The endoscope-assisted supraorbital “keyhole” approach for anterior skull base meningiomas: an updated meta-analysis

Introduction The gold-standard treatment for symptomatic anterior skull base meningiomas is surgical resection. The endoscope-assisted supraorbital “keyhole” approach (eSKA) is a promising technique for surgical resection of olfactory groove (OGM) and tuberculum sellae meningioma (TSM) but has yet to be compared with the microscopic transcranial (mTCA) and the expanded endoscopic endonasal approach (EEA) in the context of existing literature. Methods An updated study-level meta-analysis on surgical outcomes and complications of OGM and TSM operated with the eSKA, mTCA, and EEA was conducted using random-effect models. Results A total of 2285 articles were screened, yielding 96 studies (2191 TSM and 1510 OGM patients). In terms of effectiveness, gross total resection incidence was highest in mTCA (89.6% TSM, 91.1% OGM), followed by eSKA (85.2% TSM, 84.9% OGM) and EEA (83.9% TSM, 82.8% OGM). Additionally, the EEA group had the highest incidence of visual improvement (81.9% TSM, 54.6% OGM), followed by eSKA (65.9% TSM, 52.9% OGM) and mTCA (63.9% TSM, 45.7% OGM). However, in terms of safety, the EEA possessed the highest cerebrospinal fluid leak incidence (9.2% TSM, 14.5% OGM), compared with eSKA (2.1% TSM, 1.6% OGM) and mTCA (1.6% TSM, 6.5% OGM). Finally, mortality and intraoperative arterial injury were 1% or lower across all subgroups. Conclusions In the context of diverse study populations, the eSKA appeared not to be associated with increased adverse outcomes when compared with mTCA and EEA and offered comparable effectiveness. Case-selection is paramount in establishing a role for the eSKA in anterior skull base tumours. Electronic supplementary material The online version of this article (10.1007/s00701-020-04544-x) contains supplementary material, which is available to authorized users.


Introduction
The gold standard treatment for symptomatic anterior skull base meningiomas is complete surgical resection-if possible to do so without causing significant morbidity. Although the traditional microscopic transcranial approach (mTCA) has proven to be effective at removing such tumours [84,86], minimally invasive surgical approaches may offer the possibility of reducing brain exposure and manipulation, and therefore increasing safety [105]. However, these less invasive techniques are often technically challenging with steep learning curves [105]. Factors influencing case-by-case surgical decision-making include the preservation of olfaction and vision, tumour size and location, the involvement of neurovascular structures, surgical experience, and patient choice [24,86].
A previous comprehensive meta-analysis comparing the traditional mTCA and the expanded endoscopic endonasal approach (EEA) found similar gross total resection (GTR) and mortality rates, with more favourable visual outcomes but higher cerebrospinal (CSF) leak incidence with EEA [84]. This generally corroborates with findings from other systematic reviews in the field [24,59,110]. However, a third approach-the endoscope-assisted supraorbital "keyhole" approach (eSKA)-has yet to be compared with mTCA and EEA in the context of existing literature. This approach includes multiple variations (such as the medial supraorbital, basal supraorbital, and lateral supraorbital approaches) that are unified by the principle of achieving surgical control of a deep-seated lesion whilst minimizing iatrogenic injury to the brain (via exposure, retraction, and manipulation) [102,107]. This is achieved through using smaller craniotomies with smaller dural openings and may theoretically reduce postoperative complications and length of stay, whilst improving cosmesis, patient satisfaction and carrying lower CSF leak rates than the EEA [102,104,105,107]. However, important limitations of the eSKA include (a) difficult visualization and orientation of deep structures, (b) difficult (almost co-axial) instrument control owing to instrument size and the fulcrum effect (requiring specialized instruments), and (c) limited and predefined surgical corridors which require extensive preoperative planning [102,107]. Endoscope assistance provides a high light intensity with wider viewing angles distal to the craniotomy, allowing high-resolution visualization of deeper tissues. Indeed, combined with image-guidance systems and intra-operative adjuncts (e.g. ultrasound, MRI), endoscopes facilitate surgical orientation and resection during keyhole approaches [102,107].
Therefore, we updated a previous systematic review and meta-analysis comparing mTCA with EEA and extended this review with the eSKA for the management of olfactory groove (OGM) and tuberculum sellae meningiomas (TSM).

