Browse the corpus

Angiographic Features of Meningiomas Predicting Extent of Preoperative Embolization. BACKGROUND AND OBJECTIVES: Preoperative embolization is used as an endovascular adjunct to surgical resection of meningiomas. However, there is no standardized system to assess the efficacy or extent of embolization during the embolization procedure. We sought to establish a purely angiographic grading system to facilitate consistent reporting of the outcome of meningioma embolization and to characterize the anatomic and other features of meningiomas that predict the degree of devascularization achieved through preoperative embolization. METHODS: We identified patients with meningiomas who underwent preoperative cerebral angiography and subsequent resection between 2015 and 2021. Demographic, clinical, and imaging data were collected in a research registry. We defined an angiographic devascularization grading scale as follows: grade 0 for no embolization, 1 for partial embolization, 2 for majority embolization, 3 for complete external carotid artery embolization, and 4 for complete embolization. RESULTS: Eighty consecutive patients were included, 60 of whom underwent preoperative tumor embolization (20 underwent angiography with an intention to treat but ultimately not embolization). Embolized tumors were larger (59.0 vs 35.9 cc; P = .03). Gross total resection, length of stay, and complication rates did not differ among groups. The distribution of arterial feeders differed significantly across tumors in a location-specific manner. Both the tumor location and the identity of arterial feeders were predictive of the extent of embolization. Anterior midline meningiomas were associated with internal carotid (ophthalmic, ethmoidal) supply and lower devascularization grades ( P = .03). Tumors fed by meningeal feeders (convexity, falcine, lateral sphenoid wing) were associated with higher devascularization grades ( P < .01). The procedural complication rate for tumor embolization was 2.5%. CONCLUSION: Angiographic outcomes can be graded to indicate the extent of tumor embolization. This system may facilitate consistency of reported angiographic results. In addition, arterial feeders vary in a manner predicted by tumor location, and these patterns correlate with typical degrees of devascularization achieved in those tumor locations.

Our institutional procedural database was accessed to identify all patients with meningiomas who underwent preoperative cerebral angiography with or without embolization before craniotomy for resection between 2015 and 2021. Institutional review board approval was obtained with a waiver of informed consent, because of the retrospective nature of the study (Institutional Review Board Serial Number: 23-01539). Demographic, clinical, imaging, and perioperative data were collected and included sex, age, tumor volume, days from embolization to craniotomy, World Health Organization (WHO) grading, tumor location, arterial feeders, embolization grade, gross total resection according to surgeon assessment and as determined by postoperative MRI, length of stay, and complications. The use of n-butyl cyanoacrylate (n-BCA) glue for tumor embolization was off-label and investigational as of the time of this study.

ing, tumor location, arterial feeders, embolization grade, gross total resection according to surgeon assessment and as determined by postoperative MRI, length of stay, and complications. The use of n-butyl cyanoacrylate (n-BCA) glue for tumor embolization was off-label and investigational as of the time of this study. To facilitate analysis, all meningiomas were organized according to their location, in the following categories: Anterior fossa (lateral), anterior midline (including olfactory groove, planum sphenoidale, and tuberculum sellae), clinoidal/clival, convexity, cerebellopontine angle, foramen magnum, parasagittal/falcine, petrous apex, sphenoid wing, and tentorial (an additional category was used to capture all other regions, including the spine). All radiographic parameters, including anatomic location, were assessed by 2 radiologists, with any discrepancies decided by consensus review with a third radiologist. Arterial feeders were organized into the following categories: External carotid artery (ECA): meningeal (for all meningeal feeders, including middle meningeal artery), other internal maxillary artery feeders, occipital artery, ascending pharyngeal artery, and other ECA feeders; Internal carotid artery (ICA): ophthalmic, anterior cerebral artery (ACA), middle cerebral artery, and extradural ICA feeders (meningohypophyseal trunk and inferolateral trunk); and Posterior circulation: vertebral artery, anterior inferior cerebellar artery, superior cerebellar artery, and posterior cerebral artery. Figure 1 illustrates the anatomic locations defined in this study.

ry (ACA), middle cerebral artery, and extradural ICA feeders (meningohypophyseal trunk and inferolateral trunk); and Posterior circulation: vertebral artery, anterior inferior cerebellar artery, superior cerebellar artery, and posterior cerebral artery. Figure 1 illustrates the anatomic locations defined in this study. Anatomic classification of meningiomas in this study. Illustrations of the skull base from 2 orientations, A, from above and B, from a lateral position, highlight the anatomic regions used in this study to classify the meningioma location. Used with permission from ©Mount Sinai Health System. To allow for relatively objective assessment of the angiographic images, we introduced an angiographic devascularization grading scale as follows: grade 0 for no embolization, grade 1 for “partial” embolization (defined as <50% reduction of initial tumor blush), grade 2 for “majority” embolization (50%-99% reduction of initial tumor blush), grade 3 for complete embolization of ECA circulation supply with possible residual internal carotid or posterior circulation supply, and grade 4 for complete embolization with no residual internal, external, or posterior circulation supply.

grade 2 for “majority” embolization (50%-99% reduction of initial tumor blush), grade 3 for complete embolization of ECA circulation supply with possible residual internal carotid or posterior circulation supply, and grade 4 for complete embolization with no residual internal, external, or posterior circulation supply. Embolized and nonembolized tumors were compared in several ways. Frequencies and percentages described categorical variables. Continuous variables were assessed for normality using the Shapiro-Wilk Normality. Normally distributed means reported means and SD, whereas non-normally distributed variables reported medians and IQRs. Categorical variables were compared between groups using the χ2 test. Normally distributed outcome metrics were compared between groups using the independent 2-sample t-test. Non-normally disquieted time and outcome metrics were compared between groups using the Mann-Whitney U (Wilcoxon rank-sum) test. Fisher's exact test was used to evaluate angiographic vs embolization treatment outcome according to tumor location. A sample test of equal proportions with Holm-Bonferroni–adjusted P-values was used to compare the distribution of each feeder vessel by tumor location. P-values less than .05 were used as a threshold for statistical significance throughout the study. All statistical analyses were performed using R software version 4.1.2 (R Foundation for Statistical Computing).

