Warning: mkdir(): Permission denied in /home/virtual/lib/view_data.php on line 87 Warning: chmod() expects exactly 2 parameters, 3 given in /home/virtual/lib/view_data.php on line 88 Warning: fopen(/home/virtual/e-kjs/journal/upload/ip_log/ip_log_2024-10.txt): failed to open stream: No such file or directory in /home/virtual/lib/view_data.php on line 95 Warning: fwrite() expects parameter 1 to be resource, boolean given in /home/virtual/lib/view_data.php on line 96 Spinal Schwannoma Classification Based on the Presumed Origin With Preoperative Magnetic Resonance Images

Spinal Schwannoma Classification Based on the Presumed Origin With Preoperative Magnetic Resonance Images

Article information

Neurospine. 2024;21(3):890-902
Publication date (electronic) : 2024 September 30
doi : https://doi.org/10.14245/ns.2448468.234
1Department of Neurosurgery, Seoul National University Hospital, Seoul, Korea
2Department of Neurosurgery, Seoul National University College of Medicine, Seoul, Korea
3Department of Neurosurgery, Hallym University Dongtan Sacred Heart Hospital, Hwaseong, Korea
4Neuroscience Research Institute, Seoul National University Medical Research Center, Seoul, Korea
Corresponding Author Chun Kee Chung Neuroscience Research Institute, Seoul National University College of Medicine, 103 Daehak-ro, Jongno-gu, Seoul 03080, Korea Email: chungc@snu.ac.kr
Received 2024 May 4; Revised 2024 July 16; Accepted 2024 July 18.

Abstract

Objective

Classification guides the surgical approach and predicts prognosis. However, existing classifications of spinal schwannomas often result in a high ‘unclassified’ rate. Here, we aim to develop a new comprehensive classification for spinal schwannomas based on their presumed origin. We compared the new classification with the existing classifications regarding the rate of ‘unclassified’. Finally, we assessed the surgical strategies, outcomes, and complications according to each type of the new classification.

Methods

A new classification with 9 types was created by analyzing the anatomy of spinal nerves and the origin of significant tumor portions and cystic components in preoperative magnetic resonance images. A total of 482 patients with spinal schwannomas were analyzed to compare our new classification with the existing classifications. We defined ‘unclassified’ as the inability to classify a patient with spinal schwannoma using the classification criteria. Surgical approaches and outcomes were also aligned with our new classification.

Results

Our classification uniquely reported no ‘unclassified’ cases, indicating full applicability. Also, the classification has demonstrated usefulness in predicting the surgical outcome with the approach planned. Gross total removal rates reached 88.0% overall, with type 1 and type 2 tumors at 95.3% and 96.0% respectively. The approach varied with tumor type, with laminectomy predominantly used for types 1, 2, and 9, and facetectomy with posterior fixation used for type 3 tumors.

Conclusion

The new classification for spinal schwannomas based on presumed origin is applicable to all spinal schwannomas. It could help plan a surgical approach and predict its outcome, compared with existing classifications.

INTRODUCTION

Gross total resection is the primary treatment for spinal schwannomas, with various surgical strategies aiming for minimal damage to adjacent structures [1]. The complexity in tumor-structure relationships necessitates diverse classification systems to select the optimal surgical approach [2].

Eden classification, introduced in 1941 for spinal dumbbell tumors, is most widely accepted classification for spinal schwannomas, but relies solely on x-ray findings, omitting modern imaging like computed tomography (CT) or magnetic resonance imaging (MRI) [3]. This limitation could affect its current effectiveness in correlating pathology, symptoms, and surgical treatments.

Numerous classifications for spinal schwannomas have emerged since Eden’s, considering factors like size, volume, vertebral segment involvement, and meningeal relationships to suggest optimal surgical methods [1,2,4-8].

Some tumors defy current classification systems, complicating preoperative planning. The limitations of these systems are outlined as follows.

Using Eden classification, tumors in the vertebral body or foraminal zone could not be classified (Fig. 1). Sun and Pamir classification struggled with extradural tumors limited to foraminal or extraforaminal zones (Fig. 2). Modified Sridhar classification failed for tumors only in the extraforaminal region (Fig. 3).

Fig. 1.

Examples of T1 contrast magnetic resonance images in cases that are difficult to classify using Eden classification. (A) It is challenging to classify tumors confined to the foramen, particularly epidural tumors limited to this region. (B) Additionally, it could not classify schwannomas involving the vertebral body.

Fig. 2.

Examples of T1 contrast magnetic resonance images in cases that are difficult to classify using Sun and Pamir classification, which has a limitation in distinguishing (A) between intradural and extradural tumors and (B) in addressing paravertebral extensions.

Fig. 3.

Examples of T1 contrast magnetic resonance images in cases that are difficult to classify using modified Sridhar classification, which has a limitation in dealing with mixed intradural and extradural tumors confined within the spinal canal (A) and those confined to the extraforaminal region (B).

We propose a new classification system for spinal schwannomas based on preoperative MRI, identifying the tumor’s origin for safe removal. By starting resection at the origin and carefully separating adjacent structures, complete tumor removal is possible with minimal damage to important surrounding tissues, improving function by reducing compression.

We compared our new classification to existing Eden, Sun, Pamir, and modified Sridhar systems, using a prospective surgery database. We evaluated if our classification guides surgical approaches, predicts resection extent, and anticipates complications.

