Introduction

The primary treatment for meningiomas remains surgical resection.1 Often, such tumors tend to be highly vascularized and intraoperative bleeding may render resection challenging. As such, neoadjunctive embolization of such tumors has become more commonplace over the past several decades with the underlying aim of reducing intraoperative bleeding and consequently making surgical resection safer.2–4 Successful embolization has been reported with the use of multiple embolic agents, most commonly polyvinyl alcohol (PVA) particles and n-butyl cyanoacrylate (nBCA).5 However, preoperative embolization carries risks including cranial nerve deficit, intratumoral hemorrhage, ischemic stroke, infection, post-embolization edema, and angiographic complications.6 7 The presence of high-risk external–internal carotid artery anastomosis and targeting of pial supply also substantially increases the risk of ischemic complications.8

Prior retrospective studies have yielded mixed results, which can be attributed in part to confounding factors such as meningioma size, tumor location, involvement of vascular structures and cranial nerves as well as inherent selection bias.9 There is variability in correlation of surgical outcomes with degree of embolization. Some studies show degree of embolization correlating with a specific surgical outcome, while others show no difference for the same outcome.10–13 In this study we aimed to retrospectively review the anatomic locations and vascular supply of meningiomas to determine the relationship between these characteristics and successful preoperative embolization.

Methods

Results

Of 139 meningioma patients undergoing preoperative angiography in the study period, 78% (108/139) were successfully embolized (partial+complete, tables 1 and 2). A total of 31% (33/108) of patients achieved complete angiographic embolization. As shown in table 3, significant multivariate predictors of partial or complete embolization were convexity/parasagittal locations (OR 5.15, 95% CI 0.93 to 28.54, p=0.060), MHT (OR 4.65, 95% CI 1.63 to 13.23, p=0.004), MMA (OR 10.89, 95% CI 3.43 to 34.64, p<0.001), and APA (OR 9.96, 95% CI 1.88 to 52.73, p=0.007) arterial supply. Significant predictors for complete embolization were convexity or parasagittal locations (OR 4.79, 95% CI 1.66 to 13.84, p=0.004) and an embolized APA supply (OR 6.94, 95% CI 1.90 to 25.39, p=0.003). Multiple arterial supply was predictive of unsuccessful embolization (OR 0.38, 95% CI 0.15 to 0.98, p=0.05). The WHO grade of the resected meningioma did not correlate with preoperative partial or complete embolization.

There were three procedure-related complications, as shown in table 4. One involved a middle meningeal artery branch perforation during embolization which required coiling. Another was catheter-induced vasospasm, which self-resolved by the end of the angiogram. There was also one instance of cervical internal carotid artery (ICA) dissection which required aspirin administration. There were no temporary or permanent neurologic deficits associated with any complications. Due to the small number of complications observed, it was not possible to identify factors predicting complications in a statistically valid fashion.

Discussion

The goal of this research was to identify specific meningioma characteristics most conducive to successful preoperative embolization. The benefits of preoperative meningioma embolization are widely reported in the literature; however, the question of whether preoperative angiography with the intention to embolize should be performed for most intracranial meningiomas remains controversial.10 15–17 Factors such as anatomic location, arterial supply, size, embolic material, endovascular skill/technique, interval to resection, and tumor grade likely contribute to this variance. Our research shows that parasagittal and convexity meningiomas and those with APA supply were most likely to achieve complete embolization.

Studies have analyzed the safety and efficacy of different particulate and liquid embolic agents. Our center prefers PVA particles as the embolic agent, while others advocate Embospheres (trisacryl gelatin),18 liquid embolic agents such as Onyx (ethyl vinyl alcohol copolymer) or nBCA.19–21 Studies have not shown a significant correlation between different embolic agents and degree of angiographic devascularization.11 Manaka et al reported combining the particulate approach with proximal feeder vessel coiling to prevent recanalization of embolized vessels, theoretically reducing the risk of intratumoral hemorrhage.18 Complete angiographic devascularization was achieved in 18.7% of their patients, slightly lower than our center’s rate of 30%. However, their series of 75 embolizations had two post-embolization intratumoral hemorrhages requiring emergent craniotomy for tumor resection. Both of these cases were partially embolized sphenoid ridge meningiomas with major supply from the MMA, neither of which had APA feeders. Despite these complications, the success of embolization was rated as ‘excellent’ or ‘moderate’ in 99% of patients, based on the surgeon’s subjective assessment of the facilitated resection by softening and whitening of the tumor. Surgeon-reported outcome of intraoperative ‘ease of resection’ is subjective and difficult to standardize and quantify across studies, but is the only consistently positive outcome identified across preoperative embolization studies that have reported this variable.18

