Difficulties of Differentiating Primary Histiocytic Sarcoma of the Central Nervous System from Glioblastoma in Older Patients: Methylation Analysis and Review of the Literature

Ryota TAOMOTO; Mikiko AOKI; Toshiyuki ENOMOTO; Noriko HIRAO; Hiroshi ABE; Junji SHIBAHARA; Koichi ICHIMURA; Kaishi SATOMI; Makoto HAMASAKI

doi:10.2176/jns-nmc.2024-0335

Abstract

Primary histiocytic sarcoma of the central nervous system is an extremely rare malignancy, with only 43 cases reported to date. It often presents diagnostic challenges due to its rarity and similarities with other central nervous system tumors, particularly glioblastoma. We report a case of primary central nervous system histiocytic sarcoma in a 72-year-old woman, one of the older patients among the reported cases. The patient presented with gait disturbance and left hemiplegia. Initial imaging and intraoperative findings suggested a high-grade glioma. However, comprehensive pathological examination, including immunohistochemistry, electron microscopy, and genetic analyses, led to the diagnosis of histiocytic sarcoma. Notably, the tumor exhibited CDKN2A homozygous deletion and a unique methylation profile distinct from gliomas. Despite treatment with surgery, radiation, and chemotherapy, the patient experienced meningeal dissemination 8 months post-surgery and died 11 months after diagnosis. This case highlights the importance of comprehensive diagnostic evaluation in challenging central nervous system tumors and provides valuable insights into the molecular characteristics of central nervous system histiocytic sarcoma, particularly regarding CDKN2A deletion and methylation patterns. Awareness of this rare entity is crucial for accurate diagnosis and appropriate management of central nervous system tumors.

Introduction

Histiocytic sarcoma (HS) is an invasive malignant tumor with histiocyte-like morphology and an immunophenotype. Common sites of occurrence include the gastrointestinal tract, skin, soft tissues, and lymph nodes; central nervous system (CNS) involvement is rare, with only 43 cases reported to date, according to our PubMed search (Table S1).¹^-³⁸⁾

Histologic findings in primary HS of the CNS include a high cell density, proliferation of highly atypical histiocytes with eosinophilic cytoplasm and well-defined nucleoli, mitotic figures, and necrosis.¹⁾ Tumor cells are highly pleomorphic and may include multinucleated giant cells with spindle-shaped or rhabdoid morphology. Some reports suggest that primary CNS tumors tend to show inflammatory infiltrates and abscess formation.³⁴⁾ Differential diagnoses included lymphoma, meningioma, glioblastoma (GBM), malignant Langerhans cell histiocytosis, metastatic melanoma, and dendritic cell sarcoma. Immunostaining is essential in many cases, as a diagnosis is difficult to make based on tissue morphology alone.³⁶⁾

We report a case of HS in older person, initially identified as GBM based on histologic findings.

Case Report

A 72-year-old woman visited her local doctor with a complaint of gait disturbance. A neoplastic lesion was found on the parietal side of the right frontal lobe. An endoscopic tumor biopsy was performed, but due to the hardness of the tumor, it was not possible to obtain sufficient tissue. Consequently, a diagnosis could not be reached, and the patient was discharged home. One month after discharge, the patient's left hemiplegia worsened, and she was readmitted to the hospital.

Medical history

Appendicitis surgery at age 16, coronavirus disease 2019 infection at age 71.

Present on admission

Clear consciousness, left hemiplegia (manual muscle testing: 1/5).

Imaging findings

Head magnetic resonance imaging (MRI) showed a 30 x 40 mm tumor lesion on the parietal side of the right frontal lobe. The tumor showed an equal signal on T1-weighted image (T1WI) and a heterogeneous high signal on T2-weighted image (T2WI). There were edematous changes around the tumor and some hemorrhagic changes inside the tumor. Gadolinium contrast showed heterogeneous internal enhancement, and positron emission tomography-computed tomography showed highly abnormal accumulation (Fig. 1A-C). Differential diagnoses on imaging included high-grade glioma, malignant lymphoma, and metastatic tumor.

