Bogdan: Hello, everyone and welcome to a new Novalis Circle webinar focusing on modern radio surgical practices in the treatment of intracranial metastases. My name is Bogdan Valcu. I'm the director of the Novalis Circle and today I have the pleasure to welcome five speakers who will detail a broad spectrum of topics pertaining to intracranial radiosurgery. To begin, we have Dr. Rupesh Kotecha who is an associate professor and the chief of radiosurgery at Miami Cancer Institute. Dr. Kotecha will review the current state of clinical evidence and highlight the value of radiosurgery for patients with more than four tumors and set the general guidelines for what a modern radiosurgical practice requires in order to have an efficient solution for patients who develop advanced disease proliferation.
To follow his talk, we have Dr. Zachary Seymour is an assistant professor at William Beaumont and Dr. Seymour will highlight the necessary physician steps in order to have an efficient radiosurgical program. Time to therapy from diagnostic MRI as well as implications of technology such as distortion correction will be reviewed in Dr. Seymour's presentation. To review those symmetry considerations, we have Dr. Giuseppe Minniti, who is the chair of radiation oncology at the University of Siena, as well as an associate professor at UPMC Hillman Cancer Center Hospital, San Pietro in Rome. Dr. Minniti will address the dose symmetric implications for technique selection, as well as review the role of PTV margins for both adequate tumor coverage as well as toxicity implications in the normal brain.
For a holistic review on how modern technologies are impacting radiation oncology practices, we have a lecture from Dr. Nzdhe Agazaryan, who is a professor of radiation oncology and the chief of Clinical Medical Physics and Asymmetry at UCLA. Dr. Agazaryan will review both physician and physicist supervision changes when employing a single lysis into multiple targets technology, summarize time and financial implications, and articulate the value of a dedicated IGRT solution as a complement to a smaller margins direction in treatment planning. To complete the lectures, we have Dr. Dheerendra Prasad, who is both a professor of radiation medicine and neurosurgery at Roswell Park in Buffalo, New York. Dr. Prasad will discuss the benefits of dedicated response assessment technologies as modern tools to identify earlier when patients are failing therapy and may require a new course of treatment.
As always, we are pleased to offer CE credits if you are in need of CAMPEP, MDCB, or ARST credits. Upon completion of this webinar, please email us at firstname.lastname@example.org to obtain further details. And as we have all become accustomed by now, please utilize Google Chrome or Safari to access the webinar. Utilize the chat line to send us questions, we will answer those questions upon completion of the five lectures, review the polling section for questions that we might like to ask you, and should you follow us on social media please utilize the #NovalisCircle. With that, I'd like to turn it over to Dr. Kotecha for the first lecture.
Dr. Kotecha: Well, thank you to the Brainlab team for inviting me here to speak today. I'm gonna discuss the modern role of stereotactic radiosurgery for patients with brain metastasis. These are my disclosures. So when a patient is diagnosed with brain metastasis, there are a number of management considerations that are taken into consideration or we're thinking about treatment options, including the patient performance status, the number of lesions, the size, the location. Does the patient have neurologic deficits, age, what is the primary tumor extent disease status and stage? Is there extracranial disease and has it progressed in their current line of therapy? But increasingly, patient input is also driving these treatment decisions.
Now, there are similar variety of treatment strategies that can be used in the modern era for patients who have brain metastases. For example, patients who have a poor performance status, extensive systemic disease, they may just be managed with medical therapy with corticosteroids alone. Combination approaches have been used in a variety of clinical trials. But increasingly, the role of stereotactic radiosurgery as a primary treatment option, really without whole-brain radiation therapy in the modern era or systemic therapy alone are becoming increasingly recommended.
Now traditionally, whole-brain radiation therapy was used for anyone who had brain metastases. And that would be even in patients who have favorable features such as excellent performance status, stable extracranial disease or few intracranial lesions are those who had unfavorable features. However, more recent evidence has shown us that there are significant neurocognitive consequences with the use of this approach. And therefore, we've transitioned to use of stereotactic radiosurgery for patients who have these more favorable features.
Now, there are a variety of advantages to the use of primary stereotactic radiosurgery for patients who have brain metastases. And here I basically summarize a number of retrospective and clinical trials and prospective cohort studies that have been performed in this patient population. But essentially, SRS has been demonstrated to be an effective local therapy. And it's even for patients who have radio-resistant disease. There are multiple series for example, in patients who have melanoma, or renal cell cancer brain metastases showing excellent local control rates. It can also be used to target surgically inaccessible areas of the brain, is associated with a shorter recovery time than resection of brain metastasis, which has been shown in a number of recent series.
It can be used for patients who have limited lesions traditionally considered less than three or four, or now multiple brain metastases more than four, which will be the focus of this talk later. It can be performed in between treatment cycles for patients who are receiving standard cytotoxic chemotherapy regimens. Or often if a patient is diagnosed initially with brain metastasis, we've actually treated them with radiosurgery prior to their initiation of treatment, or even before they have a port placed. It's an effective and safe treatment for patients who have metachronous brain metastases, something that will actually spend time on today. And it also maintains quality of life and leads to less neurocognitive deterioration compared to whole-brain radiation therapy. And finally, in the modern era, as patients are receiving targeted therapies, or immunotherapies, with CNS penetration, there actually may be a synergistic benefit to the use of radiosurgery over approaches like whole-brain radiation therapy.
Now, a question has always been how many metastases can we safely treat with stereotactic radiosurgery? And the technology has significantly developed over this period of time. From a dose of metric or a physics standpoint, if you would consider a whole-brain radiation therapy equivalent dose, you could treat 177 lesions with stereotactic radiosurgery that are less than four millimeters in size, or 40 lesions of mixed sizes. Now, this is really not the upper extent, we don't recommend treating that many lesions in a patient. But just to tell you that this is physically possible. So we really have to develop the evidence from a clinical standpoint to support this.
