Surgical Management of Cranial and Spinal Arteriovenous Malformations

Definition and Pathogenesis

Cerebral AVMs are vascular lesions consisting of tangled anastomoses of blood vessels in which blood is shunted directly from the arterial system to the venous system, with no intervening capillary bed. AVMs typically occur sporadically; however, they can also occur in patients with diagnosed hereditary disorders and in those with familial forms. ^, Although the exact pathogenesis of cerebral AVMs continues to be investigated, recent literature has implicated somatic mutations in the KRAS gene, a signaling protein in the MAPK/ERK pathway, as a key feature of sporadic AVMs. Activating KRAS mutations in endothelial cells triggers dysregulation of angiogenic genes and downstream proteins previously identified as contributing to the development and maintenance of AVMs, such as vascular endothelial growth factor (VEGF) and Notch. Indeed, VEGF and Notch signaling both have been shown to affect vessel wall morphology and architecture through the regulation of pericytes, and this effect has been suggested to play a role in the biology and clinical course of AVMs.

Natural History

Cerebral AVMs account for between 1% and 2% of all strokes and for most hemorrhagic strokes in children and young adults. The incidence of cerebral AVMs has been approximated as 1 per 100,000 persons per year in unselected populations, with a prevalence of approximately 18 per 100,000 persons. ^, Individuals with AVMs in whom the nidus has not been completely obliterated are subject to a life-long risk of hemorrhagic stroke, and the risk of mortality is 10%, and that of morbidity is 30%–50% with each episode of AVM-related hemorrhage. Significant efforts have been made to characterize the long-term overall risk of hemorrhage in patients with untreated AVMs. Most studies report annual rates of intracranial hemorrhage for patients with AVMs to be approximately 2%–4% per year; however, reported rates have varied widely, ranging from 2%–32.6%. A recent systematic review and meta-analysis published an annual hemorrhage rate for cerebral AVMs of 3%. The rate of initial rupture was approximately 2%, with a re-rupture rate of 4.5%. Re-hemorrhage rates in the first year after presentation with rupture were reported to be between 6% and 15%. However, the risk of hemorrhage for each AVM is variable and depends on multiple factors, including location, morphology, and angioarchitectural characteristics of the lesion.

Many variables have been identified as being associated with an increased risk of hemorrhage. These variables include a deep or infratentorial AVM location, a periventricular location, deep venous drainage, small AVM size, impaired venous drainage, a single draining vein, presence of an intranidal aneurysm or multiple associated aneurysms, elevated feeding artery pressure, and blood supply from the posterior circulation. ^, A recent meta-analysis of the literature found that significant risk factors for hemorrhage included prior hemorrhage, deep AVM location, exclusively deep venous drainage, and associated aneurysms. Contrary to prior published studies, there was no increased risk associated with small AVM size (<3 cm) or older patient age.

Clinical Presentation

Most patients with cerebral AVMs (52%) present with hemorrhage. These hemorrhages typically result in acute onset of severe headache and can include loss of consciousness, nausea, vomiting, neurologic deficit, and seizure. Patients with cerebral AVMs may also present with epilepsy, headaches, focal progressive neurologic deficits, or, in newborns, congestive heart failure. Symptoms from unruptured AVMs can be produced from mass effect, inflammation, altered hemodynamics (e.g., steal phenomenon), or recruitment of dilated perinidal capillary networks. Advances in neuroimaging techniques, such as improved image resolution, have permitted greater detection and evaluation of cerebral AVMs. A greater number of AVMs are now incidentally discovered due to the greater availability of computed tomography (CT), magnetic resonance imaging, and magnetic resonance angiography. Because of the large number of possible presentations for these lesions, it is important to identify which patients will benefit from intervention compared to conservative management.

Lesion Classification

Many different systems have been proposed to classify AVMs based on their surgical risk, but the most common system for risk stratification has been the Spetzler-Martin grading system. Originally proposed in 1986, the Spetzler-Martin grading system classifies cerebral AVMs based on size, eloquent location, and deep venous drainage ( Table 74.1 ). Spetzler-Martin grades range from grades I to V, and as the grade increases, so does the surgical risk. Both retrospective and prospective studies have validated the Spetzler-Martin grading system as a clinically applicable and reliable classification system.

