More Videos

Grand Rounds-Resection of Arteriovenous Malformation: Subtypes and Surgical Technique

Michael Lawton

November 16, 2012


- Hello, ladies and gentlemen and thank you for joining us. We're very lucky to have with us one of our special guests, Dr. Michael Lawton from University of California, San Francisco. He has extensive experience with surgical treatment, arteriovenous malformation that can be some of the most daunting, technically challenging lesions to handle. He's gonna share with us his technical nuances and Michael, please take it away.

- Well, thank you, Aaron and let me just begin by thanking you for inviting me to this forum and allowing me to share with you some of my thoughts on AVMs. The topics I'll cover are preoperative factors for minimizing the risk and some of the intraoperative techniques for doing the same. So let me just jump right into it here. I think when we consider AVM pathology, there's a spectrum of the types of lesions that you'll see. There's a spectrum of the patient's presentations that you'll see. There's a spectrum of the surgical difficulty, and you'll really see a variety of different outcomes. So, to deal with all these things really requires that you, develop a sense of who to pick for surgery. I think that's really one of the secrets to good results. We've all become accustomed to using grading systems because they predict the risk and they're applicable at the bedside, they take a very complex decision and help simplify it. And if you rely on these systems, they can help guide you through the selection process. So, at least when I see a new AVM patient, I always have my thoughts geared around these grading systems. This is the Spetzler-Martin grading system that I think everybody is familiar with. The variables are size, venous drainage and eloquence and you can see the point assignments described here. I think I don't need to spend any time on that because most of us are familiar with that. That system will break down AVMs into these five different grades. You can see that the grade III is the largest of the groupings. There are four different types. The others have either three or one in each of them. This I think is the real, it has become a real important part of the language of AVMs. Recently, Spetzler and his colleagues have simplified this into these three different classes, taking the grade I and II's and lumping them into a class A and the upper right of the spectrum, classing these grade IV and V AVMs into this class C. And the motivation for doing this was to help guide some of these management decisions, recommending surgery for the I's and II's and recommending no surgery or no treatment for the high-grade AVMs. That's basically this new classification and here's just a chart showing that when you look at the distribution, we've gone to even distributions now, four in each different class. So it's a little bit more balanced and a little bit more evenly distributed. The thing I don't like about it, if I can go back a slide, is it just doesn't really give you a lot of insight into, for example, the grade III's. What does this really mean multimodality treatment, does that mean embolization and surgery? Does it mean surgery and then radiosurgery? Does it mean radiosurgery then surgery? There's a lot of ambiguity there. And then there are these exceptions here in the no treatment category for class C that I think also helped to weaken that way of thinking a little bit. So, what we've done is we've actually gone the opposite direction and we've tried to enhance the grading by offering up a new grading system. This slide just shows you this concept of diffuseness, which a lot of AVMs are compact, the majority of AVMs are compact. Meaning that that tangle of arteries and veins is very tight. The opposite of that is when they're diffused, where that tangle is sort of pulled apart or lacy. And you can see different examples where in this one, there's just one kind of ragged edge here. This is another example where the whole nidus is sort of loosely pulled apart. Here's an example of the diffuse AVM with these large feeding vessels. And you can see this one here has this whole diffuse little area above it. So, there are many different patterns of diffuseness and the trouble with diffuseness is that there's normal brain that's intermixed with these arteries and veins. So it requires that you get into the brain a little bit more than you would otherwise with the compact AVM. One of the first groups to incorporate this idea compactness into a grading system was the Toronto Group and this was a grading system that they proposed looking at venous drainage, eloquence and compactness. Assign these different point numbers based on that assignment. And you can see that it's a little bit hard to remember the two, the four, and the three. It's also requiring that you reallocate venous drainage and eloquence to a different grading system. So, I think for those reasons, this Toronto grading system never really gained traction, but the concept that diffuseness was important, I think is the point. Another