Fig. 1

Imaging findings. A 40 × 30 mm large neoplastic lesion is seen on the parietal side of the right frontal lobe. MRI T2WI showed a heterogeneous high signal (A), and MRI Gd contrast showed heterogeneous internal enhancement (B). PET-CT showed abnormal accumulation (C). The tumor was grayish-white in color and very firm, with a clear border with the surrounding brain tissue (D).

MRI: magnetic resonance imaging; PET-CT: positron emission tomography-computed tomography; T2WI: T2-weighted image

Intraoperative findings

A craniotomy was performed. The tumor was grayish-white in color, very firm, and clearly demarcated from the surrounding brain tissue (Fig. 1D). The tumor was carefully resected, preserving the area near the motor cortex.

Pathological findings

Intraoperative diagnosis revealed a highly cellular tumor with necrosis (Fig. 2A). Although palisading necrosis and microvascular proliferation were not evident, cellular atypia, together with preoperative imaging findings, led to the diagnosis of a high-grade glioma.

Fig. 2

Pathological findings. Intraoperative rapid diagnosis showed high cell density. Strong nuclear atypia and necrosis were seen (A: H&E staining, ×200). Permanent specimens with high cell density showed disorganized growth of tumor cells with eosinophilic cytoplasm and well-defined nucleoli, as well as a fission picture (B: H&E staining, ×400). Necrosis was observed (C: H&E staining, ×100), and an image of erythrocytes incorporated into the sporangia of atypical cells (arrows) (D: H&E staining, ×400). In areas rich in stromal components, collagen proliferated (E: H&E staining, ×100) (F: MT staining, ×100).

H&E: hematoxylin and eosin; MT: Masson-trichrome

Formalin-fixed specimens showed highly cellular areas and stromal component-rich areas. In the highly cellular areas, tumor cells with eosinophilic cytoplasm and well-defined nucleoli were randomly proliferating. Conspicuous nuclear atypia was high. Numerous mitotic figures and foci of necrosis were observed (Fig. 2B and C). Emperipolesis of erythrocyte-like findings was seen in some areas (Fig. 2D). In areas rich in stromal components, collagen stained blue by Masson-trichrome staining was observed (Fig. 2E and F).

Immunohistochemically, the tumor cells were positive for histiocyte markers (CD68 and CD163; Fig. 3A and B) and S100, while all 30 remaining markers (including glioma, lymphoma, myeloid cells, dendritic cells, epithelial cells, and melanoma) were negative. Immunohistochemically, isocitrate dehydrogenase 1 (IDH-1) was negative. We did not consider isocitrate dehydrogenase 2 (IDH-2) mutation because the patient was 72 years old, and other glial markers were negative. Based on immunostaining results, a diagnosis of HS was made.

Fig. 3

Tumor cells were positive for CD68 (KP-1) (A: ×200) and CD163 (B: ×200) FISH (C). Red signals indicate 9p21, and green signals indicate chromosome 9 kinetochore. Cells lacking red signals and showing only green signals had CDKN2A homozygous deletions. In total, 79.2% (95/120) of the cells showed CDKN2A homozygous deletions. Electron microscopy of reconstructed formalin-fixed, paraffin-embedded sections showed nuclei with slits, and lysosomes and rough endoplasmic reticulum were conspicuous (D).

FISH: Fluorescence in situ hybridization

Fluorescence in situ hybridization (FISH) analysis showed CDKN2A homozygous deletion (95/120, 79.2%; Fig. 3C). Electron microscopy of reconstructed formalin-fixed, paraffin-embedded sections showed nuclei with slits, numerous ribosomes, and rough endoplasmic reticulum, indicating histiocytic differentiation (Fig. 3D).