Well, the NCCN guidelines do support the treatment of primary stereotactic radiosurgery for patients who have multiple brain metastases, as you can see here. In fact, historically, if you go back to when this change happened in 2014, the panel actually recently added radiosurgery alone as a primary treatment option for multiple metastatic lesions. And by 2016, basically made that change confirmatory by saying it can be used as a primary therapy for those who have multiple metastatic lesions. And this has been used consistently since. So essentially five years of NCCN backing this.
Well, where's the data to support this? Well, this is an initial series that we published previously at Cleveland Clinic out of 64 patients who were treated to at least 5 lesions. And although this is a pilot study developed in an early treatment era, it essentially showed that radiosurgery was an effective option for patients with multiple brain metastases with really a stratification between those who had more than eight lesions versus eight or fewer lesions. So historically, this then helped to dictate our practice. Then the group in Japan actually published their prospective series for patients with up to 10 anesthetic lesions. And as you can see here, those who had 2 to 4 or 5 to 10 tumors essentially have similar overall survival, therefore promoting the use of radiosurgery alone in this patient population, really regardless of number of lesions. We had done a similar analysis I presented at ASCO in 2017 in which we looked at our large series. Now, this is over 6000 patients treated to a number of lesions. And as you can see here, the survival curves look very similar from the Japanese analysis, but more so if you stratify by lesion size, which was done in their prospective cohort study, you see that the curves become more similar.
Now very recently, we have randomized data that supports the use of primary stereotactic radiosurgery in patients with more than 4 and now up to 15 previously untreated brain metastasis. This was a randomized control trial at a 1-to-1 ratio, 50 patients in each initial arm. And that was radiosurgery alone using the RTOG 9005 dosing schedule. So all these patients were treated with single fraction versus whole-brain radiation therapy. This study was designed prior to the use of hippocampal avoidance, that's why I did not complete its accrual. But essentially with this study, they showed that there was a decrease neurocognitive decline seen in patients who received whole-brain radiation therapy both at four months. And then also if you look at the one month and at the six months, and that was worse than patients who received whole-brain radiation therapy compared to stereotactic radiosurgery.
Now, there have also been institutional series about using repeat courses of radiosurgery and deferring whole-brain radiation therapy. This is a 95 patient series in which patients were treated to a median of 2 courses of radiosurgery ranging from as few as 2 up to 14 in a single patient to a total of 652 brain metastases. This is a very heavily treated patient population. As you can see the outcomes here, the risk of distant brain failure at 2 years was 54%. But notably, the risk of adverse radiation effect was only 8%. So again, this was considered a safe treatment.
The number of brain metastases may not matter in certain populations of patients, and therefore, you may continue to do radiosurgery in patients who have specific molecular alterations such as ALK rearrangement or EGFR mutation. This is a series that we put together from Cleveland Clinic, we initially looked at 1000 patients who are treated with non-small cell lung cancer brain metastasis over almost a 15-year period, pulled out the 322 patients with mutational analysis. As you can see here, as you increase lesion number, you actually don't see a decrease in survival. In fact, it actually trends the opposite direction. On the other hand, for non-small cell lung cancer wild type brain metastases, you clearly see a difference in survival for those that have more than one lesion compared to those that only have one lesion. So, therefore, the number of brain metastases may actually not have any impact in these specific molecular cohorts.
Now, finally, for the use of salvage radiosurgery in patients who develop metachronous brain metastasis, this is a series that we put together of 59 patients treated to a median of 3 courses of radiosurgery, ranging from 3 to 8 to a total of 765 brain metastases. Again, a very heavily pretreated patient population. We actually obtained prospective quality of life measures in this patient population is your patient-derived, and as you can see here, over time, actually their patient quality of life remains very stable. And that's even picking out the patients who only had three to four courses or patients who had more than four courses of radiosurgery, as you can see here with the blue and the red line. So overall, in this series, we also saw a radiation process risk of 17%. Again, higher given that very heavily treated patient population. But otherwise, these patients received multiple courses of radiosurgery, modest treatment, late toxicity, and they appeared to have a similar performance status, and we were able to preserve their quality of life.
Now, the field of radiosurgery is rapidly evolving and changing. As you can see here, there are multiple areas where we can continue to update our practice principles and potentially integrate the use of AI in this space. There are important principles to understand, for example, the time interval between treatment planning MRIs to radiosurgery. And we've actually recently published this analysis looking at tumor size, volume position, and all of these dynamics between when a patient is planned for radiosurgery and then ultimately treated. And finally, we do need to integrate neurocognitive biomarkers into our modern management of brain metastases, especially with patients who have a number of lesions, as you can see here, and this coordinate grid of a patient treated to approximately 30 lesions. We have an ongoing study to prospectively collect electronic neurocognitive outcomes in these patients.
So in conclusion, there are several patient-related and disease-related considerations that factor into the multi-disciplinary management of patients with brain metastasis. And now we need to know the molecular profile, not just the number of lesions. There's increasing indications for primary stereotactic radiosurgery and several advantages compared to whole-brain radiation therapy. There are evolving retrospective evidence and clinical trials that demonstrate favorable outcomes when you're treating patients to more than four lesions. And also salvage radiosurgery can be considered for those patients who experience a distant intracranial failure with modest treatment-related toxicity. Thank you.
Bogdan: Thank you, Dr. Kotecha, for that excellent review, and let's move on now to Dr. Seymour for the subsequent talk.
Dr. Seymour: Hi, I'm Dr. Zachary Seymour. I'm here to discuss modern radio surgical practices in the treatment of intracranial metastases. As we know more about how fractionation works in terms of improving both tumor control rates and also reducing the rate of radionecrosis in tumors that are at least 2 centimeters or 4cc in size, in order to optimize treatment for these patients with moderate to large tumors, we are going to have fractionate. And in order to do that you need two things, you need a frameless system, then you need an established workflow to monitor and maximize the efficacy of that system.
