Clinical features and prognostic factors in spinal meningioma surgery from a multicenter study

Meningiomas are benign tumors that are treated surgically. Local recurrence is likely if the dura is preserved, and en bloc tumor and dura resection (Simpson grade I) is recommended. In some cases the dura is cauterized and preserved after tumor resection (Simpson grade II). The purpose of this study was performed to analyze clinical features and prognostic factors associated with spinal meningioma, and to identify the most effective surgical treatment. The subjects were 116 patients (22 males, 94 females) with spinal meningioma who underwent surgery at seven NSG centers between 1998 and 2018. Clinical data were collected from the NSG database. Pre- and postoperative neurological status was defined using the modified McCormick scale. The patients had a mean age of 61.2 ± 14.8 years (range 19–91 years) and mean symptom duration of 11.3 ± 14.7 months (range 1–93 months). Complete resection was achieved in 108 cases (94%), including 29 Simpson grade I and 79 Simpson grade II resections. The mean follow-up period was 84.8 ± 52.7 months. At the last follow-up, neurological function had improved in 73 patients (63%), was stable in 34 (29%), and had worsened in 9 (8%). Eight patients had recurrence, and recurrence rates did not differ significantly between Simpson grades I and II in initial surgery. Kaplan–Meier analysis of recurrence-free survival showed that Simpson grade III or IV, male, and dural tail sign were significant factors associated with recurrence (P < 0.05). In conclusion, Simpson I resection is anatomically favorable for spinal meningiomas. Younger male patients with a dural tail and a high-grade tumor require close follow-up. The tumor location and feasibility of surgery can affect the surgical morbidity in Simpson I or II resection. All patients should be carefully monitored for long-term outcomes, and we recommend lifelong surveillance after surgery.

Simpson grade II resection (macroscopically complete removal of the tumor and its visible extensions with coagulation of its dural attachment) is also a consensus if the anatomy of the meningioma does not allow Simpson I removal. However, Simpson grade II resection has been found to have a recurrence rate of 1-8% 6,17,18 .
Given the range of recurrence rates in previous reports, there is a need for a study of long-term outcomes to determine the optimum treatment strategy. The purpose of this study was to analyze clinical features and prognostic factors of spinal meningioma treated at several centers to identify the characteristics and appropriate treatment for this tumor.

Materials and methods
Patients. Data from the Nagoya Spine Group (NSG) database were used in the study. Between 1998 and 2018, 116 patients with spinal meningioma underwent surgery at seven NSG centers. Age, gender, symptom duration, sagittal and axial location, histological type, preoperative symptoms, dural tail, neurological examination results, resection classified by Simpson grade 15 , and tumor recurrence were obtained for these cases. The tumor location was classified into dorsal, lateral, ventral, and dumbbell types depending on the location of dural attachment to the tumor 16 . Preoperative magnetic resonance imaging (MRI) was routinely performed using T1W, T2W, and T1 with gadolinium (Gd)-enhanced sequences to investigate anatomical details of the lesions.
All patients in the series were examined neurologically preoperatively and postoperatively at final follow-up. Simpson grade was evaluated based on the description of the surgeon and MRI performed during the first postoperative week in selected cases or 3-6 months after surgery. In preoperative MRI, the dural tail sign was defined using Gd MRI, as described by Goldsher et al. 19 . Surgery was performed by surgeons certified by the Japanese Spine Surgery and Related Research (JSSR) society at each facility. The study protocol was approved by the Ethics committee of Nagoya University Hospital. All procedures involving human participants were conducted in accordance with the ethical standards of the institutional research committee and the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all subjects or, if subjects are under 18, from a parent and/or legal guardian.
Neurological assessment. A preoperative neurological grade was assigned using the modified McCormick scale (grade I = normal gait, II = mild gait disturbance not requiring support, III = gait with support, IV = assistance required, and V = wheelchair needed) 20 . McCormick grades I and II were defined as stable gait with no support required for walking, indicating independent gait ability 21-23 . Surgical technique. All surgeries were performed using the conventional posterior midline approach and laminectomy beyond the rostral and caudal ends of the tumor to minimize kinking of the spinal cord by laminotomy and to expose the proximal and distal subarachnoid space to allow dissection to begin 14,18,24 . After the dura was opened, an extraarachnoid tumor was identified as the meningioma. For Simpson grade I resection, the tumor origin (dura) was resected after tumor removal. The defect was reconstructed with an artificial dura that covered the defect extradurally. Depending on the facility, lumbar drainage was often performed to prevent postoperative CSF leakage with subcutaneous CSF accumulation in cases in which artificial dura mater was used 25 . For Simpson grade II surgery, the dura was coagulated to the extent of the dural tail sign after complete removal of the tumor with an ultrasonic aspirator. Then, curettage was performed at the tumor origin, which was thoroughly coagulated. In a case in which calcification was found on preoperative CT, meningioma was strongly suspected and an intraoperative pathological examination was not performed. At the discretion of the surgeon, this examination was performed in cases that did not show calcification on preoperative CT. Statistical analysis. Results were analyzed by Student t test after calculating means and standard deviations. Chi-square and Fisher exact tests were used for analysis of contingency tables. A one-way ANOVA Bonferroni multiple comparison test was used to compare differences between groups. Kaplan-Meier survival analysis was performed to examine the recurrence-free survival period. Data were analyzed using SPSS Statistics 25 (IBM Corp., Chicago, IL, USA), with P < 0.05 considered to be significant.
Extent of resection and postoperative outcome. Peri-and postoperative data are shown in Table 2.
The mean follow-up period was 84.8 ± 52.7 months (range 18-252 months). An intraoperative pathological examination was performed in 20 cases. Complete resection was achieved in 108 cases (94%), including 29 with Simpson grade I and 79 with Simpson grade II resection. Of the other 8 cases, 4 each (4%) underwent Simpson grades III and IV resection. The relationship of axial tumor location and Simpson grade is shown in Fig. 3. Adjuvant radiotherapy was performed in 2 cases in WHO grade II, and the overall rate of cases treated with adjuvant therapy was 2%. These cases underwent external-beam radiotherapy (48 Gy, 54 Gy) after subtotal resection. Chemotherapy was not performed. There were no long-term side effects of radiation therapy. During followup, the patients had recurrence, but are alive with disease. There were 3 cases with postoperative CSF leakage www.nature.com/scientificreports/ in which subcutaneous accumulation (including superficial and deep layers) of CSF continued for more than 3 months, and these cases required reoperation for dural and fascia repair due to headache and nausea. Wound complications due to infection also occurred in these 3 cases.
Results for pre-and postoperative neurological status are shown in Fig. 4. As of the last follow-up, neurological function had improved in 73 patients (63%), was stable in 34 (29%), and had worsened in 9 (8%) ( Table 2). The characteristics of cases treated with Simpson II, III or IV resection (that is, not Simpson I) are shown in Table 3.

