The most common tumors to involve the skull base are meningiomas and the most common vascular tumors are paragangliomas (glomus tympanic and glomus jugulare lesions) and juvenile nasopharyngeal angiofibromas. Adequacy and safety of tumor resection will depend upon the extent of disease, arterial supply, and the ability to perform safe preoperative embolization. Radical resection may rarely also require sacrifice of the internal carotid or vertebral artery. Reduction in morbidity associated with sacrifice of the internal carotid artery is achieved by balloon test occlusion.
Key wordsSkull base – tumor – vascular – angiography – balloon test occlusion
2 Evaluation of Tumor-Involved Vasculature (Including Balloon Test Occlusion)
2.1 Key Learning Points
Digital subtraction angiography remains the gold standard for vascular assessment of skull base tumors and their suitability for embolization.
Six vessel angiography and assessment of collateral blood flow are often required.
Diagnostic digital subtraction angiography can help identify dangerous anastomoses and venous drainage patterns.
Internal carotid artery sacrifice without prior balloon occlusion testing carries significant morbidity/mortality.
The skull base is anatomically complex, involving osseous, soft tissue, neural, and vascular components which give rise to a diverse range of benign and malignant tumors. Tumors arising from the extracranial head and neck may also involve the skull base by direct extension. Paragangliomas (glomus tumors) (Fig. 2.1, Fig. 2.2, and Fig. 2.3), after skull base meningiomas, are the most frequent vascular tumors arising directly from the skull base, and juvenile nasal angiofibromas (JNAs) are the most frequent vascular tumors to involve the skull base, extending from their site of origin in the pterygopalatine fossa/sphenopalatine foramen (Fig. 2.4). Other vascular skull base tumors include hemangiomas, hemangiopericytomas, esthesioneuroblastomas, endolymphatic sac tumors, and vascular metastases.
2.3 Arterial and Venous Anatomy
Although variations are expected, the commonest skull base tumor types have well-described typical arterial supply patterns.
Glomus jugulare tumors (Fig. 2.1, Fig. 2.2, and Fig. 2.3) arise from paraganglia located along either Jacobson’s or Arnold’s nerve, around the jugular bulb and the vascular supply is well reported. 1 , 2 The inferior tympanic branch of the ascending pharyngeal artery supplies the inferomedial portion of the tumor at jugular foramen. The posterolateral component derives a supply from the stylomastoid artery, a branch of either the occipital or posterior auricular arteries. Anteriorly, the tumor is supplied by the anterior tympanic branch of the internal maxillary artery and the superior compartments from the middle and accessory meningeal arteries. Cerebellar arteries contribute to the intradural extension of the tumor.
Juvenile nasopharyngeal angiofibromas (Fig. 2.4) are benign but locally aggressive tumors commonly arising from the nasopharynx in the sphenopalatine foramen/pterygopalatine fossa. They expand anteriorly into the nasal cavity, ethmoid and maxillary sinuses, and laterally into the infratemporal fossa. Further expansion along skull base foramina or by direct bone erosion of the greater sphenoid wing or lateral wall of the sphenoid sinus results in intracranial extension.
The primary arterial supply is from the distal internal maxillary artery, 3 , 4 mainly the sphenopalatine artery, descending palatine arteries, and ascending pharyngeal artery but with accessory supply from the anterior and posterior deep temporal arteries and accessory meningeal artery. As the tumor grows, it recruits additional supply from the ascending palatine branch of the facial artery, ethmoidal branches from the ophthalmic artery, and branches of the internal carotid artery (ICA) (mandibulovidian, pterygovaginal artery, inferolateral trunk, and meningohypophyseal trunk). Arterial supply may be bilateral even if the tumor is lateralized.
The arterial anatomy of skull base meningiomas is given in ▶ Table 2.1. Hemangiopericytomas may be confused with meningiomas but are even more vascular, have corkscrew-like vessels, and receive their dominant supply from the internal carotid or vertebrobasilar circulation. Endolymphatic sac tumors are supplied by the ascending pharyngeal and stylomastoid arteries.
