Clinical significance of NF2 alteration in grade I meningiomas revisited; prognostic impact integrated with extent of resection, tumour location, and Ki-67 index

NF2 alteration is the most commonly–found genetic abnormality in meningiomas and is known to initiate events for aggressive-type meningiomas. Whereas the prognosis of meningiomas differs depending on their epigenomic/transcriptomic profile, the effect of NF2 alteration on the prognosis of benign meningiomas is not fully elucidated. This study aimed to probe the importance of NF2 alteration in prognosis of WHO grade I meningiomas. A long-term retrospective follow-up (5.3 ± 4.5 years) study involving 281 consecutive WHO grade I meningioma patients was performed. We assessed tumour recurrence in correlation with extent of resection (EOR), histopathological findings, tumour location, and NF2 alteration. “NF2 meningioma” was defined as meningiomas with presence of NF2 mutation and/or 22q loss. Overall, NF2 meningioma per se was not a predictor of prognosis in the whole cohort; however, it was a predictor of recurrence in supratentorial meningiomas, together with EOR and Ki-67. In a striking contrast, NF2 meningioma showed a better prognosis than non-NF2 meningioma in infratentorial lesion. Supratentorial NF2 meningiomas had higher Ki-67 and forkhead box protein M1 expression than those of others, possibly explaining the worse prognosis in this subtype. The combination of NF2 alteration, high Ki-67 and supratentorial location defines subgroup with the worst prognosis among WHO grade I meningiomas. Clinical connotation of NF2 alteration in terms of prognosis of WHO grade I meningioma differs in an opposite way between supratentorial and infratentorial tumors. Integrated anatomical, histopathological, and genomic classifications will provide the best follow-up schedule and proactive measures.

. In general, a 2nd surgery or radiotherapy are common complementary treatment practices for recurrent meningiomas in addition to surgical resection, especially in case of WHO grade II or III meningiomas [3]. However, WHO grade I meningiomas are also known to recur, with a sobering report of around 24-60% recurrence rate with long-term follow-up [1,[3][4][5][6][7][8]. Despite these reports, observation with or without imaging follow-ups remains commonplace following gross total resection (GTR) and subtotal resection (STR) in WHO grade I meningioma [1]. Thus, for the therapeutic management of WHO grade I meningiomas, identification of cases that are at a higher risk of and of the indicators of recurrence is crucial for making a right decision at the right timing in performing the complementary treatment. According to previous reports, the extent of resection (EOR), Ki-67 index, and WHO grades based on histological findings are known as predictors of meningioma recurrence [4][5][6][7][8][9][10][11][12][13]. Many of these studies have focused on the clinical characteristics of atypical or malignant meningiomas, while recurrence of WHO grade I meningiomas has not yet been fully elucidated [13,14].
Recent molecular analyses have shown that meningiomas are not genomically homogenous but have various genetic, epigenetic, and transcriptomic profiles [16][17][18][19][20][21][22][23][24][25]. For example, driver gene mutations are known to be highly dependent on the anatomical location of meningiomas [16][17][18][19][20][21]. Neurofibromatosis 2 (NF2) alteration is the most common genetic driver in sporadic meningiomas and is known to initiate events for aggressive-type meningiomas. However, the effect of driver gene mutations, especially NF2 alteration, on the prognosis of meningiomas has not been clearly established, except a few reports stating that their prognosis in all WHO grade meningiomas differs depending on the epigenomic/transcriptomic profiles [26][27][28][29][30][31]. Furthermore, tumour location has to be taken into account in considering recurrence of WHO grade I meningiomas because the location of the tumour is also related to its epigenetic profile [24,32].
In this study, we aimed to investigate the clinical significance of NF2 alteration in the prognosis of WHO grade I meningiomas. We performed a long-term follow-up study to validate the effect of EOR, the significance of NF2 alteration, and tumour anatomical location on meningioma recurrence.

Patient population
Patients who underwent surgical resection of sporadic meningioma between 2000 and 2019 were retrospectively queried using an institutional database. The study protocol was approved by our Institutional Review Board (G10028), and informed consent was obtained from all the patients. Patients with incomplete clinical or genetic data, any previous history of meningioma treatment, or any history of radiation therapy for the remaining tumour immediately after the first surgery, were excluded. Patients with WHO grade I meningioma were included in this study (Additional file 1: Figure S1).

