Transcript
Bogdan: Hello, everyone, and welcome to a new Novalis Circle Case of the Month. My name is Bogdan Valcu. I am the director of Novalis Circle. And today, I have the privilege to introduce you to our clinical team from Thomas Jefferson University in Philadelphia, Pennsylvania, who will review a neurovascular case presenting both an aneurysm and an AVM, the latter being treated with stereotactic radiosurgery. Presenting the case today, we have Dr. Pascal Jabbour, who is the Angela and Richard Clark Distinguished Professor in the Department of Neurological Surgery at Thomas Jefferson University. He is also the chief of the division of neurovascular and endovascular neurosurgery. We also have Dr. Wenyin Shi, who is the medical director in the Radiation Oncology Department at Thomas Jefferson. He is also the co-director of the Jefferson Brain Tumor Center and the radiation oncology director at the Jefferson Cutaneous Lymphoma Center at Thomas Jefferson University. To conclude the clinical team, we have Dr. Haisong Liu, who is the director of radiosurgery physics for the division of stereotactic radiosurgery at Thomas Jefferson University.
The case of the month review that you are about to see covers everything from clinical presentation to the practical aspects of generating a treatment plan. And we are privileged to offer continuing education credits upon successful completion of this course. Should you require CEs, please email us at info@novaliscircle.org and specify which credits you would like to receive. As always, please remember to utilize Google Chrome or Safari. Utilize the chat line to send us questions. We will answer those questions upon completion of the lectures. Check the polling sections for questions that we may ask you. And if you choose to follow us on social media, please use the hashtag #NovalisCircle. And now, I'd like to hand it over to Dr. Jabbour who will begin with the case presentation.
Dr. Jabbour: Hello, everyone. It's a great pleasure to be here today. We're going to be talking about the case of a frameless stereotactic radiosurgery for a brain AVM. Those are our disclosures. This is a case of a 31-year-old female who had a sudden onset of headaches and a left-sided facial droop. Past medical history, she has hypertension and a history of drug abuse, no past surgical history, and on exam, she has a left-sided peripheral 7th nerve palsy House-Brackmann IV. A CAT scan and CTA was done. So this is the CT head showing some calcification in the left lateral medullary area. And this is the CTA showing some abnormal vessels going deep all the way to the fourth ventricle.
So here it is. We're playing the movie. So as you see here, you see some flow voids on vessels in the lateral medullary area, and then you see some vessels going deep to the fourth ventricle, as you see the cluster of vessels here. And then you see maybe a possible dilation of a venous varix or aneurysm and then going deep to the ventricles. So when we saw that, we right away suspected arteriovenous malformation. So the next step is we did an angiogram. This is a coronal view of the angiogram showing exactly on the left lateral medullary area those abnormal vessels going deep into the fourth ventricles. Probably, those are veins draining deep. And a possible aneurysm here, there's a dilation. So we did an angiogram. This is an AP view, left vertebral artery injection. You see here, there is an early draining vein, and there is, you see, a nidus, so this is an AVM. On the lateral view, it's fed by the left AICA, and you can see a feeding vessel aneurysm. As you see here, the aneurysm, and then you see the nidus. And then it's draining into the superficial cerebellar veins. This is a lateral view showing the AVM, early draining veins. And this is a 3D angiogram showing the nidus and showing this aneurysm, so the feeding artery aneurysm here.
So as I said, this was a 1.8 by 1-centimeter size nidus of an AVM in the left lateral medullary area, Spetzler-Martin III, and fed by left AICA. Deep drainage into the fourth ventricle, this is what we saw on the CTA, and into the clival venous plexus, and some superficial drainage also into the cortical cerebellar veins. There is a feeding artery aneurysm, 5.9 by 4.6 millimeters.
So always, in AVMs, we try to look for high-risk features, and one of the high-risk features in this case is the presence of an aneurysm. So our priority was to treat the aneurysm first. Most likely, the patient was symptomatic from a mass effect or from this partially thrombosed aneurysm, either by an ischemia or just some mass effect and edema. So the aneurysm needs to be treated first. What are the options here? By looking at the angiogram, we have options for the AVM and options for the aneurysm.
So let's talk a little bit about the AVM. So microsurgery is always an option when it's feasible in a young patient. That's what we should do because that's the fastest way of curing the patient from the AVM and protecting the patient from a brain hemorrhage. But we need to evaluate the risk of surgery compared to the natural history of the AVM to decide if surgery is a good option. In this case, it is feasible, but it's not straightforward. This needs a far lateral approach and part of the nidus is deep and digging in the medulla. So this would be a higher risk for microsurgery. Embolization. Whenever we're talking about trying to cure an AVM, we should try to evaluate if our embolization will be able to possibly cure the AVM. But if we cannot just embolize just because we're able to inject some liquid embolic and that's it, I don't think we're doing any benefit or any good for the patient.