We've looked at the workflow and why it matters, in terms of minimizing delays from imaging to treatment delivery. So as we went back and looked at our CyberKnife experience when I was at UCSF, we found that treatment delays from MRI to delivery of more than 14 days were associated with substantially reduced level control. Looked like there was actually worse control even earlier than 14 days. However, on multivariable analysis, that was the threshold at which we found substantially reduced local control. Also, there was some reduced survival, however, that's likely more of a concomitant factor. But when we went back and looked at things, we found that there was a roughly for tumors that were moderate size, that you would probably are estimated based on buying doubling time lung tumors, we would need a 2-millimeter margin.
We then went and looked at our Gamma Knife experience for tumors treated with an MRI within a month of their actual Gamma Knife MRI. And what we found is that the pre-SRS scan, basically for patient selection, to the time at which they actually had their Gamma Knife treatments stand when their head frame or head was placed in a frame, we found is that treatment, that the tumor volume at time of treatment was actually 1.35 times the actual size if there was a 14-day delay, and this roughly equated to a 0.2-millimeter growth per day of those tumors. And what we found is that there were specific risks via features that increased the likelihood of growth during this time, one was the certain subtypes of tumors. In lung cancer, we've seen sarcoma mass grow much faster, and brain mets, sarcoma doesn't go to the brain that much, we saw that melanoma grew much faster. In order to actually fully encompass the tumors, we would have needed increasing margin based on the actual buying size of the tumors in order to encompass your target. For instance, if you were going to do a pre-plan, as we now can, what we found is that once tumors were at least 2cc in size, you needed a margin of almost 4 millimeters in order to truly encompass that tumor at a time of treatment. So that is far too much margin than you would ever want to add.
We also found that patients that were not receiving systemic therapy at the time, we no longer really stopped that, we just interdigitate our radiosurgery between cycles, if they're getting cycles, or we just have them continue their systemic therapy in general if it's just oral therapy, but if they were not having systemic therapy, they had much higher rates of growth. It's always nice to get corroboration from other institutions. Another institution evaluated their pre-planning where they would get an MRI, they would generate a pre-plan, here they generated in 34 patients and 59 targets. They found that if the MRI they're using their pre-plan for was more than 7 days old when they got an MRI, at the time of treatment, they found that they had to change their plans 78% of the time. If there was at least 7 days, if it was less than or equal to 7 days, they still had to change it 41% of the time. So as we start to get down to the nitty-gritty, eventually, it reaches a point where you start to go, how much of this is real volumetric change over growth within that first week and how much of that is potentially just changes due to distortion between one magnet on one day and one magnet another day in the same patient?
So when we first established our clinical workflow for our ExacTrac system that we installed, we set slots for both M stimulation MRI and first treatment delivery for all the patients. So we did have three slots every week that we could rapidly simulate image and deliver those treatments. So in order to get that done, what we had is as soon as a physician were to select a patient, then it would go in the work QCL for both the administrative people to get the MRI and CT scan set up with the physicists and therapists to evaluate the scans, block the appropriate amount of time on the linear accelerator, and also for the biller because the last thing you want is not having authorization as soon as you want to deliver it.
For our simulation, we've always had some people ask me so I just include this slide, but effectively we use a 1-millimeter slice. The MRI is always done within one week of treatment delivery. We prefer our Brainlab mask, there are some limitations due to habitus, however, there are more limitations if you use the different masks for setup, which like a head stepper because that limits some of your non-coplanar angles. As for planning workflow and actually estimating the amount of time you would need on your linear accelerator, we developed an algorithm to maximize so that we would know exactly how much time we would need on our linear accelerator on a given day for a given patient.
And part of our algorithm was based off of tumor size. If the target size including PTB is greater than or equal to 8 millimeters, then we would often treat with a cone. We have a Bursa HD with a 5-millimeter leaf size. At 1.5 times the leaf size is where you're gonna see increasing gradients. That doesn't necessarily mean that the plan is not quality, it just means that why can't we be better and that's why we brought in cones. But effectively, then after you have that your target sizes, then you have the number of targets, whether it's near our organ at risk, and the distance from the isocenter in order to have an outfall plan and whether or not you would then generate one plan with single isocenter multi-target or whether or not you would peel off a tumor for a second plan.
Also, within our treatment planning, we automated both boarding passes and templates within our treatment planning software within the Elements system. So that as soon as we wanna generate a plan for one, three, or five fractions, everything is there. So all the organs at risk that we would ever want were automatically generated. In addition to that, their thresholds were already in our planning system,so we're always optimizing for appropriately clinically relevant dose levels to optimize a plan.
In addition to that, once you have a rapidly acquired MRI that is good quality for stereotactic delivery, then the next question is how much are you going to do to correct for that MRI and make sure that your treatment is truly on target? We have looked at whether or not the value of additional distortion correction. On every MRI, there's some degree of distortion correction at this point, but we went back and we looked at it as soon as we were able to acquire those metrics, the dose distortion correction algorithm through the element that Brainlab offers. When we went back and looked at the 42 cranial targets that we treated before we had this element to see whether or not it would have really added value in terms of our delivery, what we found is that there were substantial worsening in terms of both clinically relevant conformity, as well as those symmetric factors in terms of just coverage over target, whether and even down to roughly the 95% isodose to the GTV, even when we used 1 to 1.5 millimeters of margin.
In addition to that, what we found was that when I applied the distortion correction algorithm, as long as we used the appropriate histogram-based calculation to optimize [inaudible 00:23:05.541] level, it actually looked like the image was slightly sharper and that our GTV size actually slightly reduced. So it wasn't as if we were just creating a fuzzy image and then drawing this fuzzy image. In addition, all these contours are drawn blindly. So I had no idea whether or not what I was drawing then corrected was truly going to overlap more or less with the original target than what I had drawn. Again, here you can see substantially reduced minimum doses to the GTV, reduced relative coverage of our PTV, and worsening conformity index. We're currently validating this between CT-based software and our MRI and without this distortion correction. In contrast, enhanced, both CT and MRI.
