Transcript
I'd just like to thank BrainLab, particularly the UK and Ireland branch of BrainLab for inviting me. I think AVM management, as you've heard from the previous speakers, it's probably one of the most variable across different centers around the world. And if we have a good registry, that will probably show the variation in practice between the different centers. But what all the speakers I think would agree is the way we target and the accuracy of targeting of an AVM is very important. Here's an AVM at surgery, and you can basically see the large draining vein. You can see the hemosiderin staining in the brain around it indicating the previous hemorrhage. Here's an AVM on an angiogram, and you can see the nidus, the draining vein. And what we want to target has been shown before is the nidus, rather than a lot of the venous drainage or arterial feeders. And basically, my talk is on the accuracy of targeting on that. Here's an AVM that's hemorrhaged in a young girl.
So you've heard some of this already, the risk of rupture according to ARUBA is about 2% per year. But if you have previous hemorrhage, or associated aneurysms, or deep venous drainage, that risk goes up. You've heard about different options. Medical therapy just means really conservative management and managing the symptoms. At our centers, I would say the vast majority of AVMs either get microsurgical resection or radiosurgery. And we really have moved away from nidal penetration of AVMs with embolizations, so very, very specific targeted goal for embolization.
Now, SRS for AVMs is very well established. It was one of the first indications for radiosurgery and has been around since the late '70s and '80s. And we have seen a transition from the frame-based techniques to the more frameless techniques. We use the Novalis at our centers. Now, the imaging modalities that we use for targeting an AVM, you've got on the left a 2D digital subtraction angiogram. So typically, you get your AP and your lateral images, and typically you would just identify the best image you can get on a still image on an AP and lateral digital subtraction angiogram.
Over the years, obviously, the 3D imaging has become available either using MRI or using CT. So CT angiography, or the rotational angiography using DSA, or MR angiography. Now, we have to somehow incorporate these together. Now, one of the first questions is, "Do we actually still need DSA to target an AVM? Can we not just use the MRA or the CTA, which shows us the AVM very nicely on a T2, you'd see the flow void on the MRA. Top MRA you would see the AVM. So do we still actually need the DSA?" So my view would be yes. DSA is certainly needed to target very small AVMs that you wouldn't see very clearly on an MRA or CTA. Some AVMs have a very diffused SNIS. It's quite difficult to know where to start and end, particularly on MR Imaging. If you've had previous embolization or previous hemorrhage, it may be quite difficult to target it on MRI alone or CTA alone. But I actually think all AVMs, do you do benefit in terms of targeting the nidus with a DSA? There are some examples here, you can see on your left a very small posterior thalamic AVM, which is quite difficult to see on the MRI. And it's better seen on the DSA. Here's a patient with a hemorrhage from AVM, and you can see the cavity where the hemorrhage occurred. And on the T2 weighted imaging, it's difficult to target that AVM whereas on the DSA, you can clearly see it here.
There are very few studies done on this, but there are two that I would mention. One was published in neurosurgery in 2008, and they looked at 34 AVMs with targeting using 3D imaging alone with MRA versus targeting of with MRA and DSA. And I think their major finding is that of the 34 cases, 19 of them target volume was larger if you didn't use DSA to help with the planning. And similarly in 2007 in the "Red Journal", there is another paper. And then again, AVMs delineated on MRA alone tended to be significantly larger and more subject to inter clinician variation. So that means your patient has come to you with an AVM, whether ruptured or unruptured, and almost all of those that we would see in our clinic, our next step if we are gonna treat is a diagnostic DSA. So we put our patient through one DSA simply for diagnosis and getting some more information about the angioarchitecture before our MDT meeting to see how we gonna excise it, or SRS it, or embolize it.