So in this case, embolization was feasible to try to treat the aneurysm because, as I said, the dangerous feature needs to be treated first. And then, for the AVM, stereotactic radiosurgery would be the best option, I think, with risk/benefit, would be good in this case. The AVM is not large. It's a compact nidus. It's not a ruptured AVM. In general, in ruptured AVMs, we'd rather not treat it with stereotactic radiosurgery because, as you know, it's going to take at least two years for the AVM to obliterate. During that time, the patient is not protected from any rebleed or re-hemorrhage. So what we decided to do here is...this is the super-selective injection showing the aneurysm. So the aneurysm was partially thrombosed. So this is a microcatheter that I drove all the way in the feeder to the aneurysm. And this says we used Onyx, which is a liquid embolic. You see the Onyx being injected here. So it's Onyx 18. I embolized the aneurysm with 0.3 cc of Onyx. You see the black Onyx here coming in. This is another view showing the injection, and you see here, that's the aneurysm, and that's the inflow here that you see. So all this is the aneurysm. So we blocked the inflow. And this is a run after the embolization. You see stasis of contrast and complete obliteration of the aneurysm, and we still see the nidus of the AVM. This is just a shot, a single shot, showing the cast of Onyx in the aneurysm. And we decided to do a frameless Gamma Knife on this patient. So we brought the patient back, did an angiogram, and this is something that...that's the 3D angiogram, and those are the native 3D images that we're going to use as part of the fusion. You're going to hear more in details about that from my colleagues here. That's the 3D that we did for the stereotactic radiosurgery, the frameless stereotactic radiosurgery. As you see here, you don't see the aneurysm anymore, but you see the nidus of the AVM.
Bogdan: Thank you for the clinical introduction, Dr. Jabbour. And we'll switch now to Haisong Liu who will cover the technical aspect of preplanning and integration for both Gamma Knife and Novalis radiosurgery solutions.
Dr. Liu: Good afternoon, everyone. My name is Haisong Liu. I'm a physicist. I'm going to continue Dr. Jabbour's talk about this case. And so let's look at the workflow using the Brainlab Elements for the AVM. So it's including selecting all the imaging data, and in this case, we have MRI, CTA, and the 3D angio, 2D angio. And also, all the image fusion will be done to all the 3D imaging datasets, including the CT, MRI, and the pseudo-CT from the 3D angio. And then we will do image registration between the 2D...the DSA pair, and the 3D imaging, including those CTA and MRA. And also after that, the AVM was defined by the neurosurgeon, Dr. Jabbour, the DSA pair, and also the CIP, which stands for color intensity projection, is a color-coded time information based on the DSA image. So after the contour, the AVM can be overlaid on any of the fused 3D images for editing and fine-tuning. And also, after the contouring, the treatment can be done on either a LINAC-based or a Gamma Knife-based frameless radiosurgery system.
So after a couple of cases done at Jefferson, we have a workflow written up. So the tasks that could be done prior to the procedure day, including the MRI, which is a contrast 3D T1 thin slice, and MRA, is a time of flight sequence, and the optional CTA. And then if we decide to use LINAC procedure, then we make BrainLAB mask and acquire planning CT, or if we select the Gamma Knife, then we use the Gamma Knife dedicated mask and the Gamma Knife planning cone-beam CT.
So on the procedure day, the patient will go into the angio switch, IR switch first, and we take AP and lateral 2-dimensional angio, the DSA. And each run, we use greater than three frame-per-second, the movie load. So each angio series contains about 20 or 30 images. And we also, after the first...the 2D view, we also acquire a 3D angio run. It's about 7 seconds, and the angio machine, the gantry will take about 7 seconds for about 200-degree and then acquire a cone-beam CT-like dataset. So after these procedures, all the image fusion will be done on all the 3D datasets and also the 2D angio pair with 3D angio CBCT. And we use a SmartBrush to segment AVM in the co-registered space. And the treatment planning in either...using Brainlab Cranial SRS Element if we use the LINAC SRS, or we can transfer all the structures to GammaPlan for a Gamma Knife radiosurgery planning.
So this is the image fusion tree for this case, and we can see, the top one is the cone-beam CT we acquired from the Gamma Knife Icon unit. And then, so the yellow one here is the pseudo-CT that we acquired from the 3D angio. And also, we have the prior CTA and MRI that both fused to the SET. So after the fusion, all the 3D datasets are kind of in the same coordinate space.
And this is the image fusion between the prior CTA and MRI, and both imaging are the pre-embolization procedure, about three weeks prior to our procedure date. An MRI is very helpful for segmentation of the critical organ at risk. In this case, it's the brainstem or medulla. So the direct fusion between the MRI and the pseudo-CT is also possible. Here is a movie that we can display to show you the image fusion between the pseudo-CT from the 3D angio and the MRI images. So it will show you the evaluation of axial, coronal, and sagittal view, each view, very carefully.
So the next, I'm going to show you a video, displaying the image fusion registration between the 2D angio pair and the 3D angio...the CBCT. So the first step is to use the window and level function to get optimal window and level for the 3D datasets. So the optimal, which means, you can see, the red is the vessels from the 3D angio, and now, as we're moving the arrow down here, we can select the best frame for the 2D angio so that they are kind of matched on the image. And then we are going to use the manual adjustment to make the images closer on each view, lateral view and the frontal view.