Once you're moving on to actual treatment planning, we evaluated a number of plans because we have Gamma Knife, we also have our ExacTrac system. So we wanted to see are there really patients who are optimized better on one planning system or over another. We found is that, by and large, there really isn't a huge difference between these systems. Even for patients, for instance, and this one, which probably may explain that and if anything looks a little less conformal, but what we actually find is that improved conformity index is when we use the multiple metastasis element regardless of tumor size.
However, gradients, once you went down to that super small-size tumors were slightly worse. However, when you actually look at the whole brain V10, there really wasn't a substantial difference, particularly when we're averaging out over five tumors, very little difference. However, the amount of treatment time was substantially higher with Gamma Knife. And again, this is largely limited probably to the gradient index of the fact that we have a leaf size of 5 millimeters. If you didn't have a leaf size, of 5 millimeters, then your V10 would be probably substantially lower and your gradient indices would also be much lower on those smaller tumors.
Moving on from version 1.5 to version 2, what we actually see is much-improved crosstalk between close tumors. What we see is that a tumor that...This was an example where we had treated a multifocal glioma. The multifocal glioma here just by basically putting into version 2.0, putting an optimization structure between the two targets, and suddenly there's no crosstalk with our 50% isodose line.
Moving on to once you deliver the treatment, how do you assess these patients in follow-up? Oftentimes, when we see worsening...This was an example of a patient who we have found real value in the contrast clearance element. What we had done was, this was a patient who had oligometastatic metastatic non-small cell lung cancer had treated to multiple metastases in the brain, had a CR to all of them except for this one. And then when we followed them, they had worsening swelling, they became steroid dependent. We got our contrast clearance study. What we found was that there appeared to be effectively necrosis in the very center of the worsening enhancement, basically that maps out exactly to where our treatment delivery was, and surrounding that there was recurrent tumor. We had just retreated this patient, and they're no longer steroid dependent. And we'll be getting their new MRI to assess the volumetric reduction in the enhancement here shortly.
But overall, basically already managed to convert this patient with this short follow up to being no longer steroid dependent. If we can find these tumors earlier when they are recurrent, then we can probably reduce the actual side effect profile that's associated with retreatment. Because we know that doing radiosurgery for a second time, the risk of them developing radionecrosis doubles. So, earlier treatment like from the salvage setting would be likely to reduce the actual side effect profile of early salvage therapy. Thank you very much.
Bogdan: Interesting presentation, Dr. Seymour. And let's move now to Dr. Minniti for the follow-up talk.
Dr. Minniti: Good evening, I am Guiseppe Minniti. I'm a radiation oncologist in Italy, working in the University of Siena in Rome at UPMC Hillman Cancer Center. And first of all, let me thank you, Brainlab, for this invitation. And I will focus on our initial experience with the single-isocenter radiosurgery for targeting multiple lesions, specifically on the optimal target volume margin strategy when we use these techniques. These are my disclosure and this is the outline. So basically, I will touch on briefly some issues that are still manageable when we use the single-isocenter multiple target radiosurgery.
Basically, we have two different techniques that also reflects two different commercially available software. The first one, which is the single isocenter techniques using the VMAT, so volumetric modulated arc therapy, and this Aegon that use the singular isocenter techniques with dynamic profile hacks, ATP is the techniques that we use and this is used by the Brainlab software. So this refers to our initial experience, it was done...we'll start with a couple of years ago, actually more than 2 years ago on 32 patients with up to 10 brain mets. These are the characteristics of the brain mets. We target all these mets with a single fraction radiosurgery. But what is more important when we look at those imaging parameters that we still use, just to define a good radiosurgery, when we look at the conformity index, which was 1.3 or the gradient index which was 3.9, the V95, which is the volume, the volume which is covered by 95% of the prescribed isodose, it was 99%.
So this means, and there are a lot of publication looking at that, that when we use single isocentric techniques, in terms of those symmetric parameters is still very good, even better than when we use single target radiosurgery. We use the 1-millimeters margin from GTV to PTV for these techniques. Margins are very, very important. So the question is why we should use very tight margins? So, the answer is yes, if we consider some spherical lesions, for example, a lesion of 1-centimeter diameter, we know that the GTV is 0.5cc. Now when we put around a GTV 1-millimeter margin to generate the PTV, we have a double volume. And we have even a tribal volume if we use 2 millimeters. So, the reason for using, like, margins is because they reveal a very strong correlation between the risk of radionecrosis and the volume the normal brain volume irradiated ay high doses. Specifically, we use the V12, which means the brain, which is covered by 12 gray, and as you see there is a correlation, we have a risk of up to 24% when we have this bit well between 60cc and 11cc and even more when we have a V12 more than 11cc.
In clinical practice, this means, for example, for these small, tiny lesions which GTV was 1.3cc, which means 12 millimeters lesions. If we bought 1-millimeter PTV we have a V12 of 0.7cc, but if we but 2-millimeters margin to generate the PTV, we have an 8.4cc of V12. If we look at this figure, there is a risk of 20% if you use 2 millimeters and the risk of less than 5% we if we use 1 millimeter. So this is the reason because margins are very important when we use radiosurgery. On the other hand, when we apply the single isocenter techniques, we have to take into account that we use a single isocenter for all lesions and the very small translational and especially rotational shift of the isocenter can translate in very large amount of shift for lesions, especially as you see in this very well done study, especially for lesion that are eway from isocenter, more than 4 or 5 centimeters from isocenter, and especially from small lesions.