To facilitate analysis, all meningiomas were organized according to their location, in the following categories: Anterior fossa (lateral), anterior midline (including olfactory groove, planum sphenoidale, and tuberculum sellae), clinoidal/clival, convexity, cerebellopontine angle, foramen magnum, parasagittal/falcine, petrous apex, sphenoid wing, and tentorial (an additional category was used to capture all other regions, including the spine). All radiographic parameters, including anatomic location, were assessed by 2 radiologists, with any discrepancies decided by consensus review with a third radiologist. Arterial feeders were organized into the following categories: External carotid artery (ECA): meningeal (for all meningeal feeders, including middle meningeal artery), other internal maxillary artery feeders, occipital artery, ascending pharyngeal artery, and other ECA feeders; Internal carotid artery (ICA): ophthalmic, anterior cerebral artery (ACA), middle cerebral artery, and extradural ICA feeders (meningohypophyseal trunk and inferolateral trunk); and Posterior circulation: vertebral artery, anterior inferior cerebellar artery, superior cerebellar artery, and posterior cerebral artery. Figure 1 illustrates the anatomic locations defined in this study. Anatomic classification of meningiomas in this study. Illustrations of the skull base from 2 orientations, A, from above and B, from a lateral position, highlight the anatomic regions used in this study to classify the meningioma location. Used with permission from ©Mount Sinai Health System.

To allow for relatively objective assessment of the angiographic images, we introduced an angiographic devascularization grading scale as follows: grade 0 for no embolization, grade 1 for “partial” embolization (defined as <50% reduction of initial tumor blush), grade 2 for “majority” embolization (50%-99% reduction of initial tumor blush), grade 3 for complete embolization of ECA circulation supply with possible residual internal carotid or posterior circulation supply, and grade 4 for complete embolization with no residual internal, external, or posterior circulation supply.

Embolized and nonembolized tumors were compared in several ways. Frequencies and percentages described categorical variables. Continuous variables were assessed for normality using the Shapiro-Wilk Normality. Normally distributed means reported means and SD, whereas non-normally distributed variables reported medians and IQRs. Categorical variables were compared between groups using the χ2 test. Normally distributed outcome metrics were compared between groups using the independent 2-sample t-test. Non-normally disquieted time and outcome metrics were compared between groups using the Mann-Whitney U (Wilcoxon rank-sum) test. Fisher's exact test was used to evaluate angiographic vs embolization treatment outcome according to tumor location. A sample test of equal proportions with Holm-Bonferroni–adjusted P-values was used to compare the distribution of each feeder vessel by tumor location. P-values less than .05 were used as a threshold for statistical significance throughout the study. All statistical analyses were performed using R software version 4.1.2 (R Foundation for Statistical Computing).

A total of 80 consecutive patients were identified, 60 (75%) of whom underwent preoperative tumor embolization (20 underwent angiography without embolization). Baseline characteristics, including female sex (40 [66.6%] in the embolization group vs 12 [60%] in the angiogram group, P = .79), mean (SD) age (59.9 [12.5] vs 57.5 [16.5], P = .56), and WHO grading, were comparable in the 2 groups (Table 1). Median time from embolization to craniotomy was shorter than median time from angiogram to craniotomy (2 [2] days vs 4 [3] days, P = .003). In addition, embolized tumors were larger compared with nonembolized tumors (59.0 vs 35.9 cc; P = .054). Regarding outcomes, gross total resection as assessed intraoperatively by the surgeon and postoperatively by MRI, length of stay, and rate of postoperative complications did not differ between groups (Table 1). Baseline Characteristics and Outcomes OR, odds ratio; WHO, World Health Organization. Normally distributed variables for which the independent two-sample t-test was used. Non-normally distributed variables for which the Wilcoxon rank-sum test was used. Categorical data for which Fisher’s exact test was used. Summary data from the cohort of (n = 80) consecutive patients who underwent embolization before meningioma resection, including patient demographics, tumor characteristics, and patient outcomes. The rows in bold reflect arterial branches with large enough sample sizes to enable further statistical analysis.