[00:17:29]
[silence]
[00:17:46]
Dr. Liu: So after the initial placement, we're going to use the auto fusion function. So the software will try to find the best match. So because of this 3D angio reconstructed CBCT and the 2D angio, they are done just in separation of a couple of minutes, and with the same [inaudible 00:18:15] position so the image fusion is much easier compared to the 2D angio pair fused with the prior CTA or MRA. And then we can use the blending function to evaluate the fusion so that, now, we're seeing, we can make the [inaudible 00:18:39] the transparency of the 3D, the red, to match with the...to compare with the black vessels. So after the registration, we are going to do the evaluation with other 3D imaging modalities. So now, you can see the 2D pair are fused with the pseudo-CT from 3D angio. However, if we edit the pair, we can select either CTA, or MRA, or even the cone-beam CT to make the evaluation, to look at...to review the fusion between the other 3D images to the 2D angio. So on the left side is the overlaid with prior MRI, and we can change the frame of the 2D angio to just view the proper sequence of the...to compare with the MRI. And also, on the right side is the overlaid with prior CTA images.
And we can also evaluate the registration by using a surrogate, such as the embolization material, it's the Onyx 18, as Dr. Jabbour was saying. So this one, we display the Gamma Knife cone-beam CT so that CT, the embolization material is here, and that's with a threshold Hounsfield units over 3,000. So using the software, we can select the contouring mode as threshold and select a proper volume of increase and set up 3,000 Hounsfield units. And then, when we apply, the high-density material is automatically detected. And then it's overlaid on top of the 2D angio. So when we review, it happens right on top of the 2D images.
So we can also evaluate the registration by using the surrogate, such as the major vessels. So when we select the 3D angio pseudo-CT and also set up the proper volume of interest, set up the proper threshold, the injection can be just automatically segmented. And as you can see here, this is the major vessel, and it's overlaid on top of the 2D angio perfectly.
And next, we're using the SmartBrush Angio Element for the AVM segmentation in the co-registered space. So we can segment either these DSA images or the CIP images. And here is a video that can show you the contouring in this co-registered workspace.
Once this is contoured and accepted, it will overlay on top of any 3D images and can also be edited in the 3D and it will also, real-time, reflected on the 2D angio. So this is the segmented AVM volume that overlaid on the pseudo-CT, as we see here, and also on the MRI. So the MRI is useful here for the contouring of the brainstem and medulla. And so next, I'm going to invite Dr. Shi to talk about the SRS plan and radiation considerations.
Bogdan: Thank you for the technical review, Haisong. And we'll now switch to Dr. Shi who will discuss the radiosurgery treatment planning for this AVM patient.
Dr. Shi: Hello. This is Wenyin Shi. I would like to discuss on this radiosurgery considerations first, then I'll review the radiation plan for this patient. So first of all, as we discussed earlier, radiosurgery for AVM is very well established. The first utilization started in the early '70s and pure on the Gamma Knife system. Subsequently, the sigma technique was established using LINAC-based equipment and it showed a similar result. Radiosurgery is ideal for small lesions and also when a surgical intervention may be difficult due to anatomic location or the vessel anatomy. And then, in general, AVM is a good target for radiosurgery because the AVM nidus is well defined and is usually compact with more abnormal vasculatures. The normal brain tissue is actually being displaced so it's very, very suitable for radiosurgery approach.
The main mechanism of action for radiosurgery is endothelial damage and chronic inflammation process, which ultimately result in smooth muscle proliferation, as well as deposits of hyalin, calcium, as well as collagen. And this ultimately leads to the thrombosis of the vasculature and the necrosis of AVM vasculatures. So this process takes some time, as, you know, we mentioned earlier. Radiosurgery is not immediately effective like surgery. There's a latency period. The latency period varies anywhere from one to three years. The median time is about 20 months. And even though the latency period can be several years, but based on a large retrospective series, during the latency period, the bleeding risk is reduced.
Radiosurgical options are mainly established based on the Gamma Knife platform. As mentioned, LINAC platform is very commonly used of the radiological platform that's equally suitable to deliver radiosurgery for AVM. And then the modern techniques are really aimed to improve the workflow and improve the patient experience and comfort. And some of the literature I report here, just showing that the platform really have a similar outcome. So either using Gamma Knife-based or LINAC-based really have very similar clinical outcome for the radiosurgery.
And there's many factors we need to take in consideration regarding the ultimate success of radiosurgery for AVM, most importantly probably is radiosurgery dose. This is the plot generated based on extensive clinical outcome information from the University of Pittsburgh. As you can see, there's a clear dose-response curve. And the peripheral dose less than 15 Gy is associated with a poor success rate, and the inflection point is around 16 Gy and probably a plateau around 18 to 25 Gy, with ultimate success rate with high dose around 98%. It's of note that if patient got a prior embolization of the AVM nidus, this actually had been reported as a negative factor. There's a couple of reason. One is due to recanalization of the embolized area related to some of the embolization material used, but this is reported as high as 10% or more. And also, when we have a rather large nidus, embolization may potentially create multiple segments that result in the radiosurgery target inaccuracy. And the embolization material itself may potentially make the radiosurgery target delineation really more challenging. So as you can see, as a result, most of the cases when we see patients with prior embolization, they end up with out-of-field failure, indicating that's probably due to recanalization or target delineation issues. There are obviously some AVM factor itself that can potentially impact the overall outcome, such as large nidus, location, draining pattern, as well as the nidus density. Many of them are related to radiation dose used because of the anatomic restriction.