So, what we looked at was exactly to see the symmetric variation when we treated the patient with up to 10 brain mets. So this is what we do in clinical practice. We use the ExacTrac X-RAY, ExacTrac for the setup of our patients, then we repeat ExacTrac before starting the treatment. Our pole is in our centers, we use also the CBCT just for looking, just to be sure that both systems are in agreement before starting radiosurgery and then we use an ExacTrac during the treatment. So, what is very important is before starting the treatment that we consider the tolerance for translational movements and rotational movements of 0.5 millimeters and 0.5 degree. So using these techniques, we look at the impact of the residual errors on the geometric deviation and dosimetric variation of each lesions. We use this software which is velocity. Then we look at the foliage lesions at the impact of these residual shifts after correction with ExacTrac. And these are the results.
So I don't want to go to all these details. I want to just to show that when we look at the main dosimetric parameters, which is the V95, again, the volume, the volume of the lesion covered by 95% of the dose, you see the dose is 1.3% median dosimetric variation. But what's more important then, when we look at the variation less than 95%, which means just a surrogate of, I can see significant variation of target coverage, we see this variation in 22 targets. So 22 targets which means that by 20% of our targets. What is very, very important was that using the 1 millimeter of GTV to PTV margins, we could see an excellent target coverage that was maintained for all targets, this is very important. So, we can say that 1 millimeter, very small tight margins, we could maintain the perfect dosiisometric coverage of our targets.
This is the correlation as expected with the volume of the lesions, which means the small lesions may have largest variation of the V95. It is the correlation with the distance from isocenter. The large distance correlates with isometric variation, but occasionally is better with this table where we look at the distance from isocenter, more or less 4 centimeters for the volume of the GTV 0.4cc, which means more or less 89 millimeters in diameter sub-lesions, you see that most of the variation occurred just for small lesions away more than 4 centimeters from isocenters.
So, this is the important message I can say that we don't see any variation of V95 for lesions more than 1 centimeters, which are at least less than 4 centimeters from isocenter. And also, we don't see any significant variation for lesions more than 50 millimeters once it meets the synapse more than 4 centimeters from isocenters. I want to show you what means this. This means that we still use 1-millimeter margins for all lesions, as you see in the yellow arrows. But we can use 0 millimeters margins, in some cases, on large lesions more than 50 millimeters more than 4 centimeters away from isocenters, or small lesions, more than 1 centimeter close to isocenters. This is very important because this is exactly what happens now in the clinical practice.
So you see, for this patient with the lesions we have, for example, these lesions are very close to each other in the modern erea, as you see here in the brainstem. So we still have a very, very perfect coverage, V95 of more than 99%. But then when we use at the margins, so of course we use different doses for the brainstem lesions. So 15 graty versus 20 gray for the other lesions. Now you see that when we look at the margins, so the margins was 1 millimeters for all lesions, but was 0 millimeters for lesions in the brainstem just because they were very, very close to the isocenters. So now we started to use different margins based on our results.
Another important issue is just to monitor with ExacTrac during the treatment because there are still a significant amount of shifts due to tiny movements of the head during the treatment. So I think it's very important to follow up with ExacTrac patients and to correct the position during the treatment, especially if the treatment lasts more than 10 minutes.
So what's the next? Just, I need another three, four minutes. What's the next in our research? So the first question is, can we treat more than 10 brain mets with the same accuracy that I showed? And I can say yes. This is another paper. Basically, it's a clinical paper where we have treated patients with a median of 13 lesions and were more than 500 lesions. And when we look at the isometric parameters, the same I showed before performance index, gradient index, and V95, you see even if when we treat patients with up to 20, 25 lesions, we don't see difference in the dosimetric parameters. But this was a clinical paper someone adjusted to show some very excellent thing in clinical results. When we looked at some neurocognitive function of these patients, specifically memory function, we didn't see any significant decline, even if we treat it with the radiosurgery single isocenter techniques, this patient up to 25 lesions.
Finally, just one minute to say that I have replanted, we have replanned all these plans, element plans with [inaudible 00:40:45.141], which is the software they use PMAP techniques. And there are several papers that they have already in the literature comparing these two techniques with different results depending on the size of lesion, the different dose fractionation. Especially in all these papers, they never looked at people, at patients with more than 10 mets. So what we did is adjusted to extend these experience, this comparison to patients with more than 10 lesions, you see at 36 patients. But a significant proportion of these patients have more than 10 lesions. These are the characteristics of the GTV, the PTV. These are typical plans and it was also replanned with [inaudible 00:41:36.414].
And I'm going to just show this. So in terms of conformity index and gradient index, we'll look at some slight but significant superiority of dynamic conformal arc singular isocentive techniques. Even if I were to say that for both systems when we treat more than 15 lesions, there is a problem, a worse, slightly worse of conformity index. And again, when we look at the gradient index, which say more or less our normal volume irradiated, our shape is that those fall off, there is again a superiority of dynamic conformer therapy to be mapped. And these are also a few figures which look at the difference in V12, V10, and V8 between the two techniques.
So always the difference between the two system was in favor of the DCAT versus VMAT in both of V12, V10, and V8 gray but also in terms of dose, min and maximum dose of hippocampi, which you know are correlated with probably with some neurocognitive decline of patients. And when we look in forms of papers, you see what we can consider, the V12, but we have always a decrease in the normal brain irradiated at 12 gray with conformal arc therapy versus VMAT. And the same when we consider also the dose of hippocampi. So this is one of our patients with the 25 lesions. And when we look at the V12, so you see, especially for lesions that were very close to each other, there is a superiority in dose distribution with DCAT versus VMAT. So this is this paper is submitted. So hopefully, will be publish in a while so we can see better these results.
So in summary, I can say that single isocenter multiple targets formal arc radiosurgery is a fast and effective approach for patients up to 25 brain mets. The use of ExacTrac imaging system is very important because still we're able to correct the position at setup and during the treatment to avoid the negative impact of rotational translational shift on the target coverage. Finally, we can use, it is safe to use 1 millimeter to avoid any risk of radionic necrosis and also to avoid the loss of target. Anyway, in a specific situation, as I say large lesions or small lesions very close to isocenter, we can reduce for the margins. Thank you for your attention.