Categorical data for which Fisher’s exact test was used. Summary data from the cohort of (n = 80) consecutive patients who underwent embolization before meningioma resection, including patient demographics, tumor characteristics, and patient outcomes. The rows in bold reflect arterial branches with large enough sample sizes to enable further statistical analysis. There were 2 procedural complications in this cohort. One patient experienced a superficial access site infection that resolved without the need for antibiotics. One patient experienced an oculomotor nerve palsy after embolization of extradural branches of the ICA supplying a medial sphenoid wing meningioma. Figures 2-8 illustrate angiographic and anatomic features of meningiomas in the most common anatomic locations in this study, in the context of tumors embolized with each devascularization grade from 0 through 4.13

There were 2 procedural complications in this cohort. One patient experienced a superficial access site infection that resolved without the need for antibiotics. One patient experienced an oculomotor nerve palsy after embolization of extradural branches of the ICA supplying a medial sphenoid wing meningioma. Figures 2-8 illustrate angiographic and anatomic features of meningiomas in the most common anatomic locations in this study, in the context of tumors embolized with each devascularization grade from 0 through 4.13 Angiographic Tumor Grading Scale: grade 0 (clival and planum sphenoidale meningioma with extradural internal carotid supply, cerebellopontine angle meningioma with posterior circulation supply). This composite figure shows exemplary lateral angiograms and the corresponding magnetic resonance images of tumors evaluated for potential preoperative embolization, but for which it was decided that safe embolization was not feasible (Grade 0 in our tumor devascularization scale). The first row shows a Grade 0 case of a meningioma predominantly supplied by extradural branches of the left internal carotid artery, including the meningohypophyseal trunk (A, lateral angiogram, left internal carotid artery injection), involving the medial sphenoid wing, clivus, and planum sphenoidale (contrast-enhanced B, sagittal, C, axial, and D, coronal T1-weighted MRI). The second row shows a Grade 0 case of a meningioma supplied by small posterior circulation branches, including branches from the right anterior inferior cerebellar artery (E, lateral angiogram, right vertebral injection), located in the right cerebellopontine angle (F, contrast-enhanced sagittal, G, axial, and H, coronal T1-weighted MRI).

a Grade 0 case of a meningioma supplied by small posterior circulation branches, including branches from the right anterior inferior cerebellar artery (E, lateral angiogram, right vertebral injection), located in the right cerebellopontine angle (F, contrast-enhanced sagittal, G, axial, and H, coronal T1-weighted MRI). Angiographic Tumor Grading Scale: grade 1 (clival meningioma supplied by inferolateral trunk). This composite figure shows exemplary lateral angiograms and the corresponding magnetic resonance images of a meningioma that underwent partial preoperative embolization (Grade 1 in our tumor devascularization scale, defined as devascularization amounting to <50% of the external carotid supply). The tumor was coil-embolized through 2 branches of the right inferolateral trunk. The first row shows the meningioma involving the medial sphenoid wing, clivus, and planum sphenoidale (left to right: contrast-enhanced A, sagittal, B, axial, and C, coronal T1-weighted MRI). The second row shows that the tumor is predominantly supplied by extradural branches of the right internal carotid artery, including the inferolateral trunk (D, lateral angiogram, right internal carotid artery injection, pre-embolization, showing distinct tumor blush; E, lateral angiogram, right internal carotid artery injection, post–coil embolization, showing no residual tumor blush from the internal carotid supply). The tumor was also supplied by multiple dural branches from the external carotid artery, including branches of the middle meningeal artery (F, anteroposterior angiogram, right external carotid artery injection, showing residual tumor blush post–coil embolization). The third row shows views similar to those shown in the first row, but from an MRI performed after embolization, before tumor resection. The pattern of contrast enhancement after coil embolization of the extradural internal carotid supply to the tumor shows minimal change (left to right: contrast-enhanced G, sagittal, H, axial, and I, coronal T1-weighted MRI).

shown in the first row, but from an MRI performed after embolization, before tumor resection. The pattern of contrast enhancement after coil embolization of the extradural internal carotid supply to the tumor shows minimal change (left to right: contrast-enhanced G, sagittal, H, axial, and I, coronal T1-weighted MRI). Angiographic Tumor Grading Scale: grade 2 (transosseous convexity meningioma with internal and external carotid supply). This composite figure shows exemplary anteroposterior and lateral angiograms and the corresponding MR images of a meningioma that underwent Grade 2 preoperative embolization, defined as 50%-99% devascularization of the external carotid supply. The top row shows pre-embolization images, whereas the bottom row shows postembolization preoperative images. All MR images are contrast-enhanced, T1-weighted. The tumor in this case was a transosseous convexity meningioma (we have previously reported this case in Matsoukas, Rapoport and Colleagues, “Combined transarterial and percutaneous preoperative embolization of transosseous meningioma,” Interventional Radiology 2022) extensively involving the frontoparietal regions bilaterally and the superior sagittal sinus, with both intra- and extracranial components.17 Top row (left to right): A, sagittal pre-embolization MR image, B, lateral pre-embolization angiogram (left common carotid artery injection), C, coronal pre-embolization MR image, D, anteroposterior pre-embolization angiogram (left common carotid artery injection). Bottom Row (Left to Right): E, sagittal postembolization (preoperative) MR image, F, lateral postembolization angiogram (left common carotid artery injection), G, coronal postembolization (preoperative) MR image, H, anteroposterior postembolization angiogram (left common carotid artery injection). Comparing the pre- and postembolization angiograms, it is apparent that the tumor blush from the extracranial part of the tumor in the pre-embolization angiograms is almost completely obliterated in the postembolization images, but small collaterals from the anterior and middle cerebral arteries continue to supply the cortical surface of the tumor, supplying residual tumor blush.

rent that the tumor blush from the extracranial part of the tumor in the pre-embolization angiograms is almost completely obliterated in the postembolization images, but small collaterals from the anterior and middle cerebral arteries continue to supply the cortical surface of the tumor, supplying residual tumor blush. The postoperative MR images show loss of contrast enhancement in the more superficial regions of the tumor embolized through external carotid branches (n-BCA glue and Onyx 18 and 34 through the left middle meningeal, occipital, and superficial temporal, as well as percutaneous injection), whereas there is persistent contrast enhancement in the deeper regions of the tumor that retain collateral supply from the internal carotid circulation. MR, magnetic resonance.