So there's some formulas we can use to predict the success of obliteration without complication. This is known as the radiosurgery-based AVM grading scale, and this is also established from University of Pittsburgh. This data, also being subsequently validated using LINAC data. As you can see, with increase in AVM score, there's a pretty significant decrease in the overall obliteration rate.
AVM, as a benign disease, it often affects rather younger patient population. It's critically important for us to evaluate the robustness of the radiosurgery plan to make sure it's safe. So the long-term complication risk is really key in evaluating the radiosurgery plan. And obviously, the area of focus, we're looking at a normal parenchyma brain tissue and also the eloquent area of cranial nerves. And for brain parenchyma tissues, there are a lot of data reporting the risk of radionecrosis, and the parameter of v12 Gy volume has been shown to be very reliably associated with the risk of radionecrosis. In order to keep the radionecrosis rate in the reasonable, acceptable range, the ideal range for the V12 cc should be less than 5 to 10 cc. However, if the location is in the more eloquent area, the criteria need to be more stringent.
And besides the brain parenchyma, also, many of the organ of risk need to be carefully evaluated, such as brainstem, optic nerve, [inaudible 00:29:21] brainstem is involved. So there's more stringent criteria that need to be observed to avoid radionecrosis to other organ at risk. For example, for one fractions, the ideal dose for optic nerve should be 10 Gy or less, and the brainstem should be 15 Gy or less. As you may be aware, some of the literature started reporting that brainstem radiation tolerance could be higher than 15. But keep in mind, some of the data, most recent data, are based on brain metastasis patients, and those patients are affected for shorter survival. So none of the data generally from brain metastasis patients are truly not suitable to be applied to AVM patients. In general, for AVM patient, we need to use a higher level of scrutiny to evaluate cranial nerve or brain structure tolerance. And in order to achieve a more beneficial radiation dose, we can also look into fractionation schedule. When we're using three fraction or five fraction, we are able to raise the brainstem or optic nerve tolerance to make it more closer to our prescription goal.
So for this particular case, as we've already seen, this is in the medullar area, showed brainstem is really a dose-limiting organ. And as we mentioned earlier, if you prescribe a single fraction dose less than 15 Gy, then your obliteration success rate is actually pretty low. So based on that, we decided to prescribe 7 Gy times 3 to a total dose of 21 Gy. We're able to keep the maximum dose of the brainstem on 23 Gy, which is in the recommended tolerance range. And because we are adopting this new frameless system and really make this multi-fraction radiosurgery feasible, with a frame-based system, this is not quite practical to do fractionated radiosurgery. And if we're using a low dose single fraction, then the success rate obviously will be significantly compromised.
And as mentioned earlier, many platform we can use to perform radiosurgery, and they really have very similar clinical outcome. This is just a demonstration of different plan generated with different radiosurgery platform, including the LINAC system on the top and the Gamma Knife system on the bottom. And the LINAC system particularly using iMark TV map planning tend to give a better conformity index. And the Gamma Knife, due to clinical conservation of delivery time, often, the plan needs to be balanced with delivery time versus the conformity, so sometimes the conformity is somewhat inferior.
But nonetheless, if you look at the key parameters for the plan quality, look at the target coverage, as you can see on the left panel. All the plan get excellent plan coverage. And for Gamma Knife plan, due to the prescription isodose around 50% or so, the hot spot is a lot higher. For this particular LINAC plan, the hot spot is relatively low, and some of the radiosurgery plan for LINAC-based system tend to prescribe around 70% to 80% isodose line. So the center dose are somewhat different than the Gamma Knife, but peripheral dose are very much similar. And when we look at the brainstem deviation, you can see, they pretty much overlay each other, indicating that both system can safely protect OAR at risk and deliver a safe plan.
So I would hand this back to Dr. Jabbour. He's going to give some summary of this case. Thank you.
Dr. Jabbour: Thank you. So I hope we were able to, in this quick presentation, show a unique case of frameless stereotactic radiosurgery for an AVM by choosing the right patient. So in general, in AVMs, it's very important to always look for high-risk features, as I said earlier. Try to treat the associated aneurysm first, because this is going to be the higher risk of rupture. Evaluate eloquent versus non-eloquent brain to assess the microsurgical risk to be able to give to the patient to decide what's the best treatment. Evaluate the arterial feeders and the drainage pattern on angiogram to be able to maybe embolize the AVM but to keep in mind always what's the ultimate goal of your embolization. And if both options are not viable at the time, stereotactic radiosurgery would be a great option, keeping in mind the delayed fashion how an AVM is going to be cured. And I think at this point, if you have the choice between going with the frame or frameless, frameless is the way to go. If we are able to demonstrate that we are as accurate as we could be with the frame, at the time, it's much easier on the patient, and it's faster, and it's more user-friendly. And in those cases, we are able, as Dr. Shi said, we are able to do fractionation much easier.
