Preoperative and Postoperative Imaging Evaluation of Arteriovenous Malformations

8 Preoperative and Postoperative Imaging Evaluation of Arteriovenous Malformations

Michaelangelo Fuortes, Joseph Gastala, Adam Liudahl, Minako Hayakawa, and Colin Derdeyn

Abstract

Imaging features of brain arteriovenous malformations (AVMs) are critical for pretreatment planning and posttreatment decision-making. This chapter will describe the important, clinically relevant imaging features of brain AVMs as well as the relative advantages and disadvantages of the widely available imaging modalities: computed tomography (CT), magnetic resonance imaging (MRI), and digital subtraction angiography (DSA). These techniques often yield complementary information. Critical features include the size, location, and pattern of venous drainage, as well as the presence of intranidal or flow-related aneurysms. Imaging is critical for confirming cure of these lesions, particularly after stereotactic radiosurgery. Finally, we will review new and emerging applications of imaging for patients with brain AVMs, including the assessment of microhe-morrhage as an indicator of increased bleeding risk for clinically asymptomatic brain AVMs.

Keywords: computed tomography, magnetic resonance imaging, digital subtraction angiography, hemorrhage, aneurysms, nidus

Key Points

Imaging evaluation of arteriovenous malformations.
Preoperative planning.
Postoperative imaging of complications.

8.1 Introduction

A brain arteriovenous malformation (AVM) is a cerebrovascular pathology composed of afferent arteries and draining veins connected by a nidus of arteriovenous fistulas without an intervening capillary bed.¹^,² Management of brain AVMs is complex and requires a multidisciplinary approach, with treatment options including surgical resection, endovascular embolization, stereotactic radiosurgery, or various combinations thereof. Treatment decisions are based on risk estimates of rupture and studies of outcomes of invasive treatment. Diagnostic imaging plays a critical role in the identification, pretreatment classification, and posttreatment follow-up of brain AVMs.

8.2 Materials and Methods

A Medline search was performed utilizing PubMed for original research publications, meta-analyses, guidelines, and consensus statements discussing clinical diagnosis, imaging, and clinical management of AVMs. The search employed various combinations of the following keywords: arteriovenous malformation (AVM), digital subtraction angiography (DSA), computed tomography angiography (CTA), magnetic resonance angiography (MRA). The review was limited to English-language literature and human studies. Additional articles were selected by reviewing the reference lists of pertinent publications with identification of relevant authors. The authors performed a critical review of the identified article titles and abstracts followed by review of the full text in relevant articles. The senior authors’ personal experience contributed to multiple topics covered in this chapter, more specifically regarding those topics that are sparsely covered in other resources.

8.3 Results

8.3.1 Imaging Diagnosis

The clinical scenario influences the role of diagnostic imaging in the evaluation of brain AVMs with the focus varying from primary diagnosis to clinical follow-up of a known lesion. The initial diagnosis of a brain AVM most commonly occurs in one of three clinical settings: (1) diagnostic work-up of an intracranial hemorrhage; (2) diagnostic work-up for a neurological symptom (e.g., headache, seizure, focal neurological deficit); or (3) imaging evaluation for an unrelated clinical presentation with incidental discovery of a brain AVM. The distinctions between unruptured brain AVMs and those diagnosed following a hemorrhagic episode have important potential clinical implications. The established literature has frequently reported that around 50% of brain AVMs are identified on diagnostic work-up following presentation with intracranial hemorrhage. The increasing utilization of advanced medical imaging in the emergency and outpatient clinical settings may lead to a greater proportion of incidentally discovered brain AVMs in future practice.

In the setting of nontraumatic intracranial hemorrhage, the recommended initial imaging evaluation is a noncontrast head CT. An initial CT study is quick, widely available, and can provide rapid assessment of an intracranial hemorrhage. This diagnostic approach is well validated by the literature, which demonstrates > 90% sensitivity of noncontrast CT for detection of acute subarachnoid hemorrhage.³ Additional findings on the noncontrast head CT that may suggest an underlying vascular etiology include dilated and/or calcified vessels adjacent to the hemorrhage, hyperdense foci suggesting a vascular nidus, and atypical location of an intraparenchymal hemorrhage suggesting a secondary cause such as a vascular malformation rather than a primary hypertensive hemorrhage (► Fig. 8.1).⁴ There may be localized brain atrophy surrounding an AVM, or the adjacent brain parenchyma may be hypodense on CT reflecting ischemic changes or gliosis.

Fig. 8.1 (a) Noncontrasted head CT showing large acute parietal hemorrhage with abnormal calcification posteriorly. The lobar location and the calcification suggest an underlying vascular malformation. (b) Source image from the CTA showing a tangle of abnormal vessels consistent with a brain AVM.

CT angiography (CTA), magnetic resonance imaging (MRI), and MRA are frequently used in the initial work-up of AVMs and provide valuable anatomical information. The diagnostic imaging approach varies among individuals and institutions, with some advocating starting with CTA and others proposing MRI and MRA as the initial examinations. Both CTA and MRI/MRA are noninvasive alternatives to catheter angiography and likely provide safety benefits as the initial diagnostic tests.

