Abstract
The histological diagnosis of malignant peripheral nerve sheath tumor (MPNST) is challenging because of the wide morphological spectrum and suboptimal performance of conventional immunohistochemical markers. MPNST arising primarily in the bone is exceptional, and its definitive diagnosis, particularly out of the neurofibromatosis type 1 (NF1) context, is even more problematic. Recurrent inactivation of EED or SUZ12 in a majority of MPNSTs results in a complete loss of trimethylated histone H3 at lysine 27 (H3K27me3) immunoreactivity, making it a highly specific biomarker of MPNSTs. In this article, we report a case of sporadic MPNST of the proximal femur that showed complete loss of H3K27me3. The patient was treated with limb-sparing surgery and postoperative radiotherapy. He developed multiple lung and bone metastases 4 months after surgery. Our case confirms the utility of H3K27me3 immunohistochemistry to yield a definitive diagnosis of sporadic MPNST in a rare primary site.
Introduction
Malignant peripheral nerve sheath tumor (MPNST) is an uncommon sarcoma that accounts for less than 5% of sarcomas. MPNST typically develops from the soft tissues, and development primary to the bone is exceptional. Primary MPNST of the bone has been most commonly reported in the mandible and spine,1,2 likely associated with the mandibular and spinal nerves that transgress the skeleton, respectively. In contrast, only 15 cases of primary osseous MPNSTs have been reported in the extremities, such as the femur, ulna, and humerus.3-15 Although approximately 50% of MPNSTs arise in patients with neurofibromatosis type 1 (NF1), 16 most MPNST of the bone is sporadic. However, the accurate diagnosis of MPNST of the bones in sporadic settings remains challenging due to its wide morphologic spectrum and the lack of specific markers. Recently, a loss of function somatic alterations in components (EED or SUZ12) of the polycomb repressive complex 2 (PRC2) was identified in MPNSTs. 17 Because PRC2 is essential for establishing and maintaining the trimethylation at lysine 27 of histone H3 (H3K27me3), MPNSTs with PRC2 inactivation show complete loss of H3K27me3 on an immunohistochemical analysis, and this pattern of immunoreactivity has been shown to be specific to MPNST.17-20 In our recent immunohistochemical study, 20 we identified an MPNST of the femur that lacked H3K27me3 expression. In the present report, we document the clinicopathological details of that exceptional but educational case to illustrate the utility of this new biomarker in the difficult diagnostic setting.
Case Presentation
The patient provided his written consent for this case publication. The patient was a 54-year-old man who presented with pain in his left hip joint for 2 weeks. He had no remarkable medical or family history and specifically lacked the signs of NF1. A plain radiograph showed an osteolytic lesion with a thinned cortex in the proximal femur (Figure 1A). Computed tomography (CT) revealed a mass with destruction of the posterior cortex and an extraosseous component that lacked calcification and ossification (Figure 1B). The patient subsequently experienced a fall, which resulted in acute pain in the left hip that kept him from ambulating. A plain radiograph showed a pathological fracture of left proximal femur (Figure 1C), and magnetic resonance imaging (MRI) revealed inflammatory changes surrounding the fracture as an enhanced lesion in gadolinium-enhanced T1-weighted images, and abnormal intensity as a mass lesion in the bone marrow of the proximal femur (Figure 1D). Fluorodeoxyglucose positron emission tomography-computed tomography revealed an osteolytic lesion in the right proximal femur around the fracture with a maximum standardized uptake value of 4.47. There was no evidence of distant metastasis.

Plain radiograph of the left femur (A) shows an osteolytic lesion with the cortical bone thinning (arrow). CT (B) shows the destructive changes in the cortical bone of the posterior femur and an extraosseous mass (arrowheads). Plain radiograph shows pathological fracture of the proximal femur (C). A coronal section of T1-weighted fat-suppressed postcontrast MRI show a well-enhanced lesion around the pathological fracture and in the bone marrow of the proximal femur (D).
A needle biopsy was performed, and it revealed a sarcoma comprising fascicular spindle cell proliferation without any osteoid formation. We performed limb-sparing wide resection with endoprosthetic replacement of the proximal femur (Figure 2A). We excised the quadriceps femoraris, adductors, and abductors, to obtain negative surgical margins (Figure 2B).

Postoperative radiograph of the left femur shows endoprosthetic replacement after wide resection of the tumor (A). A coronal section of the excised tumor shows a grey-white cutting surface around the fracture site (arrowheads) (B). Histologically, the tumor shows fascicular growth of malignant spindle cells. The tumor cells tend to concentrate around blood vessels (C; hematoxylin and eosin). The tumor cells show complete loss of H3K27me3 immunoexpression (D; H3K27me3 immunohistochemistry).
