Transcript
Host: Hello, everyone, and welcome to another Novalis Circle clinical webinar. Today we are visiting Thomas Jefferson University Hospital in Philadelphia and I have the distinct pleasure of introducing Dr. Wenyin Shi, associate professor and co-director of the radiosurgery program, and Dr. Haisong Liu, associate professor and medical physicists. They will present today their clinical experience utilizing our Elements multiple brain Mets SRS. Thomas Jefferson University Hospital was the first customer in North America to utilize this technology a few years ago and since then, they have treated more than 100 patients utilizing our software. As always, please feel free to submit questions through the chat interface. At the end of the lectures, we will answer those questions. So, here are the presentations.
Dr. Shi: Thank you, everybody, who join us on this webinar. I'm Wenyin Shi and I will kick up this program to introducing some clinical background of treating multiple brain metastases, then we will transition to discuss the Elements software, and after the presentation, we're going to go over some cases to show some examples. We don't have any disclosure for this activity. As we all aware, the incidence of brain metastases are increasing over time, there's multiple reasons that lead to this phenomenon. First of all, there's improvement of systemic therapy, so the survival of patients increased over time. Unfortunately, when a patient had longer survival, they have the risk of developing distant metastasis, including brain metastases are also increased.
And also we have an aging population and more patients were at risk. And we have imaging technology with high-resolution MRI scan, so we're able to identify much smaller brain metastases, which was usually lost anytime before. And it is really having a paradigm shift towards more focused radiation treatments, radiosurgery or fractional video surgery, for patients with brain metastases. Currently, there's high-level evidence to support using radiosurgery alone without the whole-brain for the patients who has a limited number of metastases, particularly one to four metastasis. As you can see here as a summary table, I included one of the key randomized trials indicating that there's no compromise to the overall survival if we omitting the whole-brain upfront.
More importantly, is many of the studies also addressed the new cut function and primary endpoints and it was shown very clearly, that by omitting a whole brain, there's a significant improvement of preservation or neurological function in those patients, which is clinically significant. As a result, radiosurgery is considered a high level of evidence by the NCCN guideline for patient with limited brain metastases, and Astro "Choose Wisely" initiative also included do not routinely add adjuvant whole-brain radiation treatment to stereotactic radiosurgery for patient with limited to brain metastases. There's also efforts looking at treating patient with more brain metastases with radiosurgery alone. And several studies have been shown that the number of brain metastases does not really predict outcome, but it is what's the total volume of the brain metastases.
The highest evidence using radiosurgery alone for patient with multiple metastases coming from Dr. Yamamoto's experience. In this prospective radiosurgery study, we have over 1,200 patients with 1 to 10 brain metastases, and they all were treated with radiosurgery alone with gamma knife. And this select patient who has individual lesions less than 10 cc and the total brain met volume less than 15 cc, and the prescription dose is 22 gray for small lesion and 20 gray for larger lesions. As you can see on the survival curve, compare the patients with 2 to 4 brain metastases and 5 to 10, there's no difference in survival. As a result, it supports the use of radiosurgery alone for patient with up to 10 brain metastases.
The subsequent study by Dr. Yamamoto was looking at the patient that has even more brain metastases, even more than 10 This is a case-matched study looking at patients with 2 to 9 brain metastases as well as more than 10. As a result, you can see, there's no difference in overall survival or neurological deaths free survival, so that also supports the practice of treating radiosurgery for more than 10 brain metastases. However, the neurocognition impact of using radiosurgery alone for patients with more than four brain metastases is less known. And this is a study by the same group, which using the mini-mental status, which showing that there's no deterioration, however, mini-mental status test is not as sensitive neurocognitive testing instruments.
The more comprehensive neurocog testing relies on Hopkins Verbal Learning Test, Trail Making Tests, and Controlled Oral Word Association test which has been utilized in most RTOG and RG trials. And there's also some development using computer assistant or iPad software to make this test more user-friendly, and some apps I showed here, including the Cognition by Brainlab was one of the examples. Currently, in our institution, we select patients for radiation treatment for brain metastasis in the multidisciplinary clinic. And in general, the patient who has one to four brain metastases, radiosurgery alone is the preferred option unless you have very sensitive histology such as small cell or lymphoma. And the patient who has 5 to 10 brain metastases, we still consider radiosurgery as a first choice and if they prefer, we put the patient on trial, which is an in-house trial focusing on evaluating the neurocognitive function for the patient after radiosurgery to 1 to 10 brain metastasis.