Clinical data
All clinical data were collected through a retrospective chart review. The collected clinical endpoints included patient age, sex, and radiological follow-up. Preoperative and postoperative radiological data including tumour size, anatomical location, EOR, and the presence/absence and timing of recurrence were noted. Patients were followed-up with contrast-enhanced MRI (CE-MRI) within 2 days, around 6 months, and 1 year after surgery. If there was no tumour recurrence, follow-up was continued annually by CE-MRI. We defined tumour recurrence as apparent enlargement of the residual tumour on CE-MRI, by blind inter-observer agreement between the neuro-radiologists and neurosurgeons in charge.
The EOR was categorized as either GTR, which included Simpson grades I, II, and III, or STR, which included Simpson grades IV. The EOR was determined based on the postoperative imaging and operation records.

Histopathological data
The pathological diagnosis was established by the expert neuropathologist at our institution based on the 2016 WHO classification of tumours of the central nervous system. Formalin-fixed paraffin-embedded tissue was used for immunohistochemistry (IHC) in all cases. IHC was performed using whole-slide sections for forkhead box protein M1 (FOXM1) and Ki-67. The following antibodies were used: anti-Ki67 rabbit polyclonal (30-9; Ventana Medical Systems, Tucson, AZ) and FOXM1 rabbit monoclonal (EPR17379). Ki-67 index was calculated as the highest value in areas of maximum cellular density, identified by visual inspection. We define 4% as the cutoff of Ki-67 in reference to previous publications about Ki-67 and the recurrence of meningioma [33,34]. FOXM1 IHC was quantified as the highest number of positive nuclei / high power field (HPF) within each meningioma using ImageJ (U.S. National Institutes of Health). "High" FOXM1 expression was defined as more than 3 nuclei/HPF, and "low" as otherwise.

Sanger sequencing and microsatellite analysis
Tumour samples were stored at −80 °C after tumour resection until genomic analysis. Tumour DNA was obtained from frozen samples using a DNA Extraction Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. Mutation analysis was performed as previously described, including direct Sanger sequencing and microsatellite analysis [35].

Anatomical groups
Tumour location was defined as the location of the dural attachment of the tumour based on intraoperative observation as well as preoperative imaging. All meningiomas were classified into two groups, namely supratentorial or infratentorial lesions. Supratentorial lesions included meningiomas at the convexity, falx, parasagittal, sphenoid ridge, anterior fossa, middle fossa, clinoid, tuberculum sellae, and cavernous sinus. Infratentorial lesions included meningiomas at the clivus, petro-clivus, petrous, cerebellopontine angle, cerebellar convexity, jugular foramen, and foramen magnum. Tentorial meningiomas were divided into two groups depending on the dominant tumour protrusion direction, as evaluated using sagittal MRI images.

Subgroups encompassing driver gene mutation and tumour location
"NF2 meningioma" was defined as meningiomas with the presence of NF2 mutation and/or 22q loss [14]. Based on the driver gene mutation profile and the tumour location, meningiomas of all enrolled patients were categorized into the following four subgroups: "Supratentorial NF2", "Infratentorial NF2", "Supratentorial non-NF2", and "Infratentorial non-NF2". For each group, we evaluated clinical, radiological, and histopathological data, and progression-free survival (PFS).

Statistical analysis
Statistical analyses were performed using R version 3.1.2 (R Core Team, http:// www.R-proje ct. org). Numerical variables were expressed as the mean and standard deviation. Categorical data were compared between the subgroups using Fisher's exact test. PFS was evaluated using the Kaplan-Meier method, followed by the log-rank test for each variable. The Mann-Whitney U test was used to compare two non-parametric continuous variables.
The hazard ratio (HR) of all variables for the presence of recurrence as endpoints was analyzed using the univariate Cox proportional hazards model. Variables associated with endpoints in univariate analyses (p < 0.05) were included in a backward, stepwise manner using p-value multivariate analysis.
All reported p-values were two-sided, and in all comparisons, p-values of less than 0.05, were considered significant. Post-hoc pairwise comparisons between subgroups and prognostic grades were adjusted using the Bonferroni method.

Data availability
The authors confirmed that the data supporting the findings of this study will be shared upon request from any qualified investigator.

Clinical characteristics
A total of 343 patients who underwent surgical treatment for sporadic meningioma were enrolled in this study at the University of Tokyo Hospital between 2000 and 2019 (Additional file 1: Fig. S1). The remaining 281 patients with WHO grade 1 meningioma were eligible for subsequent analyses. The average follow-up period after surgery was 5.3 ± 4.5 years (Table 1). Supratentorial and infratentorial meningiomas accounted for 188 (66.9%) and 93 (33.1%) patients, respectively (Table 1, Table 2). The detailed tumour locations are shown in Additional file 1: Table S1.

