Small cell osteosarcoma with Ewing sarcoma breakpoint region 1 gene rearrangement detected by interphase fluorescence in situ hybridization

Annals of Diagnostic Pathology 17 (2013) 377–382

Contents lists available at SciVerse ScienceDirect

Annals of Diagnostic Pathology

Small cell osteosarcoma with Ewing sarcoma breakpoint region 1 gene rearrangement detected by interphase fluorescence in situ hybridization Ema Dragoescu, MD a,⁎, Colleen Jackson-Cook, PhD a, Gregory Domson, MD b, Davis Massey, MD, PhD a, William C. Foster, MD b a b

Department of Pathology, VCU Health System, Richmond, VA Department of Orthopaedic Surgery, VCU Health System, Richmond, VA

a r t i c l e Keywords: Small cell Osteosarcoma EWSR1 gene FISH

i n f o

a b s t r a c t Because of its characteristic morphologic appearance, small cell osteosarcoma (SCO) can be confused with other small round cell malignancies of the bone, most importantly with Ewing sarcoma, making this distinction difficult. A specific tool used in separating SCO from Ewing sarcoma has been the detection of Ewing sarcoma breakpoint region 1 (EWSR1) gene rearrangements in Ewing sarcoma and their absence in SCO. However, there are rare case reports that have documented the existence of EWSR1 gene rearrangement in SCO. In this report, we describe another case of SCO with an EWSR1 gene rearrangement detected by interphase fluorescence in situ hybridization. Our finding adds support to the existing evidence that SCO is a tumor that can be characterized by EWSR1 gene arrangements. Therefore, we caution the pathology community not to rely solely on molecular studies in distinguishing SCO from Ewing sarcoma. © 2013 Elsevier Inc. All rights reserved.

1. Introduction Small cell osteosarcoma is an uncommon variant of high-grade osteosarcoma (1.3%) [1] characterized morphologically by smallsized, uniform tumor cells with at least focal osteoid production. Although this variant of osteosarcoma has clinical features and a skeletal distribution similar to conventional osteosarcoma [1], due to its characteristic morphologic appearance, it can be confused with other small round cell malignancies of the bone, most importantly with Ewing sarcoma. As the treatment protocols and chemotherapy regimens for high-grade osteosarcoma and Ewing sarcoma are different (based on National Comprehensive Cancer Network guidelines), reaching a correct interpretation in challenging cases becomes paramount. The correct diagnosis of small cell osteosarcoma cannot always be achieved based on the morphologic appearance of the tumor alone, as the presence of osteoid in small cell osteosarcoma is frequently focal and can be overlooked. Immunohistochemical stains have been developed to help distinguish small cell osteosarcoma from Ewing sarcoma [2], but some of them have proven to be nonspecific. More recently, a specific tool used in separating small cell osteosarcoma from Ewing sarcoma has been the detection of Ewing sarcoma breakpoint region 1 (EWSR1) gene rearrangements in Ewing sarcoma and its absence in small cell osteosarcoma. However, even EWSR1 ⁎ Corresponding author. Department of Pathology, VCU Health System, P.O. Box 980662, Richmond, VA 23298–0662. Tel.: +1 804 628 0354; fax: +1 804 828 8733. E-mail address: [email protected] (E. Dragoescu). 1092-9134/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.anndiagpath.2012.08.004

gene rearrangements have not proven to be entirely specific for Ewing sarcoma, as these rearrangements have been described in several other unrelated soft tissue lesions [3]. Despite these new reports about the nonspecifity of EWSR1 gene rearrangement, small cell osteosarcoma has not been considered a tumor to harbor this particular gene rearrangement. Yet, to date, 3 case reports have documented the existence of EWSR1 gene rearrangement in small cell osteosarcoma [4-6]. In this report, we describe another case of small cell osteosarcoma with an EWSR1 gene rearrangement detected by interphase fluorescence in situ hybridization (FISH). Our finding adds support to the existing evidence that small cell osteosarcoma is another tumor that can be characterized by EWSR1 gene rearrangement. Therefore, we caution the pathology community not to rely solely on molecular studies in distinguishing small cell osteosarcoma from Ewing sarcoma.