Discussion
Spinal meningiomas generally respond favorably to surgical excision and have a low recurrence rate of 3-15.3% after excision 5,6,23 . In our series, in postoperative follow-up of a mean of 84.8 months, the recurrence rate was 6.9% (8/116 cases). The first choice of treatment for spinal meningiomas is complete surgical resection. The recurrence rate associated with subtotal resection is significantly higher than that with total resection 6,24,26,27 , and complete removal of the attached dura (Simpson I) may therefore be the most effective surgical method. In our series, in 29 patients with Simpson grade I resection, there was no tumor recurrence, and only 2 (3%) of 79 patients with Simpson grade II resection developed recurrence. Simpson II resection is also acceptable in spinal meningioma surgery 39 and results in low recurrence rates 14,27 . Therefore, it is uncertain if Simpson grade  T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 41 and Nakamura et al. 16 found dural invasion in 19/25 (76%) and 15/43 (35%) cases, respectively. These results suggest that late-developing tumor recurrence after Simpson grade II resection was attributable to the presence of residual tumor cell infiltrates between the inner and outer dural layers. In our series, we found no significant difference between Simpson grade I and grade II cases using Kaplan-Meier survival curves (Fig. 6). In our series, there was a high rate of WHO grade 1 histology among spinal meningiomas (97%). This could be due to selection bias of larger and more symptomatic tumors for surgical resection, as many small and indolent meningiomas are not resected. However, in the spinal cord, a much smaller size threshold is used to indicate surgical resection, given the narrower anatomy. This may explain the higher than expected rate of WHO grade 1 meningiomas and the high female-to-male ratio in our cohort.
Multiple clinical factors have been associated with increased recurrence rates. Maiti et al. found that males are more likely to have recurrence, despite the lower tumor incidence 30 . In younger patients, Cohen-Gadol et al. showed that extradural extension was associated with a higher recurrence rate 28 . Klekamp and Samii found that arachnoid scarring was significantly associated with increased recurrence 27 . Regarding the tumor location, ventral meningiomas have a higher risk of recurrence than dorsal or lateral meningiomas 16 26 , with the grade being an independent predictor of recurrence. The three WHO grade II cases in this study had a dumbbell shape tumor with foraminal extension, and waveform deterioration in intraoperative transcranial motor evoked potential (Tc-MEP) spinal cord monitoring during tumor resection that did not recover. Therefore, Simpson grade I resection could not be performed and       (Table 5). Given the high recurrence rate for WHO grade II tumors, complete resection should be attempted when possible. The postoperative neurological outcome for spinal meningioma is generally favorable, with a clear possibility of recovery even in patients with deep functional preoperative impairment 23,43 . In our series, neurologically independent gait ability (McCormick I and II) at the last follow-up was achieved in 83% of cases (96/116), and among all patients, 92% (107/116) had improved or stable postoperative neurological function. These results are as favorable as those in previous reports. However, large tumors with ventral attachment causing spinal cord signal changes are associated with poor functional outcomes and increased risk of spinal cord traction and surgical damage. In our series, severe preoperative impairment (McCormick grades III, IV and V) was significantly Table 3. Characteristics of cases treated with Simpson II, III or IV resection (n = 87).   Total resection is possible for most spinal meningiomas 29,36 . Dorsal or dorsolateral lesions are easier to resect, but for ventral lesions sequential debulking and dissection may be helpful. Regarding postoperative complications, the role of resection of the dural attachment is controversial, and contrasting results have been reported. Nakamura et al. found a lower recurrence rate for Simpson grade I resection than for Simpson grade II 16 . In contrast, King et al. reported a low recurrence rate even when dural resection was not performed 37 . Originally, resection of the dural attachment with suturing of a patch graft was advocated 30 . Dura-splitting dissection was also suggested to be useful in cases in which the margin of the tumor is excised in continuity with the inner layer of the dura, and preservation of the outer layer of the dura minimizes the post-resection dural defect, which prevents CSF leakage 44 . However, complete resection of the tumor and associated dura (Simpson I) is limited by postoperative complications, especially for ventral meningiomas. However, in our series, Simpson grade 3 resection was performed in some cases with lateral lesions if waveform deterioration in intraoperative Tc-MEP monitoring occurred without recovery. In such cases, we were forced to interrupt the procedure without achieving total resection.
In our series, Simpson grade I resection was less frequently performed for ventral meningiomas (20%, 13/64). When the tumor was located ventrally, dural resection could not be attempted in most cases. Simpson grade I removal can be technically challenging, because of neurological complications, pseudomeningoceles or CSF leakage. CSF leakage and wound complication are most common, occurring in 0-4% and 0-6% of cases, respectively 30,44 . In our series, reoperation due to CSF leakage was required in three patients, even after Simpson grade I removal. Given these results, Simpson grade I resection should be planned for dorsal lesions, and Simpson grade II resection might be an alternative in a case at high risk for postoperative complications.
Spinal meningiomas are generally isointense to the spinal cord on T1-and T2-weighted images, and show enhancement after administration of contrast agent 26 . A dural tail or linear enhancement of the adjacent dura may appear after Gd administration, similarly to intracranial meningiomas, and its presence often distinguishes meningiomas from other intradural extramedullary lesions 19,45 . Nakamura et al. 16 also found that preoperative Table 5. Factors associated with postoperative neurological function and prognostic factors for recurrence. Significant difference *P < 0.05, **P < 0.01.