In meningiomas (Fig. 2.6), an early arterial phase blush persisting into the late venous phase is characteristic and the presence of arteriovenous shunting has been associated with more aggressive tumors. 7 It has been demonstrated that skull base meningiomas change the pattern of venous circulation by exerting mass effect and obstructing the surrounding venous structures, causing drainage to be diverted toward collateral venous routes. 8 , 9
This is particularly relevant in petroclival meningiomas (PCMs) as particular patterns of drainage of the superficial middle cerebral vein (SMCV), such as drainage into the pterygoid plexus through a sphenobasal vein or into the transverse sinus (TS) through a sphenopetrosal sinus, are more prevalent in this population and have been associated with intraoperative and postoperative venous complications. 10 , 11
Invasion of the dural sinuses by the tumor can cause a similar pattern of dural sinus occlusion and alternative pattern of venous drainage on the cerebral angiogram. Magnetic resonance (MR) is invaluable in distinguishing simple mass effect on the venous structures from invasion of the dural sinus.
Therefore, detailed assessment of the drainage pattern will impact on outcomes and chosen surgical approach. For example, if an anterior petrosal approach is considered, the presence of a sphenobasal vein should prompt consideration of strategies to avoid injuring this vein in the region of the foramen ovale. 12 If there is drainage through a sphenopetrosal sinus, a tentorial incision that preserves drainage during a transpetrosal approach are needed. 11
Venous drainage characterization is also important in anterior clinoid meningiomas. If drainage of the SMCV into the pterygoid plexus or directly into the cavernous sinus (CS) is present, a tailored surgical strategy is required. 13
2.4 Imaging of Skull Base Vascular Tumors
Clinical assessment is limited by the inaccessible location of most skull base neoplasms and, as such, evaluation is predominantly radiological. Computed tomography (CT), including high-resolution bone reconstructions and CT angiography (CTA) and venography (CTV), and magnetic resonance imaging (MRI), including postcontrast sequences and magnetic resonance angiography (MRA) and magnetic resonance venography (MRV), are the primary imaging modalities for the initial evaluation and diagnosis of skull base pathologies. Cross-sectional imaging allows for the assessment of the primary lesion as well as involvement of adjacent tissue, bone infiltration, perineural spread, and distortion or direct invasion of vascular structures. Although vascular anatomy, gross tumoral vascularity, and the patency of involved blood vessels are adequately evaluated using CTA/CTV and MRA/MRV, these modalities cannot accurately assess the arterial feeding vessels, venous drainage, collateral blood flow, and complex vascular architecture of the lesion. Dynamic CTA can provide noninvasive high-resolution four-dimensional radiographic information of tumor vasculature but currently its clinical use is yet to be defined and does not yet replace formal angiography. 14
Digital subtraction angiography (DSA) remains the “gold standard” for preoperative evaluation of tumor-involved vasculature. Comprehensive angiography of the internal and external carotid arteries, vertebral arteries, and (depending upon tumor location and extent) the thyrocervical and costocervical trunks provides essential preoperative information regarding the tumoral vascular anatomy, identifies the dominant feeding arteries that may be considered for preoperative embolization, and assesses the anatomy of and adequacy of the circle of Willis (COW), the degree of collateralization, and adequacy of the posterior circulation and venous patency and dominance. Head and neck tumors may parasitize regional pial blood supply as they enlarge and extend intracranially. Arterial displacement, distortion, and encasement may be identified (Fig. 2.6). It is also important to define the arterial supply to the retina as preoperative embolization may involve arterial connections with the central artery of the retina. Dangerous anastomoses, however, may not reveal themselves on an initial angiogram but may only become obvious when changes in blood flow occur during embolization. These are, therefore, more appropriately discussed elsewhere.
