Distinguishing Characteristics of Benign Versus Malignant Intraosseous Schwannomas: A Comparative Study

ABSTRACT

Objectives

Benign and malignant intraosseous schwannomas are rare, and primarily documented in case reports. This study aims to elucidate the differences in clinical features and imaging manifestations between these tumors. This will help clinicians identify malignant lesions at an early stage, reliable guide treatment decisions, and accurately predict outcomes.

Methods

Eighteen patients who underwent surgery and got pathological examinations in our hospital from 2012 to 2023 were retrospectively reviewed. Among them, 14 cases were found benign with 4 malignant. In the benign group, patients underwent curettage followed by bone grafting, whereas the malignant group was treated with extensive resection or amputation. Patients' demographics and radiographic features, including gender, age at diagnosis, symptom duration, tumor location, tumor margin, and the ratio of sclerotic margins were documented and compared between these tumors. All imaging was reviewed by two fellowship‐trained musculoskeletal radiologists, who also quantified the sclerotic margin ratio. The intraclass correlation coefficient was used to determine inter‐observer agreement. The Mann–Whitney U test was applied for continuous clinical variables, and the chi‐square test or Fisher's exact test for categorical variables.

Results

In our series, the mean age of these patients was 43.1 ± 14.0 years, six patients were male and 12 were female. Pain was the predominant preoperative symptom. The average duration from symptom onset to initial physician visit was 28.5 ± 25.3 months for benign schwannomas and 8.3 ± 4.3 months for malignant schwannomas (p = 0.012). On plain radiographs, 13 (13/14) of benign schwannomas exhibited well‐defined margins of bone destruction, compared to 1 (1/4) of malignant schwannomas (p = 0.019). Furthermore, benign schwannomas had a significantly higher sclerotic margin ratio (75.5%) than malignant ones (16.7%) (p = 0.001). No statistically significant difference was found between the two groups in terms of cortical bone destruction (p = 1.0). On MRI, both tumors demonstrated intermediate to slightly hypointense signal intensity on T1‐weighted images and heterogeneous high signal intensities on T2‐weighted images.

Conclusions

Despite their rarity, benign and malignant intraosseous schwannomas should be considered in the differential diagnosis for patients presenting with painful and radiographically lytic bone lesions, especially in the mandible, sacrum, and vertebrae. The ratio of sclerotic margins, which we proposed for the first time, in combination with symptom duration and the clarity of tumor margins, provide valuable diagnostic clues for distinguishing the malignancy of the tumors.

Benign and malignant intraosseous schwannomas are exceedingly rare, which is easily overlooked in the initial differential diagnosis, resulting in diagnostic delays that may reach several years. In this paper, we elucidate the differences in clinical features and imaging manifestations between these tumors. Through this article, we aim to contribute our experience to the scattered reports in the literature, thereby enhancing awareness of these tumors and facilitating future clinical diagnoses and treatment. To the best of our knowledge, the submitted manuscript is the first study to provide a detailed description and comparative analysis of these two types of tumors.

1. Introduction

Schwannomas, also known as neurilemmomas, are benign tumors originating from Schwann cells of the nerve sheath. They commonly occur in the skin or subcutaneous soft tissues, particularly in the head and neck regions [1]. Intraosseous schwannomas, which involve the bone, are extremely rare, accounting for approximately 0.2% of benign bone tumors [2, 3, 4]. The malignant form (malignant schwannoma), known as malignant peripheral nerve sheath tumor (MPNST), is even rarer as a primary lesion in the bone [5, 6, 7, 8, 9, 10, 11, 12]. To date, the relevant literature available is mostly case reports. Their rarity often leads to overlook during the initial differential diagnosis, resulting in diagnostic delays that may reach several years [13]. As is well known, such delays and misdiagnoses of the disease may promote tumor growth, complicate surgical interventions, and ultimately threaten patients' quality of life [5, 12, 14, 15].

In this study, we retrospectively reviewed 18 cases of benign and malignant intraosseous schwannomas treated in our hospital. The purpose of our study is to characterize the differences in clinical features and imaging appearance between these two types of tumors. We aim to contribute our experience to the scattered reports in the literature, thereby enhancing awareness of these rare tumors to aid in earlier diagnosis and treatment.