MATERIALS AND METHODS

Between March 2003 and June 2022, the authors collected data from a patient surgery database. We specifically selected 497 individuals without a history of other neurological disorders for this study, after preoperative MRI evaluations. Ultimately, 482 patients were included, comprising 250 men (51.87%) and 232 women (48.13%), with an average age of 49.59 years. Preoperative MRI evaluations were conducted for all. The tumor was in the cervical (121 patients, 25.1%), cervicothoracic (8 patients, 1.7%), thoracic (120 patients, 24.9%), thoracolumbar (36 patients, 7.5%), lumbar (165 patients, 34.2%), lumbosacral (15 patients, 3.1%), and sacral spine (17 patients, 3.5%) (Table 1).

Characteristics of the study population and level of surgery (n=482)

The study was approved by the Institutional Review Board (IRB) of Seoul National University Hospital (IRB No. H-2308-079-1458). The study adhered to the guidelines specified in the Declaration of Helsinki, and the manuscript followed the STROBE checklist.

Neuroradiologists’ reports based on MRI images from the first postoperative day determined the extent of tumor removal and were used to identify recurrence during the follow-up period.

This study carefully analyzed patients’ preoperative MRI scans to identify spinal schwannomas’ origins, noting the lack of consensus on their localization using imaging techniques, yet offering insights to infer these origins.

Schwannomas, including spinal tumors, arise from Schwann cells with the NF2 gene mutation [9]. The tumor may be originated from all components of spinal nerves, including ventral/dorsal roots, spinal nerve rami, and sympathetic structures, highlighting the relevance of spinal nerve anatomy in understanding these tumors.

The size and shape of schwannomas are crucial for identification of the origin site, which is usually the largest part of the tumor, although some tumors tend to extend beyond their main mass. A study by Sohn et al. [10] found that spinal schwannomas grow at a rate of approximately 1.0±4.4 mm/yr, a figure close to the 1.2 mm/yr growth rate of vestibular schwannomas from a meta-analysis of 26 clinical series. The research also indicated that extraforaminal tumors grow slightly faster than intradural or intraosseous types, though the difference is not significant [10,11]. Regardless of growth rate differences based on location, the most critical factor to consider in delineating the tumor’s origin is where the largest part is, even with multiple extensions.

The presence of the tumor cyst aids in identifying the origin site, since cystic degeneration is indicative of previous active cell division, with Antoni Type B tissue patterns [9,12]. This suggests cystic changes in tumors can help pinpoint their initial growth regions [9,12]. Based on these findings, we propose a new classification system for spinal schwannoma, detailed in Fig. 4, which we then compared with existing classifications by Eden, Sun and Pamir, and a modified version of Sridhar. Our comparison focused on the ‘unclassified’ categories within each system, using data from our prospectively collected patient surgery database [1,3,4]. Using the anatomical segmentation of the nerve, the most significant part of the tumor, and the location of a cystic portion – we classified spinal schwannomas into 9 types:

Fig. 4.

A new classification of spinal schwannomas based on the presumed origin.

Type 1: Tumor presumed to originate from the dorsal root

Type 2: Tumor presumed to originate from the ventral root

Type 3: Tumor presumed to originate from the dorsal root ganglion

Type 4: Tumor presumed to originate from the ventral ramus

Type 5: Tumor presumed to originate from the dorsal ramus

Type 6: Tumor presumed to originate from the sympathetic ganglion

Type 7: Tumor presumed to originate from the gray and white rami communicantes

Type 8: Tumor presumed to originate from the sinuvertebral nerve

Type 9: Tumor presumed to originate from the intramedullary spinal cord (medulla)

Preoperative MRI was used to determine the type of tumor, and all available T1, T2, and T1 contrast images were utilized. Sagittal, coronal, and axial cuts were all used to determine the type.

We categorized patient tumor data into types using criteria from previous studies [1,3,4]. ‘Unclassified’ refers to the failure to categorize schwannoma patients by Eden, Sun, Pamir, and modified Sridhar criteria using preoperative MRI images.

The modified Sridhar classification proposed by Park et al. [13] was used for the Sridhar classification. The volume measurement formula proposed by Sun and Pamir was followed for the Sun and Pamir classification [1].

Tumor volume=4/3π×(craniocaudal length/2)×(transverse diameter/2)2

This formula was used to measure the volume of all tumors, and the correlation between volume and gross total removal rate was also measured.

In our study, we analyzed surgical approaches based on a new classification, aiming for en bloc resection while minimizing nerve damage. We used nerve stimulators to identify nerves during surgery, avoiding damage by only removing the tumor mass from within the perineurium, preserving surrounding fascicles. If the tumor and a nerve were inseparable, the nerve was coagulated and cut for tumor removal. Dura mater openings were directly repaired with 6-0 polypropylene for watertight closure. Postsurgery, if facetectomy or significant bone removal led to instability, we employed surgical instruments for stabilization at the affected level.

In our study on surgical techniques, we categorized them into 5 types: laminectomy/-otomy, facetectomy, anterior corpectomy, posterior corpectomy, and no bone work. Laminectomy/-otomy covers all posterior surgeries excluding facetectomy and corpectomy, including laminectomy and laminoplasty procedures, and laminectomy/laminotomy after anterior dissection. Facetectomy involves surgeries removing facets, but not corpectomy. No bone work refers to surgeries avoiding spine bone alteration, like video-assisted thoracoscopic surgery, non-laminectomy tumor removal, and anterior or lateral noninvasive approaches, including access through neck muscles.