thing that we found important is the presence of hemorrhage on presentation. It's an important predictor of future risk for the natural history, but also surgically speaking, it really can facilitate surgical resection. Here's an example of looking down the Sylvian fissure, you can see this arterialized draining vein. And when you go to the insular cortex, you can see how having this underlying hematoma really helps you get around the AVM and remove this nidus. So hemorrhage is an important factor. And so we put together these three factors, age, bleeding and compactness. But I like to think of as the ABCs of AVMs. And we've put together this grading system in a way that's completely analogous to Spletzer-Martin. So, for these three age breakdowns, you get one point for pediatric patients, if you're 20 to 40, in the prime of life, that's two points, and if you're older than 40, that's when the co-morbidity start to creep in and that's three points for that age group. Bleeding, we look at ruptured or unruptured status. And then an unruptured patient, generally has no deficits. So you're more likely to cause some sort of morbidity in those patients, so they get a point in the risk scale. And then for compactness and diffuseness, again, compactness is a favorable factor, diffuseness is an unfavorable factor, and so you get a point on that for diffuseness. Again, like Spetzler-Martin, you add up the points at the end of it all, and you arrive at a grade that ranges from I to V. This is just the distribution in a sample of 300 above patients. Looking at Spetzler-Martin and looking at supplementary grades. And you can see a nice, fairly even distribution with a peak at around the grade III. So, more importantly though, is the outcome. If you look at the outcomes, according to supplementary grade, you can see that the morbidity steadily rises as your supplementary grade increases from a very low 3% here, all the way up to 50% for grade V. And it also increases in a way that I think is a little bit more stepwise than Spetzler-Martin. We looked at the statistics of the predictive power and this graph just shows you that in Spetzler-Martin grade here in red, gives you an area under the curve that's less than what you get with the supplementary grade. And what you're looking for, the greater the area under this curve, the more predictive power. So for Spetzler-Martin, the area under the ROC curve is 0.6, for the supplementary grading system it's 0.76. So greater predictive accuracy with this supplementary scale. Here, are couple of examples just to show you. This is an AVM that you can see here in the frontal lobe, if you graded on Spetzler-Martin, it's a grade IV, or if you use the Spetzler-Ponce Class, it's a class C. Patients 60 years old presented with seizure. This is one that I think most of us would feel very comfortable recommending observation. 'Cause the patient just had a single seizure, did not rupture and the morbidity associated with the grade IV is significant. So, on the supplementary scale, this would also be a grade IV because the patient's elderly. It has not bled and it is compact, so it's a grade IV. So, it's a IV Spetzler-Martin and a IV supplementary, so a combined score of eight. This is one that we would recommend management that is conservative. Here's the exact same AVM, except this time, it's in a 14 year old kid and it's presented with rupture. So you can see the bleeding here, down to the ventricle, here's the AVM. Spetzler-Martin stays the same, it's a grade IV. But if you do the supplementary grade, it's a grade I. A point for age, now it's bled, and it's still compact. So it's a supplementary grade I. If you combine these, you've got a IV plus a I, it's a combined grade of five. This is one that I think would be a good one to take out surgically, and we did in this case. So it just shows you how the clinical circumstances as they change your decision can also change. And the supplementary grading system will guide you through that. So that case example shows you an example of mismatch between Spetzler-Martin and supplementary. And what we found is that if you have matched risk prediction, in other words, the Spetzler-Martin and the supplementary are both low, then those patients really do quite well. The combined M and M for those is 15%. On the flip side, if you have high Spetzler-Martin and high supplementary, you have a 50% morbidity. So when things are matched, the decisions are really fairly straightforward because the risks are so different here. But what happens when the risk is mismatched? In other words, low Spetzler-Martin, but high supplementary. But what you see is that you have a higher than expected morbidity than what you would expect just from a low Spetzler-Martin grade. So these are ones that you have to be careful of, and you have to look at that supplementary score to see of that temperature decision. On the flip side, you have a high Spetzler-Martin AVM that you might otherwise manage conservatively. If that patient has a low supplementary score, you see that the risk is actually lower than what you