That then means that after our MDT meeting, if we decide to go for radiosurgery, we now need a planning stereotactic DSA, which in the frame-based techniques would need the frame and a fiduciary box and an angiogram and the frameless, the mask, the box, and another, so you're basically subjecting your patient to a second angiogram. Now, catheter tech... so you obtain your DSA with the fiduciary box. We use a Novalis so we import it into iPlan and then you match these fiduciary lines together to get rid of that distortion that you get from your 2D DSA, and then fuse it with your CT scan for planning and if an MRI, if you're gonna use that. In that standard way of doing it, what you then get is a static AP and lateral image from the DSA and an overlap of the target on that. You can't run through the DSA as you would normally. So that's your additional planning DSA. DSAs nowadays with catheter technology has improved a lot, so the risks are very low, but it's not zero. So you certainly have patient discomfort. Some of my patients where I take an AVM out through a craniotomy tell me that the angiogram afterwards was worse than having the craniotomy. So it does have some discomfort. It has a low but significant risk of stroke, arterial dissection, hematoma, sometimes femoral pseudoaneurysm, potential for contrast reaction, its additional radiation dose, and certainly cost. It does mean another day-case angiogram for your patient and a highly skilled team that you need to have to do that.
So I'm gonna talk about Smartbrush Angio, which is basically part of the elements package. And here, instead of using your fiducials to fuse the DSA with your MRA, you're basically using the vessel tree to do that. So there's no need for that additional stereotactic DSA because you're gonna use your original diagnostic DSA which you had and fuse that with your MRA. So here it is. You import your images there and then it moves on to showing you the DSA and then it overlaps the arterial tree from the MRA onto the DSA. You can move it around in any direction that you want to try and match those two as best as you can, and then click Fuse and the machine... the software essentially fuses your MRA with your 2D DSA. And on the next screen, you can essentially outline your region of interest under your AP and lateral, except that here, I can run through the whole DSA as I would normally do. And when we go forward, it will automatically generate the AVM nidus for you, which you can view on your MRA and your DSA side by side. And you can view your MRA in coronal, in sagittal, or axial planes, and you can have any of your MR sequences that you want, T2. And you can see here the accuracy if I'm running the cursor over the left internal carotid, and on the other side, I can see that it runs over the carotid artery on the angiogram, and similarly over the superior sagittal signs.
So we've got our AVM, and you can adjust your AVM nidus in using the software. And as you do that, it automatically shows you happening on the DSA. And here it is on the T2. The current version allows you to do the changes on MRI only, but on the next version, you can also make changes on the DSA and see your changes on the MRI. And the next step would be to export this out of elements. At the moment you export it into iPlan and fuse it with your stereotactic DSA that's with your stereotactic CT and then you do your dose planning. So when we got this software originally in, kind of, late 2013, as always we don't trust anything, so we had to do our own validation and our own physicist made a phantom with tubes inside the phantom. And we filled it with the appropriate contrast for CT for MRA, for DSA, and we imaged the phantom, and we compared the MRA DSA fusion with the CT scans. And I won't for a minute to pretend to understand this. This is done by Chief Physicist Paul Davenport, but essentially, the fusion accuracy was excellent. It was in part... we use an upper loop where the accuracy was less than 0.4 millimeter and the lower loop, less than 1.2 millimeters. And then we did three patients where we did the old-fashioned way with the fiducials and the new way without the fiducials using Smartbrush. We generated vessel trees, we outlined different parts of the vessel tree, and again, the mean accuracy was less than one millimeter.
So we then went ahead and treated our patients using Smartbrush. And we've treated almost 50 patients, 20 of whom have more than two years follow-up. And in those 50 patients so far, we have two radiation-induced edemas. Professor Yamamoto in 1980s reported the T2 signal changes that you see after AVM radiosurgery. So we've had two of those which were symptomatic with seizures both controlled with dexamethasone and some extra anticonvulsant medication, we've had one patient hemorrhage and die from that in-the-latency period, and we have the 19 patients with two or more year follow-up, of whom 12 have had complete obliteration and four have had significant reduction and I think they will continue to obliterate. I just want to emphasize that our doses are between 16 and 20 greys. We don't use the 24 and higher greys that are often reported in the literature.