We're using the patient with one to four brain metastases as a control and compare the patient who has 5 to 10 brain metastases who receive radiosurgery and look at the impact on neurocognitive function. For patients with more than 10 brain metastases, we more routinely consider whole brain radiation treatment and in particular, hippocampus-spearing whole brain treatment if those patients are eligible. And if the patient progress after whole brain radiation treatment, then we will consider radiosurgery and those could be more than 10 metastases. And there are obviously a lot of technical challenges when we are moving from single targets or a few targets of brain metastases to multiple brain metastases.
There are multiple solutions currently being utilized in clinics including gamma-based systems, LINAC-based systems. But we can summarize this approach into two major categories, one is multi-isocenter techniques, basically, we treat each individual brain metastases individually and this will be a perfect example for gamma knife or CyberKnife approach as well as some of the LINAC approach. The drawback of this is very cumbersome delivery in the patient who has multiple brain metastases, particularly more than five, the treatment time could be long. And if we use the LINAC system, then the treatment planning complexity can be particularly challenging when we're treating multiple brain metastases with multiple isocenters.
Other solutions including mono-isocenter techniques, particularly using VMAT techniques, and it has been shown that we can achieve very similar plan quality when we compare to gamma knife plans, maybe a slightly more low dose wash in the dose less than 3 gray or so range. However, it's really highly complex inverse planning and heavily dependent on the user experience, and requires complex patient individual QA. So, this is the challenge we're facing in the clinic and the Brainlab Elements software is a perfect solution for this challenging clinical problem. And I will have Dr. Liu to take over and explain this...to introduce the software.
Dr. Liu: Hello, everyone. This is Haisong Liu and I'm going to introduce the Brainlab Elements for a multi-mets SRS planning. So, Brainlab multi-mets Elements, it's one isocenter non-coplanar dynamic arc technique, treating multiple metastases and with the current version, it can treat up to 15 mets in one session. So, this is the workflow of the multiple mets SRS and starting from import all the images into the software including CT MRI, and image fusion is basically a rigid fusion between CT and MRI and it's fully automatic, it takes seconds to fuse the CT and MRI, and the evaluation always showing very good results. And the next one is called anatomical mapping, basically, identify or automatic contour all the organs at risk including both eyes, both nerves, optic chiasm, and brainstem, and also the whole brain.
So, for the cases that we have done, all these automatic contouring really need no or very rare human intervention. And the next one is called a SmartBrush, it's their kind of tool to doing the contouring. And instead of in the traditional planning system, we have to contour slice by slice on each axial slice, the SmartBrush automatically picks the higher density in the T1 contrast MRI images and you can just draw the circles on one axial view and one of the sagittal view or coronal view, basically, two brushes that create a three-dimensional volume for you. And that is a good starting point and then you can just go through the slice and manually make revisions to make it perfect. And the next one, they call it "Object Manipulation," and it has several Boolean operations there. But most of the time, we're using this to generate margins for the smaller brain mets.
So, in our clinical practice, we usually generate a one-millimeter margin for most of the mets, but when the math is very small, for example, the one that we show here, the GTV is only 0.06 cc, so in this case, we generated 2-millimeter rounded. And then we're going into the multiple brain mets planning itself. And this is a planning interface and basically, you choose from two lists of protocols, one is the protocol determines the dose, the prescription dose. So, we have several levels, say for example here, all the mets treated to 15, 18, 21, or 24 gray, or three-level doses depend on the size or volume. But this is just a starting point, you can feel free to select any mets to change it to, say, even 17 gray or 16.5 gray. And then, the next column is the setup.
So, the setup, basically, the arc geometry to either use five table angles. So, the first one is coming with the default installation, it has five table angles starting with couch angle at 0 and 40-degree arc separation all the way to 160 table angle. So, here we listed a couple of different arc geometry. So, these is the different templates we make, we have a total of 10 of these templates and this is an example of the default template, 5 table angle, 40-degree arc separation, and this one has 6 table angles starting from 10 and ends at 170 and 32-degree arc separation. The most complicated, we have 9 arc, 9 table angles, and 20-degree arc separation. So, after we select the dose and the arc template, we just click "Calculation" and everything here happened in the background.