The surgical specimen showed a 7 × 5 × 4 cm tan, solid mass with hemorrhaging and necrosis in the femoral metaphysis. The tumor was extensively sampled using a standard mapping procedure, and 28 tumor sections were submitted for a microscopic examination. A histological analysis revealed swirling fascicles of malignant spindle cells forming alternating hypercellular and hypocellular areas, with the cellularity being often increased around blood vessels (Figure 2C). The tumor cells were long, spindled, and mildly to moderately pleomorphic and exhibited numerous mitoses (31 per 10 high-power fields at hot spots). Geographic necrosis was present. There was no heterologous component, such as bone or cartilage. The tumor was not associated with any identifiable nerve fibers or neurofibroma components. Immunohistochemically, the tumor cells were focally positive for AE1/AE3, epithelial membrane antigen (EMA), CDK4, CD34, and SOX10, while they were negative for S100, desmin, myogenin, and MDM2. H3K27me3 staining using Tri-Methyl-Histone H3 (Lys27) rabbit monoclonal (C36B11, dilution 1:200; Cell Signaling Technology, Danvers, MA) showed a complete loss of nuclear reactivity against an intact staining of the endothelial cells as an internal positive control (Figure 2D). A fluorescence in situ hybridization analysis using a Vysis SS18 Break Apart FISH Probe Kit (Abbott Molecular, Abbott Park, IL) showed no evidence of SS18 rearrangement. The histopathological, immunohistochemical, and molecular cytogenetic findings thus supported the diagnosis of MPNST primary to the proximal femur. Postoperative radiotherapy (60 Gy/30 Fr) was performed to reduce the risk of tumor contamination after fracture.
Two months after surgery, the patient was able to ambulate with 2 crutches. However, 4 months after surgery, he developed multiple pulmonary and bone metastases. Despite systemic chemotherapy using doxorubicin, denosumab, and radiotherapy to the spinal metastases, the metastatic lesions progressed. The patient is currently alive under palliative care with metastatic disease, but without local recurrence, 1 year after surgery.
Discussion
According to a few reports of MPNST of the bone, the most common primary sites were the facial bones and spine. Fifteen cases of MPNST of the bone in the extremities have been reported, and the clinical data are summarized in Table 1. Interestingly, only 2 of the 15 cases were associated with NF1.7,10 The peak incidence of MPNST in the long bones is in the second decade of life, with a preference for males. Twelve cases had surgical resection, and 7 cases were treated by amputation. Five of the 12 developed local recurrence even after the surgery. Three cases received multi-agent chemotherapy,7,8 and neoadjuvant radiotherapy was given to 1 case. 15 As there were only a small number of cases treated with chemotherapy or radiotherapy, the effects of these adjuvant therapies are difficult to evaluate for MPNST of the bone. Five cases developed distant metastases, and the most common site was the lung.6-8,11,13 Three patients died of the disease7,8,11 (Table 1).
Summary of Reported Cases of MPNST of the Bone in Extremities.
Abbreviations: MPNST, malignant peripheral nerve sheath tumor; NF1, neurofibromatosis 1; NED, no evidence of disease; NA, not available; DOC, death of other causes; MAID, mesna, adriamycin, ifosfomide, dacarbazine; DOD, death of disease.
In the literature, the histological diagnosis of MPNST of the bone was often made based on the characteristic histological features and immunoreactivity to Schwann cell markers. MPNSTs often harbor swirling fascicles of spindle cells, relatively monotonous tapering nuclei, geographic necrosis, numerous mitoses, hyaline bands, “marbled” heterogeneity in cellular density, and perivascular condensation of tumor cells. However, these characteristic features are not entirely specific to MPNSTs. Similarly, although 9 of 11 tumors reported were positive for S100 protein (Table 1), S100 expression is not sensitive for the MPNST diagnosis, as it is expressed in only 30% to 39% of cases21 -23 and often only in a small fraction of tumor cells. Sox10, another Schwann cell marker, was positive in 27% to 67% of MPNST cases,21-23 but this may also suffer from insufficient specificity. 21 The clinical history of NF1 may be helpful, as approximately 50% of MPNSTs arising in the soft tissue were associated with NF1. 16 However, only 2 cases of MPNSTs of bone in the extremities have been reported in association with NF1.7,10 Given the above, the diagnosis of MPNST of the bone in sporadic settings can be very difficult. Recent studies have shown that a complete loss of H3K27me3 labeling by immunohistochemistry is highly specific for the diagnosis of MPNST, despite moderate sensitivity, both in syndromic and sporadic cases.17-20 Most articles on this topic have agreed on the high specificity of this marker to MPNST.17 -20,24 Although one study 25 reported a less-specific performance, this discrepancy may be due to differences in staining interpretation, as the latter study 25 equated weak immunoreactivity to the “loss” of H3K27me3. Because MPNST of the bone is not clearly defined in the present World Health Organization classification, there can be a trend for the underdiagnosis or at least hesitation to make a definitive diagnosis of MPNST of the bone. The introduction of H3K27me3 immunohistochemistry into practice should enhance the recognition and diagnostic confidence of this rare subset of MPNSTs.
In conclusion, we present the first case of primary MPNST arising in the femur that showed a complete loss of H3K27me3 immunohistochemistry. This case highlights the significant contribution of this new staining target to the definitive diagnosis of MPNST of the bone outside NF1.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