Bogan: Thank you for sharing your most recent results, Dr. Minniti, and let's move on to Dr. Agazaryan for the subsequent talk.
Dr. Agazaryan: Thank you, Bogdan, for the introduction. I will be presenting on the UCLA experience with the single isocenter treatment technique for multiple brain metastases. And I will be primarily focusing on the practice improvements, motion management, and margin reductions. These are my disclosures. The implementation of the single isocenter treatment technique at our institution had a significant impact on patient experience, provider benefits, and treatment unit throughput. As shown in this one example, the benefits are pretty significant. And these benefits increase incrementally with additional lesions. Certain key IGRT technologies and methods are necessary to successfully implement this new practice, and I will cover that in the second part of my talk.
These types of improvements are an integral part of our clinical practice where with continuous patient growth, we are looking for ways to do things better and faster. Here's one specific area of time savings. Shown is the treatment planning time as a function of number of lesions planned with single isocenter versus multiple isocenters, the time savings are really significant. And in the case of the 16 isocenter plan, the difference was even more significant. The physician supervision and attendance time for SRS cases is much shorter doing a single isocenter technique. The physicist for supervision attendance time for these SRS cases follows a similar trend with the number of lesions treated.
There is also cost savings associated with the application of this treatment method. In June 2020, we have published a study in "Red Journal" where we have used time-driven activity-based costing to determine the difference in cost to provider for delivering SRS with a single isocenter versus multiple isocenters. Here in this slide, we show the cost savings when we deliver SRS with multiple isocenters versus single isocenter as a function of number of lesions being treated. Cost savings increases with the number of lesions treated, obviously.
With all that said, there are challenges with the single isocenter treatment technique associated with rotational alignments. Here we simulate a three-degree rotation for demonstration purposes and the rotation is around the isocenter. As seen in this animation, rotations have larger impact for targets that are further away from the isocenters. Hence, many institutions are considering or already using different margins depending on the distance from the isocenter. Displayed on this graph are translational misalignments that can occur with rotations as a function of distance from the isocenter. 0.5-degree rotation can result in 0.5-millimeter misalignment for a target that is 6 cm away from the isocenter.
For motion management, we use ExacTrac and we use ExacTrac prior to each arc and we utilize 0.5-millimeter and 0.5-degree tolerances. We also recommend 1 millimeter or less margins based on the distances from the isocenter. With the use of the tolerances mentioned, 0.5 millimeters and 0.5 degrees, the repositioning rate for patients is 57%. So this means 57% of the time when the therapists are obtaining ExacTrac images, they have to reposition the patient. So there is some residual motion. And if you look at the ExacTrac data from these patients, it shows that with the use of ExacTrac prior to each arc, and using 0.5-millimeter and 0.5-degree tolerances, there are some residual motions and 98 percentile of those motion is 0.7 millimeters with a mean of approximately 0.2 millimeters.
The 90th percentile rotations are 0.6 degrees with the mean of approximately 0.2 degrees. So all in all, if you use ExacTrac prior to each arc, then the residual motion after the arc for translations is mean of 0.2 millimeters and the rotations mean is 0.2 degrees. The reason that's measured is because, based on that, there is some justification to use smaller margins for these treatments. It's not based on the outcome, but it's based on patient motion and technology that's being used. Let's keep in mind that using 1-millimeter or 2-millimeter margin, compared to the 0-millimeter margin, we're essentially doubling and tripling the volumes of V5, V10, V12, and V8. We also double and triple actually the treatment volume so there is a significant normal tissue that's been included in a treatment volume.
One can also use variable margins based on the distance from the isocenter. And that can also help with the reduction of the V5, V8, V10, and V12. The variable margin could be an important approach because only small percent of the targets are far away from the isocenter. In this quick study that we did here for 74 targets, it's shown that only 10% to 15% of the targets are really further away from the isocenter more than 6 cm. I should mention that using larger margins can also impact the prescription dose if one follows the consensus guidelines. Depending on the clinical prescription strategy used at your institution, about 10% to 15% of the targets may get lower prescriptions because of the use of larger margins. So for various reasons, we have decided to go with a uniform 1-millimeter margin at our institution. This is a significant improvement from previously used 2-millimeter margin. And our practice may even further improve in the future with the use of 0.5, 1-millimeter variable margin. So we are currently using this methodology and in the future, we may use that one.
Lastly, our institution has shown great clinical outcomes with the retrospective study. We show excellent local control and freedom from radionecrosis using single isocenter technique, we label that as a simultaneous SRS. And the results are very much comparable to the results coming from the conventional SRS, which means using multiple isocenters. In summary, the single isocenter treatment technique enables faster and efficient treatment planning, enables faster and efficient treatments, requires shorter supervision from the physicians and a physicist alike. It provides financial benefits to the provider. It requires dedicated technology for safe and accurate treatments. And with this dedicated technology, one can use smaller margins. There is also an evidence that single isocenter treatment technique results in similar clinical outcomes compared to the multiple isocenter technique. With that, thank you for your attention.
Bogdan: Great overview Dr. Agazaryan, and it's refreshing to see how modern radiosurgery has transformed your practice. Let's now move on to the final presentation from Dr. Prasad.
Dr. Prasad: Good day, everybody. I will be the last talk of this session. And I wanted to throw out a provocative title to get your attention. We'll talk about tumor or trouble. Essentially the use of contrast clearance analysis in analyzing the outcomes of radiosurgery, and managing these patients clinically. No work is done by one person, and neither is this. This is the work of many of my collaborators who work with me every day at the Gamma Knife Center at Roswell Park, where we've created more than 1000 cases I've been doing this for too long, more than 12,000 cases personally, and all indications as you can see. We are a busy center, last year we treated nearly 700 patients, which is, I think, for a single Gamma Knife, a very, very high volume. And you will see how we have integrated Brainlab elements and contrast currents analysis into the daily management of our patients.