Postoperative complications encompass all issues arising after surgery until any follow-up, excluding pre-existing symptoms that improved or remained unchanged postsurgery.

Based on our surgical approach selection criteria that categorized surgical procedures for preganglionic, ganglionic, or postganglionic types of spinal schwannomas, we determined the surgical approach [14].

In managing spinal tumors, the surgical approach varies by tumor type and location. For type 1 and 2 tumors (originating from dorsal and ventral roots within the thecal sac), a posterior midline approach with laminotomy is preferred (Fig. 5). Type 3 tumors (dorsal root ganglion origin) involve an external approach to the dura mater, with potential facet joint removal and spinal stabilization as needed (Fig. 6). Types 4 through 7 tumors, originating from various nerve roots and ganglia, are primarily extraforaminal, requiring external vertebral approaches (Fig. 7); thoracoscopic-assisted surgery for thoracic, and anterolateral or extraperitoneal for cervical and lumbar tumors, respectively. Type 8 tumors (sinuvertebral nerve origin) necessitate a posterior approach with possible partial corpectomy. Lastly, type 9 (intramedullary origin) tumors are addressed posteriorly with circumferential separation of the tumor from the spinal cord (Fig. 8).

Fig. 5.

Magnetic resonance images and intraoperative gross photos of new classification. (A–C) Type 1 (dorsal root origin). (A) A sagittal T1 contrast image. (B) An axial T1 contrast image. (C) After dural and arachnoid opening, exposed type 1 schwannoma. (D–F) Type 2 (ventral root origin) tumor. (D) A sagittal T1 contrast image. (E) An axial T1 contrast image. (F) Exposing the tumor ventral to nerve fibers while undergoing resection of type 2 schwannoma.

Fig. 6.

Magnetic resonance images intraoperative gross photos and postoperative x-ray images of new classification of type 3 (dorsal root ganglion origin) tumor. (A) A sagittal T1 contrast image. (B) An axial T1 contrast image. (C) Tumor exposed after facetectomy. (D) Removing the tumor. (E) The origin fascicle of the tumor is identified. (F) A postoperative x-ray image showing screw fixation after facetectomy.

Fig. 7.

Magnetic resonance images and intraoperative gross photos of new classification continued. (A, B) Type 4 (ventral ramus origin). (A) Coronal T1 contrast image. (B) Axial T1 contrast image. (C, D) Type 5 (dorsal ramus origin). (C) Axial T1 contrast image. (D) An intraoperative microscopic photo depicting the tumor exposed after paraspinal muscle dissection. (E, F) Type 6 (sympathetic ganglion origin). (E) Axial T1 contrast image. (F) Sagittal T2 image. (G, H) Type 7 (communicantes origin). (G) Axial T1 contrast image. (H) An intraoperative gross microscopic depicting the pulling out of the type 7 schwannoma with 2 forceps.

Fig. 8.

Magnetic resonance images and intraoperative gross photos of new classification continued. (A–C) Type 8 (sinuvertebral nerve origin). (A) An axial T1 contrast image. (B) A sagittal T1 contrast image. (C) An intraoperative microscopic photo depicting the remaining capsule of type 8 schwannoma after removing the tumor mass through the capsular incision. (D–F) Type 9 (intramedullary origin). (D) A sagittal T1 contrast image. (E) An axial T1 contrast image. (F) An intraoperative microscopic photo depicting an intramedullary type 9 schwannoma after an arachnoid incision.

Statistical analysis utilized IBM SPSS Statistics ver. 25.0 (IBM Co., Armonk, NY, USA), presenting means, standard deviations, and ranges. Student t-tests compared variables, while Pearson correlation assessed continuous variable relationships. Significance was set at p<0.05.

RESULTS

Our study classified spinal schwannomas in our surgical patient database through various systems, including Eden, Sun and Pamir, modified Sridhar, and a newly proposed classification. Eden classification showed type 1 as the most prevalent (70.1%), with type 2, type 4, and type 3 following, and 4.6% of cases unclassified (Table 2). Sun and Pamir classification found type 1 leading (62.2%), with 3.5% unclassified (Table 2). Modified Sridhar highlighted type 1a as most common (54.4%), with a 7.1% unclassified rate (Table 2). Contrarily, our novel classification eliminated “unclassified” cases, with type 1 (61.4%) being the most frequent, indicating dorsal root origin tumors, followed by types 3 and 4, and hybrid forms comprising 13.7% of cases, predominantly type 1+type 3 (Table 3). This new system showed a comprehensive categorization of tumor origins, including multiple or hybrid origins, enhancing the understanding of spinal schwannomas.

Our surgical database classified by previous classifications (n=482)

Classification of our surgical database by new classification (n=482)

Additionally, a correlation between tumor volume using Sun and Pamir formula and the gross total removal rate revealed significant differences in volumes between patients with gross total removal (14.31±48.1 cm3) and those with suboptimal removal (46.61±112.1 cm3), indicating the practical relevance of tumor size in surgical outcomes (p=0.034).

A statistical analysis of surgical methods for 482 patients revealed that ‘Laminectomy/-otomy’ was the predominant procedure in 382 cases (79.3%), especially for type 1, 2, 3, and 9 tumors. ‘Facetectomy’ was the primary method for 36.6% of type 3 tumors. Types 4, 6, and 7 surgeries did not involve bone work (Table 4). Among hybrid forms, type 1+3 was common, with laminectomy/-otomy and facetectomy being the main surgical approaches in 40.5% and 28.6% of cases, respectively (Table 4).