would get with just a high Spetzler-Martin. So it's important to look for those mismatches and you have to sort of allow the supplementary score to recalibrate your decision. So what we do is we look at the combined grade, we add up the Spetzler-Martin with supplementary, and we draw this line right at six. So anything, six or less, has a very favorable risk associated with it. Anything seven or greater has a much higher risk, and we tend not to operate those. So this sort of summarizes how we approach an AVM patient. We look at Spetzler-Martin, we look at the supplementary grid, we add them together, six or less, generally fair game to consider surgical resection. Now it doesn't mean that you're gonna ultimately decide that, but it puts that AVM in play from a surgical standpoint. This is just another slide showing risk prediction. When you look at this supplementary, Spetzler-Martin score in red here, it gives you the greatest risk prediction on this ROC curve. This receiver operating characteristic curve. You just look at the others. You look at the traditional Spetzler-Martin or these other Toronto grading scales. You see that they're much less predictable overall. Here's just another example to make a couple of comments about the supplementary grade. It changes over time. It becomes dynamic because age will change. Their bleeding status can change. And sometimes even the compactness of the AVM can change. This is maybe one patient that presented with this diffuse large AVM. We treated with Gamma Knife radiosurgery, and you can see dramatic reduction in the volume of the AVM and the patient present the hemorrhage. So the compactness score changed here. The bleeding status changed here. The age remained the same, and you can see the Spetzler-Martin grade remained the same. So the supplementary grade changes with the patient, and in this case, it sort of pushes you towards a surgical posture. And that's in line with your clinical intuition. Just a quick comment about grade III AVMs, if you look carefully at the different types of grade III's. And what we found was that there are differences in the morbidity, depending on which subtype of AVM you're dealing with for a small grade III the risk was quite low, for an eloquent medium-sized grade III, the risk was much higher. So I think it's very important to look at what subtype within the grade III's you're dealing with to help you finalize your decision. This is just a chart showing that if you look at the small grade III, it actually has a risk that's like what you see in other surgeon's series for the grade II. So this one really behaves like a grade II, and so I tend to be more aggressive with the smaller grade III. On the flip side, if you look at the eloquent medium-sized grade III, the risk of that AVM is very much like what you see with grade IV's. And so that one, I tend to be a little bit more conservative with. So we did propose this a number of years ago, this modified Spetzler-Martin, which really breaks down the grade III's into these grade III minus, grade III plus, and this grade III asterisk, which I'll talk about in just a second. These are the subtypes, three minus is the small one with deep venous drainage. The grade III plus is the medium-sized one with eloquence. Three has medium-size and deep venous drainage. And this three-asterisk is one that you'll see very rarely because it's very rare to have some AVMs big enough to be six centimeters, but not have either a deep training vein or eloquence. This is a grade III asterisk they do exist. This is in the temporal lobe. You can see that it goes from temporal pole all the way back to mid temporal lobe. So it is greater than six centimeters and it had no deep venous drainage or eloquence. So they do exist. They're just exceptionally rare. The only other place where you'll get it is in the nondominant frontal lobe. But those two sites you can't see. So this is some data from my experience, 15 years worth of AVM patients, total brain AVM's 573. And you can see that about a third of these are the grade III. So they're very common, they're the biggest sub-group, and you should really carefully consider the subtypes as you go through. These are just a breakdown of where they're located really in all areas of the brain. The approaches to these deep AVMs really varies. I think, depending on whether you're in insular territories, basal ganglia territories with thalamus, these will dictate different surgical approaches, but in particularly dealing with the small, deep AVMS, it's a matter of finding the right surgical approach to get you there safely. And when you do that, you can achieve some good results despite these deep locations. And I'll just show you two examples. This is a thalamic AVM. You can see it's presented with hemorrhage. You can see the vessels on the medial wall of the thalamus and geographically, you can see it right here. The approach for this is a contralateral transcallosal that takes you into... Through the corpus callosum, into the lateral ventricle and down to the thalamus. Sorry, not contralateral. It's an ipsilateral