ches (n-BCA glue and Onyx 18 and 34 through the left middle meningeal, occipital, and superficial temporal, as well as percutaneous injection), whereas there is persistent contrast enhancement in the deeper regions of the tumor that retain collateral supply from the internal carotid circulation. MR, magnetic resonance. Angiographic Tumor Grading Scale: grade 3 (convexity meningioma with middle meningeal artery and pial collateral supply). This composite figure shows exemplary anteroposterior and lateral angiograms and the corresponding MR images of a meningioma that underwent Grade 3 preoperative embolization, defined as complete devascularization of the external carotid supply with residual supply from the internal carotid or posterior circulation. The top row shows pre-embolization MR images, whereas the bottom row shows postembolization preoperative images. All MR images are contrast-enhanced, T1-weighted. The tumor in this case was a right convexity meningioma embolized using 150- to 250-micron polyvinyl alcohol particles through the right middle meningeal artery. Top row (left to right): A, coronal, B, sagittal, and C, axial pre-embolization MR images. Middle row (left to right): D, anteroposterior pre-embolization angiogram (right external carotid artery injection), E, anteroposterior postembolization angiogram (right external carotid artery injection), and F, anteroposterior postembolization angiogram (right internal carotid artery injection). Bottom Row (Left to Right): G, coronal, H, sagittal, and I, axial postembolization preoperative MR images. The pre-embolization angiogram shows supply to the superficial aspect of the meningioma from the right middle meningeal artery. The corresponding tumor blush has been obliterated in the corresponding postembolization angiogram. The corresponding internal carotid artery injection shows persistent supply to the pial surface of the tumor by small right anterior and middle cerebral artery branches. The postoperative MR images show loss of contrast enhancement in the more superficial regions of the tumor embolized through the middle meningeal artery, whereas there is persistent contrast enhancement in the deeper aspect of the tumor that retains collateral supply from the internal carotid circulation. MR, magnetic resonance.

erative MR images show loss of contrast enhancement in the more superficial regions of the tumor embolized through the middle meningeal artery, whereas there is persistent contrast enhancement in the deeper aspect of the tumor that retains collateral supply from the internal carotid circulation. MR, magnetic resonance. Angiographic Tumor Grading Scale: grade 3 (orbital roof meningioma with middle meningeal artery and ophthalmic artery supply). This composite figure shows exemplary anteroposterior and lateral angiograms and the corresponding MR images of a meningioma that underwent Grade 3 preoperative embolization. The top row shows pre-embolization MR images, whereas the bottom row shows postembolization preoperative images. All MR images are contrast-enhanced, T1-weighted. The tumor in this case was a right orbital roof meningioma embolized using 100-micron polyethylene glycol microspheres through the right middle meningeal artery and right internal maxillary artery. Top Row (Left to Right): A, coronal, B, axial, and C, sagittal pre-embolization MR images showing the tumor. Middle Row (Left to Right): D, anteroposterior pre-embolization angiogram (right middle meningeal artery injection), E, anteroposterior postembolization angiogram (right middle meningeal artery injection), and F, lateral postembolization angiogram (right internal carotid artery injection). Bottom Row (Left to Right): G, coronal, H, axial, and I, sagittal postembolization preoperative MR images. The vascular tumor blush of the pre-embolization angiogram demonstrates supply to a region of the meningioma from the right middle meningeal artery. This tumor blush has been obliterated in the corresponding postembolization angiogram. The internal carotid artery injection shows significant arterial supply to the tumor from branches of ophthalmic artery. The postoperative MR images show loss of contrast enhancement of a region of the tumor embolized through the middle meningeal artery, whereas there is persistent contrast enhancement in other regions that seem to retain collateral supply from branches of the ophthalmic artery. MR, magnetic resonance.

ophthalmic artery. The postoperative MR images show loss of contrast enhancement of a region of the tumor embolized through the middle meningeal artery, whereas there is persistent contrast enhancement in other regions that seem to retain collateral supply from branches of the ophthalmic artery. MR, magnetic resonance. Angiographic Tumor Grading Scale: grade 4 (lateral sphenoid wing meningioma supplied by middle meningeal artery). This composite figure shows exemplary anteroposterior and lateral angiograms and the corresponding MR images of a meningioma that underwent Grade 4 preoperative embolization, defined as angiographically complete devascularization with no residual internal or external carotid artery supply. The top row shows pre-embolization MR images, whereas the bottom row shows postembolization preoperative images. All MR images are contrast-enhanced, T1-weighted. The tumor in this case was a left lateral sphenoid wing meningioma embolized using 100-micron perfluorinated polymer microspheres through the left middle meningeal artery. The patient had parotid surgery and had an occluded left internal maxillary artery, with distal reconstitution through other external carotid collaterals. Top row (left to right): A, sagittal, B, coronal, and C, axial pre-embolization MR images showing the tumor. Middle row (left to right): D, lateral pre-embolization angiogram (left external carotid artery injection), E, anteroposterior pre-embolization angiogram (left external carotid artery injection), and F, anteroposterior postembolization angiogram (left common carotid artery injection). Bottom Row (Left to Right): G, sagittal, H, coronal, and I, axial postembolization preoperative MR images. The vascular tumor blush of the pre-embolization angiogram shows supply apparently to the entire meningioma from the reconstituted left middle meningeal artery. This tumor blush has been obliterated in the corresponding postembolization angiogram. The postoperative MR images show loss of contrast enhancement in almost the entire tumor. MR, magnetic resonance.