Methods
In order to identify studies reporting on outcomes of surgically treated TSMs and OGMs, we adapted our previous methodology [84], expanding our search to include eSKA and updating our search to include mTCA and EEA articles published after our last search.

Search strategy
This review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement [81]. A search strategy was created using the keywords "Meningioma," "Tuberculum Sellae," "Olfactory Groove," and synonyms (Appendix A). Studies were included if they reported on (1) patients with olfactory groove (OGM) or tuberculum sellae (TSM) meningiomas; (2) patients undergoing surgery using the mTCA, EEA or eSKA approaches; and (3) surgical outcomes and complications. Exclusion criteria included case reports, commentaries, congress abstracts, reviews, animal studies, studies describing a combined surgical approach (for example EEA + mTCA), studies in paediatric patients (< 18 years old), re-operations, and cadaveric studies. A date filter was applied, with articles from 2004 to 2020 being included-reflecting a period of the contemporary adaptation of endoscopic and keyhole approaches and the continuous improvement of traditional microsurgical approaches for the relevant pathologies [11,15,32,106].
Both PubMed and Embase databases were searched on 19 April 2020. Duplicates were removed using Endnote X9. Independent title and abstract screening of updated results was performed in duplicate by two authors (DZK, HJM). Review of full-text articles ensued according to the inclusion/exclusion criteria. Any discrepancies in selection were settled out by discussion and mutual agreement.

Data extraction
Data points extracted from the included articles comprised of patient characteristics (age, sex distribution), tumour characteristics (surgical approach used, sum of sample, tumour grade, tumour diameter or volume, follow-up length), outcomes (GTR, visual improvement), and complications (CSF leak, 30-day mortality, intra-operative arterial injury). World Health Organization (WHO) grading included recording the proportion of WHO Grade 1 tumours [72]. Of note, the grading system was revised in 2016 to include brain invasion as a criterion to upgrade Grade 1 tumours to Grade 2 [72]. Therefore, if any studies pre-2016 reported brain invasion amongst Grade 1 tumours, the respective tumours were upgraded accordingly [72]. Gross total resection (GTR) referred to Simpson Grades 1 and 2 as per our original methodology [84,116]. Visual improvement was in the context of those with preoperative visual problems only. Mortality (within 30 days after resection) was recorded on an all-cause basis.
Owing to the not uncommon reporting of follow-up time as a median number of months (as opposed to mean), the estimated mean number of months was calculated as per recommendations of Hozo et al. [53]. Of note, in sample sizes greater than 25, the sample's median follow-up is presented as the best estimate of the mean [53].
Importantly, studies that did not report specific outcomes for OGM/TSM and approach combination were excluded from the final meta-analysis. These studies were considered for qualitative analysis if the relevant the tumour (TSM or OGM) and approach (mTCA, EEA, or eSKA) combination of interest was > 50% of the study population [9,48,98,105,111]. Similarly, articles that grouped TSM cases with planum sphenoidale meningiomas [3,92] were considered for qualitative review only (owing to the similarity of these tumour groups) but not included in the final meta-analysis.

Risk of bias assessment
Study quality was assessed with a modified New-Castle Ottawa Scale (mNOS), which assesses two domains: sample selection and outcome reporting. The modification made to the original NOS was the exclusion of the "comparability" domain as this is not applicable to case series [126]. The scale is scored out of 6 (3 for selection domain, 3 for outcome domain). Publication bias was assessed using Begg's tests [8] and by generating funnel plots with and without trimand-fill method [34].

Meta-analysis
A meta-analysis was conducted using R 3.6.1 (The R Foundation, Austria) applying the "meta" package. Pooled incidence (using the random-effect model method of DerSimonian and Laird [33]) was calculated for each approach (eSKA, mTCA, EEA), tumour (TSM, OGM), and outcome (GTR, arterial injury, visual improvement, CSF leakage, and mortality) combination. Study heterogeneity was assessed by calculating I-squared values [52] (I 2 > 50% considered significant) and Cochran's Q test (p < 0.10) [36,52]. Sensitivity analysis was performed by running the above analyses on a low risk of bias sub-group (mNOS score greater than or equal to 4).
Additionally, a univariate meta-regression was performed to explore the effect of mean age (continuous variable) and male percentage (continuous variable) on each approach, tumour, and outcome combination. Meta-regression was only performed if 8 or more studies were available for the outcome/ approach/tumour combination being explored. This threshold was chosen (a deviation from the standard threshold of 10) on a pragmatic basis, to reflect the relative paucity of literature from the newer eSKA approach [51]. This threshold was also applied to the performance of Begg's test, trim-and-fill analysis, and the generation of funnel plots.