2. Methods

The research was approved by the Institutional Review Board of Beijing Jishuitan Hospital (No. 2024267‐00) and written informed consent was obtained from each patient.

The inclusion criteria included: (i) the treatment was conducted between January 2012 and December 2023; (ii) the diagnosis of benign and malignant schwannomas was confirmed by postoperative pathological examinations; (iii) the bone destruction was observed on x‐ray or computed tomography (CT).

The exclusion criteria included: (i) the lesions involving the skin or subcutaneous soft tissues; (ii) incomplete medical records; (iii) incomplete follow‐up information.

During the study period, 1555 patients with benign and malignant schwannomas underwent surgical treatment in our hospital. Two fellowship‐trained musculoskeletal radiologists reviewed all relevant imaging. Ultimately, 18 (1.2%) patients were considered that the bone was destructed by the tumor. Clinical, pathological, and radiographic information of these patients were retrospectively retrieved.

2.1. Evaluation Indicators

Clinical information included age at diagnosis, sex, duration of symptoms, preoperative and postoperative symptoms, history of hereditary disorders, treatments, and follow‐up results. Radiographic imaging was performed using x‐ray, CT, and magnetic resonance imaging (MRI). Imaging information included tumor location, tumor size, tumor margin, sclerotic margin, signal intensity, and contrast enhancement.

To quantitatively characterize the sclerotic margins, the optimal CT slice that clearly delineated the tumor boundary was selected. The tumor border was manually outlined and the perimeter was calculated automatically using ImageJ software. Subsequently, we outlined the sclerotic margins and measured their total length (Figure 1). The ratio of sclerotic margins was defined as the proportion of the sclerotic margin length to the tumor perimeter. Each measurement was repeated three times, and the average value was utilized for accuracy.

FIGURE 1.

The calculation of the ratio of sclerotic margins. (A) The perimeter of the tumor was delineated manually in blue colored line and automatically calculated using ImageJ software. (B) The length of sclerotic margins was depicted in green colored line. The ratio of sclerotic margins was defined as the length of the sclerotic margins divided by the perimeter of the tumor.

2.2. Treatment Strategy

For most cases of benign intraosseous schwannoma, curettage followed by bone grafting was the preferred approach. In contrast, all malignant cases underwent extensive resection or amputation. Subsequently, stabilization therapy with instrumentation was performed to maintain the stability following tumor removal.

2.3. Statistical Analysis

The intraclass correlation coefficient (ICC) was used to determine inter‐observer agreement. A score < 0.40 indicates poor agreement, 0.40–0.59 fair, 0.60–0.74 good, and 0.75–1.00 excellent agreement. The normality of continuous variables was assessed using the Shapiro–Wilk test. Data with a normal distribution was expressed as mean and standard deviation (SD), and for those that did not, the median (range) was used. The categorical variables were summarized as frequency and percentage. The Mann–Whitney U test was applied for continuous clinical variables, and the chi‐square test or Fisher's exact test for categorical variables, with statistical significance set at p < 0.05.

3. Results

3.1. Demographic Features

Eighteen patients with benign and malignant intraosseous schwannomas who underwent surgery in our hospital between 2012 and 2023 were retrospectively reviewed. Of these, 14 were benign and 4 were malignant. The characteristics of the 18 patients are presented in Table 1. The age of these patients at presentation ranged from 20 to 64 years (43.1 ± 14.0 years). There were 12 females and 6 males. In the benign group, the mean age was 42.8 years, with a male‐to‐female ratio of 5:9. For the malignant group, the average presenting age was 44.3 years, and the ratio of male to female was 1:3. The average duration from symptom onset to initial physician visit was 28.5 ± 25.3 months for benign schwannomas, contrasting with 8.3 ± 4.3 months for malignant ones (p = 0.012) (Table 2) Pain was the most common preoperative symptom (15/18). One patient was asymptomatic and incidentally discovered during a routine medical examination, and two patients presented with bowel or bladder incontinence.

TABLE 1.

Clinical and imaging features of the patients.