Surgical outcome
The overall recurrence rate of WHO grade I meningiomas was 16.0% in the follow-up period (Table 1  & Table 2), and the average time for recurrence was 4.1 ± 3.7 years (Additional file 1: Table S2). 5-year PFS was 83.4% (CI 77.3-88.1), and GTR was achieved in 84.0% of patients (Tables 1, 2).

Discussion
Recent genetic and clinical studies have evaluated clinical/molecular characteristics of meningiomas including all WHO grades [22][23][24][25][26][27][28][29][30][31]; however, recurrence of WHO grade I meningiomas, especially after GTR has not been well-studied based on the assumption that benign meningioma with perfect resection should be of paramount reassurance [3,11]. In the present study, we performed a long-term follow-up study (5.3 ± 4.5 years) for 281 WHO grade I meningiomas. The study revealed the significance of NF2 mutation and/or 22q loss in WHO grade I meningiomas in predicting recurrence, which was intriguingly reinforced by tumour anatomical locations. By integrating driver gene mutation, tumour location, and Ki-67 index, we now present a novel prognostic three-class category to predict tumour recurrence for WHO grade I meningiomas following GTR.

Association between driver gene mutation and recurrence in WHO grade I meningiomas
According to the recent study about the association between subgrouping by driver gene mutation and recurrence of meningiomas [14], only the Hedgehog group remained a significant genetic predictor of tumour recurrence by multivariate analysis in all WHO grade meningiomas. Furthermore, regarding WHO grade I meningiomas, the PFS of the group with the Hedgehog and the tumour necrosis factor-receptor associated factor 7 (TRAF 7) were shorter than that of the NF2 group [14]. In our results, the prognosis of the meningiomas was different mutation-wise (5-year PFS: 77.9% in NF2 vs 90.3% in non-NF2, p = 0.04 (Table 2, Fig. 1C) and univariate analysis identified between worse prognosis and NF2 meningioma (HR:1.9, confidence interval [CI] 1.01-3.58, p = 0.04)", however, the statistical significance was marginal. Corroborating their results, NF2 meningioma per se was not a significant predictor by multivariate analysis for PFS in WHO grade I meningiomas in the current study. This finding is counterintuitive, considering that NF2 alteration are known to initiate events for aggressive-type meningiomas [36]. Actually, the latest trends of the molecular subgroups for meningioma focus on whether NF2 meningioma or not according to the data of single-cell RNA sequencing in meningiomas [25,37,38]. Addressing this pathophysiological question, the present study probed the connotation of driver gene mutation status in association with tumour location and revealed that the clinical significance of the NF2 mutation or 22q loss differed remarkably or inversely with tumour location. The presence of NF2 alteration was found to have a significantly worse prognostic impact in supratentorial meningiomas, while it has a tendency to have a better prognostic effect in infratentorial meningiomas. More specifically, "Supratentorial" and "NF2" was the worst combination as a predictor of PFS in subgroups. And this was echoed by a significantly higher Ki-67 index and FOXM1 expression in this subgroup compared with others.
Recently, meningioma development was reviewed from an embryological perspective by Fountain et al. and Boetto et al. [39,40]. Collaborating these reports and our result, we speculate that the difference in the prognostic implication of NF2 alteration between supra-and infratentorial space may be attributed to the spatial distribution of the tumour cell of origin (neural crest vs mesoderm). Correspondingly, an integrated study surrounding location, phenotype, genotype, and meningeal embryology might be meaningful study in the future.