I think it's a multidisciplinary neurovascular team that will be able to choose the best patients and to choose the best treatment to be able to offer the patient a good estimate of the risks and the expectations. Stereotactic radiosurgery remains highly effective in the treatment of small AVM with a favorable toxicity and side effect profile. And we believe that frameless technology for stereotactic radiosurgery will further improve the patient's experience and reduce side effects and complications.
So I would like to acknowledge the whole team. This couldn't have happened without all the team that we have, from the radiation oncology, from neurosurgery, and the Brainlab team. And thank you very much.
Bogdan: This concludes our AVM case of the month review, and we can now go to a live question and answer session.
The case of the month review that you are about to see covers everything from clinical presentation to the practical aspects of generating a treatment plan. And we are privileged to offer continuing education credits upon successful completion of this course. Should you require CEs, please email us at info@novaliscircle.org and specify which credits you would like to receive. As always, please remember to utilize Google Chrome or Safari. Utilize the chat line to send us questions. We will answer those questions upon completion of the lectures. Check the polling sections for questions that we may ask you. And if you choose to follow us on social media, please use the hashtag #NovalisCircle. And now, I'd like to hand it over to Dr. Jabbour who will begin with the case presentation.
Dr. Jabbour: Hello, everyone. It's a great pleasure to be here today. We're going to be talking about the case of a frameless stereotactic radiosurgery for a brain AVM. Those are our disclosures. This is a case of a 31-year-old female who had a sudden onset of headaches and a left-sided facial droop. Past medical history, she has hypertension and a history of drug abuse, no past surgical history, and on exam, she has a left-sided peripheral 7th nerve palsy House-Brackmann IV. A CAT scan and CTA was done. So this is the CT head showing some calcification in the left lateral medullary area. And this is the CTA showing some abnormal vessels going deep all the way to the fourth ventricle.
So here it is. We're playing the movie. So as you see here, you see some flow voids on vessels in the lateral medullary area, and then you see some vessels going deep to the fourth ventricle, as you see the cluster of vessels here. And then you see maybe a possible dilation of a venous varix or aneurysm and then going deep to the ventricles. So when we saw that, we right away suspected arteriovenous malformation. So the next step is we did an angiogram. This is a coronal view of the angiogram showing exactly on the left lateral medullary area those abnormal vessels going deep into the fourth ventricles. Probably, those are veins draining deep. And a possible aneurysm here, there's a dilation. So we did an angiogram. This is an AP view, left vertebral artery injection. You see here, there is an early draining vein, and there is, you see, a nidus, so this is an AVM. On the lateral view, it's fed by the left AICA, and you can see a feeding vessel aneurysm. As you see here, the aneurysm, and then you see the nidus. And then it's draining into the superficial cerebellar veins. This is a lateral view showing the AVM, early draining veins. And this is a 3D angiogram showing the nidus and showing this aneurysm, so the feeding artery aneurysm here.
So as I said, this was a 1.8 by 1-centimeter size nidus of an AVM in the left lateral medullary area, Spetzler-Martin III, and fed by left AICA. Deep drainage into the fourth ventricle, this is what we saw on the CTA, and into the clival venous plexus, and some superficial drainage also into the cortical cerebellar veins. There is a feeding artery aneurysm, 5.9 by 4.6 millimeters.
So always, in AVMs, we try to look for high-risk features, and one of the high-risk features in this case is the presence of an aneurysm. So our priority was to treat the aneurysm first. Most likely, the patient was symptomatic from a mass effect or from this partially thrombosed aneurysm, either by an ischemia or just some mass effect and edema. So the aneurysm needs to be treated first. What are the options here? By looking at the angiogram, we have options for the AVM and options for the aneurysm.
So let's talk a little bit about the AVM. So microsurgery is always an option when it's feasible in a young patient. That's what we should do because that's the fastest way of curing the patient from the AVM and protecting the patient from a brain hemorrhage. But we need to evaluate the risk of surgery compared to the natural history of the AVM to decide if surgery is a good option. In this case, it is feasible, but it's not straightforward. This needs a far lateral approach and part of the nidus is deep and digging in the medulla. So this would be a higher risk for microsurgery. Embolization. Whenever we're talking about trying to cure an AVM, we should try to evaluate if our embolization will be able to possibly cure the AVM. But if we cannot just embolize just because we're able to inject some liquid embolic and that's it, I don't think we're doing any benefit or any good for the patient.
So in this case, embolization was feasible to try to treat the aneurysm because, as I said, the dangerous feature needs to be treated first. And then, for the AVM, stereotactic radiosurgery would be the best option, I think, with risk/benefit, would be good in this case. The AVM is not large. It's a compact nidus. It's not a ruptured AVM. In general, in ruptured AVMs, we'd rather not treat it with stereotactic radiosurgery because, as you know, it's going to take at least two years for the AVM to obliterate. During that time, the patient is not protected from any rebleed or re-hemorrhage. So what we decided to do here is...this is the super-selective injection showing the aneurysm. So the aneurysm was partially thrombosed. So this is a microcatheter that I drove all the way in the feeder to the aneurysm. And this says we used Onyx, which is a liquid embolic. You see the Onyx being injected here. So it's Onyx 18. I embolized the aneurysm with 0.3 cc of Onyx. You see the black Onyx here coming in. This is another view showing the injection, and you see here, that's the aneurysm, and that's the inflow here that you see. So all this is the aneurysm. So we blocked the inflow. And this is a run after the embolization. You see stasis of contrast and complete obliteration of the aneurysm, and we still see the nidus of the AVM. This is just a shot, a single shot, showing the cast of Onyx in the aneurysm. And we decided to do a frameless Gamma Knife on this patient. So we brought the patient back, did an angiogram, and this is something that...that's the 3D angiogram, and those are the native 3D images that we're going to use as part of the fusion. You're going to hear more in details about that from my colleagues here. That's the 3D that we did for the stereotactic radiosurgery, the frameless stereotactic radiosurgery. As you see here, you don't see the aneurysm anymore, but you see the nidus of the AVM.