So, the software will automatically select the isocenter based on the center of mass of all selected PTVs and use the pre-configured table angles and arc lengths to calculate which arc to select. So, each arc may deliver a dose to a subgroup of metastasis, so if two mets lined up in the same direction of the leaf motion, then the software will add another arc in the reverse direction treating the other met in line. So, this concept leads to optimized conformal treatment combined with low dose spread to normal brain tissue. ANd MLC margins, the MLC will conform to each individual PTV with up to one-millimeter margin, and it can vary between zero to one and it also automatically select which PTV to treat for each arc. So, after the selection matched with MLC with each individual met, there's the weight optimization of each arc to achieve the best conformality index, it's based on the inverse Paddick calculation for conformity index, and then the final dose calculation based on pencil beam algorithm.
A planning time for a simple case like three or four mets, it should take about 1 minute and 10 seconds, and for the most complicated 11 or 12 mets with 7 arcs, this may take up to four minutes. So, we're going to show you several planning cases and how we evaluate the plan in the next...in the last slides, but I'm going to jump into the patient QA part. So, for patients-specific QA, we don't do patient-specific measurements, rather we do an independent calculation with independent software, which is Eclipse planning system. So, after the MME planning, DICOM export CT, structure, and plan to Eclipse, and then we will recalculate the plan with both AAA and the Acuros dose grid with one-millimeter dose grid. And then we compare the mean dose of each PTV, so for our first 56 patients and 233 targets, the percentage difference between the MME and AAA is 2.7% and between MME and Acuros is 3%.
And also, we observe that the smaller the PTV volume, the larger the percentage difference. And our Eclipse calculation or Eclipse planning system is not configured for small fields, so our smallest output factors PTV and dose profile end at the field size three centimeter by three centimeter. So, that's why when we go to the smaller fields, we have larger discrepancies and that's our assumption. So, during the summer, we also work with Mobius. We sent a couple of our cases to Mobius and they use their Mobius3D and/or Mobius Fraction Check with our Truebeam trajectory files to calculate the TPS calculated dose and Mobius calculated dose and also the real delivered dose coming from the projection calculation from the Truebeam trajectory file. And one of the good things I like about this software is other than the windows, it also provides us 90% coverage or even probably 95% coverage as you defined, so that is a more meaningful comparison.
So, for the other several tests that we did for initial commissioning, one is the hidden target test. We already do a hidden target test for the BB and isocenter, but for this one, we put the BB at the off-center location. In this case, it's almost nine centimeters in 3D, far, far away from the isocenter. And then we treat the Rando phantom with patient treatment protocol and IGRT tolerance at the same as we treat the patient, and then we acquire EPID images at each start and stop angle if available. So, in this example, this is arc number one, gantry at 270, table at 0. So, you can see this is the intended-to-treat, the BB here, and this is the EPID image we got, and really, it's at the center of this MLC aperture. And this is arc number two with a table at 70, gantry at 160, and you can see the other two mets are here and this is the BB here, and that's still at the center of this MLC aperture. And for quantitative analysis, we import all these...we acquire actually a total of 10 EPID portal in this test case, and we import them into the RIT analysis software.
And you can see the results, on the lateral direction, we have two parts or one part over one millimeter, and on the longitudinal, we have two parts a little bit over one millimeter, and the other part is within plus-minus one millimeter. And another test, we use our Delta4 Phantom which is a 3D diode array phantom. We intentionally put all 10 mets on the tiles and this is the...these tiles have 10 mets, range from 1 centimeter to 1.3-centimeter in diameter, 0.3 to 0.4 cc, and we use 6MV Flattening Filter Free mode to treat these 10 mets. And the middle picture is showing the dose profile along the blue lines here and you can see the red line is the planned dose and the green dots is a diode measurement, so it's matched with very good results.
And for a 3% dose deviation, one millimeter DTA, we get a gamma passing rate of 99% and that's based on almost 400 diodes. So, we also did a 3D gel phantom and this whole process, we did with a Greece company called RTSafe. We sent anonymized patient CT to them and they build this skull and internal bony structures phantom using 3D printer, filled it with gel, and sent it back to us. So, we treat the gel phantom as if treating a patient and then use their provided protocol to do a post-irradiation MRI and they will provide all the analysis for us. So, this is the plan that we make, there's nine targets planned, the smallest volume is 0.1 cc and the largest is 0.5 cc, and 24 gray prescription, about 30 gray at each target center.
And this is the compare between the planning CT versus the post-irradiation MRI, and after the fusion, we can see that the colored circle here are the intended-to-treat targets and the black hole is what we irradiated. And RTSafe provide us the dose profile on one dimensional and the two-dimensional gamma index map with isodose overlay. And cognitively, we have about 0.2 to 1.1-millimeter three-dimensional targeting displacement for the nine targets, and D95 for the nine targets range from minus 2 to 6.5%, and the gamma passing rate is 99%. So, Dr. Shi is going to continue with our current clinical experience and plan evaluations.
