64 Supratentorial Cavernous Malformations

10.1055/b-0038-162193

64 Supratentorial Cavernous Malformations

Michael Lang, Ricky Medel, Aaron S. Dumont, and Peter S. Amenta

Abstract

The supratentorial compartment represents the predominant location for cavernous malformations of the brain. These lesions may be found incidentally with the ubiquitous use of advanced imaging or be found in patients presenting with seizures, hemorrhage or focal neurological deficits. Asymptomatic lesions can be monitored with clinical and radiological follow-up while symptomatic lesions are treated based on multiple factors including presentation, location, associated deficits, seizure localization and other patient and lesion specific factors weighing the risks of treatment against the anticipated benefits. Lesions presenting with hemorrhage or lesion-related seizures/epilepsy amenable to resection can be treated via craniotomy and resection with microsurgical technique. Lesions presenting with an initial hemorrhage in a deep or eloquent region may be observed, with surgery or radio-surgery reserved for recurrent hemorrhage depending upon patient and lesion-specific considerations. Pre-operative advanced imaging including functional and diffusion-tensor imaging with tractography, neuronavigation, meticulous microsurgical technique and intraoperative monitoring can make surgical resection safer and more effective with lesions located in deep and/or eloquent locations. An algorithm is presented to facilitate the management of patients harboring supratentorial cavernous malformations.

Introduction

Cerebral cavernous malformations (CCM) are rare vascular malformations occurring in approximately 0.2 to 0.4% of the population and represent 8 to 15% of intracranial vascular malformations. These lesions occur throughout the central nervous system in a volume-based distribution, with supratentorial location in 70 to 80% of those affected. Presentation varies based on location, presence or absence of hemorrhage, and epileptogenicity. New-onset seizures represent the most common symptomatic presentation, though widespread use of magnetic resonance imaging (MRI) has increased the rate of asymptomatic detection.

Major controversies in decision making addressed in this chapter include:

Is treatment indicated for a given CCM?
Timing of intervention.
How does eloquent or deep location affect indication and timing of surgical intervention?
What is the role of non-microsurgical treatment of CCM?

Whether to Treat

Management decisions regarding CCMs are largely based on whether the patient is symptomatic or asymptomatic at the time of presentation ( 1 , 2, 3 in algorithm ). Despite their rarity, introduction of MRI has increased the rate of detection of asymptomatic CCMs. Incidentally discovered CCMs have a low rate of first-time hemorrhage (0.25–6%). The risk of symptomatic hemorrhage increases with age at time of diagnosis (representing lifetime cumulative risk), female gender, and eloquent or infratentorial location. Hemorrhage risk related to location is related to the risk of new or worsening neurological deficits even with small hemorrhages or minimal growth.

**Algorithm 64.1** Decision-making algorithm for supratentorial cavernous malformations.

Given the risks inherent to surgical or radiosurgical intervention, clinically asymptomatic lesions are followed conservatively with repeat MRI at one- or two-year intervals ( 4 in algorithm ). Cavernous malformations are dynamic lesions, with imaging and histologic appearance consistent with chronic microhemorrhage. As such, CCMs may be incidentally discovered with presence of small intracerebral hemorrhage (ICH), leading to inconsistent classification of hemorrhagic presentation in the literature.

Symptomatic presentation occurs in two-thirds of patients, including (in descending frequency) seizures, headache, ICH, and non-hemorrhagic focal neurological deficit (FND). Patients presenting with seizures undergo evaluation for surgical resection of the offending CCM and surrounding gliotic tissue. Patients are evaluated with phase I video electroencephalographic (EEG) monitoring to confirm that the seizure focus localizes to the anatomical location of the CCM. With incongruent phase I monitoring or multiple CCMs, consideration should be given to intracranial monitoring to precisely localize the epileptogenic focus ( 12-17 in algorithm ).

Surgical resection is typically considered with hemorrhagic presentation in superficial non-eloquent cortex after the first ICH. Surgical resection of CCMs in deep or eloquent locations after a second ICH should be considered depending on the degree of neurological impairment ( 5-11 in algorithm ). Enlarging lesions with progressive FND should be considered for resection if surgically accessible ( 18 in algorithm ).

Anatomical Considerations

Eloquent location and pial/ventricular presentation are the most relevant considerations in CCM surgery. In the supra-tentorial space, eloquent location typically warrants expectant management, particularly following first hemorrhagic presentation. However, lesions presenting to a pial or ventricular surface create natural surgical corridors that permit earlier surgical intervention.