2. Report of a case A 17-year-old adolescent boy complained of right knee pain for approximately 2 months before developing a pathologic fracture through his right proximal fibula following minimal physical activity. On physical examination, he had a firm, nonmobile, painful mass over his right proximal fibula. He had a full range of motion of his right knee and ankle with no instability and had normal sensation to touch throughout his right lower extremity. His coordination and deep tendon reflexes were within normal limits, and he demonstrated a 2+ dorsalis pedis pulse.

378

E. Dragoescu et al. / Annals of Diagnostic Pathology 17 (2013) 377–382

Fig. 3. Focally, malignant cartilaginous matrix is present in the biopsy material (medium power). Fig. 1. Axial T2-weighted magnetic resonance imaging demonstrates the large soft tissue mass circumferentially around the right proximal fibula.

Conventional x-rays showed a permeative lesion associated with a pathologic fracture in the right proximal fibula with magnetic resonance imaging confirming the presence of a large soft tissue mass circumferentially around the right proximal fibula (Fig. 1). The patient underwent incisional biopsy under fluoroscopic guidance of the right proximal fibula mass. The pathologic specimen from the incisional biopsy consisted of a 2.2 × 1.3 × 0.5 cm aggregate of pink-tan, soft tissue. The histologic sections showed a malignant neoplasm composed of relatively uniform tumor cells with increased nuclear-to-cytoplasmic ratio, minimal cytoplasm, and round-oval nuclei with coarse chromatin and 1 to 2 small nucleoli. Mitotic figures and apoptotic bodies were easily found. The tumor cells were arranged in nests separated by fibrovascular stroma. Tumor necrosis was observed in the center of the tumor nests. Pink, lace-like osteoid was focally identified percolating between the tumor cells (Fig. 2). This focal finding was better appreciated at higher magnification. A single focus of malignant cartilaginous matrix was identified (Fig. 3). Tumor cells were strongly and diffusely positive for CD99, with a distinct membranous pattern as

Fig. 2. The tumor consists of uniform, small-sized cells with oval, hyperchromatic nuclei and scant cytoplasm. Frequent mitotic figures are present. Pink, lace-like osteoid is focally noted between the tumor cells (high power).

well cytoplasmic staining (Fig. 4). A diagnosis of small cell osteosarcoma was rendered. Tissue from the incisional biopsy that was fixed in 10% neutral buffered formalin and paraffin embedded was submitted for interphase FISH using a dual color-labeled break-apart probe specific for the EWSR1 gene, which is localized to 22q12 (Abbott Molecular, Abbott Park, IL, USA). Following hybridization, 103 (41%) of 250 cells analyzed had an atypical FISH pattern consistent with the presence of a rearrangement of the EWSR1 probe. Specifically, 2 fusion signals and 1 green signal were seen (Fig. 5). Interestingly, one of the fusion signals was smaller than the other, suggesting that the breakpoint for the rearrangement was localized to the telomeric (3’) portion of the probe. In the remaining cells, either a normal pattern or a random abnormal pattern was present. To determine if this rearrangement might involve a translocation between chromosomes 11 and 22, a dual color fusion probe for FLI1 and EWSR1 (Cytocell, Cambridge, United Kingdom) was evaluated (Fig. 5). The FISH pattern seen with this probe was within normal limits (no fusion signals seen). However, because the breakpoint for the rearrangement in this tumor is distal (telomeric) to those often most described in patients having a t(11;22), the construct of the fusion probe may be uninformative for this case (breakpoint may lie outside the region targeted by the probe) (Fig. 5). Thus, one cannot

Fig. 4. Tumor cells are strongly and diffusely positive for CD99 with a distinct membranous pattern as well as cytoplasmic staining (high power).

