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).
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).
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:
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).
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.
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).
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).
Tumor volume differences between gross and subtotal removal were significant overall (p=0.034) but not by individual types (Table 6).
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.
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.
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.
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.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