ization angiogram shows supply apparently to the entire meningioma from the reconstituted left middle meningeal artery. This tumor blush has been obliterated in the corresponding postembolization angiogram. The postoperative MR images show loss of contrast enhancement in almost the entire tumor. MR, magnetic resonance. Angiographic Tumor Grading Scale: grade 4 (middle cranial fossa meningioma supplied by middle meningeal artery). This composite figure shows exemplary lateral angiograms and the corresponding MR images of a meningioma that underwent Grade 4 preoperative embolization. The top row shows pre-embolization MR images, whereas the bottom row shows postembolization preoperative images. All MR images are contrast-enhanced, T1-weighted. The tumor in this case was a large left lateral sphenoid wing meningioma embolized using 200-micron polyethylene glycol microspheres and Onyx 18 through the left middle meningeal artery. Top row (left to right): A, axial, B, sagittal, and C, coronal pre-embolization MR images showing the tumor. Middle row (left to right): D, lateral pre-embolization angiogram (left external carotid artery injection), E, lateral pre-embolization angiogram (left middle meningeal artery selective microcatheter injection), and F, lateral postembolization angiogram (left external carotid artery injection). Bottom row (left to right): G, axial, H, sagittal, and I, coronal postembolization preoperative MR images. The vascular tumor blush of the pre-embolization angiogram shows supply apparently to the entire meningioma from the left middle meningeal artery. This tumor blush has been obliterated in the corresponding postembolization angiogram. The postoperative MR images show loss of contrast enhancement in most of the tumor. MR, magnetic resonance.

ush of the pre-embolization angiogram shows supply apparently to the entire meningioma from the left middle meningeal artery. This tumor blush has been obliterated in the corresponding postembolization angiogram. The postoperative MR images show loss of contrast enhancement in most of the tumor. MR, magnetic resonance. Table 2 presents a cross-tabulation of tumor locations and associated arterial feeders. Both variables were assessed by 2 staff neuroradiologists and by 2 independent reviewers (SM, BIR), and consensus was reached in each case. Certain tumor locations were consistently associated with specific patterns of arterial feeders (arterial feeders were not homogeneously distributed across tumor locations). In particular, meningeal feeders (including the middle meningeal artery) were dominantly associated with convexity, parasagittal (falcine), and lateral sphenoid wing meningiomas (P < .001); extradural ICA feeders (meningohypophyseal and inferolateral trunks) were mostly present in medial sphenoid wing and clinoidal or clival meningiomas (P < .001); and posterior circulation feeders were present in foramen magnum, petrous apex, and tentorial meningiomas (P < .001). Several additional feeding patterns were noted but did not reach statistical significance when corrected for multiple tests. In particular, unadjusted significance was found in meningeal and other internal maxillary artery feeders that supplied anterior midline and medial sphenoid wing meningiomas (P = .04); other ECA feeders were present in anterior midline, convexity, and some sphenoid wing meningiomas (P = .05); and ACA supply was present in meningiomas of the anatomic midline (parasagittal falcine and anterior fossa midline) (P = .02).

eders that supplied anterior midline and medial sphenoid wing meningiomas (P = .04); other ECA feeders were present in anterior midline, convexity, and some sphenoid wing meningiomas (P = .05); and ACA supply was present in meningiomas of the anatomic midline (parasagittal falcine and anterior fossa midline) (P = .02). Tumor Location and Its Relationship to Tumor Vascular Supply ACA, anterior cerebral artery; ECA, external carotid artery; Extradural ICA: MHT, meningohypophyseal trunk or ILT, inferolateral trunk; ICA, internal carotid artery; IMAX, internal maxillary artery; MCA, middle cerebral artery; MMA, middle meningeal and accessory meningeal arteries; PCA, posterior cerebral artery. Meningioma location is tabulated here for the associated vascular supply. The P values reported represent significance from a test of equal proportions with Holm-Bonferroni–adjusted P-values. Statistically meaningful vascular supplies are given in bold.

ACA, anterior cerebral artery; ECA, external carotid artery; Extradural ICA: MHT, meningohypophyseal trunk or ILT, inferolateral trunk; ICA, internal carotid artery; IMAX, internal maxillary artery; MCA, middle cerebral artery; MMA, middle meningeal and accessory meningeal arteries; PCA, posterior cerebral artery. Meningioma location is tabulated here for the associated vascular supply. The P values reported represent significance from a test of equal proportions with Holm-Bonferroni–adjusted P-values. Statistically meaningful vascular supplies are given in bold. Tables 3 and 4 present results related to the extent of embolization. Among all the tumors in this series, 20 (25%) received no embolization (Grade 0), 8 (10%) were assessed as Grade 1, 24 (30%) as Grade 2, 12 (15%) as Grade 3, and 16 (20%) as Grade 4. Convexity meningiomas were associated with higher grades of embolization (P = .019) than nonconvexity meningiomas, with convexity location accounting for a pseudomedian difference of 1 level in the grading system. On the other hand, anterior midline meningiomas were associated with lower grades of embolization (P = .003), with a pseudomedian difference of −2 between anterior midline and nonanterior midline locations. Other locations, including sphenoid wing and parasagittal, and embolization degrees did not differ statistically significantly (Table 3). Table 4 provides a cross-tabulation of arterial feeders and the corresponding grade of embolization. The analysis reveals that tumors with meningeal artery supply are significantly more likely than tumors without meningeal artery supply to be amenable to embolization, with median devascularization one grade higher than meningiomas without meningeal artery supply (P < .001). Supply from other ECA branches was not independently predictive of the degree of devascularization. On the other hand, meningiomas receiving vascular supply from an ophthalmic or ACA branch trended toward less extensive embolization, with median devascularization one grade lower than meningiomas without ophthalmic or ACA supply, although the sample sizes in these subgroups were not powered for statistical significance.

