broad preliminary histologic differential diagnosis, including Ewing sarcoma, Ewing-like tumor (BCOR-CCNB3, CIC-DUX4 fusions), non-Hodgkin’s lymphoma, small cell or osteoblastic osteosarcoma, and solid alveolar rhabdomyosarcoma.

Discussion

This report is an important addition to the scarce literature on primary DSRCT of bone. In the molecular frontier of diagnostic bone and soft tissue pathology, this case and those prior illustrate some contextual pitfalls even with careful integration of available evidence.

DSRCTs show a unique and hallmark divergence into epithelial (EMA, cytokeratins), neural (NSE), and mesenchymal/myogenic (vimentin, desmin) phenotypes. Originally speculated to be of “mesothelioblastic” differentiation by virtue of their predominantly intraabdominal and serosal localization and immunophenotypic similarities to fetal mesothelium, they are also documented to arise in a variety of nonmesothelial sites, including pleura, subcutis, and orbit.^12–15 Morphologically - in addition to its descriptive eponym – DSRCT can have significant cellular pleomorphism (up to frank anaplasia), perinucleolar clearing, spindled or signet-ring cellular features, “zellballen”-like or papillary architecture, and conspicuous absence of desmoplasia.^16–20

The canonical EWSR1-WT1 fusion in DSRCT is inconsistent at the transcript and protein levels due to different breakpoints in the involved genes, corresponding to variable functional levels of the expressed protein.^16,21–25 The most frequently reported genomic alteration in DSRCT (of any site) is a fusion of exon 7 of EWSR1 to exon 8 of WT1, rendering a composite of the N-terminus of the EWS protein to the C-terminus of WT1. The latter explains nuclear immunoreactivity exclusively with antibodies to the C-terminus of the WT1 protein. (As such the “N-terminus” antibody employed at most academic institutions is typically not applied towards this diagnosis.) As we illustrate here, nonspecific cytoplasmic or paranuclear positivity with the N-terminus antibody should not confound the diagnosis.

Alternative breakpoints for this fusion span the breadth of the EWSR1 and WT1 coding regions with consequent variation in the gene segments included in the chimeric transcript. In keeping with this, our case harbored in-frame breakpoints within exon 7 of each gene, a heretofore unreported variation. Although the more common EWSR1 exon 7 breakpoint was detected via both break-apart FISH and RNA Oncoplus fusion panel, these inherent molecular deviations could represent cryptic rearrangements undetectable by standard molecular methods (due to lack of primer or epitope coverage).

The past two decades are peppered with reports of primary intraosseous DSRCT (Table 1). Of the documented five (including ours), four were males with ages spanning 6 to 33 years. Prior history of trauma was specifically denied in a majority. Three arose in the diaphysis of the long bones (tibia, femur) with two of these extending to the metaphysis and one continuing into the epiphyseal plate. The two others arose in the flat bones (ilium, calvarium). All save one showed cortical breach; three of these had adjacent soft tissue extension. Maximum tumor size on imaging (when provided) ranged from 7 to 17 cm. All appeared aggressive with a prebiopsy radiologic differential including (most commonly) osteosarcoma, Ewing sarcoma, and lymphoma. Only two others, (in addition to ours), had metastatic disease at presentation (one widespread, the other with pulmonary nodules). Molecular confirmation in all cases (save ours) was performed via reverse transcriptase polymerase chain reaction. The only specified transcript in this context was for the calvarial DSRCT reported by Kachaturov, et al, in which exon 10 of EWSR1 was fused to exon 8 of WT1. All were treated with a variable combination of resection (save the calvarial tumor) and neoadjuvant/adjuvant chemotherapy (most frequently doxorubicin, cyclophosphamide, etoposide, ifosfamide, and/or vincristine). Treatment for the femoral and calvarial tumors included myeloablative and focused radiation, respectively. Overall, the behavior and outcome of these intraosseous DSRCT appear similar to those from the abdominal or extra-abdominal soft tissues.

OPEN IN VIEWER

Table 1.

Review of the literature to date on primary DSRCT of bone.