Bogdan: Thank you for the clinical introduction, Dr. Jabbour. And we'll switch now to Haisong Liu who will cover the technical aspect of preplanning and integration for both Gamma Knife and Novalis radiosurgery solutions.
Dr. Liu: Good afternoon, everyone. My name is Haisong Liu. I'm a physicist. I'm going to continue Dr. Jabbour's talk about this case. And so let's look at the workflow using the Brainlab Elements for the AVM. So it's including selecting all the imaging data, and in this case, we have MRI, CTA, and the 3D angio, 2D angio. And also, all the image fusion will be done to all the 3D imaging datasets, including the CT, MRI, and the pseudo-CT from the 3D angio. And then we will do image registration between the 2D...the DSA pair, and the 3D imaging, including those CTA and MRA. And also after that, the AVM was defined by the neurosurgeon, Dr. Jabbour, the DSA pair, and also the CIP, which stands for color intensity projection, is a color-coded time information based on the DSA image. So after the contour, the AVM can be overlaid on any of the fused 3D images for editing and fine-tuning. And also, after the contouring, the treatment can be done on either a LINAC-based or a Gamma Knife-based frameless radiosurgery system.
So after a couple of cases done at Jefferson, we have a workflow written up. So the tasks that could be done prior to the procedure day, including the MRI, which is a contrast 3D T1 thin slice, and MRA, is a time of flight sequence, and the optional CTA. And then if we decide to use LINAC procedure, then we make BrainLAB mask and acquire planning CT, or if we select the Gamma Knife, then we use the Gamma Knife dedicated mask and the Gamma Knife planning cone-beam CT.
So on the procedure day, the patient will go into the angio switch, IR switch first, and we take AP and lateral 2-dimensional angio, the DSA. And each run, we use greater than three frame-per-second, the movie load. So each angio series contains about 20 or 30 images. And we also, after the first...the 2D view, we also acquire a 3D angio run. It's about 7 seconds, and the angio machine, the gantry will take about 7 seconds for about 200-degree and then acquire a cone-beam CT-like dataset. So after these procedures, all the image fusion will be done on all the 3D datasets and also the 2D angio pair with 3D angio CBCT. And we use a SmartBrush to segment AVM in the co-registered space. And the treatment planning in either...using Brainlab Cranial SRS Element if we use the LINAC SRS, or we can transfer all the structures to GammaPlan for a Gamma Knife radiosurgery planning.
So this is the image fusion tree for this case, and we can see, the top one is the cone-beam CT we acquired from the Gamma Knife Icon unit. And then, so the yellow one here is the pseudo-CT that we acquired from the 3D angio. And also, we have the prior CTA and MRI that both fused to the SET. So after the fusion, all the 3D datasets are kind of in the same coordinate space.
And this is the image fusion between the prior CTA and MRI, and both imaging are the pre-embolization procedure, about three weeks prior to our procedure date. An MRI is very helpful for segmentation of the critical organ at risk. In this case, it's the brainstem or medulla. So the direct fusion between the MRI and the pseudo-CT is also possible. Here is a movie that we can display to show you the image fusion between the pseudo-CT from the 3D angio and the MRI images. So it will show you the evaluation of axial, coronal, and sagittal view, each view, very carefully.
So the next, I'm going to show you a video, displaying the image fusion registration between the 2D angio pair and the 3D angio...the CBCT. So the first step is to use the window and level function to get optimal window and level for the 3D datasets. So the optimal, which means, you can see, the red is the vessels from the 3D angio, and now, as we're moving the arrow down here, we can select the best frame for the 2D angio so that they are kind of matched on the image. And then we are going to use the manual adjustment to make the images closer on each view, lateral view and the frontal view.
[00:17:29]
[silence]
[00:17:46]
Dr. Liu: So after the initial placement, we're going to use the auto fusion function. So the software will try to find the best match. So because of this 3D angio reconstructed CBCT and the 2D angio, they are done just in separation of a couple of minutes, and with the same [inaudible 00:18:15] position so the image fusion is much easier compared to the 2D angio pair fused with the prior CTA or MRA. And then we can use the blending function to evaluate the fusion so that, now, we're seeing, we can make the [inaudible 00:18:39] the transparency of the 3D, the red, to match with the...to compare with the black vessels. So after the registration, we are going to do the evaluation with other 3D imaging modalities. So now, you can see the 2D pair are fused with the pseudo-CT from 3D angio. However, if we edit the pair, we can select either CTA, or MRA, or even the cone-beam CT to make the evaluation, to look at...to review the fusion between the other 3D images to the 2D angio. So on the left side is the overlaid with prior MRI, and we can change the frame of the 2D angio to just view the proper sequence of the...to compare with the MRI. And also, on the right side is the overlaid with prior CTA images.