Primary surgical approaches used for each new classification types (n=482)

In our database, the gross total removal rate for spinal schwannomas was 88.0%. By our classification, rates varied: 95.3% (type 1), 96.0% (type 2), 65.9% (type 3), 83.9% (type 4), 80.0% (type 5), 60.0% (type 6), 66.7% (type 7), 0% (type 8), and 75.0% (type 9) as detailed in Table 5. The common hybrid, type 1+3, had an 83.3% removal rate (Table 5).

Gross total removal rate by types of new classification (n=482)

Tumor volume differences between gross and subtotal removal were significant overall (p=0.034) but not by individual types (Table 6).

Tumor volume distribution and relationship between tumor volume and GTR rate by surgical level (n=482)

Preoperative symptoms in 81.5% of patients either improved or remained unchanged without significant complications. Improvement or stability rates varied by type: 83.8% in type 1, 72% in type 2, 80.5% in type 3, 87.1% in type 4, 60% in type 5, 100% in type 6 and 9, 83.3% in type 7, and 0% in type 8 (Table 7). Complications across types included hypesthesia or paresthesia (11.2%), weakness (1.7%), urinary or ejaculation issues (2.1%), cerebrospinal fluid leakage and wound complications (2.3%), epidural hematoma (0.4%), among others (Table 7). The most common hybrid, type 1+3, had a 78.6% rate of symptom stability or improvement (Table 7). The limited patient numbers in certain groups limited further statistical analysis.

Outcomes of surgery by type in new classification

DISCUSSION

We developed a novel classification system for spinal schwannomas, aimed at addressing the limitations of existing classifications by Eden, Sun and Pamir, and modified Sridhar. Utilizing a prospectively collected patient surgery database, we identified the rate of ‘unclassified’ cases—where spinal schwannomas could not be categorized—within each classification: 4.6% for Eden, 3.5% for Sun and Pamir, and 7.1% for modified Sridhar. Our system, in contrast, classified all cases without any ‘unclassified’. Our classification offers a comprehensive framework that ensures no case remains unclassified, enhancing both clinical assessment and research capabilities.

Compared to Eden paper [3], our series has significant differences in tumor type distribution: 70.1% type 1, 16.2% type 2, 2.9% type 3, and 6.2% type 4, with 4.6% unclassified. Unlike Eden, who included all dumbbell-shaped spinal tumors, such as neurofibromas and meningiomas, our study analyzed schwannomas specifically, leading to a refined classification and highlighting the evolution in understanding tumor pathology since Eden time. Also, Eden classification relies solely on x-ray findings, omitting modern imaging like CT or MRI, which are incorporated into our classification system.

Our study revealed that our classification system is effective in guiding the surgical approach with expected total removal rates, and postoperative complications. Specifically, type 1 schwannomas, which represent a significant portion of cases, were primarily treated with laminectomy, achieving a 95.3% total removal rate and a 16.2% complication rate. Complications included hypesthesia or paresthesia (9.8%), weakness (1.7%), urinary or ejaculation difficulties (2.4%), and wound issues or cerebrospinal fluid leakage (2.4%).

Sridhar classification first noted the vertebral body-tumor relationship, defining tumors invading the vertebral body as giant invasive tumors [4]. Schwannomas of the vertebral body, presumed to originate from the sinuvertebral nerve, are extradural despite their size and invasiveness, facilitating surgery without neurological risk [4]. This led to the definition of type 8 (sinuvertebral nerve origin) in our classification, with surgeries performed accordingly. Although Xin et al. [2] highlighted the membrane-tumor relationship, in large spinal tumors, it is difficult to differentiate the pia mater and arachnoid from the dura mater using MRI, limiting preoperative planning.

From a pathological perspective, schwannomas grow from one nerve fascicle and form a mass within the perineurium. Therefore, theoretically, gross total removal is possible by only removing the true mass inside the perineurium without removing all surrounding nerve fascicles.

The GTR rate in our series is 88.0%, aligning with published literature of between 80% and 100% [1,2,4] (Table 8). For type 1 (dorsal root origin) and type 2 (ventral root origin) tumors located within the thecal sac, though our data shows that 97.6% of type 1 and 96% of type 2 tumors necessitated laminectomy or laminotomy. Type 3 tumors, originating from the dorsal root ganglion, are primarily extradura [15], often necessitating foraminal zone access for removal. This approach may require facet joint removal. If spinal stability is compromised, instrumentation should be considered [14]. In our series 53.7% underwent laminectomy or laminotomy only, and 36.6% facetectomy with posterior fixation.

Comparison of gross total removal (GTR) rates from published studies

Types 4 through 7 tumors, which originate from nerve rami and sympathetic ganglia, typically reside in the extraforaminal region. Ribet and Cardot [16] noted neurogenic tumors in paravertebral regions associated with sympathetic chains and rami communicantes, including gray and white rami. These are approached from outside the vertebra, generally avoiding spinal canal exploration [14]. Thoracoscopic removal is an option for thoracic tumors, but the risk of sympathetic chain-related complications, like Horner syndrome, must be discussed preoperatively. Our records note Horner syndrome in patients with type 3 and type 3+type 6 hybrid tumors at the C5–6 level.