the other hand, meningiomas receiving vascular supply from an ophthalmic or ACA branch trended toward less extensive embolization, with median devascularization one grade lower than meningiomas without ophthalmic or ACA supply, although the sample sizes in these subgroups were not powered for statistical significance. Tumor Location and Its Relationship to Devascularization Grade Statistical significance at the P < .05 level. (A) Angiographic embolization grade is tabulated here for the tumor location. The shaded rows reflect tumor locations with large enough sample sizes to enable further statistical analysis: anterior midline, convexity, parasagittal/falcine, and sphenoid wing. (B) Further analysis of the 4 bold tumor locations with largest representations in our series using a Wilcoxon rank-sum test. The analysis confirms that convexity meningiomas are more likely (P = .019) than nonconvexity meningiomas to be amenable to embolization, with a median devascularization one grade higher than nonconvexity meningiomas. By contrast, anterior midline meningiomas are significantly less likely to be amenable to embolization, with median devascularization 2 grades lower than nonanterior-midline meningiomas. Sphenoid wing and parasagittal location alone are not predictive of differences in devascularization grade without more information about the vascular supply. Tumor Vascular Supply and Its Relationship to Angiographic Outcome

(A) Angiographic embolization grade is tabulated here for the tumor location. The shaded rows reflect tumor locations with large enough sample sizes to enable further statistical analysis: anterior midline, convexity, parasagittal/falcine, and sphenoid wing. (B) Further analysis of the 4 bold tumor locations with largest representations in our series using a Wilcoxon rank-sum test. The analysis confirms that convexity meningiomas are more likely (P = .019) than nonconvexity meningiomas to be amenable to embolization, with a median devascularization one grade higher than nonconvexity meningiomas. By contrast, anterior midline meningiomas are significantly less likely to be amenable to embolization, with median devascularization 2 grades lower than nonanterior-midline meningiomas. Sphenoid wing and parasagittal location alone are not predictive of differences in devascularization grade without more information about the vascular supply. Tumor Vascular Supply and Its Relationship to Angiographic Outcome ACA, anterior cerebral artery; ECA, external carotid artery; ICA, internal carotid artery; IMAX, internal maxillary artery; MCA, middle cerebral artery; MMA, middle meningeal artery; PCA, posterior cerebral artery. Statistical significance at the P < .05 level.

Tumor Vascular Supply and Its Relationship to Angiographic Outcome ACA, anterior cerebral artery; ECA, external carotid artery; ICA, internal carotid artery; IMAX, internal maxillary artery; MCA, middle cerebral artery; MMA, middle meningeal artery; PCA, posterior cerebral artery. Statistical significance at the P < .05 level. (A) Angiographic embolization grade is tabulated here for the tumor vascular supply. The shaded rows reflect arterial branches with large enough sample sizes to enable further statistical analysis: meningeal (middle and accessory) arteries, other external carotid artery branches, ophthalmic artery, anterior cerebral artery, and extradural internal carotid branches (meningohypophyseal trunk and inferolateral trunk). (B) Further analysis of the 5 bold arterial branches with largest representations in our series using a Wilcoxon rank-sum test. The analysis reveals that tumors with meningeal artery supply are significantly more likely than tumors without meningeal artery supply to be amenable to embolization, with median devascularization one grade higher than meningiomas without meningeal artery supply. Supply from other external carotid artery branches was not independently predictive of the degree of devascularization. Meningiomas receiving vascular supply from an ophthalmic or anterior cerebral artery branch trended toward less extensive embolization, with median devascularization one grade lower than meningiomas without ophthalmic or anterior cerebral artery supply although the sample sizes in these subgroups were not powered for statistical significance.

scular supply from an ophthalmic or anterior cerebral artery branch trended toward less extensive embolization, with median devascularization one grade lower than meningiomas without ophthalmic or anterior cerebral artery supply although the sample sizes in these subgroups were not powered for statistical significance. Finally, in a multivariate logistic regression analysis, increasing age was associated with slightly higher chances of grade 3/4 embolization, after controlling for sex, WHO classification, tumor volume, tumor location, and the existence of meningeal feeders (adjusted odds ratio: 1.1, 95% CI: 1.02-1.23, P = .028).

In this study, we carefully classify meningiomas undergoing preoperative angiography, with the intent of embolization, according to their intracranial location and arterial supply. We correlate these properties of the tumors with the degree of embolization achieved. To date, there is no consensus on or standardized approach to assess the efficacy of embolization. Many previous studies of the efficacy of meningioma embolization have used estimated blood loss at the time of surgical resection as a measure for this purpose.3-5 However, estimated blood loss has limitations as a measure of embolization efficacy, including that intraoperative blood loss is notoriously difficult to measure accurately (in practice, reported estimates are highly subjective) and that a significant amount of blood may be lost in the operative exposure of the tumor even when the tumor itself bleeds minimally during resection. Other measures used to assess preoperative embolization efficacy include totality of tumor resection and operative time, which again are subject to bias and can be affected by many variables unrelated to embolization.3-5,14