Pat	Sex/age (years)	Tumor	Location	Symptom	Size (cm)	Radiology	Herit. dis.	F/U
1	F/41	Benign	R capitate	Dull pain	2.2 × 2.5 × 1.3	Lytic, cortical disruption	No	NR
2	F/39	Benign	R ilium	Pain	NA	Lytic, sclerotic rim	No	NR
3	F/36	Benign	L femur	Intermittent pain	3 × 1.5 × 1.5	Lytic, thick sclerotic rim	No
4	F/48	Benign	C5 vertebra	Shoulder paresthesia and pain	4.2 × 2.6 × 2.2	Enlarged C5‐C6 foramina	No	NR
5	F/62	Benign	L4 vertebra	Pain	NA	Lytic, sclerotic rim	No	NR
6	F/32	Benign	T11 vertebra	Pain following low‐impact fall	3.9 × 4.6 × 7.0	T11 vertebral body to left pedicle lytic lesion had marginal sclerosis	No	NR
7	F/64	Benign	L4 vertebra	Incidental finding	NA	Lytic, sclerotic, cortical disruption	No	NR
8	M/63	Benign	Sacrum	Pain following low‐impact fall	NA	Sacral foramens expanded with expansile lytic bone destruction	No	NR
9	M/32	Benign	Sacrum	Intermittent pain	9.0 × 6.2 × 2.4	Sacral foramens expanded with expansile lytic bone destruction	No	NR
10	M/29	Benign	Sacrum	Low back pain with radiation	6.5 × 4.5 × 7.3	Lytic lesion with sacral foramina enlargement	No	NR
11	M/41	Benign	Sacrum	Intermittent pain	NA	Lytic lesion with cortical destruction	No	NR
12	F/55	Benign	Sacrum	Incontinence and numbness	NA	Expansile lytic lesion with cortical destruction	No	NR
13	M/28	Benign	Sacrum	Intermittent dull pain, frequent urination, and tenesmus	8.8 × 9.1 × 9.8	Lytic lesion, soft tissue mass	No	NR
14	F/29	Benign	Sacrum	Low back pain	4.5 × 5.1 × 3.7	Well‐defined lytic lesion with sclerotic rim	No	NR
15	M/40	Malignant	L metacarpal	Increasing pain	NA	Poorly defined osteolytic destruction	No	LR
16	F/57	Malignant	T7‐T8 vertebra	Back pain	9.0 × 7.0 × 5.0	Poorly defined osteolytic destruction	No	NR
17	F/20	Malignant	R fibula	Intermittent pain	NA	Expansile lytic, nonsclerotic	No	NR
18	F/60	Malignant	R humerus	Shoulder dull pain	6.0 × 4.0 × 4.0	Expansile lytic lesion	NF‐1	NR

Abbreviations: C, cervical; F, female; F/U, follow‐up; Herit. dis., hereditary disorders; L, left; L, lumbar; LR, local recurrence; M, male; NA, not available; NF‐1, neurofibromatosis type 1 syndrome; NR, no recurrence; Pat, patient; R, right; T, thoracic.

TABLE 2.

Benign versus malignant intraosseous schwannoma.

Demographic	Benign (n = 14)	Malignant (n = 4)	p
Female (n)	9	3	1.00
Age (years)	42.8 ± 13.3	44.3 ± 18.4	0.86
Symptom duration (months)	28.5 ± 25.3	8.3 ± 4.3	0.012 a
Well‐defined margin (n)	13	1	0.019 a
Sclerotic margin ratio (%)	75.5 ± 11.2	16.7 ± 8.4	0.001 a
Cortical destruction (n)	12	4	1.00

^a Statistically significant.

Benign intraosseous schwannomas predominantly occurred in the sacrum (7/14), followed by the lumbar vertebrae (2/14), and less frequently in the thoracic vertebra (Figure 2), cervical vertebra, femur, ilium, and capitate, each with 1/14 cases. The four malignant cases were located in the humerus, metacarpal bone, thoracic vertebra, and fibula.

FIGURE 2.