Favorable outcomes for patients with Spetzler-Martin grade I lesions have been reported in 92%–100% of patients. Approximately 95% of patients with grade II AVMs achieve a good or an excellent outcome. ^, Reported outcomes for patients with Spetzler-Martin grade III lesions vary greatly; good or excellent outcomes have been reported in 68%–96% of patients. ^, For patients with Spetzler-Martin grades IV and V AVMs, good or excellent outcomes are reported in 71%–75% and 50%–70% of patients, respectively. ^, Based on the great variability in surgical outcomes for patients with grade III lesions, attempts have been made to subclassify these lesions with the hopes of better defining surgical risk. The first subclassification consisted of grades IIIA and IIIB. Grade IIIA included all AVMs greater than 6 cm in diameter occurring in noneloquent cortex without deep venous drainage, and grade IIIB included all other traditional Spetzler-Martin grade III lesions. This modified classification scheme reported fair or bad outcomes in 4.5% of grade IIIA and 30% of grade IIIB AVMs after surgery. These results prompted the authors to recommend microsurgical resection, or embolization followed by microsurgical resection, for grade IIIA lesions and radiosurgery for grade IIIB lesions. A second modification of the Spetzler-Martin grade III AVM category favored deconstructing each combination of AVMs within the grade according to size (S), venous drainage (V), and eloquence (E) of the surrounding brain. The risk of a new neurologic deficit or death due to surgical resection was 2.9% for small AVMs (S1V1E1), 7.1% for medium/deep AVMs (S2V1E0), and 14.8% for medium/eloquent AVMs (S2V0E1). Despite including a consecutive series of 174 AVM patients, the author did not treat any S3V0E0 AVMs in the study.

Spetzler and Ponce published a revision of the Spetzler-Martin classification system in 2011, recommending the distillation of the grades into three classes. Class A consisted of Spetzler-Martin grades I and II AVMs, class B consisted of Spetzler-Martin grade III AVMs, and class C consisted of Spetzler-Martin grades IV and V AVMs ( Fig. 74.1 ). Microsurgical resection is recommended for all class A AVMs, and preoperative embolization is discussed as a treatment option but not recommended based on the current risk profile for endovascular treatment. Surgical morbidity for class A AVMs has been reported to be from approximately 1% to 2.5%, which tends to decrease with longer follow-up. ^{, , ,} Class B AVMs represent a heterogeneous group, as evidenced by prior attempts at subclassification. An individualized approach to intervention for class B lesions is recommended by the authors, who suggest consideration of multimodality therapy. Patients with class C AVMs have surgical morbidity and mortality rates ranging from 20% to 30% under the best circumstances. ^, Conservative management is recommended for class C AVMs with few exceptions. In patients with repeated hemorrhage or progressive neurologic deficit, treatment may be considered. Aneurysms associated with class C AVMs should also be considered for microsurgical or endovascular treatment.

Additions have also been made to the Spetzler-Martin grading scale in attempts to improve predictive accuracy. Lawton et al. published a large series of microsurgically treated AVMs and compared the predictive accuracy of the Spetzler-Martin grading scale to a supplementary scale. The supplementary scale included the patient’s age, diffuseness of the AVM, presence of deep perforating artery supply, and whether the patient presented with a rupture (see Table 74.1 ). The authors found that the supplementary scale provided better stratification of surgical risk and could alter treatment decisions in cases of mismatched Spetzler-Martin and supplementary grades. Low-grade lesions (supplementary grades 2–4) were associated with excellent postoperative outcomes (91%–100% neurologically unchanged or improved), while supplementary grade 5 and 6 lesions had slightly worse outcomes (78.9% neurologically unchanged and 72.9% improved). Supplementary grades higher than 7 were associated with significantly worse outcomes, leading the authors to recommend supplementary grade ≤6 as a cutoff point for guiding treatment decisions.

Although the supplementary scale is slightly more cumbersome to translate to the bedside due to its increased complexity, it does highlight several critical features of AVM treatment and outcomes. Additionally, it has been validated in a multicenter cohort as a reliable predictor of neurologic outcomes after surgery, further supporting its use when treating patients with this complex lesion.

Surgical Considerations