he operative exposure of the tumor even when the tumor itself bleeds minimally during resection. Other measures used to assess preoperative embolization efficacy include totality of tumor resection and operative time, which again are subject to bias and can be affected by many variables unrelated to embolization.3-5,14 Some surgeons question the utility of preoperative embolization, arguing that early surgical devascularization (including coagulation of dural attachments) can obviate the potential role of embolization. Our surgical experience has suggested that preoperative embolization may be especially helpful in meningiomas whose dural attachments and associated vascular supplies are not easily accessible during the initial stages of tumor resection, such as meningiomas of the falx or sphenoid wing.3-5 In addition, many surgeons have noted that some embolized tumors seem to necrose after devascularization or undergo tissue changes that mechanically facilitate tumor removal. Both because of the unreliability of surgical blood loss and because blood loss alone may not adequately capture the usefulness of tumor embolization, we chose not to use surgical parameters to assess the extent of embolization. Instead, we introduced and validated a preoperative tumor embolization grading scale based on a verifiable parameter obtained at the time of embolization, the reduction in angiographic tumor blush.

apture the usefulness of tumor embolization, we chose not to use surgical parameters to assess the extent of embolization. Instead, we introduced and validated a preoperative tumor embolization grading scale based on a verifiable parameter obtained at the time of embolization, the reduction in angiographic tumor blush. Meningiomas recruit regional vasculature as they grow. It makes sense that the intracranial anatomic location of a tumor would predict its vascular supply, with some variability. In this study, we characterize the common patterns of vascular supply for the classic meningioma locations. Several patterns emerged from the analysis, which we discuss here; these are reflected in Table 2. In understanding these patterns, it is helpful to recall some aspects of neurovascular embryology, including the nature of the major anastomotic networks of the intracranial and craniofacial arteries. The network formed by the arteries around the orbit is of special interest in surgery for meningiomas of the anterior skull base and sphenoid wing. The region receives supply from external carotid branches, including the middle meningeal and other branches of the internal maxillary artery, and from internal carotid branches, including the ophthalmic artery. In the absence of a vascular or other lesion, distal branches of the ophthalmic artery may anastomose with distal branches of the middle meningeal artery. However, in the presence of a vascular lesion in the region of the anterior skull base, including the orbital roof or sphenoid wing, blood flow through such anastomoses increases. A vascular lesion recruiting blood supply from both pedicles may open a high-flow anastomosis between the pedicles. It is essential for procedural safety to understand these features, which are well known to angiographers. Less well appreciated is how this type of pathologic vascular anatomy affects tumor embolization efficacy.

r lesion recruiting blood supply from both pedicles may open a high-flow anastomosis between the pedicles. It is essential for procedural safety to understand these features, which are well known to angiographers. Less well appreciated is how this type of pathologic vascular anatomy affects tumor embolization efficacy. We observe that the extent of embolization of sphenoid wing meningiomas is correlated with the mediolateral location along the skull base (Figure 9). Lateral sphenoid wing meningiomas may be supplied almost exclusively by the middle meningeal artery and have favorable odds of high extent of embolization. Many such tumors act like parasagittal or falcine meningiomas and convexity meningiomas in other regions, in that they may be supplied entirely by meningeal arteries originating from branches of the ECA. Such tumors are amenable to extensive devascularization through preoperative embolization. The result in optimal cases is a soft, necrotic, often white, and completely devascularized mass that yields almost entirely to suction intraoperatively (Figure 10).15 Medial sphenoid wing meningiomas, like clinoidal and clival meningiomas, may be supplied by the extradural internal carotid branches, through which embolization poses higher risk of stroke or cranial neuropathy (although some groups have demonstrated safety and efficacy of tumor embolization through these branches).16 Between these extremes, meningiomas along the sphenoid wing and orbital roof recruit blood supply from both the ophthalmic artery and the middle meningeal artery, and the extent of embolization of such tumors is limited by the robustness of the ophthalmic contribution. In some situations, an operator may judge that it is safe to embolize through branches of the ophthalmic artery distal to those supplying the retina, but more commonly, tumor embolization through the ophthalmic artery is avoided to minimize the likelihood of ischemic visual compromise.

s of the ophthalmic contribution. In some situations, an operator may judge that it is safe to embolize through branches of the ophthalmic artery distal to those supplying the retina, but more commonly, tumor embolization through the ophthalmic artery is avoided to minimize the likelihood of ischemic visual compromise. Patterns of sphenoid wing meningioma vascular supply. This composite figure illustrates several features observed with consistency in the combined endovascular and operative management of sphenoid wing meningiomas. The vascular supply and extent of embolization achievable in these tumors vary with the mediolateral position of the tumor along the sphenoid ridge. All magnetic resonance images are contrast-enhanced, T1-weighted. Top row: Images from a meningioma of the medial sphenoid wing, clivus, and planum sphenoidale. This tumor and meningiomas in this location tend to recruit dural vascular supply from the extradural internal carotid artery and, in part for this reason, tend to be poorly amenable to embolization. From left to right: A, Sagittal view of the tumor; B, lateral angiogram, right internal carotid artery injection, showing tumor blush arising from the inferolateral trunk; C, coronal view of the tumor; D, anteroposterior angiogram, right external carotid artery injection, showing faint tumor blush in the region of the sphenoid wing. Middle rows (blue rectangle, figures E to L): The middle 2 rows show images from a meningioma of the mid-sphenoid wing. This tumor and meningiomas in this general location tend to recruit dural vascular supply from the middle meningeal artery-ophthalmic artery anastomotic network. The balance of supply from the meningeal and ophthalmic sides of this network decides the degree to which these tumors admit extensive embolization. Upper middle row, left to right: Pre-embolization images: E, sagittal pre-embolization view of the tumor; F, lateral pre-embolization angiogram, right internal carotid artery injection, showing tumor blush arising from branches of the ophthalmic artery; G, coronal pre-embolization view of the tumor; H, anteroposterior pre-embolization angiogram, right external carotid artery injection, showing tumor blush arising from the middle meningeal artery.

