Imaging Predictors for Rupture

Introduction

For a substantial proportion of patients with intracranial arteriovenous malformations (iAVMs)—variously reported at between 37% and 77%—the initial presentation is hemorrhagic. The annual risk of hemorrhage has been reported to be as low as 0.9% for superficial unruptured AVMs and as high as 34% for previously ruptured deep AVMs with exclusively deep venous drainage. The mortality rate for patients with iAVMs who present with hemorrhage is approximately 10%, and the prevalence of iAVMs is between 15 and 18 per 100,000 adults, Once an iAVM has hemorrhaged, it carries a 6.7% average annual rate for hemorrhage, with a first-year rehemorrhage rate as high as 15% in some series.

The sequelae of AVM hemorrhage are not limited to initial mortality; the long-term neurological morbidity and mortality for iAVM bleeding have been reported to be as high as 29% and 35%, respectively. Thus there is value in understanding imaging determinants predisposing iAVMs to hemorrhage when such imaging determinants suggest the need for expediting treatment. In considering whether iAVM treatment is justified based on imaging characteristics, it is helpful to understand the natural history of iAVMs as related to imaging findings.

The ARUBA (A Randomised Trial of Unruptured Brain Arteriovenous Malformations) investigators reported an overall spontaneous AVM rupture rate in patients with untreated and unruptured AVMs of 2.2% per year. The trial was stopped early due to the increased rate of stroke and death seen in patients undergoing single or multimodality interventions; ARUBA showed a 12.32 per 100 patient-years risk of stroke or death following intervention as compared to 3.39 per 100 patient-years in the conservatively treated arm. This treatment risk in ARUBA outweighed potential gains, particularly when measured against the spontaneous rupture rate. As with any study, there were limitations related to this trial. For example, 56% of participating centers enrolled 5% or fewer of all their AVM patients evaluated during the study period. The 226 enrolled ARUBA patients therefore represented a small fraction of AVM patients treated at the study sites. Moreover, two-thirds of the patients enrolled in ARUBA had Spetzler-Martin grade I or II AVMs, even though open neurosurgical monotherapy, a mainstay primary therapy for low-grade AVMs, was undertaken in only five ARUBA cases. This suggests an apparent disinclination for open surgery.

Treatment options for iAVMs include open surgery, endovascular treatment, and radiation therapy/radiosurgery; however, neither single-modality nor multimodality treatment ensures permanent AVM cure. Residual AVMs may exist after treatment and be an unsecured source of future bleeding. In a systematic review of 137 observational single-modality and multimodality treatment studies, posttreatment fatality rates, complications, long-term risk of hemorrhage, and successful brain AVM obliteration rates were assessed. This study included 13,689 patients over 46,314 patient-years and demonstrated an averaged all-modality case-fatality rate of 0.68 per 100 person-years—1.1 following open neurosurgery, 0.50 following radiosurgery, and 0.96 following embolization. The corresponding posttreatment intracranial hemorrhage rates were 1.4 per 100 person-years overall and 0.18 following open neurosurgery, 1.7 following radiosurgery, and 1.7 following embolization. Successful AVM obliteration was observed in 96% of patients following open neurosurgery, 38% following radiosurgery, and 13% following embolization. Permanent neurological complications or death occurred in 7.4% of patients following open neurosurgery, 5.1% following radiosurgery, and 6.6% following embolization.

Imaging Assessment

Angiographic imaging assessment of iAVMs is directed at evaluating AVM location, AVM size, arterial feeders, associated aneurysms, AVM nidus morphology, and venous drainage and morphology. The categories used to assess these features are summarized in the following list. Key considerations are discussed in subsequent sections.

• Location. The location of an iAVM may be classified as superficial or deep. Superficial includes groupings such as frontal, parietal, occipital, insular, or temporal lobes and locations. Deep includes groupings such as the cerebrospinal fluid ventricles and related nuclei, thalami, diencephalon, corpus callosum, basal ganglia, and the posterior fossa contents, including the brainstem and cerebellum.

• Size. Intracranial AVMs are categorized as small (<3 cm in maximum diameter), medium (3–6 cm), or large (>6 cm).

• Arterial supply. Arterial supply is categorized as deep, superficial, or combined as well as perforator or arborizing. Deep arterial supply includes choroidal arteries, arteries of the posterior fossa, and deep perforator vessels, which are small groupings of arteries arising directly from a large parent vessel. Superficial arterial branches are cortical branches of the anterior, middle, and posterior cerebral arteries that are prominent arborizing vessels directly feeding portions of the nidus. Studies have examined the number of feeding branches and collateral extraterritorial arterial recruitment.

• Associated aneurysms. Aneurysms associated with iAVMs are categorized as being on feeding vessels, intranidal, or remote.

• Nidus morphology. The nidus morphology is categorized as plexiform, fistulous, or mixed. Plexiform morphology refers to a fine network of multiple indistinct supplying and draining vessels. A fistulous nidus possesses discrete supplying and draining vessels.

• Venous drainage. Venous drainage is classified as deep, superficial, or combined. Other drainage characteristics, including venous stenoses or dilations, may also be noted. Deep drainage as per the Spetzler-Martin grading system refers to angiographic visualization of the internal cerebral veins, basal veins, or precentral cerebellar veins. Venous morphology refers to distinguishing morphologic venous characteristics such as focal venous stenoses of greater than 50% or focal venous dilations at least twice the average diameter of the draining veins. Venous assessment also includes recognizing delayed venous draining, which could imply thrombosed or otherwise occluded venous routes of egress, and unexpected direction of venous drainage, also implying the same possibility.

Detailed angiographic assessment is not available for all iAVMs since typically only patients who are selected for endovascular treatment undergo endovascular subselective angiography and exploration. Such detail is not available for patients who may be triaged for surgical and/or radiation treatment without selection for endovascular embolization. The subselective angiography performed on patients undergoing embolization allows an otherwise unavailable detailed assessment of vascular anatomy and nidal subcompartment flow characteristics, including such variables as nidal aneurysms, venous stenoses, and venous ectasias. Though the majority of iAVM patients do not undergo subselective angiography outside the context of endovascular embolization, conventional diagnostic angiography, where catheters remain below the skull base, provides critical diagnostic and treatment information for all iAVMs.

Location