Dr. Shi: Since the implementation of this Multi-Mets Elements in our institution probably a little bit over two years ago, this is quickly being embraced and we have a very, very steady flow of patients. And currently, in the past about two years or so, we have already treated over 100 patients using the software. For patients with single brain metastases, we just using the standard Brainlab iPlan system to plan. As you can see, based on our treatment guidelines, the majority of our patients are still patient with limited brain metastases, two to three brain metastases, comprise about 50% of our total volume, and we also have about 40% of patients who have between 4 to 10 brain metastases. The only very select patients that are getting radiosurgery are 10 or more brain metastases, usually for recurrence or patient who has a highly resistant tumor and without interruption for chemotherapy for other consideration.
This is just our initial evaluation for our overall outcome, local control, and intracranial control rate. As you see, we have close about 90% local control rate at six months, and we have not really seen any high-grade toxicity, and those numbers without digging into more detail about histology and GPA are in par with historical data. And I want to go through some of the planning philosophy we have here in our institution and then I will show a few cases, hopefully, to illustrate those points and make it more meaningful to you. So, first of all, currently, all the MME software came with a default template as Dr. Liu already mentioned, which is the 40-degree arc separation with 5 table angle, and this is as shown on the top right...top left-hand side.
And then we quickly learned that, you know, using one template is not sufficient and one template is not necessary the optimized template to use because the distribution, the number, and the volume of the brain metastases vary greatly from one patient to the other. As a result, we develop multiple custom templates using different table angle as well as arc separation, and we run all those different templates for each case and we select cases based on dosimetry, the ideal dosimetry. As you can see, here are the list of the 10 currently tested templates and there are a different number of arcs from 5 to 9 arcs and arc separation anywhere from 40-degree to 20-degree. And then after 100 or so cases we treated, as you can see, this is the history of distribution of which arc we would end up picking as results based on the idea of dosimetry.
So, it's relatively evenly distributed with most of the arcs...or template we picked are using around 6 table angle, between 28 to 35-degree separation as highlighted in the yellow. And the default template does being used as well but we will also see that if we using more table angles such as 7 to 9 table angle or even 10 table angle, we not necessarily getting better results, it's a diminishing return, and then when we running complex cases, more table angle doesn't necessarily help and it stresses out the planning system sometimes. And another key parameter we looking at the dose is V12. We know that the V12 is linearly related to the individual brain metastasis and the volume and we have a quick reference point to give us an idea if a brain metastasis...single brain metastasis of a certain volume, which roughly V12 range will it be to help us to evaluate a plan.
And the software also gives you pretty user-friendly features based on QVH, you can look at either each individual V12, which here as example in the highlighted area, you can see the individual V12 is 2.6 cc, and we evaluate each individual V12 and in a cluster metastasis, we'll also make sure the V12 isodose line does not converge or the converged isodose lines still have a limited number...a limited volume of V12 total volume. And this is one other example, we use individual V12 to help us with dose prescription. This is a patient with four metastases with two rather large brain metastases, 5 cc and 10 cc. And if we give 21 gray and 18 gray, both of them have a V12 number in the range of 10 cc, and given that there are four brain metastases in total, we feel that it is too high a V12 volume, so we reduce the prescription dose to 18 gray and 15 gray, that bring down the V12 to 6.7 and 7 cc, which is more clinically appropriate.
And this software also gives us a very well-detailed parameter of each lesion. For example, here, as you can see nine brain metastases, we will have the volume of the metastasis, maximum dose, minimum dose, and as well as the mean dose, and the conformity index. And they also help us to further evaluate if something that's outlined that we can potentially individually evaluate if something needs to be, you know, adjusted. And based on our current experience, we look at all our brain metastases plan with Elements software as listed in the first column, you can see the conformity index is quite much in bar as the VMAT plan or gamma knife plan based on our own institutional experience published in the prior paper. And the gradient index is actually in favor compared to VMAT planning and very comparable to the gamma knife plan. The isodose line is very similar to what we usually do with a single target, around 80% isodose line.