(Case 6, benign intraosseous schwannoma) A 32‐year‐old female had lower back pain after a fall while cycling 1 month ago. The x‐rays of the lumbar spine revealed no evidence of fracture, but an incidental finding of an abnormal morphology in the T11 vertebra was observed. (A) Coronal and (B) axial CT scan showing a lytic lesion with sclerotic margins of T11, which involves the vertebral body and its left appendages. The lesion displayed a large lobulated mass with low signal on sagittal T1‐weighted MRI (C) and relatively heterogeneous high signal intensities on sagittal T2‐weighted MRI (D). Anteroposterior (E) and lateral (F) radiographs showed satisfactory fixation and no recurrence at the 4‐year follow‐up.

3.2. Imaging Characteristics of Benign Versus Malignant Intraosseous Schwannoma

3.2.1. CT Findings

Thirteen (13/14) of benign schwannomas exhibited well‐defined margins of bone destruction, compared to 1 (1/4) of malignant schwannomas (p = 0.019). Furthermore, the ICC of inter‐observer reliability of the ratio of sclerotic margins was 0.83, indicating high consistency between the two observers. The benign schwannomas had a significantly higher sclerotic margin ratio (75.5%) than malignant ones (16.7%) (p = 0.001). No statistically significant difference was found between the two groups in terms of cortical bone destruction (p = 1.0).

3.2.2. MRI Findings

On MRI, both tumors exhibited isointense to slightly hypointense signal to muscle on T1‐weighted images and heterogeneous hyperintense signal on T2‐weighted images.

3.3. Follow‐Up

The median follow‐up duration was 12 months (range: 6 months to 5 years). Postoperatively, the symptoms of these patients were markedly alleviated. Unfortunately, one of the four malignant cases (case 15) experienced recurrence 8 months after the initial surgical resection, presenting with increasing pain in his right hand. Subsequently, a second operation was performed, which involved extensive excision of the lesion, including the proximal portions of the metacarpal bone and the distal carpal row of the right hand, and subsequent reconstruction with an autologous iliac crest graft and fixation with a titanium plate (Figure 3). To date, no further recurrence has been detected.

FIGURE 3.

(Case 15, malignant intraosseous schwannoma) A 40‐year‐old male complained of pain in his left hand for 1 year. He had no family and trauma history. (A) Anteroposterior x‐ray showing a slightly hypodense lesion at the base of the third metacarpal bone. (B) Intraoperative photograph showed the lesion was curettage with bone grafting. The pathology report suggested a low‐grade malignant peripheral nerve sheath tumor. Fourteen months after the surgery, the patient returned to our outpatient clinic for the reappearance of previous symptoms. Anteroposterior x‐ray (C) and (D) coronal CT scan demonstrated a lytic expansile lesion. The mass had an intermediate signal on the coronal T1‐weighted image (E) and hyperintense signal on the T2‐weighted image (F). Anteroposterior (G) and oblique (H) plain radiographs after 5 years of second surgery showed no sign of recurrence.

4. Discussion

In this study, we delineated the distinct clinical and radiographic characteristics that differentiate benign from malignant intraosseous schwannomas. We observed that malignant schwannomas were associated with a shorter symptom duration, poorly defined tumor margins, and a lower sclerotic margin ratio compared to their benign counterparts. These findings may be helpful for distinguishing the malignancy of the tumors.

4.1. The Clinical Features of Benign and Malignant Intraosseous Schwannoma

Osseous involvement in both benign and malignant schwannoma is exceedingly rare, with most studies limited to single case reports. de la Monte et al. suggested that this rarity may be attributed to the scarcity of intramedullary sensory nerve fibers [3]. Nonetheless, there has been an increase in documented cases, highlighting the need for further investigation. The etiology of bone involvement remains debated, but it is commonly linked to three potential mechanisms: tumors arising centrally within the medullary cavity, secondary invasion by extraosseous tumors, or originating from the nutrient canals, exhibiting a dumbbell‐shaped growth pattern [2, 3, 13]. Osseous involvement is most frequently reported in the mandible, followed by the sacrum. This is attributed to the mandibular nerve being a predominant sensory nerve and the multitude of sensorimotor nerve roots traversing the sacral foramina [16, 17, 18]. Other reported sites include the metacarpals, phalanges, proximal ulna, distal radius, and distal humerus [10, 11, 19, 20, 21, 22, 23]. In our orthopedic center, the spine was the most commonly affected area for intraosseous benign schwannoma, with half of the cases located in the lumbosacral region, consistent with previous studies [17, 24]. Interestingly, one benign intraosseous schwannoma of the capitate was included in our study, marking the first documentation of carpal involvement (Figure 4).