CTA is widely available and commonly used in the emergency setting for assessment of intracranial hemorrhage for multiple reasons. CTA is fast, minimally invasive, and has high spatial resolution that can provide rapid diagnosis of an underlying vascular malformation in the setting of intracranial hemorrhage, including AVMs.⁵ CTA may provide valuable diagnostic information about an AVM, including characterization of the angioarchitecture, localization, and presence of associated aneurysms.⁶ Comparative studies of CTA versus DSA in the diagnosis of cerebral vascular disease demonstrate high sensitivity and specificity for both imaging modalities. The main drawbacks to CTA are radiation exposure and administration of intravenous contrast agents which are associated with side effects, albeit relatively infrequently.

MRI and MRA are diagnostic techniques that do not expose the patient to ionizing radiation. Various MRI and MRA techniques may provide information about surrounding brain parenchyma and allow for noninvasive characterization of the AVM. In particular, MRI/MRA provides information about the size of the nidus and number and location of feeding arteries and draining veins. Susceptibility-weighted imaging aids in the detection of hemosiderin that suggests prior episodes of bleeding. There is recent evidence supporting the superiority of initial diagnostic evaluation with MRI and MRA compared with CTA. A Cochrane review and meta-analysis of studies investigating the diagnostic work-up of nontraumatic intracranial hemorrhage for underlying vascular causes reported estimates for CTA of 95% sensitivity and 99% specificity, as compared to 98% sensitivity and 99% specificity for MRA.⁷

The gold standard for the diagnosis of AVMs and other cerebral vascular disease is DSA, which has superior spatial and time resolution to other imaging modalities. DSA has the added benefit of providing both diagnostic information and facilitating treatment of vascular lesions. The main drawbacks of DSA are that it is invasive, time-consuming, and has associated morbidity.

Recent advances in both CTA and MRI/MRA techniques are introducing the potential for noninvasive imaging studies with functional, hemodynamic, and physiologic information rivaling DSA. Four-dimensional CTA (4D CTA) and 4D MRA protocols are being developed, which provide time-resolved angiographic information allowing for more specific lesion characterization and potential information about perfusion, early draining veins, and shunting similar to DSA. Advances have also been made in catheter angiography techniques, with a recent study reporting the development of a time-resolved 4D DSA protocol which provides detailed information about the angioarchitecture of an AVM.⁸ Currently, even with the newest imaging techniques, DSA continues to demonstrate superior temporal and spatial resolution compared to CTA and MRA.

8.3.2 AVM Grading

An important factor to consider when planning treatment of brain AVMs is the estimated risk of treatment for that patient. For surgical treatment of brain AVMs, Spetzler and Martin developed a grading system to predict the risk of surgical morbidity and mortality for individual patients. The three components of the original grading system are lesion size, pattern of venous drainage, and neurological eloquence of the brain tissue (► Table 8.1). The brain AVM size is divided into small (< 3 cm), medium (3–6 cm), and large (> 6 cm). The larger the lesion is, the higher is the score. The pattern of venous drainage is related to ease of resectability. Superficial drainage is defined as AVM drainage in the cortical venous system and cerebellar hemispheric veins into the straight sinus or transverse sinus. Deep drainage is defined as a venous drainage pattern involving the deep veins such as internal cerebral veins, basal veins, or precentral cerebellar vein. Eloquence identifies the functional importance of the brain tissue near the brain AVM. If this brain tissue is injured, the eloquence factor addresses how disabling the injury would be to the patient. This area is susceptible to dissection, retraction, and postoperative hemorrhage or edema. The regions considered as eloquent are the sensorimotor, language cortex, visual cortex, basal ganglia, hypothalamus, thalamus, internal capsule, brainstem, cerebellar peduncles, and deep cerebellar nuclei.⁹^,¹⁰ Examples of noneloquent regions are anterior frontal lobes, anterior temporal lobes, and the cerebellar cortex. The grade is the sum of these three scores. Grades I to V correlate with the operative results, respectively. A grade VI is allocated to inoperable brain AVMs such as large size in eloquent areas or a nidus in the hypothalamus or brainstem.⁹ Those brain AVMs determined to be grade I or II (low-grade AVMs) are associated with low morbidity rates and frequently managed surgically. Management of grade III AVMs is not widely agreed upon.¹¹ Grade IV or V AVMs (high-grade AVMs) are associated with high morbidity rates, so these lesions are frequently monitored.

Table 8.1 Spetzler–Martin scale

Parameter
Size (cm)	< 3 cm (1 point)	3–6 cm (2 points) > 6 cm (3 points)
Eloquence	Noneloquent (0 points)	Eloquent (1 point)
Venous drainage	Superficial (0 points)	Deep (1 point)
Total = size + eloquence + venous drainage