One other important feature or issue we need to recognize is the off-access effects. And this is a paper that's published in 2015 by Dr. Roper, they look at the impact of the geographic location of the metastasis based on the distance from the isocenter. As you can see here, the further away from the isocenter, the larger the displacement if there's a rotational or translational error. And if we do some simple calculation at the table here, if there's a 5 degree rotation error...0.5 degree rotation error, sorry, then the displacement at 7 cm from isocenter will be about 0.6 millimeter, and this number can increase beyond 1 to even 2-millimeter if the rotation error is more than 1 degree. So, currently, based on this calculation, we try to restrict the lesion treated more than seven centimeters away from the isocenter because the five-millimeter rotation error can lead to more than 0.5 millimeter transition disagreements.
And so, if we have multiple brain metastases and some of the lesion is more than seven centimeters, we'd prefer to use more than one isocenter to keep everything less than seven centimeters away from the isocenter. And so, this is our current planning procedure and we will consider using two isocenters if there's lesions more than seven centimeters away from the isocenters. And we usually do a one-millimeter margin for regular targets, but something that is very small, less than five-millimeter or four-millimeter in diameter, we may consider adding about a two-millimeter margin. And we generate treatment plans with different customized templates and based on the arc geometry, we will select the best plan based on the V12, dose to OAR, and the conformity index as well as the mean dose. So, right now, we run all the 10 templates for each individual plan and we compare all the parameters and we'll treat...we will choose the plan that gives us the best dosimetry.
And regarding the treatment delivery, in our institution, we use ExacTrac as IGRT solution, so the residual error for the IGRT setup is 0.5-millimeter, that should be a degree in rotation, and 0.5 millimeter in translational error. And we will reimage with ExacTrac each arc angle and if the error is exceeding our cutoff, which is 0.5 degree and 0.5 millimeter, we will apply the shifts and reimaging with ExacTrac to verify. And in the next couple of minutes, I will show you a few clinical cases and really want to just illustrate some of the points I just discussed. And first of all, this is a patient who has renal cell carcinoma, unfortunately, develops multiple brain metastases, and he has more than 10 metastases both supratentorial and infratentorial and involving both hemispheres of the brain.
Given the heavy disease burden and a high number of brain metastases, we recommend whole brain radiation treatment with hippocampus sparing techniques. So, the patient received the whole brain 30 gray in 10 fractions. Unfortunately, when she returned for follow-up after finishing the whole brain, quite a few of the lesions already progressed, so the whole brain, unfortunately, does not have a meaningful benefit for this patient who has renal cell carcinoma back, which is not surprising because the resistance nature of that. So, we get the patient for reduced surgery salvage, and as you can see, there's further 12 brain metastases and the volume is anywhere between 0.2 to 4.2 cc, and the total volume is around 15 cc.
And there's some unfavorable feature of that, one of the patient...one of the lesion is close to the optic chiasm and overall, these brain metastases have an average separation, means they're not heavily clustered but some of them are relatively close to each other. And we use a uniform prescription of 18 gray to all those lesions and we're using a 6MV Flattening Filter Free delivery mode. As you can see, I've just shown an example of the six templates, we've added actually...you know, as you can see, based on the conformality index, V12, and the chiasm max dose, we ultimately choose template number four, which give us the best conformity index, the lowest V12, and keep the chiasm maximum dose within our tolerance, which is at 10 gray in our institution. And this is the ultimate plan the patient was treated with.
And this is a summary of dosimetry using, you know, 3D illustration of the arc separation and the DBH graph as well as the isodose line. And as you can see, we have a conformal plan and there are very tight V12 clouds. And we also look at individual V12 as I mentioned earlier to see if any individual lesion or clustered brain metastasis dose need to be adjusted, and our largest brain metastasis V12 for the individual lesion is only 7.3 cc, which is reasonably well within our clinical tolerance for the patient and we receive the treatment without further plan adjustment, 18 gray to all 12 lesions in single isocenter.
And this is the post-treatment follow-up MRI scan. As you can see, two months later when he returned, we can see very marked improvements of all the brain metastasis treated, some achieve even ACR as shown here, and some of the large lesion is significantly reduced in size. So, it's a very, very favorable clinical outcome with no toxicity. And the second case is a patient with melanoma with multiple brain metastases. This patient has 11 brain metastases, they are relatively well-distributed, which means there were good separation, there's no clustered lesion. And this patient's number of lesions is slightly above our usual cut-off of 10 brain metastases, however, his melanoma histology was highly radioresistance, and also resistant to systemic. Given the burden that requires the initiation of systemic therapy quickly, we ultimately agreed to give him radiosurgery and we also treat every single one with 18 gray with 6MV Flattening Filter Free mode.