FIGURE 4.

(Case 1, benign intraosseous schwannoma) A 41‐year‐old female presented with dull pain in her right wrist for 3 years with no history of trauma. (A) Preoperative anteroposterior radiograph of the right hand showing an osteolytic expansile lesion in the capitate. (B) Preoperative sagittal, (C) coronal, and (D) axial computed tomography (CT) scan showing the destruction of the cortex of the capitate and a soft mass on its dorsal side (white arrow). (E) Intraoperative photograph showed that the extraosseous component of the tumor had a smooth surface, while the intramedullary portion exhibited a villous appearance. (F) Anteroposterior plain radiograph demonstrated lesion healing with successful integration of the autologous graft at the 6‐month follow‐up.

Pain is the most common presenting symptom of these tumors, which typically commence insidiously and may persist for years before diagnosis [2, 17]. In our study, the average duration from symptom onset to initial visit was significantly longer for benign tumors compared to malignant ones. While benign and malignant intraosseous schwannomas typically occur without gender predilection, our cohort had a slight female predominance. Neurofibromatosis type 1 (NF‐1) syndrome is reported to be the most significant risk factor for MPNSTs, with 52% of MPNSTs occurring in NF‐1 patients and 4.6% of NF‐1 patients developing malignant schwannomas in soft tissues [25]. In our series, only one malignant case was associated with NF‐1 (Figure 5).

FIGURE 5.

(Case 18, malignant intraosseous schwannoma) A 58‐year‐old woman with osteoporosis presented with a 5‐month history of pain in her right shoulder. She had a history of neurofibromatosis. Anteroposterior x‐ray (A), axial (B), and coronal (C) CT scan showing a lytic lesion in her right proximal humerus, characterized by cortical destruction, absence of sclerotic margins, and indistinct boundaries. Coronal T1‐WI (D) demonstrated a multicompartmental tumor exhibiting intermediate signal intensity, while the lesion displayed heterogeneous hyperintensity on T2‐WI (E). Anteroposterior radiograph (F) following extensive resection of the tumor and implantation of a shoulder prosthesis.

4.2. The Radiographic Characteristics of Benign and Malignant Intraosseous Schwannoma

Both benign and malignant intraosseous schwannomas typically present as expansile osteolytic lesions on plain radiographs and CT, occasionally with cortical expansion, and often exhibit a dumbbell‐shaped configuration. The bone cortex may be broken, with the lesion sometimes protruding expansively to form a soft tissue mass [10, 11, 16, 17, 25]. Consequently, differentiation between the two tumors may sometimes be challenging. In our case series, 13 of 14 benign schwannomas exhibited well‐defined margins of bone destruction, compared to 1 of 4 malignant schwannomas (p = 0.019). Additionally, the mean ratio of sclerotic margins was significantly higher in benign cases (75.5%) than in malignant ones (16.7%) (p = 0.001). This discrepancy likely reflects the more rapid and invasive growth of malignant tumors, as the sclerotic margin could be viewed as a bone remodeling to the expansion of the tumor, typically seen around the slow‐growing tumors. Thus, we believe that the delineation of tumor margins and the ratio of sclerotic margins are reliable indexes for distinguishing benign from malignant intraosseous schwannomas.

The MRI features of both intraosseous schwannomas are analogous to their soft tissue counterparts, typically presenting as intermediate to low signal intensity on T1‐weighted images and high, heterogeneous signal intensity on T2‐weighted, with heterogeneous enhancement following contrast administration [17]. Recently, diffusion‐weighted (DW) imaging has been proposed to offer more distinctive features for differentiating between benign and malignant soft tissue schwannomas [26, 27, 28]. Yun et al. reported that the mean apparent diffusion coefficient (ADC) value was significantly higher in benign peripheral nerve sheath tumors compared to malignant ones. The combination of mean ADC value and absence of split fat sign effectively distinguished between these tumors [26]. Regrettably, we did not perform DW imaging in this study to assess its utility for differentiating benign and malignant intraosseous schwannomas.