And we can also evaluate the registration by using a surrogate, such as the embolization material, it's the Onyx 18, as Dr. Jabbour was saying. So this one, we display the Gamma Knife cone-beam CT so that CT, the embolization material is here, and that's with a threshold Hounsfield units over 3,000. So using the software, we can select the contouring mode as threshold and select a proper volume of increase and set up 3,000 Hounsfield units. And then, when we apply, the high-density material is automatically detected. And then it's overlaid on top of the 2D angio. So when we review, it happens right on top of the 2D images.
So we can also evaluate the registration by using the surrogate, such as the major vessels. So when we select the 3D angio pseudo-CT and also set up the proper volume of interest, set up the proper threshold, the injection can be just automatically segmented. And as you can see here, this is the major vessel, and it's overlaid on top of the 2D angio perfectly.
And next, we're using the SmartBrush Angio Element for the AVM segmentation in the co-registered space. So we can segment either these DSA images or the CIP images. And here is a video that can show you the contouring in this co-registered workspace.
Once this is contoured and accepted, it will overlay on top of any 3D images and can also be edited in the 3D and it will also, real-time, reflected on the 2D angio. So this is the segmented AVM volume that overlaid on the pseudo-CT, as we see here, and also on the MRI. So the MRI is useful here for the contouring of the brainstem and medulla. And so next, I'm going to invite Dr. Shi to talk about the SRS plan and radiation considerations.
Bogdan: Thank you for the technical review, Haisong. And we'll now switch to Dr. Shi who will discuss the radiosurgery treatment planning for this AVM patient.
Dr. Shi: Hello. This is Wenyin Shi. I would like to discuss on this radiosurgery considerations first, then I'll review the radiation plan for this patient. So first of all, as we discussed earlier, radiosurgery for AVM is very well established. The first utilization started in the early '70s and pure on the Gamma Knife system. Subsequently, the sigma technique was established using LINAC-based equipment and it showed a similar result. Radiosurgery is ideal for small lesions and also when a surgical intervention may be difficult due to anatomic location or the vessel anatomy. And then, in general, AVM is a good target for radiosurgery because the AVM nidus is well defined and is usually compact with more abnormal vasculatures. The normal brain tissue is actually being displaced so it's very, very suitable for radiosurgery approach.
The main mechanism of action for radiosurgery is endothelial damage and chronic inflammation process, which ultimately result in smooth muscle proliferation, as well as deposits of hyalin, calcium, as well as collagen. And this ultimately leads to the thrombosis of the vasculature and the necrosis of AVM vasculatures. So this process takes some time, as, you know, we mentioned earlier. Radiosurgery is not immediately effective like surgery. There's a latency period. The latency period varies anywhere from one to three years. The median time is about 20 months. And even though the latency period can be several years, but based on a large retrospective series, during the latency period, the bleeding risk is reduced.
Radiosurgical options are mainly established based on the Gamma Knife platform. As mentioned, LINAC platform is very commonly used of the radiological platform that's equally suitable to deliver radiosurgery for AVM. And then the modern techniques are really aimed to improve the workflow and improve the patient experience and comfort. And some of the literature I report here, just showing that the platform really have a similar outcome. So either using Gamma Knife-based or LINAC-based really have very similar clinical outcome for the radiosurgery.
And there's many factors we need to take in consideration regarding the ultimate success of radiosurgery for AVM, most importantly probably is radiosurgery dose. This is the plot generated based on extensive clinical outcome information from the University of Pittsburgh. As you can see, there's a clear dose-response curve. And the peripheral dose less than 15 Gy is associated with a poor success rate, and the inflection point is around 16 Gy and probably a plateau around 18 to 25 Gy, with ultimate success rate with high dose around 98%. It's of note that if patient got a prior embolization of the AVM nidus, this actually had been reported as a negative factor. There's a couple of reason. One is due to recanalization of the embolized area related to some of the embolization material used, but this is reported as high as 10% or more. And also, when we have a rather large nidus, embolization may potentially create multiple segments that result in the radiosurgery target inaccuracy. And the embolization material itself may potentially make the radiosurgery target delineation really more challenging. So as you can see, as a result, most of the cases when we see patients with prior embolization, they end up with out-of-field failure, indicating that's probably due to recanalization or target delineation issues. There are obviously some AVM factor itself that can potentially impact the overall outcome, such as large nidus, location, draining pattern, as well as the nidus density. Many of them are related to radiation dose used because of the anatomic restriction.
So there's some formulas we can use to predict the success of obliteration without complication. This is known as the radiosurgery-based AVM grading scale, and this is also established from University of Pittsburgh. This data, also being subsequently validated using LINAC data. As you can see, with increase in AVM score, there's a pretty significant decrease in the overall obliteration rate.