So here's the 40-year old taxi driver, which in UK and Ireland, if I took his AVM out he can't drive for one year, and he has a medial-temporal AVM with lenticular strike supply. It's an unruptured AVM, but he and we decided to treat it despite ARUBA. And in 2018 you can see complete obliteration of the AVM. For the pediatric age group it's great because you can avoid the additional angiogram and the general anesthetic that goes with it. Here is the 13-year old with a ruptured AVM that you can see here with the intraventricular hemorrhage. And in 2018, it has completely disappeared. And a 48-year old with a small AVM here on the posterior aspect of the brain stem which again, we used Smartbrush Angio to outline.
So two years ago I gave this talk and really I was just concentrating on we avoid another angiogram, we avoid another angiogram. But is there more to it than that? And more recently I've been thinking, "Does it actually make AVM nidus delineation more accurate?" When we use the standard technique, we just have a static image, AP and lateral, and either your neurosurgeon or interventional neuroradiologist picks what they feel is the best image. You can't run through it. With this, you can. You can run through the whole AP, lateral, or even oblique views of the angiogram. And the DSA in diagnostic DSA is obtained with a cordial tilt which brings a lot of the normal vessels out of the view of the AVM, so you get a better view of the AVM. When you do your stereotactic DSA, it's a direct AP and lateral, so I think the view maybe is not quite as good. This is how we normally look at an AVM when we run through it, and we decide where does arteries and veins start and end? So it's much nicer to have the full view of the AVM.
Yep, I'm coming to an end. And here's a good example on the MR. Every time I outlined that more lateral bit, and then I looked at the DSA, there's nothing there. And the reason was when you actually go into the venous phase, there's lots of veins there which on the MR looked like part of the nidus but they're not. And the DSA is very helpful in outlining that. So, I think with Smartbrush Angio, you get increased safety and comfort for your patient, it reduces the cost of the overall treatment, you get better visualization, I think, of the AVM nidus for the reasons I said. It also gives you some flexibility in your workflow. You don't have to do everything in the same day because the Angio is your diagnostic angiogram, and you can separate the different steps of your care pathway and clinical outcomes so far. And I know our follow-up is short and our numbers are small, but so far is at least as good as frame-based systems. Thank you.
So you've heard some of this already, the risk of rupture according to ARUBA is about 2% per year. But if you have previous hemorrhage, or associated aneurysms, or deep venous drainage, that risk goes up. You've heard about different options. Medical therapy just means really conservative management and managing the symptoms. At our centers, I would say the vast majority of AVMs either get microsurgical resection or radiosurgery. And we really have moved away from nidal penetration of AVMs with embolizations, so very, very specific targeted goal for embolization.
Now, SRS for AVMs is very well established. It was one of the first indications for radiosurgery and has been around since the late '70s and '80s. And we have seen a transition from the frame-based techniques to the more frameless techniques. We use the Novalis at our centers. Now, the imaging modalities that we use for targeting an AVM, you've got on the left a 2D digital subtraction angiogram. So typically, you get your AP and your lateral images, and typically you would just identify the best image you can get on a still image on an AP and lateral digital subtraction angiogram.
Over the years, obviously, the 3D imaging has become available either using MRI or using CT. So CT angiography, or the rotational angiography using DSA, or MR angiography. Now, we have to somehow incorporate these together. Now, one of the first questions is, "Do we actually still need DSA to target an AVM? Can we not just use the MRA or the CTA, which shows us the AVM very nicely on a T2, you'd see the flow void on the MRA. Top MRA you would see the AVM. So do we still actually need the DSA?" So my view would be yes. DSA is certainly needed to target very small AVMs that you wouldn't see very clearly on an MRA or CTA. Some AVMs have a very diffused SNIS. It's quite difficult to know where to start and end, particularly on MR Imaging. If you've had previous embolization or previous hemorrhage, it may be quite difficult to target it on MRI alone or CTA alone. But I actually think all AVMs, do you do benefit in terms of targeting the nidus with a DSA? There are some examples here, you can see on your left a very small posterior thalamic AVM, which is quite difficult to see on the MRI. And it's better seen on the DSA. Here's a patient with a hemorrhage from AVM, and you can see the cavity where the hemorrhage occurred. And on the T2 weighted imaging, it's difficult to target that AVM whereas on the DSA, you can clearly see it here.