And the mets are pretty small, the total volume of the brain metastases is only about five cc or so. And this is the dosimetry illustration here, as you can see, very nice DVH, very conformal dose distribution, all the mets were getting the prescription dose and the average of the conformity index is very good at 1.3. The total V12 is only 12 cc with 11 mets, and the mean brain dose is 1.6 gray also is favorable. And when we look at the individual V12 as we do at the early, the largest individual V12 is only 2.2 cc, so, again, no further dose plan adjustment is made. The next case I want to show you is a case with two large brain metastases, the patient was with stage 4 non-small cell lung cancer and he was found to have two large brain metastases. One is in the brainstem, another one is resected, so it was irradiated by radiosurgery. For resected lesions, we usually give two-millimeter margin based on a Stanford experience to reduce the local failure rate.
And when we're doing these two large volume brain metastases, you can see the total volume of the lesion is quite high, about 17 cc. And even though they have good separation, there is also one lesion very, very close to the brainstem, so as a result, we decided to use a fractional regiment. So, the software also allow you to plan fractional radiosurgery, so you don't have to do everything in one delivery, you can do three, four, five, or whatever fractionation size you prefer. So, we end up using the 8 gray times 3 to the left temporal one and 7 gray times 3 to the brainstem one to observe the brainstem tolerance. And this is a four-arc plan as illustrated here. And this is the DVH, as you can see, we have an excellent conformity index of 1.2, and the brainstem 0.5 cc dose is very, very favorable at 21.5 gray and the patient was delivered with three fractions over three different days.
The last case I want to show you is one with complex distribution and this is a patient also who has brain metastases from melanoma and required a radiosurgery to all those lesions. And as you can see, he has one dominant lesion, which is quite sizable, it's about 13 cc in size, and other lesions are pretty small, usually in subcentimeter, and then some lesions are relatively close to each other but overall has reasonable separations. And this is a couple of arcs we evaluated, as you can see that when we have a very uneven distribution of the size of the lesion as well as the geographic variable in the location, the templates make a huge difference. As you can see just for the six temporary that we show you here with different arcs separation as well as the number or arcs, you look at the conformity index vary from 1.4, which is very, very favorable, to 4.2, and the total V12 vary from 77 cc to as high as 177 cc by the default template.
So, ultimately, the patient, because of the large V12 and the high number of brain metastases, we using a fractionation regimen and as everybody can see, we use Template 3 which has a good conformity index of 1.4 and the V12 cc is 77.7. And this is the illustration showing you that the distribution as well as the DVH. Because the large lesion of the brain metastasis, the V12 is 39 cc, however this patient is fractionated, so that's a very reasonable and accepted V2. So, as a summary, so we feel that radiosurgery alone is a very appropriate option for patients with multiple brain metastases and that's really being adopted globally and the treatment for patients who have multiple brain metastases with radiosurgery, we expect it to continue to rise. However, further prospective study is really needed to better understand the neurocognitive impact from radiosurgery for patients who has more than four metastases, and prospective trials will address this.
And the Multi-Mets Elements software is highly efficient for treatment planning for patients who has multiple brain metastases, it really minimized inter-operator variation. And the plan quality is highly comparable to multi-isocentric plans, as well as gamma knife plans. The delivery time is very short and that's really helped with the patient experience as well as clinical flow. And lastly, I will just show an acknowledgment for the people who help this practice available and feasible, including our neurosurgeons, our radiation oncologists, our dedicated dosimetrist and many physicists, administration support, as well as Brainlab collaboration with us. Okay. All right. So, we'll save the last few minutes to answer some of the questions.
Dr. Shi: So, we got some questions and the first one is, "Do you use fine-tuned AAA or Acuros model for this reason, and that's patient-specific QA independent calculation with Eclipse?" And the answer is the only tuned part for AAA or AXB is the source...what they call it? I forgot the term but we're using variant-recommended source size, which is for both AAA and AXB, variant published a technical bulletin, they have a study showing that what type of...either 1 millimeter or 0.5 millimeter or even 1.5 millimeter, they have a couple of choices, and they have recommendations to the source size to use in those AAA and AXB model and that is what we used. And as I mentioned in the talk that the beam model is not dedicated for small field dosimetry, it's a general radiotherapy planning system. And the next question is how long did it take us to commission the multi-mets software? And I'd say it's between two to three months and we did the in-house measurement with [inaudible 00:46:59] for measurement, that takes about one month, and then the RTSafe gel phantom is about two months because of the shipping between U.S. and Greece does take some time. And what's the next question?