Pathophysiology and Natural History

Cavernous malformations are dynamic low-flow vascular malformations. Once considered to be strictly developmental anomalies, they have been shown to occur de novo in both sporadic and familial populations, especially after cranial radiation. Histologic examination demonstrates interconnected venous channels resembling hepatic sinusoids, lined by single layers of endothelium separated by collagenous matrix. Sinusoidal CCM vessels do not possess typical features of cerebral vasculature such as tight junctions, elastin, or smooth muscle. Calcifications or cystic regions with evidence of blood products at various ages of breakdown are common. Hemosiderin staining is a constant feature, varying only in degree, serving as evidence of regular microhemorrhage. Cadaveric series suggest a 14 to 30% co-incidence of CCM with developmental venous anomaly (DVA). Classification of these lesions is based on MRI appearance (see below).

Retrospective population-based studies have shown hemorrhage rates between 0.25% and 2.3% per patient year, and 0.1% to 1.4% per lesion-year (though these calculations presumed risk present from birth). Reports based on prospectively collected data have shown a 0.8% to 3.8% hemorrhage rate. Asymptomatic cavernomas have an annual rupture risk between 0.39 and 4.2%. When symptomatic, the rebleed rate has been reported at 7% to 8.9% per patient-year. However, a feature of CCM-related ICH is temporal clustering (first reported by Baker et al), such that patients experience an increased rate of hemorrhage for the first two years after first hemorrhage with a return to or below baseline thereafter, peaking at 2.1% per month initially ( 1 , 2, 5 in algorithm ).

Hemorrhage rates are influenced by the proximity of symptomatic lesions to critical structures. Porter et al reported on 170 patient-years of observation, in which 4.1% of deep lesions were observed to bleed versus 0% of superficial lesions. Cavernomas located in eloquent anatomy are also associated with a low likelihood of full spontaneous recovery following hemorrhage.

Workup

Clinical Evaluation

Focused neurological examination should identify signs and symptoms related to a lesion of interest. Although rare, a large ICH secondary to a ruptured cavernoma is treated much like any other acute ICH. For patients presenting with FND or seizures, particularly in the setting of multiple CCMs, clinical localization should be concordant with imaging findings. Continuous EEG with video monitoring is utilized to confirm electrographic and semiologic concordance with the identified structural lesion. Consideration for resective surgery for epilepsy includes multi-disciplinary evaluation.

In the setting of multiple CCMs, genetic testing for known associated mutations is warranted. Familial CCM is defined as the occurrence of CCMs in at least two first-degree family members or the presence of multiple CCMs in a patient with confirmed germline mutation. Three autosomal dominant genes have been identified with familial CCM, known as CCM1 (KRIT1–chromosome 7p), CCM2 (MGC4607–chromosome 7q), and CCM3 (PDCD10–chromosome 3q) with increased frequency in Hispanic populations.

Imaging

Computed tomography (CT) demonstrates limited sensitivity for the identification of asymptomatic or small lesions. Large or calcified CCMs are identified as well-circumscribed hyperdense or mixed density intraparenchymal lesions. Acute hemorrhage is often first identified on CT, and early repeat imaging is recommended to ensure hemorrhage stability. Surrounding edema is suggestive of recent hemorrhage. CCMs are angiographically occult lesions, but an associated DVA may be observed in the venous phase of catheter angiography.

MRI is the modality of choice for evaluation of CCM, with sensitivity and specificity nearing 100% for identification of clinically significant lesions. The appearance of CCMs on MRI is particularly dependent on dephasing of local magnetic fields (susceptibility effect) based on the oxidation state of heme iron oxidation state in various breakdown products. The susceptibility effect results in a faster-than-predicted decay of the T2-signal, known as T2*, which is pronounced on Gradient Recalled Echo (T2*GRE) using gradient re-focusing instead of an inversion pulse. Susceptibility-weighted imaging (SWI) utilizes this effect similarly, with the added advantage of being able to differentiate calcification from blood products.

Appearance on MRI can be used to classify CCMs into four distinct categories (▶ Figs. 64.1 – 64.4 ), which have an association with predicted clinical course. Type I lesions are characterized by hyperintensity on both T1- and T2-weighted sequences (methemoglobin) representing subacute hemorrhage, often with a hypointense ring (hemosiderin/ferritin). Type II lesions demonstrate the classic “popcorn” appearance, with a reticulated core and mixed signal characteristics on both T1- and T2-weighted images. T2 and GRE sequences demonstrate a prominent hypointense rim that represents a hemosiderin-stained gliotic capsule. These lesions have a predilection for clinical progression and recurrent symptoms. Type III lesions are characterized by iso- to hypo-intense signal on T1-weighted sequences and T2 hypointensity that is representative of chronic hemorrhage. These lesions are prominent on GRE due to hemosiderin deposition. Clinically, these lesions are often asymptomatic at the time of imaging due to a lack of acute hemorrhage. Type IV lesions are punctate hypointense foci seen exclusively on GRE sequences. They are seen frequently in familial subtypes, and may represent early-stage CCMs. Typically, all CCM imaging classes do not enhance with gadolinium administration.