When methylation analysis was performed, 12.8% of the brain classifier results did not find any matches. Similarly, in the Sarcoma classifier, v.12.3, no matches were found. However, in the methylation class, Langerhans cell histiocytosis was considered during differential diagnosis. In t-distributed stochastic neighbor embedding (tSNE) analysis, the sample was located near hemangioblastoma in the CNS classifier but did not match, while in the Sarcoma classifier, it was positioned in the vicinity of Kaposi sarcoma (Figure S1).

Postoperative course

The patient was treated with temozolomide and intensity-modulated radiation therapy (60 Gy/30). After discharge, the patient continued temozolomide therapy in our outpatient clinic. At 6 months postoperatively, there was no recurrence, but at 8 months postoperatively, meningeal dissemination was observed. Avastin was attempted based on GBM treatment protocols because there is no established standard treatment for HS, and due to its extreme rarity, there are few reported cases of treatment. However, her disease progressed, and she died 11 months after diagnosis.

Discussion

Forty-three cases of primary HS of the CNS reported to date were reviewed (Table 1). Their ages ranged from 17 months to 84 years; 34 of the 43 cases (79%) had adult onset, with a median age of 47 years. This case is the second oldest case reported to date. The prognosis was very poor, with a median survival of 7 months. Fourteen (33%) cases survived for more than 12 months, of whom 12 (28%) experienced sporadic cases. The median survival for solitary and multiple cases was 10 and 6 months, respectively, suggesting that total resection and radiation chemotherapy may have some effect on solitary cases. Treatment options include surgery, radiation therapy, and chemotherapy, but due to the small number of cases, standard treatment protocols have not yet been established. Because of reports of long-term prognosis with temozolomide, the patient received postoperative temozolomide and radiation therapy and had no recurrence at 6 months postoperatively; however, meningeal dissemination was noted at 8 months postoperatively. Avastin was attempted based on GBM treatment protocols, but the disease progressed, and the patient died 11 months after diagnosis. At present, considering the extremely poor prognosis of intracranial HS in older patients, transition to best supportive care prioritizing quality of life may be advisable, depending on the patient's condition.

Table 1

Summary of reported cases of CNS histiocytic sarcoma

	n = 43
1 The number of lesions exceeds 100% because multiple cases are included. 2 Eight extracranial metastases (lung, spine, iliac, mediastinum, hip, and bone marrow).
Age	Median 47 years (range 15 months –84 y)
>20 years	34 (79.1%)
<20 years	9 (20.9%)
Sex
Male	22 (51.2%)
Female	21 (48.8%)
Number of lesions
Solitary	27 (62.8%)
Multiple	16 (37.2%)
Size	Median 28 mm (range 7–77)
Site of occurrence*1
Over a tent	31 (72.1%)
Under a tent	15 (33.9%)
Overall survival	Median 7 months (range 0.6–60)
Dissemination and metastasis*2	11 (25.6%)

Primary HS of the CNS is uniformly or heterogeneously enhanced internally on gadolinium contrast when there are signs and symptoms of a malignant brain tumor, but without specific findings. Gross findings are usually white-toned soft tumors with occasional yellow-toned necrotic tissue.¹⁾ In cases with a cystic component, an abscess-like content can be seen within the cyst, and infectious disease may be a differential diagnosis.²⁸⁾ Gross intraoperative findings in this case showed that the tumor was very firm and clearly differentiated from the brain tissue. During the biopsy at the previous hospital, adequate tissue sampling was not possible due to the firm consistency of the tumor. Permanent specimens showed collagen proliferation, which could explain the stiffness of the tumor (Fig. 2E and F). There have been reports of reactive glial cell proliferation with fibrosis³²⁾ and collagen fibers on electron microscopy,²⁹⁾ but neither of these reports describes the gross findings. This is probably the first case in which histologically confirmed collagen proliferation (Fig. 2E and F) has been reported along with gross findings.