Dr. Liu: So, there's a question regarding what criteria do we use to determine the plan is clinically acceptable aside from conformity and gradient indices like V12 volume, etc. So, first of all, I think each institution should have their established guideline regarding their radiosurgery practice and that's including, you know, they take into consideration their equipment and commission data. But generally speaking, I think we'll cover some of the key parameters we evaluate for the radiosurgery plan. And obviously, the majority of data to support the evaluation of the plan for clinical consequences is based on single isocenter or a very limited number of brain metastases, such as V12 and association with renal necrosis.
Generally, we will, based on the size of the individual brain metastases, recommend a clinical targets radiosurgery dose and that's also mainly based on the old RTOG study and we usually consider a 15 gray or 18 gray or 24 gray based on the size less than one-centimeter in diameter, one to two or more than three centimeter in diameter but a patient who has multiple brain metastases, we frequently slightly reduce the dose. And when we have the plan, we will run multiple templates and we'll compare the parameters of the multiple templates, we will select the plan that will give us the most ideal conformity index as well as gradient index and the plan with the lowest V12. And based on individual V12 and the volume relationship as a graph as I showed you, we have a general estimation based on individual V12, what's the combined V12 would be like if every single isocenter is optimized in the plan and we would compare that number to make sure that is in agreement.
And usually, we were looking for the conformity index and the gradient index less than 1.5 in V12, again, as we look and pay more attention to the individual V12 than the total V12 because the number of the brain metastases is heavily variable. Other parameters we'll use when a patient who would surely have a lot of brain metastases is we're looking at the mean brain dose, that is something that based on Dr. Yamamoto's experience, he recommended looking at the total joules of the radiosurgery to the brain less than 10 joules, but that is basically mean brain dose multiply the volume of the brain. So, we'll try to get a mean brain dose less than three gray...no, less than two gray and that will be considered a safe range for patient with multiple brain metastases. Okay.
Dr. Shi: So, the next question is, "The current version focuses on conformity, but the gradient between close lesions may not be optimized. So, when you see those bridges between lesions, you just ignore it or you just look at V12 gray?" So, I would say we don't ignore it, we look at V12 gray, and if the two lesions are very close , so sometimes we have to generate helping structures to encompass the whole region and calculate V12 gray by that means. And if V12 gray over our acceptable range, sometimes we reduce the prescription dose, or sometimes we have varies cases that we have to send the whole plan to Eclipse and using VMAT technology to solve it.
Dr. Shi: I just add to that, sometimes with patient who has multiple clusters of brain metastases, all the V12 isodose converged and overall V12 isodose high, we may even consider a fraction of 8 gray times three as a solution for that as well.
Dr. Liu: The next one is, "Was there a reduction in planning time or treatment delivery time using MME versus previous treatment planning system?" The answer is yes, definitely. We used Brainlab iPlan before when we are using the multi-isocenter for multiple mets as basically a serial planning. So, after 4 or 5 mets or 20 arcs, the workstation becomes really slow and that will also eventually crash after close to 30 arcs. And with this MME, as I show in the slides, that the average planning time for a case is about two to three minutes, so the planning time is really saved. And also treatment delivery time, our treatment delivery time is about 20 minutes per isocenter. So, like before, we have to separate the treatment...we have to limit the session to one hour, in other words, three mets for a patient, to be treated in one session, because otherwise, the patient may get very anxious after an hour of treatment. That is three mets. But now no matter how many mets, the treatment time will be 20 minutes or less.
Dr. Shi: So, our institution also has 4C gamma knife, and so that's also clearly very, very favorable when we're treating multiple brain metastases, it dramatically reduced treatment delivery time. You know, as you know, the gamma knife has its own planning system and so, each isocenter requires at least one shot, and if your source decay over time, the treatment time becomes very, very long and potentially non-tolerable for patients. And the planning time is also very, very equivalent compared to the gamma knife plan, relatively straightforward, but it's clearly a significant improvement in clinical efficiency and patient experience.
Dr. Liu: And also regarding the planning time, not also the calculation planning but also the contouring. For contouring like 10 or 15 mets, if we do like the old way, you have to contour slice by slice, and that really, you know, time-consuming. But with the SmartBrush function, you just give SmartBrush two views, axial or coronal or sagittal, and it automatically constructed the 3D volume for you and that's really saved a lot of time also.
Host: Great, thank you very much to all of you for your participation and for submitting questions, and a very special thank you to Drs. Shi and Liu for their lectures and for sharing their clinical experience utilizing the Brainlab Multiple Brain Mets SRS Elements with us today. As always, the webinars are recorded and they are available on the Novalis Circle websites if you'd like to rewatch them or watch them in the future. And for those of you who participated, credits are available provided that you have a successful completion of the survey following up the webinar.