Type 8 tumors, originating from the sinuvertebral nerve, may cause vertebral body erosion. Zhang et al. [17] suggested that the origin of intraosseous schwannomas is the sinuvertebral nerve. A case involving a retroperitoneal approach with corpectomy and cage insertion has been reported [13], highlighting their extradural location yet potential spinal canal involvement. Our experience with type 8 schwannomas is limited, hindering further statistical analysis.

Type 9 (intramedullary origin) tumors are rare and challenging to distinguish from other intramedullary tumors like astrocytomas and ependymomas on MRI [18]. Their removal necessitates laminectomy or laminotomy [18].

For the practical implications of our new classification, we provide a table outlining the surgical approach recommended for each classification type (Table 9). This new classification could guide to individualize surgical approach based on tumor type and location, leading meticulous preoperative planning with expectant discussions with the patient about potential complications.

Surgical approach recommended for each type of new classification

This single-center retrospective study has limitations, including a small sample of schwannomas from sinuvertebral or intramedullary regions (type 5 to type 9), hindering complication rate analysis by surgical approach. A larger, prospective study across multiple institutions or countries is essential for validation.

CONCLUSION

We developed a classification for spinal schwannomas, informed by preoperative MRI and a study of 482 patients, contrasting it with existing classifications. This approach factors in the tumor’s neuroanatomical context, its most significant portion, and cystic features to conjecture origin. It evaluates surgical strategies, outcomes, and complications for each tumor type, offering a comprehensive framework for all spinal schwannomas, facilitating tailored surgical planning.

Notes

Conflict of Interest

The authors have nothing to disclose.

Funding/Support

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Author Contribution

Conceptualization: TSK, WTY, JH, CHL, CHK, CKC; Data curation: TSK, JHK, CKC; Formal analysis: TSK, JHK, JK, CKC; Funding acquisition: TSK, CKC; Methodology: TSK, JK, WTY, CHL, CHK, CKC; Project administration: TSK, CKC; Visualization: TSK, JK, JH, CHL, CHK, CKC; Writing – original draft: TSK, CKC; Writing – review & editing: TSK, WTY, JH, CHL, CHK, CKC.

References

1. Sun I, Pamir MN. Non-syndromic spinal schwannomas: a novel classification. Front Neurol 2017;8:318.
2. Xin Z, Orazmyradov B, Li J, et al. A novel classification and its clinical significance in spinal schwannoma based on the membranous hierarchy. Neurosurgery 2020;87:1206–22.
3. Eden K. The dumb-bell tumours of the spine. Br J Surg 1941;28:549–70.
4. Sridhar K, Ramamurthi R, Vasudevan MC, et al. Giant invasive spinal schwannomas: definition and surgical management. J Neurosurg 2001;94:210–5.
5. Safavi-Abbasi S, Bambakidis NC, Zabramski JM, et al. Nonvestibular schwannomas: an evaluation of functional outcome after radiosurgical and microsurgical management. Acta Neurochir (Wien) 2010;152:35–46.
6. Jiang L, Lv Y, Liu XG, et al. Results of surgical treatment of cervical dumbbell tumors: surgical approach and development of an anatomic classification system. Spine (Phila Pa 1976) 2009;34:1307–14.
7. Asazuma T, Toyama Y, Maruiwa H, et al. Surgical strategy for cervical dumbbell tumors based on a three-dimensional classification. Spine (Phila Pa 1976) 2004;29:E10–4.
8. Jinnai T, Koyama T. Clinical characteristics of spinal nerve sheath tumors: analysis of 149 cases. Neurosurgery 2005;56:510–5. discussion 510-5.
9. Helbing DL, Schulz A, Morrison H. Pathomechanisms in schwannoma development and progression. Oncogene 2020;39:5421–9.
10. Sohn S, Chung CK, Park SH, et al. The fate of spinal schwannomas following subtotal resection: a retrospective multicenter study by the Korea spinal oncology research group. J Neurooncol 2013;114:345–51.
11. Yoshimoto Y. Systematic review of the natural history of vestibular schwannoma. J Neurosurg 2005;103:59–63.
12. Ando K, Imagama S, Ito Z, et al. How do spinal schwannomas progress? The natural progression of spinal schwannomas on MRI. J Neurosurg Spine 2016;24:155–9.
13. Park SC, Chung SK, Choe G, et al. Spinal intraosseous schwannoma: a case report and review. J Korean Neurosurg Soc 2009;46:403–8.
14. Chung CK. Surgery of spinal cord tumors based on anatomy: an approach based on anatomic compartmentalization Springer Publishing; 2021.
15. Ozawa H, Kokubun S, Aizawa T, et al. Spinal dumbbell tumors: an analysis of a series of 118 cases. J Neurosurg Spine 2007;7:587–93.
16. Ribet ME, Cardot GR. Neurogenic tumors of the thorax. Ann Thorac Surg 1994;58:1091–5.
17. Zhang F, Lu F, Jiang J, et al. Two case reports and an updated review of spinal intraosseous schwannoma. J Korean Neurosurg Soc 2015;57:478–83.
18. Lee SE, Chung CK, Kim HJ. Intramedullary schwannomas: long-term outcomes of ten operated cases. J Neurooncol 2013;113:75–81.

Article information Continued

Fig. 1.