4.3. The Diagnosis, Treatment Strategies, and Prognosis of Benign and Malignant Intraosseous Schwannoma

The definitive diagnosis of benign and malignant intraosseous schwannomas is established based on histologic and immunohistochemical characteristics [2, 17]. Most benign schwannomas consist of Antoni A (align in palisades in highly cellular areas) and Antoni B (myxoid and less‐cellular areas). In contrast, malignant schwannomas are characterized by elongated, spindle‐shaped cells arranged in interlacing fascicles, along with high cellularity, nuclear pleomorphism, and increased mitotic activity [10, 25]. Immunohistochemical staining for S‐100 typically exhibits diffuse positivity in benign schwannomas, whereas in malignant cases, the pattern of S‐100 staining may vary [25]. Up to one‐third of malignant cases lack S‐100 immunoreactivity, or exhibit weak and focal staining. Benign intraosseous schwannomas generally have a favorable prognosis, with curettage and bone grafting being the preferred treatment [2, 24]. Conversely, malignant intraosseous schwannomas, which carry a higher risk of recurrence and metastasis, are typically managed with extensive resection and rigid internal fixation. The efficacy of adjuvant therapies, such as radiation or chemotherapy, remains controversial [10, 11, 12, 25].

4.4. Strengths and Limitations

In this study, we introduce a novel metric to quantify sclerotic margins in benign and malignant intraosseous schwannomas: the ratio of sclerotic margins, a parameter not previously mentioned in the literature. Furthermore, we identified a threshold value of 42.4%, based on the probability density distribution function within our cohort, with both specificity and sensitivity reaching 100% [29]. However, due to the rarity of intraosseous schwannomas and the small sample size in this study, the accuracy of this threshold may be limited. Therefore, future studies with larger sample sizes are imperative to validate and potentially refine this threshold. We believe this index could be valuable in differentiating other benign and malignant tumors involving bone destruction.

Several limitations of our study should be mentioned. First, it was a retrospective case series with a relatively small sample size, reflecting the rarity of the condition. Second, the majority of our cases were located in the lumbosacral region. And not all patients had undergone comprehensive imaging at the time of presentation. Third, we did not perform further subclassification of benign intraosseous schwannomas to determine which subtype exhibits aggressive behavior. At last, machine learning and deep learning models have been widely used in the medical field and have played an important role in recent years, providing a new idea for differentiating these tumors. However, the effective training and generalizability of machine learning algorithms typically require substantial datasets. Given the current small sample size, training a machine learning model may not produce generalizable outcomes. Therefore, multicenter studies with larger cohorts and extended follow‐up are essential for validating our findings.

5. Conclusion

In conclusion, intraosseous schwannomas, both benign and malignant, are rare and frequently misdiagnosed. They should be considered in the differential diagnosis for patients presenting with painful and radiographically lytic bone lesions, especially in the mandible, sacrum, and vertebrae. The ratio of sclerotic margins, which we propose for the first time, in combination with symptom duration and tumor margin clarity, can aid clinicians in identifying malignant lesions earlier, guiding treatment decisions, and predicting patient outcomes more reliably.

Author Contributions

Jiangchao Zhang: conceptualization, methodology, writing – original draft, writing – review and editing. Ge Xiong: conceptualization, funding acquisition, supervision, validation, writing – original draft, writing – review and editing. Wei Zheng: investigation, formal analysis, supervision, funding acquisition. Jing Sun: investigation, formal analysis.

Disclosure

All authors listed meet the authorship criteria according to the latest guidelines of the International Committee of Medical Journal Editors, and all authors are in agreement with the manuscript.

Ethics Statement

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Consent

The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of Beijing Jishuitan Hospital (No. 2024267‐00). Written informed consent was obtained from each patient.

Conflicts of Interest

The authors declare no conflicts of interest.