E. Dragoescu et al. / Annals of Diagnostic Pathology 17 (2013) 377–382

A

379

Normal chromosome 22

3’ 3

T590 00 UT

W-8762

5’ 5

Telomere

D22 2S5 590

D22S113

W-8 8513 3

D22 2S4 448

Centromere

Probe appearance

MTMR3

EWSR1 1100 kb

500 kb

Break-apart 139 kb

151 kb 100 kb

UT5 5900 0

3’

Telomere

B k i t? Breakpoint? W-8 8762 2

5’

D22 2S59 90

W-8 W 8513 3

500 kb

13 D22 2S11

Rearranged chromosome 22

C t Centromere

D22 2S44 48

B

Fusion

MTMR3

EWSR1

Break-apart

139 kb

151 kb

Fusion

100 kb

C

Break-apart probe

D

Fusion probe

Fig. 5. Interphase FISH studies. The genes/markers encompassed by the break-apart probe (adapted from Abbott Molecular) and the fusion probe (adapted from Cytocell) are shown for: a normal chromosome 22 (A) and the rearranged chromosome 22 (B). The breakpoint for the chromosome 22 rearrangement is not known (indicated by ?) but is conjectured to lie distal to the DNA included in the fusion probe because only 2 signals were seen for this probe (rather than 3 that would be expected if the breakpoint occurred within the probe construct). The FISH patterns observed are shown schematically (right portion of A and B) and in representative images (C and D). The interphase nuclei hybridized with the breakapart probe (C) showed an abnormal pattern having an additional green signal, 1 normal fusion signal, and 1 small fusion signal. This aberrant pattern is consistent with the presence of a rearrangement involving sequences on the 3’ telomeric (green) portion of this probe. In contrast, a normal pattern was seen for the fusion probe (green signal is the EWSR1 probe, with the red signal being the FLI1 probe from chromosome 11). Collectively, these observations suggest that a rearrangement occurred in sequences that were telomeric to those included in the fusion probe and telomeric, but closely juxtaposed, to the 3’ portion of the EWSR1 gene.

rule out the possibility that this rearrangement involves the FLI1 gene based on the results of this FISH study. To further investigate the possibility of a rearrangement involving EWSR1 and FLI1, immunohistochemical staining of the biopsy material

was performed using a FLI1 rabbit polyclonal antibody. Diffuse nuclear staining of the tumor cells with FLI1 was observed (Fig. 6). Computed tomography of the patient's chest, abdomen, and pelvis was negative for metastatic disease, and a bone scan showed no evidence of metastatic disease to other bones. The patient underwent 12 cycles of chemotherapy at another institution before coming back to our institution for wide excision of the right proximal fibula. A 16.0-cm-long segment of right proximal fibula with the biopsy tract was removed en bloc (Fig. 7). Histologic evaluation revealed only 10% tumor necrosis post chemotherapy (Fig. 8) with tumor present at the distal bony resection margin and less than 0.1 cm from anterior, medial, and lateral margins due to close proximity of the right popliteal artery bifurcation to the tumor. Right above the knee amputation was performed 2 weeks later, which revealed residual foci of tumor in the soft tissue. The patient was living and well at the 6-month follow-up visit. 3. Comment

Fig. 6. Diffuse nuclear staining of the tumor cells with FLI1 antibody; lace-like osteoid and tumor necrosis present in the background (high power).

Review of the literature reveals an inconsistency between studies regarding the existence of EWSR1 gene rearrangements in small cell osteosarcoma, suggesting the need for delineation of this potential association. We are reporting here a patient having small cell osteosarcoma with a rearrangement of the EWSR1 gene that was detected by interphase FISH. Three other similar case reports were identified in the English literature [4-6] (Table).

380

E. Dragoescu et al. / Annals of Diagnostic Pathology 17 (2013) 377–382

Fig. 7. Right proximal fibula resection with attached soft tissue showing the tumor originating in the proximal epiphysis and metaphysis of the right fibula with cortical bone destruction and extension into soft tissue.