Examples of T1 contrast magnetic resonance images in cases that are difficult to classify using Eden classification. (A) It is challenging to classify tumors confined to the foramen, particularly epidural tumors limited to this region. (B) Additionally, it could not classify schwannomas involving the vertebral body.

Fig. 2.

Examples of T1 contrast magnetic resonance images in cases that are difficult to classify using Sun and Pamir classification, which has a limitation in distinguishing (A) between intradural and extradural tumors and (B) in addressing paravertebral extensions.

Fig. 3.

Examples of T1 contrast magnetic resonance images in cases that are difficult to classify using modified Sridhar classification, which has a limitation in dealing with mixed intradural and extradural tumors confined within the spinal canal (A) and those confined to the extraforaminal region (B).

Fig. 4.

A new classification of spinal schwannomas based on the presumed origin.

Fig. 5.

Magnetic resonance images and intraoperative gross photos of new classification. (A–C) Type 1 (dorsal root origin). (A) A sagittal T1 contrast image. (B) An axial T1 contrast image. (C) After dural and arachnoid opening, exposed type 1 schwannoma. (D–F) Type 2 (ventral root origin) tumor. (D) A sagittal T1 contrast image. (E) An axial T1 contrast image. (F) Exposing the tumor ventral to nerve fibers while undergoing resection of type 2 schwannoma.

Fig. 6.

Magnetic resonance images intraoperative gross photos and postoperative x-ray images of new classification of type 3 (dorsal root ganglion origin) tumor. (A) A sagittal T1 contrast image. (B) An axial T1 contrast image. (C) Tumor exposed after facetectomy. (D) Removing the tumor. (E) The origin fascicle of the tumor is identified. (F) A postoperative x-ray image showing screw fixation after facetectomy.

Fig. 7.

Magnetic resonance images and intraoperative gross photos of new classification continued. (A, B) Type 4 (ventral ramus origin). (A) Coronal T1 contrast image. (B) Axial T1 contrast image. (C, D) Type 5 (dorsal ramus origin). (C) Axial T1 contrast image. (D) An intraoperative microscopic photo depicting the tumor exposed after paraspinal muscle dissection. (E, F) Type 6 (sympathetic ganglion origin). (E) Axial T1 contrast image. (F) Sagittal T2 image. (G, H) Type 7 (communicantes origin). (G) Axial T1 contrast image. (H) An intraoperative gross microscopic depicting the pulling out of the type 7 schwannoma with 2 forceps.

Fig. 8.

Magnetic resonance images and intraoperative gross photos of new classification continued. (A–C) Type 8 (sinuvertebral nerve origin). (A) An axial T1 contrast image. (B) A sagittal T1 contrast image. (C) An intraoperative microscopic photo depicting the remaining capsule of type 8 schwannoma after removing the tumor mass through the capsular incision. (D–F) Type 9 (intramedullary origin). (D) A sagittal T1 contrast image. (E) An axial T1 contrast image. (F) An intraoperative microscopic photo depicting an intramedullary type 9 schwannoma after an arachnoid incision.

Table 1.

Characteristics of the study population and level of surgery (n=482)

Characteristic Value
Sex
 Male 250 (51.87)
 Female 232 (48.13)
Age (yr) 49.59 ± 14.27 (16–86)
 Male 48.46 ± 14.28 (16–83)
 Female 50.81 ± 14.18 (17–86)
Level of surgery
 Cervical 121 (25.1)
 Cervicothoracic 8 (1.7)
 Thoracic 120 (24.9)
 Thoracolumbar 36 (7.5)
 Lumbar 165 (34.2)
 Lumbosacral 15 (3.1)
 Sacral 17 (3.5)

Values are presented as number (%) or mean±standard deviation (range).

Table 2.

Our surgical database classified by previous classifications (n=482)

Classification No. (%)
Eden classification
 Type 1 338 (70.1)
 Type 2 78 (16.2)
 Type 3 14 (2.9)
 Type 4 30 (6.2)
 Unclassified 22 (4.6)
Sun and Pamir classification (type number, group)
 Type 1 300 (62.2)
  Type 1A 127 (26.3)
  Type 1B 110 (22.8)
  Type 1C 63 (13.1)
 Type 2 71 (14.7)
  Type 2A 4 (0.8)
  Type 2B 13 (2.7)
  Type 2C 54 (11.2)
 Type 3 67 (13.9)
  Type 3A 1 (0.2)
  Type 3B 2 (0.4)
  Type 3C 3 (0.6)
 Type 4 27 (5.6)
  Type 4A 0 (0)
  Type 4B 4 (0.8)
  Type 4C 23 (4.8)
 Unclassified 17 (3.5)
Modified Sridhar classification
 Type 1a 262 (54.4)
 Type 1b 11 (2.3)
 Type 2 43 (8.9)
 Type 3 62 (12.9)
 Type 4a 27 (5.6)
 Type 4b 26 (5.4)
 Type 5 11 (2.3)
 Type 6 1 (0.2)
 Type 7 5 (1.0)
 Unclassified 34 (7.1)

Eden classification: type 1, intra- and extradural; type 2, intra- and extradural and paravertebral; type 3, extradural and paravertebral; type 4, foraminal and paravertebral.

Sun and Pamir classification: type 1, intradural; type 2, intradural to foramen extension; type 3, intradural to extraforaminal extension; type 4, extraforaminal/group a: volume<2 cm3, group b: 2 cm3< volume<4 cm3, group c: volume>4 cm3.