AVM, as a benign disease, it often affects rather younger patient population. It's critically important for us to evaluate the robustness of the radiosurgery plan to make sure it's safe. So the long-term complication risk is really key in evaluating the radiosurgery plan. And obviously, the area of focus, we're looking at a normal parenchyma brain tissue and also the eloquent area of cranial nerves. And for brain parenchyma tissues, there are a lot of data reporting the risk of radionecrosis, and the parameter of v12 Gy volume has been shown to be very reliably associated with the risk of radionecrosis. In order to keep the radionecrosis rate in the reasonable, acceptable range, the ideal range for the V12 cc should be less than 5 to 10 cc. However, if the location is in the more eloquent area, the criteria need to be more stringent.
And besides the brain parenchyma, also, many of the organ of risk need to be carefully evaluated, such as brainstem, optic nerve, [inaudible 00:29:21] brainstem is involved. So there's more stringent criteria that need to be observed to avoid radionecrosis to other organ at risk. For example, for one fractions, the ideal dose for optic nerve should be 10 Gy or less, and the brainstem should be 15 Gy or less. As you may be aware, some of the literature started reporting that brainstem radiation tolerance could be higher than 15. But keep in mind, some of the data, most recent data, are based on brain metastasis patients, and those patients are affected for shorter survival. So none of the data generally from brain metastasis patients are truly not suitable to be applied to AVM patients. In general, for AVM patient, we need to use a higher level of scrutiny to evaluate cranial nerve or brain structure tolerance. And in order to achieve a more beneficial radiation dose, we can also look into fractionation schedule. When we're using three fraction or five fraction, we are able to raise the brainstem or optic nerve tolerance to make it more closer to our prescription goal.
So for this particular case, as we've already seen, this is in the medullar area, showed brainstem is really a dose-limiting organ. And as we mentioned earlier, if you prescribe a single fraction dose less than 15 Gy, then your obliteration success rate is actually pretty low. So based on that, we decided to prescribe 7 Gy times 3 to a total dose of 21 Gy. We're able to keep the maximum dose of the brainstem on 23 Gy, which is in the recommended tolerance range. And because we are adopting this new frameless system and really make this multi-fraction radiosurgery feasible, with a frame-based system, this is not quite practical to do fractionated radiosurgery. And if we're using a low dose single fraction, then the success rate obviously will be significantly compromised.
And as mentioned earlier, many platform we can use to perform radiosurgery, and they really have very similar clinical outcome. This is just a demonstration of different plan generated with different radiosurgery platform, including the LINAC system on the top and the Gamma Knife system on the bottom. And the LINAC system particularly using iMark TV map planning tend to give a better conformity index. And the Gamma Knife, due to clinical conservation of delivery time, often, the plan needs to be balanced with delivery time versus the conformity, so sometimes the conformity is somewhat inferior.
But nonetheless, if you look at the key parameters for the plan quality, look at the target coverage, as you can see on the left panel. All the plan get excellent plan coverage. And for Gamma Knife plan, due to the prescription isodose around 50% or so, the hot spot is a lot higher. For this particular LINAC plan, the hot spot is relatively low, and some of the radiosurgery plan for LINAC-based system tend to prescribe around 70% to 80% isodose line. So the center dose are somewhat different than the Gamma Knife, but peripheral dose are very much similar. And when we look at the brainstem deviation, you can see, they pretty much overlay each other, indicating that both system can safely protect OAR at risk and deliver a safe plan.
So I would hand this back to Dr. Jabbour. He's going to give some summary of this case. Thank you.
Dr. Jabbour: Thank you. So I hope we were able to, in this quick presentation, show a unique case of frameless stereotactic radiosurgery for an AVM by choosing the right patient. So in general, in AVMs, it's very important to always look for high-risk features, as I said earlier. Try to treat the associated aneurysm first, because this is going to be the higher risk of rupture. Evaluate eloquent versus non-eloquent brain to assess the microsurgical risk to be able to give to the patient to decide what's the best treatment. Evaluate the arterial feeders and the drainage pattern on angiogram to be able to maybe embolize the AVM but to keep in mind always what's the ultimate goal of your embolization. And if both options are not viable at the time, stereotactic radiosurgery would be a great option, keeping in mind the delayed fashion how an AVM is going to be cured. And I think at this point, if you have the choice between going with the frame or frameless, frameless is the way to go. If we are able to demonstrate that we are as accurate as we could be with the frame, at the time, it's much easier on the patient, and it's faster, and it's more user-friendly. And in those cases, we are able, as Dr. Shi said, we are able to do fractionation much easier.
I think it's a multidisciplinary neurovascular team that will be able to choose the best patients and to choose the best treatment to be able to offer the patient a good estimate of the risks and the expectations. Stereotactic radiosurgery remains highly effective in the treatment of small AVM with a favorable toxicity and side effect profile. And we believe that frameless technology for stereotactic radiosurgery will further improve the patient's experience and reduce side effects and complications.
So I would like to acknowledge the whole team. This couldn't have happened without all the team that we have, from the radiation oncology, from neurosurgery, and the Brainlab team. And thank you very much.
Bogdan: This concludes our AVM case of the month review, and we can now go to a live question and answer session.