		Clinicopathologic Features										Imaging Features
Case #	Reference	Age (years)	Gender	Race	Clinical presentation	Location	Metastasis at presentation	Treatment	Follow-up	Unconventional histologic features	Cytogenetic rearrangment	Physis	Size (cm)	Lytic/Sclerotic	ST extension
1	Murphy et al (2005) 7	13	M	Unsp	6mo of intermittent hip and lower back pain	Iliac wing	No	Chemotherapy then resection	Widespread metastases 8mo later	None	Unspecified	Unspecified	Unspecified	Sclerotic	No
2	Yoshida et al (2008) 8	10	F	Unsp	3mo of distal thigh pain	Femur	Multiple subcentimeter pulmonary metastases	Chemotherapy	Therapy-related morbidity, died 3 years later from metastases to lung/bone	None	Unspecified	Diametaphysis, involving growth plate	15.7	Lytic	Yes, minor
3	Kachaturov et al (2015) 9	6	M	Unsp	Anterior ear pain/swelling	Right sphenoid/temporal bone	No	Unresectable; treated with chemoradiation	Unspecified	Conspicuous perinuclear clearing	EWSR1 (exon 10)-WT1 (exon 8)	NA	Unspecified	Mixed	Yes, minor
4	Xuesong et al (2020) 10	33	M	Chinese	3mo of intermittent pain in lower limb	Tibia	No	Resection then chemotherapy	NEOD 18 mo after therapy	None	Unspecified	Diaphysis	7	Lytic	No
5	Sharma et al (2021)	18	M	Hispanic	3mo of intermittent pain in upper limb	Femur	Yes, widespread	Chemotherapy and localized radiation	Undergoing treatment	WT1 (N-terminus) punctate cytoplasmic and paranuclear staining	EWSR1 (exon 7)-WT1 (exon 7)	Diametaphysis, involving growth plate	17	Blastic	Yes, minor

*Abbreviations: 'Unsp': Unspecified or unavailable; 'NEOD': No evidence of disease'; 'XR': X-ray/plain film

Of course, metastatic disease must be the primary considerations when a rare tumor is encountered in an even more rare anatomic location. The clinical presentation, namely a year-long prodrome of poorly localized leg/knee pain, was more consistent with a primary bone lesion; metastatic disease from an intraabdominal lesion would have declared itself more expediently and in a different distribution. PET uptake in the potential intraabdominal primary sites (replaced splenic lymph node and adrenal, which were not sampled pretreatment) might arguably have been lower that of the femoral tumor because they were comprised mainly by desmoplasia rather than cellular “small round blue cell” elements. This idea was entertained; however, the magnitude of the size disparity between the femoral and intraabdominal lesions, rendered such a consideration unlikely.

Intraosseous nonosteogenic/nonchondrogenic tumors are described in brief case series and defined by the WHO as very uncommon but distinct entities, and equally as aggressive as their soft tissue analogs. From ours and prior reports, DSRCT of bone can appear not only radiographically but also clinicopathologically similar to osteosarcoma and Ewing sarcoma – young patients with destructive lesions of tubular bones, frequently with periosteal reaction, cortical breach, and soft tissue extension. The significance of the soft tissue mineralization on x-ray in this case is unclear; indeed, many radiologic studies pertaining to DSRCT are understandably restricted to CT and MRI modalities. Intratumoral calcification has occasionally been reported on MRI.^26,27

On histology, small round cell tumors of bone have a standard differential diagnosis and workup that usually does not historically include DSRCT. Misconstruing a collagenous background or reactive woven bone for neoplastic osteoid production could lead to a diagnosis of conventional or small cell osteosarcoma (a distinction towards which ancillary tests are unreliable). Osteosarcomas of any subtype can be aberrantly positive for keratins, presenting another interpretive danger. Immunoreactivity for desmin in a small round cell tumor could errantly support the consideration of a solid-type alveolar rhabdomyosarcoma (also rare in bone). And although unlikely in a younger patient, EMA positivity might mimic both metastatic carcinoma and sclerosing epithelioid fibrosarcoma, appearing sclerotic on imaging as well as histology (the latter also harbors EWSR1 rearrangements).

Perhaps the greatest diagnostic hazard resides in the rather promiscuous EWSR1 rearrangement by break-apart FISH, which is the first step in molecular confirmation of a suspected Ewing sarcoma (usually followed by RT-PCR or RNA fusion sequencing to detect FLI-1 or ERG partners) in essentially the exact same clinical and radiologic context. (While most institutions now reflex to sequencing for mesenchymal neoplasms, this is still a relevant cautionary tale towards those for which it is inaccessible.) In fact, early literature on this tumor harbors three cases of DSRCT with underlying EWSR1-ERG and EWSR1-FLI1 fusions, in which the tumor was defined as DSRCT rather than Ewing sarcoma due to otherwise conventional morphologic and multiphenotypic immunohistochemical features.^24,28–33 In this or virtually any other context, these fusions would define Ewing sarcoma to those lacking molecular capabilities and/or familiarity with the tumor. The burgeoning use of Next Generation Sequencing and cases such as these emphasize how crucial it is to interpret molecular results in conjunction with other morphologic and immunohistochemical features.

Conclusion

In summary, we describe the fifth case of a primary DSRCT of the distal femur in a teenage boy, Due to the rarity of this site, DSRCT can masquerade as otherwise more common tumors of bone, underscoring future considerations regarding its cognizance and subclassification.