Modified Sridhar classification : type 1a, length<2 cm, intradural; type 1b, length<2 cm, extradural; type 2, length>2 cm; type 3, extended to foramen; type 4a, extraspinal extension<2.5 cm; type 4b, extraspinal extension>2.5 cm; type 5, body invasion; type 6, intravertebral; type 7, intravertebral and extended to foramen.

Table 3.

Classification of our surgical database by new classification (n=482)

Classification No. (%)
Type 1 (dorsal root) 296 (61.4)
Type 2 (ventral root) 25 (5.2)
Type 3 (dorsal root ganglion) 41 (8.5)
Type 4 (ventral ramus) 31 (6.4)
Type 5 (dorsal ramus) 5 (1.0)
Type 6 (sympathetic ganglion) 5 (1.0)
Type 7 (communicantes) 6 (1.2)
Type 8 (sinuvertebral nerve) 2 (0.4)
Type 9 (intramedullary) 4 (0.8)
Hybrid form (multiple origins) 67 (13.9)
 Type 1+3 42 (8.7)
 Type 1+4 8 (1.7)
 Type 1+8 2 (0.4)
 Type 2+3 1 (0.2)
 Type 3+4 7 (1.5)
 Type 3+6 1 (0.2)
 Type 3+8 2 (0.4)
 Type 4+7 1 (0.2)
 Type 4+8 2 (0.4)
 Type 5+7 1 (0.2)
Unclassified 0 (0)

Table 4.

Primary surgical approaches used for each new classification types (n=482)

Classification type Laminectomy/-otomy Facetectomy Anterior corpectomy Posterior corpectomy No bone works
Type 1 289 (97.6) 1 (0.3) 0 (0) 0 (0) 6 (2.0)
Type 2 24 (96.0) 1 (4.0) 0 (0) 0 (0) 0 (0)
Type 3 22 (53.7) 15 (36.6) 0 (0) 0 (0) 4 (9.8)
Type 4 10 (32.3) 2 (6.5) 1 (3.2) 0 (0) 18 (58.1)
Type 5 2 (40.0) 1 (20.0) 0 (0) 0 (0) 2 (40.0)
Type 6 1 (20.0) 0 (0) 0 (0) 0 (0) 4 (80.0)
Type 7 0 (0) 1 (16.7) 0 (0) 1 (16.7) 4 (66.7)
Type 8 1 (50.0) 0 (0) 0 (0) 0 (0) 1 (50.0)
Type 9 4 (100) 0 (0) 0 (0) 0 (0) 0 (0)
Total (hybrid form included) 382 (79.3) 43 (8.9) 1 (0.2) 1 (0.2) 55 (11.4)
Hybrid form (multiple origins) (n = 67)
 Type 1+3 17 (40.5) 12 (28.6) 0 (0) 0 (0) 13 (31.0)
 Type 1+4 1 (12.5) 6 (75.0) 0 (0) 0 (0) 1 (12.5)
 Type 1+8 2 (100) 0 (0) 0 (0) 0 (0) 0 (0)
 Type 2+3 0 (0) 0 (0) 0 (0) 0 (0) 1 (100)
 Type 3+4 5 (71.4) 2 (28.6) 0 (0) 0 (0) 0 (0)
 Type 3+6 0 (0) 0 (0) 0 (0) 0 (0) 1 (100)
 Type 3+8 2 (100) 0 (0) 0 (0) 0 (0) 0 (0)
 Type 4+7 0 (0) 1 (100) 0 (0) 0 (0) 0 (0)
 Type 4+8 0 (0) 1 (50.0) 0 (0) 0 (0) 1 (50.0)
 Type 5+7 0 (0) 0 (0) 0 (0) 0 (0) 1 (100)

Values are presented as number (%).

Type 1, dorsal root; type 2, ventral root; type 3, dorsal root ganglion; type 4, ventral ramus; type 5, dorsal ramus; type 6, sympathetic ganglion; type 7, communicantes; type 8, sinuvertebral nerve; type 9, intramedullary.

Table 5.

Gross total removal rate by types of new classification (n=482)

Classification type No. (%)
Type 1 282/296 (95.3)
Type 2 24/25 (96.0)
Type 3 27/41 (65.9)
Type 4 26/31 (83.9)
Type 5 4/5 (80.0)
Type 6 3/5 (60.0)
Type 7 4/6 (66.7)
Type 8 0/2 (0)
Type 9 3/4 (75.0)
Total (hybrid form included) 424/482 (88.0)
Hybrid form (multiple origins) (n = 67)
 Type 1+3 35/42 (83.3)
 Type 1+4 5/8 (62.5)
 Type 1+8 1/2 (50.0)
 Type 2+3 1/1 (100)
 Type 3+4 5/7 (71.4)
 Type 3+6 1/1 (100)
 Type 3+8 1/2 (50.0)
 Type 4+7 0/1 (0)
 Type 4+8 2/2 (100)
 Type 5+7 0/1 (0)

Type 1, dorsal root; type 2, ventral root; type 3, dorsal root ganglion; type 4, ventral ramus; type 5, dorsal ramus; type 6, sympathetic ganglion; type 7, communicantes; type 8, sinuvertebral nerve; type 9, intramedullary.

Table 6.