ion angiogram, right internal carotid artery injection, showing tumor blush arising from branches of the ophthalmic artery; G, coronal pre-embolization view of the tumor; H, anteroposterior pre-embolization angiogram, right external carotid artery injection, showing tumor blush arising from the middle meningeal artery. Lower middle row, left to right: I, Sagittal postembolization view of the tumor showing some regions of loss of contrast enhancement; J, lateral pre-embolization angiogram, right external carotid artery injection, showing tumor blush arising from the middle meningeal artery; K, coronal postembolization view of the tumor showing some regions of loss of contrast enhancement; L, axial postembolization view of the tumor showing some regions of loss of contrast enhancement. Bottom row: Images from a meningioma of the lateral sphenoid wing. This tumor and meningiomas in this general location tend to recruit dural vascular supply predominantly from the middle meningeal artery, often with minimal contribution from internal carotid branches. For this reason, these tumors often admit extensive embolization. Left to right: M, Sagittal pre-embolization view of the tumor; N, lateral pre-embolization angiogram, left external carotid artery injection, showing tumor blush corresponding to the meningioma; O, anteroposterior pre-embolization angiogram, left external carotid artery injection, showing tumor blush corresponding to the meningioma; and P, sagittal postembolization view of the tumor showing almost complete loss of contrast enhancement.

otid artery injection, showing tumor blush corresponding to the meningioma; O, anteroposterior pre-embolization angiogram, left external carotid artery injection, showing tumor blush corresponding to the meningioma; and P, sagittal postembolization view of the tumor showing almost complete loss of contrast enhancement. Intraoperative views of completely embolized parasagittal meningioma. A, Coronal preoperative MR image showing a falcine parasagittal meningioma before embolization. B, Coronal preoperative MR image showing the same falcine parasagittal meningioma after embolization. C, Anteroposterior fluoroscopic image showing the n-BCA glue cast after embolization of the falcine parasagittal meningioma shown in A and B through the bilateral middle meningioma arteries (Grade 4 in the angiographic devascularization classification described here). D, Intraoperative image of the same meningioma showing a necrotic, soft, suckable, extensively embolized meningioma. The tumor was resected 48 h after embolization. Portions of the n-BCA glue cast, seen here, were identified within the tumor and were resected together with the tumor itself. MR, magnetic resonance.

re). D, Intraoperative image of the same meningioma showing a necrotic, soft, suckable, extensively embolized meningioma. The tumor was resected 48 h after embolization. Portions of the n-BCA glue cast, seen here, were identified within the tumor and were resected together with the tumor itself. MR, magnetic resonance. Other groups have evaluated vascular anatomy of meningiomas and factors contributing to effective tumor embolization.17 In this study, we have made a particular effort to grade the degree of devascularization achieved by embolization in a systematic manner. Nevertheless, our study has certain limitations, notably the selection bias inherent in studying only those meningiomas referred for embolization during the study period and necessarily omitting from angiographic analysis those tumors treated by resection or radiation alone or in combination or tumors that have been observed without procedural interventions. In addition, as the study cohort grows, it may be fruitful to account in a multivariate fashion for simultaneous presence and absence of certain arterial branches because the presence of middle meningeal supply is more predictive of extensive embolization when concurrent internal carotid and ascending pharyngeal supply is not present, for example.

ort grows, it may be fruitful to account in a multivariate fashion for simultaneous presence and absence of certain arterial branches because the presence of middle meningeal supply is more predictive of extensive embolization when concurrent internal carotid and ascending pharyngeal supply is not present, for example. Finally, as a number of figures demonstrate, certain radiographic changes in embolized meningiomas are consistently apparent on contrast-enhanced magnetic resonance images obtained postembolization, preresection. Reduction in gadolinium contrast enhancement is often apparent on contrast-enhanced T1-weighted sequences in extensively embolized tumors when the pre- and postembolization sequences are compared. We address this topic systematically in a companion paper (“Quantifying Extent of Meningioma Preoperative Embolization Through Volumetric Analysis: A Retrospective Case Series,” Concurrently Under Review Elsewhere).

In this study, we characterize the patterns of arterial supply corresponding to common anatomic location of meningiomas. We confirm that tumor location is strongly associated with arterial supply, with some natural variability. We further describe a purely angiographic system for objectively grading the extent of tumor embolization and describe the anatomic and vascular characteristics predictive of the greater extent of embolization. In particular, convexity meningiomas are associated with the greater extent of embolization, and anterior midline meningiomas are associated with the lower extent of embolization. The extent of embolization of sphenoid wing meningiomas depends on the mediolateral location of the tumor and the presence and robustness of ophthalmic artery supply. The grading system presented here may prospectively facilitate consistency of reported angiographic results. In addition, these patterns can be instructive for identifying meningiomas susceptible to more extensive preoperative embolization and may guide selection of meningiomas for preoperative embolization in a manner that maximizes efficacy while limiting unnecessary incremental risk to the patient.