There are very few studies done on this, but there are two that I would mention. One was published in neurosurgery in 2008, and they looked at 34 AVMs with targeting using 3D imaging alone with MRA versus targeting of with MRA and DSA. And I think their major finding is that of the 34 cases, 19 of them target volume was larger if you didn't use DSA to help with the planning. And similarly in 2007 in the "Red Journal", there is another paper. And then again, AVMs delineated on MRA alone tended to be significantly larger and more subject to inter clinician variation. So that means your patient has come to you with an AVM, whether ruptured or unruptured, and almost all of those that we would see in our clinic, our next step if we are gonna treat is a diagnostic DSA. So we put our patient through one DSA simply for diagnosis and getting some more information about the angioarchitecture before our MDT meeting to see how we gonna excise it, or SRS it, or embolize it.
That then means that after our MDT meeting, if we decide to go for radiosurgery, we now need a planning stereotactic DSA, which in the frame-based techniques would need the frame and a fiduciary box and an angiogram and the frameless, the mask, the box, and another, so you're basically subjecting your patient to a second angiogram. Now, catheter tech... so you obtain your DSA with the fiduciary box. We use a Novalis so we import it into iPlan and then you match these fiduciary lines together to get rid of that distortion that you get from your 2D DSA, and then fuse it with your CT scan for planning and if an MRI, if you're gonna use that. In that standard way of doing it, what you then get is a static AP and lateral image from the DSA and an overlap of the target on that. You can't run through the DSA as you would normally. So that's your additional planning DSA. DSAs nowadays with catheter technology has improved a lot, so the risks are very low, but it's not zero. So you certainly have patient discomfort. Some of my patients where I take an AVM out through a craniotomy tell me that the angiogram afterwards was worse than having the craniotomy. So it does have some discomfort. It has a low but significant risk of stroke, arterial dissection, hematoma, sometimes femoral pseudoaneurysm, potential for contrast reaction, its additional radiation dose, and certainly cost. It does mean another day-case angiogram for your patient and a highly skilled team that you need to have to do that.
So I'm gonna talk about Smartbrush Angio, which is basically part of the elements package. And here, instead of using your fiducials to fuse the DSA with your MRA, you're basically using the vessel tree to do that. So there's no need for that additional stereotactic DSA because you're gonna use your original diagnostic DSA which you had and fuse that with your MRA. So here it is. You import your images there and then it moves on to showing you the DSA and then it overlaps the arterial tree from the MRA onto the DSA. You can move it around in any direction that you want to try and match those two as best as you can, and then click Fuse and the machine... the software essentially fuses your MRA with your 2D DSA. And on the next screen, you can essentially outline your region of interest under your AP and lateral, except that here, I can run through the whole DSA as I would normally do. And when we go forward, it will automatically generate the AVM nidus for you, which you can view on your MRA and your DSA side by side. And you can view your MRA in coronal, in sagittal, or axial planes, and you can have any of your MR sequences that you want, T2. And you can see here the accuracy if I'm running the cursor over the left internal carotid, and on the other side, I can see that it runs over the carotid artery on the angiogram, and similarly over the superior sagittal signs.