A diagnosis is made when histiocyte markers (CD68, CD163, lysozyme, CD11c, and CD14) are positive and other markers (myeloid antigens, dendritic antigens, CD30, ALK, and other lymphoid markers, glial, epithelial, melanocytic antigens, and follicular dendritic cell antigens) are negative. The histiocyte marker CD163 is said to be of particularly high diagnostic value.¹⁾ Regarding gene mutations, BRAF V600E mutation,²⁹⁾ PD-L1/PD-L2 expression,³¹⁾ loss of PTEN activity, and deletion of CDKN2A³³^,³⁵^,³⁸⁾ have been reported in the past. In particular, abnormalities in the tumor suppressor genes PTEN and CDKN2A have been reported as potentially involved in the development of CNS HS.³³⁾ The present case showed CDKN2A homozygous deletion. Previously, only 3 cases of CDKN2A homozygous deletion and one case of CDKN2A/B copy number alternation in CNS HS have been reported. The present case resulted in death 11 months after surgery. Similarly, the 3 previously reported cases of HS with homozygous deletion of CDKN2A also died within a short period of 7-16 months.³³^,³⁵^,³⁸⁾ In a report of genetic analysis of systemic HS, CDKN2A homozygous deletions were the most frequent, occurring in about 40% of patients;³⁹^,⁴⁰⁾ however, no correlation with prognosis was noted. Whether CDKN2A homozygous deletion could be a prognostic factor for CNS HS remains unclear due to the very small number of reported cases.

Immunohistochemically, this case was considered to be HS; however, due to its extremely poor differentiation, the possibility of GBM with metaplastic changes could not be completely ruled out. Moreover, methylation analysis revealed a profile entirely different from the GBM and glioma groups. DNA methylation microarray analysis was performed using the Infinium Methylation EPIC v2 BeadChip Kit (8xBead Chip Cat; WG-317-1001) (Illumina, San Diego, CA, USA), following the manufacturer's instructions. The detailed protocol has been described previously.⁴¹^-⁴³⁾ In the sarcoma analysis of this case, methylation class Langerhans cell histiocytosis was raised as a differential diagnosis, and in tSNE, it was positioned in proximity to Kaposi sarcoma. Since these tumors can be ruled out morphologically, the diagnosis of this case was considered definitive. There were only 2 other reported cases of primary CNS HS that underwent methylation analysis besides this case. In the results of the methylation analysis, similar to this case, there were no definite matches in the DKFZ sarcoma classifier. Among these, one case showed fairly high alignment with Langerhans cell histiocytosis, which was very similar to the results of this case.³⁸⁾

Conclusion

We report a case of primary CNS HS, an extremely rare tumor that can be potentially misdiagnosed as GBM. This case, involving one of the older patients reported, required extensive investigation, including immunohistochemistry for accurate diagnosis. Our analysis provides valuable insights into the molecular characteristics of CNS HS. Awareness of this rare entity and its features is crucial for correct diagnosis and appropriate treatment, highlighting the importance of comprehensive evaluation in challenging CNS tumors.

Supplementary Material

https://doi.org/10.2176/jns-nmc.2024-0335

Author Contributions

RT and MA drafted the manuscript. MA and MH supervised and reviewed the manuscript. MA and MH performed the histological examination. TE, HA, and NH treated the patient. JS, KI, and KS performed the methylation analysis. Authenticity of all the raw data has been assessed by RT, TE, and MA and MH to ensure its legitimacy. All authors read and approved the final version of the manuscript.

Ethics Approval

Not applicable.

Consent to Participate

The authors obtained written informed consent from the patient.

Patient Consent for Publication

The present case report was conducted according to the principles of the Declaration of Helsinki. All data were anonymized. Comprehensive written informed consent was obtained directly from the patient at 2 time points: prior to surgery and before initiating chemotherapy.

Availability of Data and Material

Data sharing is not applicable to this article, as no datasets were generated or analyzed during the present study.

Conflicts of Interest Disclosure

The authors have no conflicts of interest to declare.

References

責任著者(Corresponding author)

J-STAGEへの登録はこちら（無料）