The primary aim of surgery is to achieve the greatest resection with minimal complications. On occasion, total or maximal resection may only be achievable with permanent occlusion of the ICA or other parent vessel. In this context, DSA in combination with balloon test occlusion (BTO) is necessary to evaluate whether the patient will tolerate occlusion/sacrifice of the ICA. Assessing the safety of vertebral artery (VA) occlusion will depend upon VA dominance and adequacy of the posterior communicating arteries, which can also be assessed with a VA BTO.
2.4.1 Digital Subtraction Angiography (DSA)
A standard cerebral angiogram involves selective injection of the internal and external carotid arteries and vertebral arteries bilaterally (occasionally the thyrocervical and costocervical trunks as well). When embolization is being considered, super-selective injection of tumor feeding vessels can be performed using smaller microcatheters. Collateral blood flow and the adequacy of the COW can be assessed with carotid cross compression (Matas maneuver) or compression of the ipsilateral ICA during a VA injection (Allcock maneuver).
By extending each fluoroscopic “run” into the late venous phase, it is possible to characterize not only the pattern of tumor enhancement but also its venous drainage pattern, mass effect, and eventual obstruction of intracranial venous structures.
Careful characterization of the venous anatomy and existing collateral routes will help plan the best surgical approach and minimize perioperative venous complications.
DSA confers a low risk of complication, 15 , 16 , 17 which may occur at the site of puncture or within the cervicocerebral circulation. Puncture site complications include localized hematoma, vessel dissection, pseudoaneurysm formation, and retroperitoneal hematoma. Cervicocerebral complications include vessel dissection and distant emboli with the associated risk of transient ischemic attack (TIA) and stroke. The risk of neurological complication associated with DSA varies in the published literature, but the larger, more recent studies report a rate of 0 to 0.7% for transient neurological symptoms, and 0 to 0.5% for permanent neurological deficits. In experienced hands, DSA is therefore a safe procedure with a very low risk of significant associated morbidity.
The ability to achieve gross total resection or significant debulking of tumors which were previously considered inoperable has come with the advancement of skull base surgical technique over the past two decades. These neoplasms frequently involve the ICA and maximum resection can sometimes only be obtained with preoperative permanent occlusion of the vessel, if surgery carries a significant risk of vessel rupture and requires intraprocedural ICA ligation. Untested vessel sacrifice carries a significant risk of neurological morbidity secondary to immediate or delayed hypoperfusion. Historic data demonstrates a 17 to 49% 18 , 19 , 20 , 21 incidence of stroke (many fatal) following permanent occlusion of the ICA and around 28% for common carotid artery occlusion without preoperative trial of temporary occlusion.
The use of BTO in differentiating patients who will tolerate permanent vessel occlusion and those who require a bypass or vessel preservation significantly reduces the risk of stroke and the associated neurological morbidity but does not eliminate it completely. Standard BTO with clinical monitoring only will identify the cohort of patients for whom vessel sacrifice without bypass will not be tolerated, as these patients develop neurological symptoms during the procedure while the balloon is inflated and the ICA occluded. It will not identify those patients with impaired cerebrovascular reserve who are at risk of developing neurological deficits secondary to delayed hypoperfusion, which can occur hours to days following ICA occlusion. Numerous adaptations and adjuncts have therefore been developed to improve the sensitivity of BTO, but the underlying principle in all is to assess the efficacy of the collateral circulation, predominantly primary collaterals, in maintaining perfusion of the affected vascular territory. Primary collaterals include the anterior and posterior communicating arteries. Secondary collaterals from the external carotid and leptomeningeal collaterals may take longer to develop.
Approximately 10% of patients develop symptoms during the occlusion. Up to 20% who tolerate occlusion clinically will develop infarction after parent vessel occlusion (PVO); in 20% of these the onset may be after 48 hours sometimes up to 2 weeks later. 22
Adjunctive techniques to evaluate regional cerebral blood flow (CBF) and perfusion include BTO which are primarily delayed in venous phase 23 and technetium-99 m hexamethylpropyleneamine oxime single photon emission computed tomography (SPECT, Fig. 2.3 and Fig. 2.4). 24 , 25 , 26 Other methods reported include transcranial Doppler ultrasonography, 27 perfusion CT, 28 xenon CT, 29 and measurement of arterial stump pressure. 30 Although these reduce the risk of delayed cerebral ischemia down to 3 to 8%, they cannot predict embolic stroke and may also fail to identify some patients with limited reserve. No adjunctive technique is clearly superior.