So we've got our AVM, and you can adjust your AVM nidus in using the software. And as you do that, it automatically shows you happening on the DSA. And here it is on the T2. The current version allows you to do the changes on MRI only, but on the next version, you can also make changes on the DSA and see your changes on the MRI. And the next step would be to export this out of elements. At the moment you export it into iPlan and fuse it with your stereotactic DSA that's with your stereotactic CT and then you do your dose planning. So when we got this software originally in, kind of, late 2013, as always we don't trust anything, so we had to do our own validation and our own physicist made a phantom with tubes inside the phantom. And we filled it with the appropriate contrast for CT for MRA, for DSA, and we imaged the phantom, and we compared the MRA DSA fusion with the CT scans. And I won't for a minute to pretend to understand this. This is done by Chief Physicist Paul Davenport, but essentially, the fusion accuracy was excellent. It was in part... we use an upper loop where the accuracy was less than 0.4 millimeter and the lower loop, less than 1.2 millimeters. And then we did three patients where we did the old-fashioned way with the fiducials and the new way without the fiducials using Smartbrush. We generated vessel trees, we outlined different parts of the vessel tree, and again, the mean accuracy was less than one millimeter.
So we then went ahead and treated our patients using Smartbrush. And we've treated almost 50 patients, 20 of whom have more than two years follow-up. And in those 50 patients so far, we have two radiation-induced edemas. Professor Yamamoto in 1980s reported the T2 signal changes that you see after AVM radiosurgery. So we've had two of those which were symptomatic with seizures both controlled with dexamethasone and some extra anticonvulsant medication, we've had one patient hemorrhage and die from that in-the-latency period, and we have the 19 patients with two or more year follow-up, of whom 12 have had complete obliteration and four have had significant reduction and I think they will continue to obliterate. I just want to emphasize that our doses are between 16 and 20 greys. We don't use the 24 and higher greys that are often reported in the literature.
So here's the 40-year old taxi driver, which in UK and Ireland, if I took his AVM out he can't drive for one year, and he has a medial-temporal AVM with lenticular strike supply. It's an unruptured AVM, but he and we decided to treat it despite ARUBA. And in 2018 you can see complete obliteration of the AVM. For the pediatric age group it's great because you can avoid the additional angiogram and the general anesthetic that goes with it. Here is the 13-year old with a ruptured AVM that you can see here with the intraventricular hemorrhage. And in 2018, it has completely disappeared. And a 48-year old with a small AVM here on the posterior aspect of the brain stem which again, we used Smartbrush Angio to outline.
So two years ago I gave this talk and really I was just concentrating on we avoid another angiogram, we avoid another angiogram. But is there more to it than that? And more recently I've been thinking, "Does it actually make AVM nidus delineation more accurate?" When we use the standard technique, we just have a static image, AP and lateral, and either your neurosurgeon or interventional neuroradiologist picks what they feel is the best image. You can't run through it. With this, you can. You can run through the whole AP, lateral, or even oblique views of the angiogram. And the DSA in diagnostic DSA is obtained with a cordial tilt which brings a lot of the normal vessels out of the view of the AVM, so you get a better view of the AVM. When you do your stereotactic DSA, it's a direct AP and lateral, so I think the view maybe is not quite as good. This is how we normally look at an AVM when we run through it, and we decide where does arteries and veins start and end? So it's much nicer to have the full view of the AVM.
Yep, I'm coming to an end. And here's a good example on the MR. Every time I outlined that more lateral bit, and then I looked at the DSA, there's nothing there. And the reason was when you actually go into the venous phase, there's lots of veins there which on the MR looked like part of the nidus but they're not. And the DSA is very helpful in outlining that. So, I think with Smartbrush Angio, you get increased safety and comfort for your patient, it reduces the cost of the overall treatment, you get better visualization, I think, of the AVM nidus for the reasons I said. It also gives you some flexibility in your workflow. You don't have to do everything in the same day because the Angio is your diagnostic angiogram, and you can separate the different steps of your care pathway and clinical outcomes so far. And I know our follow-up is short and our numbers are small, but so far is at least as good as frame-based systems. Thank you.