The first case was a short communication by Noguera et al [4] of a small cell osteosarcoma involving the acromion process of the right scapula in a 20-year-old man. Following a 24-hour culture, standard G-banding of 25 metaphase spreads revealed the presence of a t(11;22)(q24;q12) as well as an additional isochromosome for the long arm of chromosome 1 and trisomy 8. This chromosomal translocation in this small cell osteosarcoma appears to be identical to the t(11;22)(q24;q12) seen in 85% of cases from the Ewing sarcoma/peripheral neuroectodermal tumor family. This translocation results in the formation of a EWSR1-FLI1 chimeric fusion gene [3]. A second case of small cell osteosarcoma with EWSR1 gene rearrangement was reported by Oshima et al [5], who described a 37year-old man with multiple intraabdominal masses attached to the

anterior surface of the stomach, mesentery, omentum, and peritoneum. A biopsy revealed sheets of undifferentiated small round cells and lace-like osteoid produced by the tumor cells, with a morphology that was consistent with the diagnosis of extraskeletal small cell osteosarcoma. The tumor cells were positive for CD99 and reverse transcription–polymerase chain reaction (RT-PCR) showed a “single intense band that matched the sequence of EWS-FLI1 type I gene” [5] resulting in a 418-base pair product. Recently, Debelenko et al [6] reported a case of small cell osteosarcoma with a novel EWSR1-CREB3L1 fusion transcript. Their case was that of a 12-year-old girl with multifocal small cell osteosarcoma (involving the seventh thoracic vertebra, right femoral neck, skull, left tibia, left third rib) with pulmonary nodules and a calcifying upper retroperitoneal mass. Interphase FISH demonstrated a rearrangement of the EWSR1 gene in 91% of the scored nuclei. However, RT-PCR was negative for EWSR1-FLI, EWSR1-ERG, and EWSR1-WT1 fusion transcripts. Rapid DNA amplification identified a EWSR1-CREB3L1 fusion transcript, which was confirmed by polymerase chain reaction and gene sequencing. The breakpoint was located 3’ from exon 11 of the EWSR1 gene and 5’ from exon 6 of CREB3L1 gene resulting in an 86-base pair product. These findings of EWSR1 gene rearrangements in small cell osteosarcoma are in contrast with the results from a more recent study from several investigators from the European Union, in which 10 cases of small cell osteosarcoma were analyzed using immunohistochemistry and molecular genetic techniques [7]. Machado et al [7] report that FISH did not reveal EWSR1 gene rearrangements in any of their small cell osteosarcomas and RT-PCR using primers targeting exons 7, 10, and 12 of the EWS gene or exon 6 of the FLI1 gene did not show a EWS-FLI1 gene fusion. They also did not identify any EWSR1-ERG or EWSR1-WT1 fusion transcripts in the cases that they studied. It is interesting to observe that although similar methods were used in these studies, they yielded different results. For instance, interphase FISH using the same dual color-labeled break-apart EWSR1 probe was used by Machado et al [7], Debelenko et al [6], and our group. However, different results were obtained by the various investigators. Similarly, RT-PCR was positive for a EWSR1 fusion transcript in a subset of studies. We believe that the divergence observed using this technology reflects both biological variation as well as methodological differences (different exons, primers). These 3 case reports, using different methodologies, indicate that a subset of small cell osteosarcomas demonstrate rearrangements of the EWSR1 gene, with these rearrangements involving either a EWSR1FLI1 fusion product or a novel EWSR1-CREB3L1 fusion. Our report documents the existence of another small cell osteosarcoma case with an apparent rearrangement near the 3’ telomeric portion of the EWSR1 gene. This rearrangement could result in a potential chimeric fusion product. Considering these findings, small cell osteosarcoma can be added to the ever increasing list of tumors displaying EWSR1 gene rearrangements (including Ewing sarcoma/peripheral neuroectodermal tumor, desmoplastic small round cell tumor, clear cell sarcoma, angiomatoid fibrous histiocytoma, extraskeletal myxoid chondrosarcoma, and a subset of myxoid liposarcomas) [3]. 3.1. Practical considerations in the diagnosis of small cell osteosarcoma

Fig. 8. Residual viable tumor status post chemotherapy in the resection specimen. Note the infiltration of the preexisting bony trabeculae by the tumor cells. Abundant pink, lace-like osteoid matrix is easily identified in the resected tumor as compared with the biopsy (intermediate power).