Association between other variates and recurrence in WHO grade I meningiomas
The EOR has been known as a prognostic factor for meningiomas [4][5][6][7][8][9], and in clinical practice, the Simpson grading system has been commonly utilized as a landmark to predict recurrence of meningiomas [41]. In our study, the EOR was a strong predictor of tumour recurrence. However, from a neurosurgical viewpoint, the EOR is highly and inevitably dependent on tumour location [24,32]. Furthermore, the genomic profile is strongly correlated with tumour location [16][17][18][19][20][21]. Thus, the effect of each variable, including EOR, tumour location, and molecular subgroups, should be equally considered when evaluating meningioma recurrence.
In our results, the GTR/STR rate was not statistically different between supratentorial and infratentorial tumours (Table 2). However, most of the infratentorial tumours in which GTR was not achieved were non-NF2 meningiomas (4.7% in "Infratentorial NF2" vs. 26.0% in "Infratentorial non-NF2", p = 0.04) ( Table 2). In addition, they were mainly located surrounding brainstem such as at clivus, and petro-clival region (Additional file 1: Fig. S3). Previous studies reported that one of the prognostic factors for meningioma recurrence was posterior fossa [15] and that STR was an independent predictor of recurrence in skull base meningiomas [42]. Our results showed short PFS in infratentorial non-NF2 meningiomas (Fig. 1, Table 2). This result needs careful interpretation; one explanation is that only STR was achieved in many of infratentorial non-NF2 meningiomas due to the surgical difficulty related to their proximity to the cranial nerves, vessels, and brainstem. This contributed to a higher rate of recurrence compared with infratentorial NF2 meningiomas, which are often located at cerebellar convexity.
In contrast, even if GTR was achieved, the 5-year PFS in patient with NF2 meningioma remained short in the supratentorial space (Table 2). We speculated that despite being WHO grade I, the short PFS of these meningiomas reflected the aggressive histological features such as high Ki-67 index and high FOXM1 expression. This result should weigh in our decision-making regarding the appropriate management of WHO grade I supratentorial meningioma after GTR. Actually, the prognosis of "Poor" class WHO grade I supratentorial meningioma was quite similar to that of WHO grade II or III meningiomas (Fig. 5). Facing this sobering reality, we here recommend integrated histopathological and molecular diagnosis, close follow-up, and possibly proactive treatment protocol to be implemented in cases of WHO grade I meningioma with GTR when characterized by NF2 alteration, supratentorial location, and high Ki-67 index.
Our study had several limitations that should be addressed in future investigations. First, our study suffers from retrospective, single-institution design, which restricted variables for the assessment to those included in the database. Charts were reviewed retrospectively, and thus, not all clinical/genomic data could be collected. The duration of follow-up was not ideally sufficient to evaluate the prognosis of WHO grade I meningiomas, especially considering that the average time to recurrence was only a little shorter than the median follow-up time. As NF2 mutation was predominantly the commonest mutation in meningioma, we simplified the analyses by categorizing the meningioma as NF2 or non-NF2. In the future, we hope to analyze the effects of several other driver gene mutation types in association with anatomical location in meningioma recurrence. Further external studies on a large number of cases are warranted to validate our results to a better understanding of the clinical behaviour of this disease.
Previously, NF2 alteration, multiple copy number variations (CNV; e.g. 22q loss, 1p loss, etc.), high FOXM1 expression, low immune cell infiltration, and loss of H3K27me3 were reported as characteristics of aggressive behaviour of meningiomas [28,43]. Recent reports have speculated structure variants such as multiple CNVs, especially 1p loss, as an epigenetic "second-hit" following NF2 alteration for aggressive-type meningiomas [22,24,38,45]. Although we posit that supratentorial NF2 meningiomas might have such an aggressive feature as copy number alteration in addition to Ki-67 and FOXM1, it was beyond the scope of this study. Thus, evaluation of CNV status in NF2 mutated meningiomas depending on anatomical location is also needed in the future. In terms of evaluating tumour anatomical location, while there are many ways of categorization, here we divided the cases into supra-or infra-tentorium, which possibly reflected the biological differences of the brain tumours, as evidenced by ependymomas [46]. However, more granular anatomical classification should be utilized ideally for clinical utilization. We will focus on the recurrence and the molecular status of meningiomas in each granular anatomical location.

Conclusion
In conclusion, by conducting this long-term follow-up study of a large number of patients with WHO grade I meningiomas, we demonstrated that the clinical significance of NF2 alteration status in WHO grade I meningiomas was different depending on tumour anatomical location, i.e., either supratentorial or infratentorial. By integrating driver gene mutation and tumour location, the "Supratentorial NF2" subgroup as well as the EOR and Ki-67 index, were identified as a significant predictor of recurrence of WHO grade I meningioma, clinically similar to the poor prognosis of WHO grade II/III. Correspondingly, even if GTR is achieved in WHO grade I meningiomas, this study suggested that close follow-up and proactive measures should be considered in cases characterized by NF2 alteration, supratentorial location, and high Ki-67 index. We anticipate that this integrated anatomical, histopathological, and genomic classification carries significant clinical implications and will provide the best follow-up schedule and proactive measures, as well as improve the daily clinical practice for patients with this most common brain tumour.