A study by Gruber et al reviewed 14 patients and demonstrated radiographic evidence of tumor devascularization using MRI at 6 hours and 48 hours post-embolization using sodium polymethacrylate particulate agent.22 Apparent diffusion coefficient, relative cerebral blood flow, and relative cerebral blood volume MRI sequences all showed significant reduction at 6 hours post-embolization, with no significant progression after 48 hours. The majority of these meningiomas were falcine, convexity or frontal tumors rather than skull base tumors. The extent of angiographic devascularization or radiographic ischemia on MRI did not correlate with subsequent intraoperative blood loss, degree of surgical resection, or patient functional outcome at follow-up in their study. These authors suggested meningiomas with dominant dural arterial supply are most amenable to safe and effective embolization, similar to our study results.22 Of note, aside from the use of diagnostic cerebral angiography and MRI in an attempt to assess embolization efficacy in both a qualitative and quantitative fashion, more recent efforts by Wen et al have proposed the use of semiquantitative analysis by flat-detector CT imaging to assess for degree of tumor embolization.23

Published reports on surgical outcomes associated with the degree of preoperative embolization show varying results. Iacobucci et al found a significant decrease in surgical time comparing patients with extensive (>90%) or complete embolization to patients who did not undergo preoperative embolization with the use of PVA particles.23 However, there was no difference in terms of intraoperative blood loss, measured by units of blood transfusion, and changes in preoperative to postoperative hemoglobin concentration between the two groups. Bendszus et al and Borg et al found a reduction in estimated blood loss when comparing extensively embolized (>90%) to only partially embolized tumors.11 12 Other studies have failed to demonstrate significant correlation between degree of preoperative embolization and blood loss or extent of resection.10 13 22 24 A recent study of 28 patients with skull base meningioma did not find any association between arterial supply or anatomic location and degree of devascularization.20 In contrast, our analysis demonstrates an association between ascending pharyngeal arterial supply and parasagittal or convexity location with the degree of embolization. While our study did not attempt to assess surgical outcomes, this is an important direction for future research.

The APA is well reported in the literature as a suitable option for embolization of skull base meningiomas, paragangliomas, hemangiopericytomas, and dural arteriovenous fistulas. This has been described with nBCA, PVA particles, and Onyx.25 26 Despite its smaller caliber (mean 1.54 mm) compared with other external carotid artery (ECA) branches, superselective catheterization of the vessel is possible and safe. Anastomoses with the ICA and vertebrobasilar system are crucial to identify, as is the meningeal branch, which supplies cranial nerves IX, X, XI, and XII. Arising from the ECA in approximately 80% of patients, it can alternatively branch off the ICA, occipital artery, common carotid artery bifurcation, or a common trunk with lingual and facial artery.27 With careful attention to the APA’s origin, branches, and anastomoses, it is possible to selectively embolize feeder pedicles to the target lesion and avoid both lower cranial nerve deficits and inadvertent thrombi to other locations.

The overarching goal of neoadjuvant meningioma embolization remains complication avoidance and patient safety rather than complete devascularization. The patients in our study did not experience any severe complications resulting in temporary or permanent neurologic deficits. None of the three procedure-related complications in our series of 108 embolized patients (2.8%) were associated with complete embolization. This complication rate is slightly lower than another recent series reporting 4.6% procedure-related complications overall,17 and may reflect patient selection based on diagnostic angiography prior to embolization.

A serious potential complication of preoperative meningioma embolization is blindness due to inadvertent central retinal artery (CRA) ischemia, caused by anomalous origin of the ophthalmic artery (OA) arising from the MMA. One study reported a 1.45% incidence of OA origin from the MMA stem or its branches, entering the orbit via the superior orbital fissure instead of the optic canal, among 1652 patients.28 This meningo-ophthalmic artery anomaly has several variants, including a single OA arising from the MMA, dual OAs with ICA and MMA origins, and the MMA arising from the OA. For the dual OA variant, it is critical to recognize which OA, either from ICA or ECA origin, is the parent artery for the CRA, which can occur from either trunk. The incidence of this anomaly is not insignificant, and therefore it is crucial for the treating surgeon to identify its presence during preoperative meningioma embolization to reduce the risk of CRA thrombosis and vision loss. In one study of 167 cranial base meningiomas, 2% of those patients embolized via the MMA had inadvertent CRA thrombosis, resulting in permanent visual deficits for both cases.29 Ectopic OA origin is also important to identify prior to meningioma resection, particularly for surgical approaches along the sphenoid ridge and superior orbital fissure.30

This study, similar to other contemporary published studies examining the utility of neoadjuvant embolization of intracranial meningiomas, is limited by considerations including its retrospective nature, inherent bias due to patient selection, and confounding factors such as tumor size, patient comorbidities, and surgeon experience. The focus of this research was the association of tumor characteristics with embolization outcome to inform decisions regarding cerebral angiograms for preoperative meningioma embolization. Surgical outcomes such as operative time, estimated blood loss, recurrence rate, and patient functional status on discharge or follow-up were beyond the scope of this work. Despite these limitations, this study has several strengths, including a focus on arterial supply and an in-depth analysis of anatomic and angiographic characteristics of meningiomas for preoperative embolization.

Conclusions

Although arterial supplies via the MMA and MHT are readily embolized, parasagittal or convexity location and APA supply are more likely to be associated with complete embolization. These data show that tumor characteristics, including arterial supply and anatomic location, can be used to inform decision-making regarding preoperative meningioma embolization.