Given the rarity of small cell osteosarcoma, there is limited experience among pathologists in recognizing this entity as well as a paucity of information available about prognosis/optimal treatments for patients having this type of tumor. It can thus be helpful to review morphologic aspects of this tumor as well as the differential diagnosis, usefulness of ancillary studies, and treatment implications. Small cell osteosarcoma is predominantly composed of uniform, small-sized, round-oval–shaped tumor cells (with size ranging from

E. Dragoescu et al. / Annals of Diagnostic Pathology 17 (2013) 377–382

381

Table Summary of literature review Source, y

Age, y/sex

Site

Standard G-banding

RT-PCR

Interphase FISH

Noguera et al, 1990 Oshima et al, 2004

20/M 37/M

t(11;22)(q24;q12)⁎ NA

NA Positive for EWS-FLI1 fusion transcript

NA NA

Debelenko et al, 2011

12/F

Right acromion Multiple intra-abdominal tumors Multiple bone lesions and metastases

NA

Negative for EWSR1-FLI1, EWSR1-ERG, and EWSR1-WT1 fusions transcripts

Positive for balanced EWSR(22q12) rearrangement in 91% of the scored nuclei

Machado et al, 2010

Current case

10 cases (5-48 [mean 26 y])/ 6F + 4M 17/M

Long bones (4), pelvis (3), spine (1), rib (1), knee (1)

NA

Right fibula

NA

Positive for EWSR1-CREB3L1 fusion transcript (exon 11 breakpoint) Negative for any EWS fusion transcripts (multiple primers used) (exons 7, 10, 12) NA

Negative for EWSR1(22q12) rearrangement Positive for unbalanced EWSR1(22q12) rearrangement with gain of a 3´ telomeric signal in 41% of the nuclei

Abbreviations: M, male; F, female; NA, not available. ⁎ Monoclonal karyotype: 48, XX, +i(1)(q10),+8, t(11;22)(q24;q12).

6.5-7 μm up to 12.5 μm). In a minority of cases, the tumor cells are spindle shaped [1,8-10]. Production of tumor matrix (osteoid) by these small tumor cells is mandatory for the diagnosis of small cell osteosarcoma. This morphology is in contrast to conventional osteosarcoma, which is characterized by large, markedly atypical, variable-sized, pleomorphic spindle tumor cells. However, because of its uniform, small-sized tumor cell morphology and focal osteoid production, small cell osteosarcoma is more likely to be confused with other small round cell malignancies of the bone (malignant nonHodgkin lymphoma, mesenchymal chondrosarcoma, or Ewing sarcoma). In fact, small cell osteosarcoma is a morphologic variant of conventional high-grade osteosarcoma, as it shares a similar clinical presentation (affecting more commonly young patients in the second and third decades of life), pattern of skeletal involvement (metaphysis of long bones, intramedullary location with cortical destruction and associated soft tissue mass), and roentgenographic aggressive features [1,8,9]. The presence of osteoid is very helpful in sorting out the differential diagnosis; however, in most small cell osteosarcomas (51.4%-63.6%) [1,8,11], the osteoid production is scant, usually lace like. Furthermore, distinguishing tumor osteoid matrix from hyalinized collagen, fibrin, or reactive periosteal bone formation can be difficult or subjective in some cases, especially in small biopsies [11,12]. A useful clue is the identification on plain radiologic films or computer tomography of sclerosis (mineralized tumor matrix) within the intramedullary part of the tumor or the associated soft tissue mass. This finding strongly supports the diagnosis of osteosarcoma [1]. There are some subtle morphologic details that help the pathologist in reaching a correct diagnosis. For example, recognizing a hemangiopericytomatous vascular pattern (thin-walled, dilated, and branching vessels surrounded by nests of small, uniform tumor cells) should alert the pathologist to the alternate diagnostic possibility of mesenchymal chondrosarcoma, even in the absence of low-grade hyaline cartilage in a biopsy sample [12]. This hemangiopericytomatous vascular pattern can also be seen in 33.3% to 54.5% of small cell osteosarcomas [1,8]; therefore, this feature needs to be interpreted with caution in the overall context of the case. In addition, it is important to point out that a minority of small cell osteosarcomas (5.6%-27.2%) may also have chondroid matrix in addition to the mandatory osteoid matrix, as in our case [1,8,9,11]. However, the chondroid matrix in a case of small cell osteosarcoma is morphologically different from the low-grade cartilage present in mesenchymal chondrosarcoma; specifically, it is highly cellular with marked cytologic atypia [9]. Although there is no specific immunoprofile for small cell osteosarcoma [11], there are some immunohistochemical stains that