General characteristics
The median number of patients per study was 20 (range: 3-95) for TSM and 19.5 (range: 2-129) for OGM. The average percentage of male patients was 24% for TSM and 31% for OGM. The median mean of age was 54.2 years for TSM and 54.75 years for OGM. The median mean of follow-up time for TSM was 32 months (reported in 55/67 studies) for and 44.5 months for OGM studies (reported in 39/43 studies). The modified NOS score varied between 2/6 and 6/6 amongst the TSM and OGM case series, with predominant factors affecting this variance being a description of follow-up and outcome reporting. Summary characteristics by approach (eSKA, EEA, or mTCA) are highlighted in Tables 1 and 2. Individual study characteristics are displayed in Tables 5 and 6 (Appendix B).  eSKA (I 2 = 68%, Cochran's p < 0.01) and mTCA (I 2 = 60%, Cochran's p < 0.01) groups, with Begg's test for publication bias also significant in this mTCA group (p < 0.01) ( Table 3). This impacted funnel plot asymmetry, which was most marked in the mTCA group, without any major change in summary effect using trim and fill across subgroups (Appendix D). Meta-regression suggests male sex was significantly associated with lower GTR incidence in EEA (slope − 0.05 (95% CI − 0.96-0.04)) and mTCA (slope − 0.27 (95% CI − 0.53 to − 0.01)) subgroups (Table 3).

Tuberculum sellae meningioma
Pre-operative visual impairment was reported in 6 eSKA (77 patients), 19 EEA (366 patients), and 36 mTCA (1280  However, study heterogeneity was significant within the EEA (I 2 = 88%, Cochran's p < 0.01) and mTCA (I 2 = 89%, Cochran's p < 0.01) groups. Publication bias was not evident on Begg's testing, with mild funnel plot asymmetry in mTCA and EEA groups likely due to heterogeneity. This is supported by the lack of its major change in summary effect using trimand-fill across subgroups (Appendix D). Meta-regression on age and sex did not reach statistical significance across mTCA, EEA, and eSKA groups (Table 3).

Cerebrospinal fluid leakage
Tuberculum sellae meningioma Incidence of postoperative CSF leakage was reported in 9 eSKA (108 patients), 23 EEA (473 patients), and 32 mTCA (1182 patients) studies. The pooled incidence of CSF leak ( Fig. 2; Appendix C) in the EEA group was 9.19% (95% CI 5.33-13.05), which was higher than the incidence observed among the eSKA treated group at 2.11% (95% CI 0-5.84) and mTCA treated group at 1.58% (95% CI 0.68-2.48). However, Publication bias was also evident in the mTCA group (Begg's p ≤ 0.01). The asymmetry of mTCA and EEA funnel plots is explained by the above (Appendix D), but no major change in summary effect using the trim-and-fill method is appreciable in these groups. Meta-regression on age and sex did not reach statistical significance across any group (Table 3).

Olfactory groove meningioma
Incidence of intraoperative arterial injury was reported in 9 eSKA (96 patients), 10 EEA (115 patients), and 27 mTCA (1142 patients) studies. Pooled incidence ( Fig. 3; Appendix C) was highest in the EEA group at 1.25% (95% CI 0-4.32), followed by the mTCA at 0.15% (95% CI 0-0.7) and eSKA with 0% (95% CI 0-3.52). Indeed, these results do not align with the TSM group. Across all 3 groups, study heterogeneity was not apparent; however, publication bias using Begg's test reached statistical significance in EEA (p = 0.02) and mTCA (p < 0.01) groups, mapping to funnel plot asymmetry in EEA and mTCA groups. However, trim and fill adjustment made only minor differences to the estimated summary effect (Appendix D). Again, meta-regression did not show a significant effect of age and sex on arterial injury (Table 4).