transcolossal, and that approach will get you right there. This is the view that you get surgically. I like to turn the patient's head so that the midline is horizontal. It's a flap over the sagittal sinus. You can see that head position allows gravity to retract the frontal lobe, and it gives you this very favorable vector down to the AVM. This is that view that you get. This is in the lateral ventricle, foramen of Monro, choroid plexus, thalamostriate vein here. And once we're in, sorry, this is actually the contralateral view. So in the contralateral ventricle. This is the ipsilateral ventricle. This is the contralateral ventricle. As we go in there, you can see the draining vein here. There's draining thalamostriate vein that will then follow down into the velum and repositum. Here's the AVM. You can see I'm feeding arteries from the medial posterior colateral artery. This is AVM nidus here, and we have a very nice access to the nidus and we can peel this away from the medial wall of the thalamus here and dissect it back on this venous tail of the internal cerebral vein. There's the AVM after it's been circumscribed and this is the view into the third ventricle through the floor right in front of the mammillary bodies. It's a very nice exposure. It's a contralateral transcallosal approach, and it gives us a nice view. So I'm gonna shift gears at this point and talk about the surgical technique. This slide is just to kind of set up the idea that taking out an AVM is like military combat. There's a battlefield, which you have to know backwards and forwards and that's your anatomy of the AVMs. There's the enemy, which is the AVM itself. It's important to think about not only what the AVM is, but what subtype you're dealing with. And I'll talk a little bit about that. You need a good battle plan, which are your surgical steps. The actual resection is sometimes like waging war. It's a very brutal battle and the vessels can be angry and ugly. But I think, we are like soldiers and we just have to remain in the battle and keep fighting away as we go through these tough cases. So about those steps, I'm gonna talk a little bit about temporal lobe AVMs because they were nice example of how, if you study your arterial anatomy very carefully, you'll find out exactly the supply lines that feed this AVM. It's also important to know the venous drainage because you've gotta preserve those venous outflow routes throughout the resection. And the venous anatomy is very variable depending upon where the AVM is located. And it's also important to know your battlefield, knowing where the eloquence is located. In the temporal lobe, you've got your speech and language areas here in Wernicke's area, you've got Heschl's gyrus here. You've got areas on the medial temporal lobe involved in memory function, and you need to be very careful about preserving those areas. You also have tracks, white matter tracts. And in this illustration just shows you the optic radiations, back to the occipital cortex. So this concept of subtypes is very important. There are lobes of the brain or regions of the brain that harbor AVMs, and if you think about, within those areas, you get a lot of different subtypes within them. So within the temporal lobe, there are lateral subtypes, medial subtypes, basal and sylvian subtypes. So each different region has its own different subtypes. And I think we can start to recognize various patterns of AVMs that seem to repeat themselves over time. So these are these different subtypes in these different regions, which are important to learn. Just as an example, thinking about cerebellar AVMs. This is an example of a suboccipital subtype. It's on the suboccipital surface of the cerebellum. These are approached by exposing the surface of the cerebellum. It's a direct perpendicular approach, which gives you good exposure of all the different sides, usually unilateral. The feeding supply is from all the different vessels of the posterior circulation. The veins tend to be superficial as shown here, either going up to the trochlear or transverse sinus and right in the field. And it may be eloquent depending on how deep the nidus goes down towards the nuclei. Here's an example of a different cerebellar AVM, that's on the tentorial surface, the tentorial subtype. So it's on this up sloping surface of the cerebellum. It requires a slightly different craniotomy that exposes the torcular, allows you to get access to the transverse sinus. The approach is different, it's tangential. As you're going in the supracerebellar plane, it doesn't give you that direct view that you get, that perpendicular view that you get with the suboccipital radian. They tend to be unilateral. The feeding supplies, usually the superior cerebellar artery and the venous drainage pattern can be either superficial or deeper both. These also tend to be non-eloquent, but can sometimes extend deep. Here's a different subtype of vermian AVM that's right in the midline. This also requires a torcular craniotomy. The exposure can be perpendicular or tangential. The exposure tends to be midline