This talk pertains to the multimodality management of brain mets, which as you're well aware as a group is now truly multidisciplinary. Our colleagues from medical oncology have brought to bear immunotherapy and new techniques to help these tumor patients survive longer and tumors respond well. Surgeons do remove those that are critical and producing neurological deficits or threat to life. And for all the rest, there is usually a combination of radiation tools, still, radiosurgery being the mainstay of a lot of what we do today for brain mets cases.
Despite that, the utilization of radiosurgery remains low and we have done a lot of work over the years to educate our colleagues, as well as the sources of referral to increase the utilization of radiosurgery. When you do that and you build a really busy practice and treat a large number of patients with large numbers of tumors, you'll begin to see both positive and negative effects of this technique. Brainlab Elements, at this point, is seamlessly integrated into our daily radiosurgery practice. We go from imaging to segmentation using Brainlab Elements, planning, delivering treatment, and following all these patients up with serial imaging, which is brought back into our treatment planning system as well as brought in to Brainlab Elements for contrast clearance analysis when we're suspecting radiation injury or trying to distinguish radiation injury from a residual tumor, and that can then guide us in treatment and management of these patients.
Adverse radiation effect is well known associated complication of high dose radiotherapy, and stereotactic radiosurgery is no exception. Generally, reported rates remain low. But the biggest issue with adverse radiation effect is that it involves the surrounding normal brain and can have clinical consequences as well as imaging consequences which confound the viewer looking at an MRI post-treatment, and it becomes very hard to distinguish treatment from response. It is truly an inflammatory process with [inaudible 00:56:54.510] and vascular proliferation that extends well beyond the treated field. And therefore, assuming that it is tumor progression and retreating it can only compound the problem. And we have never quite successfully been able to come up with one tool that can give us a solution to deciding between tumor progression and our radiation effect.
Contrast clearance has added to that armamentarium and provided a very, very useful visual tool for us to help make those decisions. Contrast clearance analysis depends on the rate at which contrast flows through blood vessels, tumor tissues, and damaged brain at different rates, essentially rapidly in and out of a vessel, accumulating relatively fast and out within minutes from tumor, and then slowly accumulating and building in damaged brain around a radiated field. And therefore, color coding this as blue for tumor and red for radiation effect makes a very clear-cut graphical picture. The requirement is a T1-weighted MRI done after contrast administration and then 60 to 90 minutes later.
This second exam will require patient getting back on the table. We have quite recently instituted a single session technique for all of this, wherein we will perform the contrast basically one image early in the imaging of the patient, proceed to all the other sequences that we need to acquire, including a magnetic resonance spectroscopy and DTI images that are done quite routinely in our patients, and then returning roughly about the 45, 50-minute mark to proceed and look at the clearance scan. But if you want to do it by the book, then you will take the patient off the table, wait for an hour, hour-and-a-half, put them back on and repeat just the T1 image, send this over to the Brainlab system, and everything else is pretty automated from there.
This paper, the original description of this technique was in glioblastomas, and we applied it very specifically for Gamma Knife radiosurgery, but I'll start with the GBM just to give you an example. On the left panel is a treated patient and that is their early cram study, which shows basically treatment effect on a small residual amount of blue in the center of it. So even as the MRI is getting better, you're beginning to see evidence of response. This response can be interestingly measured, even in the absence of enhancement, say in a patient receiving Avast. So that makes it very useful when you're looking at pseudoresponses to the progression. Now this patient was to then have a subsequent follow-up MRI and show you an area. As you see on the top right, the right panel on the anterior aspect of the treated area, there is new enhancement and we don't know if that's radiation effect or progression of tumor. Well, here's the contrast clearance analysis showing you in deep blue, clearly corresponding to the enhancing area, proving to us that this is a tumor, and therefore directing us to repeat radiosurgery for this patient.
You can do this for any number of patients, and you can follow them serially, and I'll show you different variations of imaging with contrast clearance after radiosurgery. So here is an example of a patient with a residual, and a patient who has some adverse effect. And you can see patient receiving Avastin for GBM. And you can identify on the imaging below that most of the central portion of this is not a still active tumor and not radiation necrosis. And that can direct retreatment for these patients. When you look at brain mets, there is less data available. And so we started using contrast clearance to help us distinguish resistant enhancement or appearing enhancement after response, as you can see in this patient who's had multiple episodes of Gamma Knife, as well as surgery. And here, the turquoise lines show you to two different Gamma Knife treatments, and the yellow and green lines are the isodoses for the final treatment, despite which we have persistent enhancement.
And so this patient was then taken to surgery. And we can't tag the biopsies that were done based on whether the areas were blue or red on contrast during analysis. And when you look at that, you see the picture here showing you a blue region in the posterior left stargate and red field all around it, which bleeds into the cystic degenerated area which also appears red. The question is at that interface, are we correctly identifying with contrast clearance, tumor versus necrosis? And here you would have tagged specimens stared, which were biopsied and labeled and sent to pathology. And as you can see, the blue area corresponds to a predominantly residual/recurrent tumor, whereas the blue-red interface shows you a clear demarcation between tumor and necrosis.
So we needed this internal validation study, we did it in a number of patients. Here's another histology map showing you that largely blue regions show predominantly tumor cells and areas that are tagged red on contrast is our fields of necrosis. That allowed us to build local intelligence and confidence in this technique. And if you do radiosurgery, in the ideal world, you'd expect the lesion to disappear. And that is what happens in this case, as you can see, top left, top right, bottom left. Serial imaging showing response and contrast gains showing no signal, no blue, no red, that is the ideal universe.
