The Dark Side

“Ependymal” dissection may be a confusing term because many arteriovenous malformations (AVMs) do not reach a ventricle, but it is meant to describe the deep dissection that comes at the end of the resection. In terms of the AVM box, the ependymal plane is the bottom, the sixth and final side. In terms of brain anatomy, the ependymal layer is the tip of the classic AVM cone tapering to ependyma. This phase of dissection, whether actually ependymal or not, is deep and diametrically opposite to the position of the neurosurgeon, and, as such, is inherently difficult to see, like the dark side of the moon. The ependymal plane is darkened and physically blocked by AVM nidus, making it the most challenging one. To make matters worse, this plane receives perforating and choroidal artery supply which can be troublesome.

I always anticipate some bleeding during ependymal dissection; I am prepared if it comes, and pleasantly surprised if it does not. Therefore, before beginning ependymal dissection, all of the parenchymal dissection should be completed, the other AVM sides should be free, and the nidus should be mobile. It is easy to get lured down to the ependymal plane and lose sight of unfinished parenchymal connections on another side. The resection cavity is lined circumferentially with Telfa strips to protect parenchyma from manipulation or retraction, but more importantly, to deal with deep bleeding. If bleeding fills the surgical corridor, Telfa strips provide extra-anatomic landmarks that are more orienting than bland parenchyma, and they are protective.

Retractors solve the problem of deep visualization. A self-retaining retractor placed on the AVM mobilizes the nidus from side to side, and importantly elevates it as well. Lifting an AVM enhances the view of the ependymal plane and places deep connections on traction. Although I try to limit fixed retractors to just one, a retractor on the brain may also angle the viewing trajectory across the ependymal plane when the walls of the cavity are steep and deep. Retractors and retraction pressure can inadvertently kink or tear superficial draining veins, and retractors can be bumped by passing instruments or resting hands. With attention focused at the cavity’s depths, occlusions of superficial draining veins may go unrecognized with bleeding consequences. Secondary draining veins that limit deep exposure are safe to sacrifice at this stage. Deep hematomas should be found and evacuated to open up the deep plane.

Perforating Artery Supply

Perforating artery supply differs from cortical artery supply. Perforators traverse deep parenchyma rather than cortical surfaces, and are smaller in caliber, thin walled, fragile, and difficult to cauterize. Bleeding can draw the neurosurgeon into critical neural structures where visualization is compromised by overlying nidus, and tolerance to bleeding and electrocautery is low. Whether from lenticulostriates to a basal ganglial AVM, insular perforators to an insular AVM, thalamoperforators to a thalamic AVM, or basilar perforators to a pontine AVM, bleeding is brisk and hazardous with consequences for the surgical outcomes.

The battle with perforators, fortunately, is usually brief. Perforator supply is typically limited to less than five arteries, and in some cases may be just one. They coalesce at the ependymal tip or the AVM’s extension into white matter. They must be controlled one by one and never packed. Thin cottonoids are very helpful to protect brain, apply suction, retract against, and define the surgical plane. However, too many cottonoids will create a tangle of strings. Perforators are difficult to see angiographically because they are small, few, and overshadowed by larger inundating arteries. However, their presence should never surprise, even with cortical, lobar, or cerebellar AVMs. Any AVM that breaches white matter can attract these perforating arteries. Furthermore, as insignificant as they might appear angiographically, they probably increase their shunt flow during the course of circumdissection as other inputs are throttled. Perforators are typically the last of the arterial supply, and I often notice that a draining vein that was bright red at the beginning of a perforator skirmish is dark blue at the end.

Choroidal Artery Supply

Choroidal supply is generally more cooperative than perforator supply because these arteries are thicker, more coagulable, and floating freely in the ventricle. Choroidal arteries run in tela choroidea and choroid plexus, making them far more visible than parenchymal perforators buried in white matter. Ventricles open widely; transcortically, transcallosally, or through a choroidal fissure; and offer anatomic orientation, working space, and proximal control. Therefore, even though they are deep, choroidal arteries are like an artery in a cistern or subarachnoid space that can be identified with clear landmarks and controlled.

Other choroidal arteries may intersect the AVM diametrically opposite to the AVM. These feeding arteries are poorly visualized, lack proximal control, and are difficult to access. These difficulties are encountered with bulkier AVMs that have a large ependymal surface rather than just an ependymal tip, and a network of choroidal supply from the anterior choroidal artery (AChA), medial posterior choroidal artery (mPChA), and lateral posterior choroidal artery (lPChA), rather than a single input. Bleeding from diametrically opposite arteries hides from view, causing unchecked intraventricular hemorrhage and, in the extreme, outward brain herniation. Microclips can be useful with choroidal arteries when access behind a bulky AVM makes them difficult to coagulate.