and bilateral because the feeding supplies from both SCAs and sometimes even PICAs. The venous drainage tends to be deep to the galenic complex and maybe eloquent as well. The tonsillar AVM is down here on the tonsil. It's typically fed by PICA, draining veins are superficial, and this is typically noneloquent AVM. And lastly, the petrosal AVM, this is on the front surface, the petrosal surface of the cerebellum, abuts the cerebellopontine angle. It's exposed through retrosigmoid craniotomy and typically it's fed by AICA. So you can see that there are a lot of different subtypes within that one entity of the cerebellar AVM. And so, as you think about how you're gonna approach the resection, I think thinking about the subtype can help you. Okay, this is the view of a Sylvian AVM, looking down the Sylvian fissure, and you can see this dilated venous anatomy, these superficial Sylvian veins. As we mobilize those veins to the temporal side, you can start to see the AVM nidus. And one of the first things we need to do is to sweep those veins and AVM nidus apart so that we get around this border and what we're looking for, the normal arterial vessels, deep to the AVM, so that we can separate those. This is the view, looking down the middle cerebral vessels, you can see normal arterial anatomy there. And by just taking down these subarachnoid connections between normal anatomy and the AVM nidus, you can sweep the nidus towards the frontal side of the fissure, and you can sweep the normal arteries to the temple side and separate the two. There's a view of the M1 as it trifurcates into the distal branches and more normal anatomy there. So at this point, we've separated the normal arteries away from the AVM nidus, the AVM nidus is over here on the frontal side of the fissure. And having done that, we can now start to make a transition from subarachnoid dissection to this deeper peel dissection. So with the bipolar cautery, we can cauterize around the margin. We can incise the pia and enter into the brain a little bit on this frontal lobe here. And as we do that, we close off all these individual feeders on the surface. So this is that pial incision, and as we go and circumscribe the AVM, we get to these deeper feeders that go through the white matter. These are colorized and divided and by circumferentially going around the different sides of the AVM, we slowly choke it down. You can see that color change and now the AVM has been completely circumscribed. What that video showed, is the different steps for resection, from exposure to subarachnoid dissection, to finding the feeding arteries, preserving that draining vein, making pial incisions around the AVM, transitioning to a parenchymal dissection, and finally going to the deep ependymal layers, in this case not all the way down to the ventricle, but finally to that deep plane, preserving hemostasis as you go and finally taking out the AVM. So there are these nine different steps that you go through with each AVM and I'll just elaborate on each one of these. In terms of exposure, I'd like to think of an AVM as a box with six sides, there'll be a superficial plane and there'll be four different sides around it and then there'll be a deep plane. And so if you can face that box perpendicularly, it gives you good exposure of all the different sides and the superficial plane, leaving only that deep plane that's on the opposite side of the AVM, where it's more challenging to see. In contrast, a tangential resection often just gives you a view of a portion of the AVM and it's very difficult to see these other sides, which are, in this example pushed out of the view. Other things to think about are free surfaces or surfaces that are normally dissected or exposed for you. Like when you first come on an AVM, there are deep fissures that often need to be open. And in planning your exposure, you should think about how to position the patient so that you can maximize these different surfaces and planes. Just to give you an example of that, these are five different AVMs that are along this kind of sagittal plane. And depending upon which quadrant you're in here, the head position changes. Here the nose is up, here things are turned to the side, and here the nose is all the way down. For these deeper locations you want the hemisphere to fall away. So you're using gravity retraction in these different AVMs. So you really will vary your position, depending on where you need to go. In terms of the craniotomy, you always wanna get as much exposure as possible or as necessary. And you really only use a unilateral simple craniotomy in about a third of the cases. About 40% require a more extensive craniotomy that crosses the midline to expose these midline planes. Skull base approaches were used in about quarter of patients. So you can see that it's really the minority of cases that use a simple unilateral convexity craniotomy. This shows how patient position is important. This is a vermin AVM in the cerebellum. The patient is prone, so the anatomy is turned upside down. Tentorium is down here, the feeder up here. And when you