But sometimes you will have adverse reactions, but no residual tumor. And that appears in this map. You see a tiny amount of blue in the back end of the target here, but the predominant result is essentially red. And that is essentially radiation effect. It has not increased dramatically, there was minimal edema associated with it. It's not someone that will require therapy. But it definitely allows you to identify what adverse radiation effects in the absence of residual tumor looks like. And if you look at the patients early on, you will still see a residual tumor. So here you see the blue spot in the middle of the target, which is still the persistent tumor cells, but the early appearance of adverse effect around them from radiation treatment with radiosurgery. And here's another example. Frontal lesion showing you similar mixed response early.
When you follow these patients over time, you will see an evolution and you can either see a persistence, enlargement, or even disappearance of the adverse effect and the slow breakdown of the blue regions of enhancement and ultimate disappearance. It's only going to be in patients where you see the adverse effect continue to progress that you will need to intervene. But you can see that both enlarging and shrinking contrast-enhancing areas on a standard MRI could be associated with contrast clearance evidence of adverse radiation effect. Partial response, as I told you, will be like this. And we know that surgeons who have gone and operated on tumors that have been treated with radiosurgery have often shown a mixed bag of necrosis and [inaudible 01:04:30.815] tumor.
This will allow you to identify areas where viable cells are most likely to be. And if you're considering boosting something like this, or repeating radiosurgery, we will specifically target the blue delineated areas on contrast clearance. Brainlab Elements allow you to delineate that as a structure, send it back over to your treatment planning system, in our case Gamma Knife, but with you, it could be a different radiosurgery platform, or it could be the Brainlab radiosurgery platform. And you can make that your new target and deliver early treatment.
Sometimes it's hard to tell what you're looking at. Here is a patient which we would have looked at and said, well, that looks like treatment, in fact, it's most likely some radiation-related adverse. When you do the contrast clearance, however, you realize that that blu ray is actually rival tumor and that is corresponding to the enhancement. You do have adverse effect and leakage of contrast for that further out. So something like this, on the left bottom panel, you see the unanalyzed, delayed image, that is the 90-minute image for the same patient. And that's what contrast clearance will take and mathematically analyzed for you. Of course, we use this to guide therapeutic intervention. And as you can see here is a patient with adverse effect on day zero. And you can see sequential imaging after a vascular started showing you progressive shrinkage in the area of the adverse radiation effect.
We will soon be publishing our experience where we have 70 patients who have been carefully analyzed out of a denominator of nearly 1400 patients who were treated over this timespan. So just give you an idea that it shows you how our incidence of radiation necrosis and adverse effect remains low. But this tool helps us make decisions which are intelligent. We have GBMs and a lot of brain mets and our mixture of all different primaries. And as you can see, 29 of these patients out of the 70 ended up requiring some retreatment. And that retreatment was based on our contrast variance analysis. So two-thirds of these were roughly tagged as adverse effect and followed or treated with other interventions. But about a third of them required retreatment because they had active visible disease on the basis of contrast clearance analysis.
And we do this sequentially, patients have had one, two, three, four, five imaging sequences with contrast clearance that is built into our standard workflow. And as you can see, when you analyze it in this manner, roughly, roughly a fifth of what you're seeing on a plane contrasted MRI will be viable tumor across the board, and the rest will be adverse effect. And that's just an odd mixture of patients. Of course, it will vary from practice to practice, but it allows us to highlight one structure over the other. And it also has guided us in retreating these patients. We also looked at whether single or multiple lesions were more likely to produce adverse effect, and we did multivariate analysis. The most important things that correlated to adverse effect were the first treatment dose, not the retreatment dose. So initial dose is the biggest correlation to adverse effect, makes sense. Also, most of the adverse effects are seen within the hydrosphere, not outside the high dose or a mixture. So we do see some early correlations of clinical parameters that predict how adverse radiation effect will occur.
But in the ultimate analysis for us, we need this. We need this why? Because if I go back and look at the radiology reports of the patients that I talked to you about, and this is just a sub-sample, you can see that the vast majority of those reports make no comment about adverse radiation effect. In fact, the treatment effect is only commented on, when commented on, was only detected about one-quarter of the time. And when we correlated the radiological interpretation versus interpretation on the contrast clearance, it becomes very clear to us that ccA increases the pickup rate for adverse radiation effect, as well as helps us identify amount of residual tumor. I see that as ultimately bringing us two pieces of valuable information.
One is, of course, how much response we're getting. Sometimes we can identify the response long before the contrast enhancement has disappeared, which is an indicator of early response or early failure, which allows us to intervene, change medical therapy, immunotherapy or repeat radiosurgery, as the case might be. And all of this is made possible because we are using one additional tool. We've been very lucky that our radiology group has accommodated us and allowed us to do these studies. They do consume some time, I'm hoping over time between our work, which we have now undertaken to do MR spectroscopy, as well as changing the minimum time that we allow to elapse between the first and second imaging sequence and running it through the contrast clearance paradigm, my intent is to do a slightly earlier scan and then do a 60 to 90-minutes scan and compare the two to see if our pickup rate is pretty good early enough and that will make this whole process a lot more integrated into an imaging system. And also compare it with MR spectroscopy. So stay tuned, that's gonna be the next thing. Roughly in a year, we should be able to share that data with you.
But what I'm happy to report is that this really gives me a very good visual tool. My whole team is educated on it. And it allows us to make smarter decisions. It allows us to take and identify those patients within our very busy practice who can benefit either from retreatment or other therapy and minimize their side effects, overall give them a better quality of life because our entire radiosurgery program, certainly our brain mets program is designed only with one intent, and that is to make the patient's quality of life better. Thank you for your attention. We will answer any questions that you might have. Thank you.
Bogdan: Thank you for your captivating talk, Dr. Prasad. I'd like to thank all our speakers for all their lectures and we can now move on to a live question and answer session.