30-day mortality
Tuberculum sellae meningioma Incidence of mortality was reported in 11 eSKA (128 patients), 23 EEA (471 patients), and 37 mTCA (1283 patients) studies. Pooled incidence of 30-day mortality ( Fig. 2; Appendix C) was 0.6% (95% CI 0-1.74) in the EEA group, followed by 0.56% (95% CI 0-1.2) in mTCA and 0.33% (95% CI 0-3.21) in eSKA in descending order. Across all 3 groups, study heterogeneity was not apparent; however, publication bias using Begg's test reached statistical significance in all three groups (p < 0.01). Resultantly, the mTCA and eSKA funnel plots are asymmetrical but are not appreciably impacted in terms of summary effect by the implementation of trim and fill (Appendix D). Meta-regression did not show a significant effect of age and sex on mortality across subgroups (Table 3).

Sensitivity analysis with low risk of bias studies
The pooled incidence of surgical outcomes of a subgroup of low-risk studies (defined as mNOS score greater than or equal to 4) is presented in Appendix E. This analysis, when compared with the total group analysis, yielded overall similar results for GTR, mortality, and intraoperative arterial injury. CSF leak incidence after EEA was apparently lower (in both OGM and TSM), and visual improvements after EEA (in the TSM group) were more marked in the lower risk of bias studies.

Principle findings
In the context of heterogeneous study populations and outcome reporting, the endoscope-assisted supraorbital "keyhole" approach appeared to be associated with similar effectiveness (GTR, visual improvement) and safety (CSF leak, 30-day mortality) compared with the mTCA and EEA alternatives based on our findings. Case selection and an understanding of relative indications are important in selecting the most appropriate approach for anterior skull base meningioma resection. As previously found, the EEA was associated with the highest rates of visual improvement across OGM and TSM groups. However, this advantage of EEA may be offset when considering the safety profile of the three approaches, with the EEA having the highest incidence of post-op CSF leak (statistically significant in the TSM sub-group). In contrast, the mTCA had a slightly higher incidence of GTR than eSKA and EEA (eSKA > EEA) across TSM and OGM groups. Interestingly, the eSKA displays intermediate results in terms of efficacy (GTR and visual improvement) and complications (CSF leak). Results for intra-operative arterial injury and 30day mortality incidences are similar and overlapping across the 3 approaches. Indeed, the eSKA, as a relatively new technique, is less well explored. When compared with the mTCA, the eSKA-as a minimally invasive technique-offers a smaller craniotomy scar, less brain exposure, and less brain/nerve retraction [105]. Thus, theoretically, it shares similar limitations to the minimally invasive EEA [92]-potentially making total resection of larger tumours or tumours with significant local invasion difficult [24,92,105]. However, when performed with the benefit of neuronavigation, neuroendoscopy (12/13 of eSKA studies in our meta-analysis), and appropriate surgical training, the eSKA has been used to resect relatively large tumours of the anterior skull base [4,41,67,105].
All 3 approaches likely have their own role in the management of anterior skull base meningiomas, with their varying safety and efficacy profiles as evidenced above. Case selection will be paramount in establishing a role for each technique/ combination of techniques [4,92,105]. Indeed, case selection of eSKA is currently considerably variable, owing to its novelty and ongoing refinement [70,92]. The selection of the preferred approach for each case must be taken in the context: (a) patientrelated factors (demographic, presentation, preferences), (b) tumour-related factors (size, consistency, extension, locationsuch as relation to optic foramen or cribriform plate), and (c) surgeon experience, surgeon preference, and surgical goals (such as GTR or STR, visual or olfactory preservation) [2,77,92,105,118]. Ottenhausen et al. presents a concise decision-making algorithm (based on tumour anatomy and resulting functional deficits), which incorporates the specific characteristics of eSKA, EEA, and mTCA approaches [92]. In this algorithm, the eSKA is suitable for TSMs with lateral extension up to the internal carotid arteries (ICA) and anterior clinoid processes (ACP), or lateral extension beyond the lamina papyracea (LP). Additionally, the eSKA is suggested for OGMs with (1) preserved olfaction and (2) disrupted olfaction without cribriform plate invasion but with significant anterior (up to the frontal sinus) or lateral extension. In contrast, the EEA is proposed for TSM without significant lateral extension (ICA/ACP/LP as above) and OGMs without significant lateral extension (where olfaction is disrupted). Finally, an mTCA or a combined EEA + eSKA approach is suggested for OGMs and TSMs with a significant anterior or lateral extension (unless there is no cribriform plate invasion, in which case, eSKA alone may be possible) [92]. Of note, other algorithms cite > 5mm sellar extension and optic canal involvement as factors favouring EEA in TSM [60]. During EEA for TSM, decompression of the optic canal from below avoids excessive vascular manipulation, can be achieved before or after tumour resection, and is well suited to tumours with extension into the inferomedial aspect of the optic canal [2,69]. Decompression of the involved optic canal is described in mTCA approaches with early decompression (before tumour resection) favoured [20,80,90]. In eSKA, studies describe both early and late bilateral canal decompression with optimum timing being less clear [16,67]. More generally, within the literature, consensus for the ideal surgical approach in various contexts is not clear [41,60,67,74,92,99]. Indeed, in light of the COVID-19 pandemic-which elucidated to the risk endonasal surgery may pose (exposing theatre staff to high viral loads and potentially serious infection)-this case selection process is likely to be a dynamic field in the near future [71,100,119].