Tumor volume distribution and relationship between tumor volume and GTR rate by surgical level (n=482)

Variable GTR patients STR patients Volume of GTR (cm3) Volume of STR (cm3) p-value
Cervical (n = 121) 99 (81.8) 22 (18.2) 10.0 ± 14.4 30.5 ± 62.9 0.144
Cervicothoracic (n = 8) 7 (87.5) 1 (12.5) 23.4 ± 29.7 14.5 -
Thoracic (n = 120) 112 (93.3) 8 (6.7) 10.9 ± 27.4 6.9 ± 14.4 0.686
Thoracolumbar (n = 36) 32 (88.9) 4 (11.1) 22.0 ± 60.3 14.3 ± 19.5 0.804
Lumbar (n = 165) 152 (92.1) 13 (7.9) 11.7 ± 51.6 79.6 ± 207.3 0.262
Lumbosacral (n = 15) 9 (60.0) 6 (40.0) 8.6 ± 10.9 78.0 ± 82.7 0.095
Sacral (n = 17) 13 (76.5) 4 (23.5) 87.4 ± 152.4 100.9 ± 92.8 0.834
Total (n = 482) 424 (88.0) 58 (12.0) 14.3 ± 48.1 46.6 ± 112.1 0.034

Values are presented as number (%) or mean±standard deviation.

GTR, gross total removal; STR, subtotal removal (GTR failed).

Table 7.

Outcomes of surgery by type in new classification

N = 482 (Patient number) Relieved or no change Developed hypesthesia paresthesia Weak-ness Urinary or ejaculation difficulty Wound problem & CSF leak Epidural hematoma cord compression Others
Type 1 (n = 296) 248 (83.8) 29 (9.8) 5 (1.7) 7 (2.4) 7 (2.4) 0 (0) 0 (0)
Type 2 (n = 25) 18 (72.0) 3 (12.0) 1 (4.0) 0 (0) 1 (4.0) 2 (8.0) 0 (0)
Type 3 (n = 41) 33 (80.5) 5 (12.2) 1 (2.4) 0 (0) 1 (2.4) 0 (0) 1 (2.4)*
Type 4 (n = 30) 27 (87.1) 2 (6.5) 1 (3.2) 0 (0) 0 (0) 0 (0) 0 (0)
Type 5 (n = 5) 3 (60.0) 2 (40.0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Type 6 (n = 6) 5 (100) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Type 7 (n = 6) 5 (83.3) 1 (16.7) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Type 8 (n = 2) 0 (0) 1 (50.0) 0 (0) 1 (50.0) 0 (0) 0 (0) 0 (0)
Type 9 (n = 4) 4 (100) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Total (hybrid form included) 393 (81.5) 54 (11.2) 8 (1.7) 10 (2.1) 11 (2.3) 2 (0.4) 4 (0.8)
Hybrid form (multiple origins) (n = 67)
 Type 1+3 (n = 42) 33 (78.6) 6 (14.3) 0 (0) 1 (2.4) 2 (4.8) 0 (0) 0 (0)
 Type 1+4 (n = 8) 6 (75.0) 2 (25.0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
 Type 1+8 (n = 2) 1 (50.0) 0 (0) 0 (0) 1 (50.0) 0 (0) 0 (0) 0 (0)
 Type 2+3 (n = 1) 0 (0) 1 (100) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
 Type 3+4 (n = 7) 5 (71.4) 1 (14.3) 0 (0) 0 (0) 0 (0) 0 (0) 1 (14.3)
 Type 3+6 (n = 1) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 1 (100)*
 Type 3+8 (n = 2) 2 (100) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
 Type 4+7 (n = 1) 1 (100) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
 Type 4+8 (n = 2) 1 (50.0) 1 (50.0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
 Type 5+7 (n = 1) 1 (100) 0 (0.0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)

Type 1, dorsal root; type 2, ventral root; type 3, dorsal root ganglion; type 4, ventral ramus; type 5, dorsal ramus; type 6, sympathetic ganglion; type 7, communicantes; type 8, sinuvertebral nerve; type 9, intramedullary; CSF, cerebrospinal fluid.

*

The relevant complication was Horner syndrome.

The relevant complication was vertebral artery injury.

Table 8.

Comparison of gross total removal (GTR) rates from published studies

Study Patients (n) Cervical (n) Cervicothoracic (n) Thoracic (n) Thoracolumbar (n) Lumbosacral (n) GTR rate (%)
Sridhar et al. [4] 10 1 - 2 - 7 80.0
Sun and Pamir [1] 82 34 - 17 - 31 98.8
Xin et al. [2] 101 47 - 16 - 38 100
New classification 482 121 8 120 36 197 88.0

Table 9.

Surgical approach recommended for each type of new classification

Classification Recommended surgical approach
Type 1 (dorsal root) Laminectomy/laminotomy
Type 2 (ventral root) Laminectomy/laminotomy
Type 3 (dorsal root ganglion) Facetectomy with posterior fixation
Type 4 (ventral ramus) No bone work; extraspinal approach including thoracoscopic-assisted surgery (thoracic tumors) and anterior/lateral approach (cervical and lumbar tumors)
Type 5 (dorsal ramus) No bone work; posterior approach
Type 6 (sympathetic ganglion) No bone work; extraspinal anterior approach
Type 7 (communicantes) No bone work; extraspinal approach including thoracoscopic-assisted surgery (thoracic tumors) and anterior/lateral approach (cervical and lumbar tumors)
Type 8 (sinuvertebral nerve) Laminectomy/laminotomy with corpectomy as needed
Type 9 (intramedullary) Laminectomy/laminotomy