can be useful, such as hematolymphoid markers. On the other hand, monoclonal antibody directed against membranous glycoprotein CD99 (O13 or 12E7), while being sensitive for Ewing sarcoma, is not specific, as it is expressed in a variety of other neoplasms, including small cell osteosarcoma [13]. Recently, Lee et al [2] reported that FLI1 monoclonal antibody was not expressed in any of the 10 small cell osteosarcomas tested in their study. This is in contrast with our finding of positive staining of small cell osteosarcoma with FLI1, an intriguing observation that needs further studies to settle this conflicting result. However, it is important to mention that 25% of Ewing sarcomas are also negative for FLI1. Therefore, a negative result needs to be interpreted with caution. In addition, excessive decalcification of bone specimens may lead to false-negative FLI1 results due to loss of antigen preservation [2]. The bone matrix glycoprotein osteonectin, although constantly present in osteogenic lesions, is expressed in small cell osteosarcoma only in 30% to 50% of the tumor cells [14]. Although there are conflicting data about osteonectin expression in other bone tumors, the data indicate that Ewing sarcoma is negative for osteonectin [14]. Because small cell osteosarcoma has a worse prognosis than conventional osteosarcoma (77% 5-year survival rate, 68% overall survival) and Ewing sarcoma (50% 5-year survival rate) [1], an accurate diagnosis is paramount. In the study from Istituto Rizzoli from 1989 [8], only 2 (18.1%) of 11 patients were alive (one undergoing treatment at the time of the study and the other one disease free at 2 years). In a review study from Mayo Clinic, the overall 5-year survival rate for small cell osteosarcoma was 28.5%, although for patients treated after 1975, when adjuvant chemotherapy was implemented for treatment of osteosarcoma, the 5-year survival rate was much higher (68.6%) [1]. The differing treatment modalities underscore the importance of distinguishing small cell osteosarcoma from Ewing sarcoma. There is no standard protocol specifically designed for small cell osteosarcoma; however, this tumor is treated in a fashion similar to that for high-grade conventional osteosarcoma. The recommended treatment for high-grade osteosarcoma based on the National Comprehensive Cancer Network (www.nccn.org) guidelines consists of multiagent neoadjuvant chemotherapy (cisplatin, doxorubicin, high-dose methotrexate) followed by wide excision and adjuvant chemotherapy. In contrast, the recommended treatment for Ewing sarcoma based on the same guidelines includes a different chemotherapy regimen (vincristine, doxorubicin, cyclophosphamide alternating with ifosfamide and etoposide) followed by wide excision and/or radiotherapy. Although small cell osteosarcoma has not been considered to be radiosensitive, there have been some reports that document a benefit when chemotherapy is combined with local radiotherapy [15].