go supracerebellar, you get to the AVM here in a way that puts the galenic complex down in this direction. And all the feeding arteries in this area here. So here the anatomy is upside down and you really don't have any benefit of gravity opening up their plane. Here's the opposite. You can put the patient in the sitting position and when gravity causes the cerebellum to sag, it really opens up this plane over the cerebellum and allows you a better view. So here's an example now with the patients sitting. This is the tentorium up here. Here's the cerebellum down here, and this is a supracerebellar infratentorial approach. You can see all the way down to the tentorial incisura, here's the incisura of the tongue. These are the draining veins up to the galenic complex. And as we deepen the dissection, we have a very good view here of the feeders as they come around from the superior cerebellar vessels and here on the opposite side. And here we are circumscribing the AVM, here's the arterialized draining vein going up to the galenic complex, and we have a nice shot at this AVM, owing our exposure to the gravity of retraction of the cerebellum. Now the second phase is the subarachnoid dissection, which is particularly important in those Sylvian AVMs, but even in a prigmore AVMs like this one here, sometimes have a nice margin around the AVM. And just by going down those sulci and really separating the nidus from the brain, you can separate the AVM very extensively. Other cases like this one, you've gotta just open up these fissures to start to visualize the nidus. Other times you're looking for these little venous landmarks. These arterialized veins that lead from the AVM that will guide you down a sulcus or down a fissure and take it to the AVM. You will typically see these whitened layers of arachnoid around AVMs that need to be opened up so that you can start to visualize the nidus. Now, this is the peel phase of dissection. There's typically an area or a point after you've maximized your subtract nodal dissection. When you have to actually enter the pia and enter some of the brains. So here we're still in the subarachnoid phase. Here, we can separate normal vessels from the AVM. We can get a nice view of the margin, but at some point that has to change. And we enter the brain a little bit on this side here, and we start to remove this or separate this from the brain. You can see darkening of the vein here and here after resection, we've got a good removal. So there is that point, where you need to transition to a pial dissection from the subarachnoid space. Finding the feeding arteries is critical. You wanna include them right at the margin of the AVM. In some cases you need to skeletonized the arteries, which means to kind of prune the vessel branch by branch, as it feeds in so that you preserve a flow that continues beyond the AVM and all the while you wanna really find your normal arteries and carefully preserve those. Some of the mistakes that can be made are, taking an artery too early, before is finished giving off branches to normal brain. Another mistake is to occlude an en passage artery that feeds the AVM, but also goes on to feed normal territory beyond it. And finally, occluding some uninvolved artery. Let me show you some examples of that. This is an anterior trans-sylvian approach, and it's a nice example of how you need to work between normal anatomy to get to the AVM. These are all normal vessels that go by the nidus and we've got to work through them, between them to get to the AVM. So this is the view, looking down the Sylvian fissure, you can see an arterialized vein here, and as we dissect the fissure, we see the M1 feeding into this trifurcation. And by working between these M2 branches, we can see this arterialized vein here. And our nidus is right here, deep to the trifurcation. So we've gotta work between these en passage arteries and carefully preserve vessels that are uninvolved with the AVM. This is an example of skeletonization. This is a large vessel that gives up a small branches. And as we open up the Sylvian fissure, we come down on the parent vessel here. And what we wanna do is we wanna get on that artery distally and follow it proximally, so that as we encounter these feeders, one by one, then go to the AVM, we can close them, but still maintain the distal flow to the artery as it goes by. So here's cauterizing that little branch, and you can see at the end, we've pruned this. These are all individual branches to the AVM that have been taken down, but the flow to the rest of the brain distally is preserved here. So there's inflow on this side, outflow on this side, and we've just taken out the feeders that go to the AVM. That's what we refer to as skeletonization. Here's an example of a mistake, occluding an artery too early. This is a posterior interhemispheric approach to a medial occipital AVM and this is our feeding artery right here. You can see the arterialized vein right here. And as we follow this along, we get closer and closer to the AVM, but we have to be patient. We have to make sure that we close that