2.5.1 Balloon Test Occlusion Protocol
Numerous protocols for performing BTO exist, 31 and the exact procedure will be dependent on the facilities and experience of individual departments. In our institution, BTO is performed under local anesthetic in the awake patient, and the adequacy of CBF evaluated by a combination of clinical assessment, venous delay, and SPECT using the metastable radio-isotope 99m Tc hexamethylpropylene-amine oxime (HMPAO).
Systemic blood pressure is measured continuously noninvasively. Sheaths are placed in the common femoral arteries bilaterally and 5,000 units of heparin administered intravenously. Due to the risk of thromboembolic events, adequate anticoagulation should be ensured with Activated Coagulation Times >250. A double lumen balloon catheter (which allows distal perfusion of the ICA, with heparinized saline, to avoid a standing column of blood) is navigated into the ipsilateral upper cervical ICA, around C1–C3 (Fig. 2.5). Inflation at a lower level, in the bulb, risks the carotid sinus reflex which can result in significant bradycardia. Common carotid artery occlusion is not recommended because it decreases perfusion pressure in the carotid sinus and, by reflex, increases arterial blood pressure and so may diminish test reliability.
The balloon is inflated under fluoroscopic control and the timer started. Occlusion is confirmed by injection of contrast through the catheter showing stasis. The patient is neurologically assessed over a 20 to 30 minutes time period, evaluating motor, sensory, memory, and analytical skills. If at any stage the patient develops a neurological deficit, the balloon is immediately deflated as the patient has failed the test occlusion. In addition to clinical evaluation, angiography is performed via the contralateral ICA and VA, assessing venous delay in the “occluded” hemisphere. Additional secondary collaterals may be assessed by injections into the external carotid artery. It is assumed that patients who have symmetry within the venous phase have sufficient collateral circulation to tolerate ICA sacrifice. The appearance of the first cortical veins is the start of the venous phase. A delay in excess of 0.5 second is considered to indicate risk of hemodynamic ischemic stroke. Toward the end of the study, approximately 20 to 25 minutes after balloon inflation, 99 m Tc HMPAO 500 to 600 MBq is injected via a peripheral intravenous cannula. After a further 2 to 3 minutes, the balloon is then deflated and the catheters removed. The patient is transferred to the nuclear medicine suite and SPECT performed, evaluating perfusion in the axial, coronal, and sagittal planes (Fig. 2.5 and Fig. 2.6).
99 m Tc HMPAO shows rapid cerebral uptake and distributes proportionately to regional CBF. The tracer is converted intracellularly to a hydrophilic compound and remains fixed in the brain for a prolonged period allowing delayed imaging after injection. Cerebral uptake of tracer reflects not only perfusion but also the metabolic status of cerebral tissue; hence, when injected during BTO, its distribution is an indicator of both regional perfusion and metabolism. 99 m Tc HMPAO SPECT scans are analyzed comparing a minimum of four hemispheric regions of interest in the occluded hemisphere and with the contralateral side. Greater than 10% reduction in activity in a region of interest is regarded as significant and may indicate suboptimal cerebrovascular reserve. Baseline SPECT scans obtained after an interval of a few days may be necessary to exclude pre-existing asymmetry.
BTO SPECT has the advantage that it avoids patient transfer to another room with the balloon in situ. Its main disadvantage is that it is only semi-quantitative.
BTO is associated with procedural complications in approximately 3.5% of patients and permanent deficits are described in 0.5 to 1.7%. 32 , 33 However, this is likely to vary significantly with operator and center experience.