382

E. Dragoescu et al. / Annals of Diagnostic Pathology 17 (2013) 377–382

In summary, an accurate diagnosis of small cell osteosarcoma is not without difficulties due to morphologic, immunohistochemical, and molecular overlap with other tumors, in particular Ewing sarcoma. Small cell osteosarcoma has biological heterogeneity at the cytogenetic level, showing EWS gene rearrangements with and without FLI1 gene partnership and not always involving the most common exons. Interphase FISH using break-apart probes may be a more effective methodology in detecting variant forms. From a practical standpoint, all aspects of a given case, from clinical and radiologic presentation to tumor morphology, need to be considered to avoid errors and to achieve a correct diagnosis. On a more theoretical level, finding EWSR1 gene rearrangements in a subset of small cell osteosarcomas is intriguing. Small cell osteosarcoma and Ewing sarcoma share many clinical and morphologic similarities, whereas other soft tissue tumors harboring EWSR1 gene rearrangements are morphologically and immunohistochemically distinct from Ewing sarcoma. It is tempting to speculate that these 2 tumors may share an early common pathway of tumor development at the cytogenetic level before the accumulation of additional genetic alterations causes their divergence. However, more data on these rare tumors are needed to test this intriguing hypothesis. Acknowledgment Authors would like to thank Mrs. Patricia Strong, Director of Writing Center and Assistant Professor in University College at Virginia Commonwealth University, Richmond, Virginia, for critical review of the manuscript and useful suggestions.

References [1] Nakajima H, Sim FH, Bond JR, Unni KK. Small cell osteosarcoma of bone. Review of 72 cases. Cancer 1997;79:2095-106. [2] Lee AF, Hayes MM, Lebrun D, et al. FLI-1 distinguishes Ewing sarcoma from small cell osteosarcoma and mesenchymal chondrosarcoma. Appl Immunohistochem Mol Morphol 2011;19:233-8. [3] Romeo S, Dei Tos AP. Soft tissue tumors associated with EWSR1 translocation. Virchows Arch 2010;456:219-34. [4] Noguera R, Navarro S, Triche TJ. Translocation (11;22) in small cell osteosarcoma. Cancer Genet Cytogenet 1990;45:121-4. [5] Oshima Y, Kawaguchi S, Nagoya S, et al. Abdominal small round cell tumor with osteoid and EWS/FLI1. Hum Pathol 2004;35:773-5. [6] Debelenko LV, McGregor LM, Shivakumar BR, Dorfman HD, Raimondi SC. A novel EWSR1-CREB3L1 fusion transcript in a case of small cell osteosarcoma. Genes Chromosomes Cancer 2011;50:1054-62. [7] Machado I, Alberghini M, Giner F, et al. Histopathological characterization of small cell osteosarcoma with immunohistochemistry and molecular genetic support. A study of 10 cases. Histopathology 2010;57:162-7. [8] Bertoni F, Present D, Bacchini P, Pignatti G, Picci P, Campanacci M. The Istituto Rizzoli experience with small cell osteosarcoma. Cancer 1989;64:2591-9. [9] Ayala AG, Ro JY, Raymond AK, et al. Small cell osteosarcoma. A clinicopathologic study of 27 cases. Cancer 1989;64:2162-73. [10] Unni KK, ed. Tumors of the Bones and Joints (AFIP Atlas of Tumor Pathology, Series 4, Vol 2), 2006. [11] Devaney K, Vinh TN, Sweet DE. Small cell osteosarcoma of bone: an immunohistochemical study with differential diagnostic considerations. Hum Pathol 1993;24:1211-25. [12] Amukotuwa SA, Choong PF, Smith PJ, Powell GJ, Thomas D, Schlicht SM. Femoral mesenchymal chondrosarcoma with secondary aneurysmal bone cysts mimicking a small-cell osteosarcoma. Skeletal Radiol 2006;35:311-8. [13] Devaney K, Abbondanzo SL, Shekitka KM, Wolov RB, Sweet DE. MIC2 detection in tumors of bone and adjacent soft tissues. Clin Orthop Relat Res 1995;310: 176-87. [14] Serra M, Morini MC, Scotlandi K, et al. Evaluation of osteonectin as a diagnostic marker of osteogenic bone tumors. Hum Pathol 1992;23:1326-31. [15] Sipos EP, Tamargo RJ, Epstein JI, North RB. Primary intracerebral small-cell osteosarcoma in an adolescent girl: report of a case. J Neurooncol 1997;32:169-74.