artery just as it gets to the AVM. And here we show the resection cavity, nice resection, but you'll see here that artery was occluded a little too early. And so some of these distal branches are not preserved. So that en passage vessel has been taken down. And the solution to that mistake is to just follow the vessel all the way to the nidus and make sure you're at the margin of the nidus before closing it down. Draining veins. These are sometimes very difficult to recognize. You have to study the wall very carefully. Looking at how thin it is, how it feels in your bipolar. Sometimes the presence of embolic agents or other things can be clues to the venous morphology, but you wanna really carefully differentiate vein from artery because they're all gonna be red. So color is no guide. You really have to look at some of these other features. I always designate what I call primary venous tail this main draining vein that will come out of the nidus. Other veins are fair game to sacrifice, but you need to designate and preserve that primary vein. That's what I call trimming if you take a secondary vein. The vein is a nice indicator of how you're progressing with the resection. You'll see the color change from red to blue, and it's a good guide. Mistakes that can be made involve injuring the vein during the drill opening or during the dissection. Misinterpreting a vein as an artery, and then closing down that artery too early. Any kind of early venous occlusion can be catastrophic. So you really wanna be careful. Whenever you see a red vessel, that's bigger than what it should normally be, you should think vein not artery. So here's an example of how, when you're going down the Sylvian fissure and you have this arterialized venous complex, it all falls down to this large dilated venous varix here. And as you go around your nidus and dissect this thing, you can see that color change. There's this dramatic darkening of the color of the vein. And it tells you exactly how you're progressing. Now, once you've done that, you've closed off your arteries, you've designated your vein. The work of getting around those parenchymal sides of the nidus side is where a lot of this resection happens. You wanna circumferentially go around that box. These are the four sides around the nidus, you wanna follow that gliotic plane. Sometimes it's very well-demarcated. You can have some red hemosiderin, you can have some scar tissue. So it sometimes makes that dissection very nice and straightforward. Other times it's right up against critical white matter tracks. So you really wanna make sure that you stay very tight against the AVM nidus, 'cause constantly this battle between staying right on the margin and minimizing your transgression into the brain. But also you don't wanna be so close to the nidus that you get the AVM bleeding. So there's that balance between staying tight and going wide. This is just an example of how, in eloquent areas like this, where the AVM is in the speech areas, we've actually done this awakened with mapping. Here you wanna be very tight on the AVM so that you really preserve all of that speech cortex that's around the AVM. In other cases like this one, this is the occipital pole. There's really no sense in being tight out of large AVM like this, where you're going to likely lose some visual field anyways. Here, you would wanna go wide and make sure that you don't get bleeding from this big nidus. And here it's more like a lobectomy with the whole occipital pole comes out with AVM. This is just a slide to show how part of what makes up that diffuse border is the deep perforator supply and when you have diffuse AVMs with lots of perforators, you can see here that the morbidity goes up. So you really wanna be careful when you see a diffuse AVM with these deep perforators. You'll encounter a lot of those perforators on that last side of the box, that deep layer or that dependable side. Everything else is done. You've done your four sides. You've done your superficial side. And now you come across the, oftentimes these deep feeders. These, what Charlie Wilson used to describe as the red devils. The AVM microclip is a nice answer to these. They're often these meticulous strike feeders or these collateral ependymal feeders that get to the deep tail of the cone of the AVM. And these clips are real nice because these vessels don't like to coagulate. The AVM clips are a nice alternative way to close them down. You wanna completely surround that ependymal tip as you go around a deep portion. And you also wanna look for a hematoma. If you can get down around that deep plane, and there's an underlying hematoma like right here, it often has defined that plaintiff separation for that deep side. Remember that this is sort of like you have to see around the whole nidus of the AVM. So it's the hardest to see. So having a hematoma, that is a huge advantage. The other thing that you often need is some retraction at this stage. I generally like to minimize retraction, but when you're trying to get around that deep plane, it helps to sometimes pull a little bit