Neurological function Recurrence
Worsened (n = 9) Improved/stable (n = 107) P value Yes (n = www.nature.com/scientificreports/ detection of tumor cell invasion between the inner and outer dura mater layers is useful for selecting the operative procedure. Tumor invasion in the dura occurred in 1/3 of spinal meningiomas and in 47% (7/15) of spinal meningiomas with a dural tail sign; therefore, histological analyses of resected dura can clarify the significance of the dural tail sign 16 . In our series, we routinely coagulated the dura to the extent of the dural tail sign in Simpson grade II resection. Our retrospective study lacks histological analysis for the resected dura, but our data show that a dural tail in MRI was significantly more common in recurrence cases. Therefore, the presence of a dural tail should be carefully analyzed in predicting recurrence, and patients with a dural tail on imaging should undergo careful long-term follow-up. Postoperative mortality in spinal meningioma surgery is usually low, ranging from 0 to 4.7% in series reported later than 1999 6,7,[38][39][40] . Radiotherapy is typically reserved for management of high-grade and recurrent spinal meningioma, and is useful in cases with macroscopic remnants of histological high grade lesions, recurrent meningiomas after subtotal primary excision, or as a therapeutic approach if surgery is unsafe for clinical or anatomic reasons 23 . Reports of chemotherapy for spinal meningiomas are limited, but the outcome seems not to be satisfactory. In our series, two WHO grade II cases received radiotherapy, and both patients are still alive.
This study has several limitations. First, the retrospective design does not offer the advantages of a prospective study. Second, although multiple reports suggest that WHO Grade II meningiomas are prone to recurrence, we had only a few patients in this category, which prevented analysis of many risk factors. In particular, the study included few WHO II meningiomas, and the low statistical power prevented significant findings. Third, a high Ki-67 index is known to correlate with recurrence and histological grade in meningioma 16,[46][47][48] , but in our series the Ki-67 index was not routinely measured. However. the series of 116 patients is relatively large, and we believe that our results provide new information on clinical features and prognostic factors in spinal meningioma surgery. A prospective study with a longer follow-up period is needed to confirm the results.
In conclusion, in our series, there was no significant difference in recurrence rate between Simpson grade I and grade II cases. However, especially in younger patients with long-term follow-up, Simpson I resection is anatomically favorable for spinal meningiomas. MRI can detect a recurrent tumor early, often before clinical symptoms develop. In younger patients, male cases with a dural tail and a histological high grade tumor require closer clinical and MRI follow-up. The tumor location and feasibility of surgery can influence the surgical morbidity in Simpson I or II resection. Even after complete resection of spinal meningioma, patients should be carefully monitored for long term outcomes, and we recommend lifelong surveillance after surgery.