Findings in the context of other syntheses
Previous meta-analyses have compared the EEA and mTCA (not eSKA) with varying results.
In terms of GTR, Muskens et al. (co-author) previously found higher GTR incidence with mTCA in OGM at 88 [115].
Regarding visual outcomes, a recent comparative metaanalysis by Lu et al. suggests improved visual outcome in OGM resection using the EEA (vs. mTCA) (OR, 0.318; p = 0.04) but not statistically significant in TSM [73]. This is slightly different from our updated findings and previous findings of Muskens et al. [84], in which the visual outcome advantage of EEA was most prominent in the TSM group. In other analyses, an early (2013) [110,115].

Limitations and strengths
The principal limitations of this study are the likely prevalent publication bias and heterogeneity of the primary literature synthesized, more specifically heterogeneity in the reporting of baseline characteristics and outcomes. This is reflected in the I 2 and Cochran Q tests highlight in Figs. 2 and 3, corroborating with respective funnel charts (Appendix B). Development of core data set, through a multi-stakeholder consensus process for example, would be useful for future pooled analysis in the field. Secondly, it is likely the study populations examined are considerably variable owing to the surgical decision-making process that informs the choice of approach. Larger, more extensive tumours may be more likely to undergo traditional open approaches (of which there are many variants) in order to achieve acceptable tumour resection [84]. This is reflected in Tables 1 and 2 where larger (diameter and/or volume) tumours are included in the mTCA group compared with the EEA group. Unfortunately, we were not able to perform meta-regression based on tumour size or grade owing to heterogeneous data reporting, potentially adding to confounding factors [18,46]. Additionally, the overwhelming majority of studies included were case series, and thus, our interpretation of our results should be tempered to reflect this. Finally, owing to the novelty of the approach, there is a relative paucity in the amount of included eSKA studies. Although overall, the results of the main analysis are similar to that of the sensitivity analysis subgroup, the number of low risk of bias studies analyzed is also limited. Future studies in the field must improve on methodological design, with an emphasis on comparative studies, in order to facilitate more robust data synthesis.

Conclusions
In the context of diverse study populations and heterogeneous case selection criteria, the endoscope-assisted supraorbital keyhole approach appeared not to be associated with increased adverse outcomes when compared with expanded endonasal and traditional transcranial approaches and offered comparable effectiveness. Case selection is paramount in establishing a role for the supraorbital keyhole approach in anterior skull base tumours. Development of standardized research databases and well-designed comparative studies that control for selection and confounding biases are needed to further delineate these selection criteria.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by DZK, ISM, RAM, AHZ, and HJM. The first draft of the manuscript was written by DZK, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding No specific funding was received for this research. HJM is funded by the Wellcome/EPSRC Centre for Interventional and Surgical Sciences (WEISS) and the National Institute of Health Research University College London Biomedical Research Centre. AEH is supported by the Cambridge Biomedical Research Centre.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval Ethical approval was unnecessary due to the nature of the study (study-level meta-analysis).
Informed consent Not applicable due to the nature of the study (studylevel meta-analysis).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.