on either the brain, bank of tissue or on the AVM nidus itself to get around that last bit. This is just a slide showing that if you look at modified Rankin scale outcomes, patients that present with ruptured AVMs tend to overall improve after therapy. Those that present with unruptured AVMs, tend to worsen after therapy. And I think the difference is the presence of that hemorrhage. Not only does it do some damage when it bleeds, but having a hematoma there, can really facilitate the resection and make things safer and easier for you. This is just an example of that ependymal plane of dissection. Here's the AVM. Here, these are the middle cerebral feeding vessels that feed into the AVM nidus. And here we've gone around the skeletonized, this feeder you can see those little interrupted branches to the AVM and here's our draining vein. There is the nidus, you can see how this ependymal surface, this went all the way down to the ventricle and having that hemorrhage in that encephalomalacia around this really made this a fairly straightforward resection. Another... This is the same case, just showing that preservation of the M1, the M2 branches as they go en passage. And this is all carried out within the system of the Sylvian fissure. Now, hemostasis is critical. You want these resections to be bloodless. You wanna maintain absolute hemostasis throughout the resection. When you get bleeding, you need to stay on it. You need to suction and cauterize. If that fails you, then you can use those micro clips, but never pack a bleeder, never leave a bleeder, just keep working at that spot where the bleeding is until you can get control. Very rarely, if things get out of control, you can do what I call the commando resection, where, if you're struggling a lot and things are getting away from you, sometimes you just need to get the AVM out quickly and then deal with the hemostasis after the nidus has been pulled out. But that's really quite rare and it's a tenuous way to go. You wanna maintain that bloodless resection throughout whenever possible. The problem is that when you lose control, your field fills with blood and you really can't see what you need to see. These feeders need to be meticulously colorized, and you need to maintain that hemostasis in order to see that. Finally, comes the gratifying part, where you can pull out the AVM, the vein is blue. You can cauterize and resect that. It's always a good idea to kind of roll the AVM out on its venous pedicle. That guarantees that you've completely severed all these connections to the AVM, and that you've completely surrounded the nidus. The mistakes come when you do that too early and you're not quite ready. This is a picture of a scarred radiated AVM. And here we've gone around the nidus and at the very end, after you've gone around these borders, it's nice to have the vein blue and to have things sort of really, have the nidus just sort of hanging on that tail. So you can pull the nidus out. You can ensure that everything's disconnected. And at that point it's safe to cauterize and divide the vein. These are long cases. It's like that movie, "The Hurt Locker" where you can find yourself in this kind of situation or this kind of situation, they get more difficult the more you go through the dissection. That last plane of dissection is the most difficult, and you have to monitor your fatigue. And if you're making errors in judgment or errors in technique, it may be that it's time to stop and stage this. This is unforgiving pathology. So you really have to, stay on it and make sure that you're up to the challenge. So these are my conclusions. These are really challenging lesions, not only at the bedside where you're forced to make decisions about whether or not to treat, but also in the where you're faced with the very unforgiving and difficult pathology. The grading systems are important and I'm a real fan of both the Spetzler-Martin and the supplementary grading system. I think they work well together. I think it's important to prepare for your surgery by studying the angiography, studying the anatomy and in your mind's eye, just getting your plan of attack in order. The difficult AVMs are unforgiving. You can get some poor outcomes and you can regret your decision to operate. So you really have to be selective. You have to tell a certain number of patients that your decision is not to operate. And that I think really takes some courage to tell a patient that you're not gonna do the surgery, but you really have to learn that selectivity. So those are my comments. I hope that wasn't too long, but I think those are sort of my pearls of wisdom after quite a number of these.

- Michael, thank you so much. That was extremely helpful. We're also gonna have a shorter session, reviewing some of the surgical videos and hopefully giving another perspective, watching the videos. Again, I'm very appreciative and all of the viewers, thanks. Thank you for your time.

Please login to post a comment.

You can make a difference: donate now. The Neurosurgical Atlas depends almost entirely on your donations: donate now.