The next Novalis Circle webinar will be a technical webinar. This actually will be on Friday, November 17th, and we will address some planning considerations that will help you improve the dosimetric output of the software when you have more complex cases, in particular answering questions to, "What can you do to reduce the bridging dose between lesions that are very close to each other?" If you join the webinar on Friday, we will teach you how to do this within the Brainlab SRS Element. Thank you to all for your participation again and for any follow-up questions, please feel free to use the form on the Novalis Circle, and we'll see you next on the next webinar.
Dr. Shi: Thank you, everybody, who join us on this webinar. I'm Wenyin Shi and I will kick up this program to introducing some clinical background of treating multiple brain metastases, then we will transition to discuss the Elements software, and after the presentation, we're going to go over some cases to show some examples. We don't have any disclosure for this activity. As we all aware, the incidence of brain metastases are increasing over time, there's multiple reasons that lead to this phenomenon. First of all, there's improvement of systemic therapy, so the survival of patients increased over time. Unfortunately, when a patient had longer survival, they have the risk of developing distant metastasis, including brain metastases are also increased.
And also we have an aging population and more patients were at risk. And we have imaging technology with high-resolution MRI scan, so we're able to identify much smaller brain metastases, which was usually lost anytime before. And it is really having a paradigm shift towards more focused radiation treatments, radiosurgery or fractional video surgery, for patients with brain metastases. Currently, there's high-level evidence to support using radiosurgery alone without the whole-brain for the patients who has a limited number of metastases, particularly one to four metastasis. As you can see here as a summary table, I included one of the key randomized trials indicating that there's no compromise to the overall survival if we omitting the whole-brain upfront.
More importantly, is many of the studies also addressed the new cut function and primary endpoints and it was shown very clearly, that by omitting a whole brain, there's a significant improvement of preservation or neurological function in those patients, which is clinically significant. As a result, radiosurgery is considered a high level of evidence by the NCCN guideline for patient with limited brain metastases, and Astro "Choose Wisely" initiative also included do not routinely add adjuvant whole-brain radiation treatment to stereotactic radiosurgery for patient with limited to brain metastases. There's also efforts looking at treating patient with more brain metastases with radiosurgery alone. And several studies have been shown that the number of brain metastases does not really predict outcome, but it is what's the total volume of the brain metastases.
The highest evidence using radiosurgery alone for patient with multiple metastases coming from Dr. Yamamoto's experience. In this prospective radiosurgery study, we have over 1,200 patients with 1 to 10 brain metastases, and they all were treated with radiosurgery alone with gamma knife. And this select patient who has individual lesions less than 10 cc and the total brain met volume less than 15 cc, and the prescription dose is 22 gray for small lesion and 20 gray for larger lesions. As you can see on the survival curve, compare the patients with 2 to 4 brain metastases and 5 to 10, there's no difference in survival. As a result, it supports the use of radiosurgery alone for patient with up to 10 brain metastases.
The subsequent study by Dr. Yamamoto was looking at the patient that has even more brain metastases, even more than 10 This is a case-matched study looking at patients with 2 to 9 brain metastases as well as more than 10. As a result, you can see, there's no difference in overall survival or neurological deaths free survival, so that also supports the practice of treating radiosurgery for more than 10 brain metastases. However, the neurocognition impact of using radiosurgery alone for patients with more than four brain metastases is less known. And this is a study by the same group, which using the mini-mental status, which showing that there's no deterioration, however, mini-mental status test is not as sensitive neurocognitive testing instruments.
The more comprehensive neurocog testing relies on Hopkins Verbal Learning Test, Trail Making Tests, and Controlled Oral Word Association test which has been utilized in most RTOG and RG trials. And there's also some development using computer assistant or iPad software to make this test more user-friendly, and some apps I showed here, including the Cognition by Brainlab was one of the examples. Currently, in our institution, we select patients for radiation treatment for brain metastasis in the multidisciplinary clinic. And in general, the patient who has one to four brain metastases, radiosurgery alone is the preferred option unless you have very sensitive histology such as small cell or lymphoma. And the patient who has 5 to 10 brain metastases, we still consider radiosurgery as a first choice and if they prefer, we put the patient on trial, which is an in-house trial focusing on evaluating the neurocognitive function for the patient after radiosurgery to 1 to 10 brain metastasis.
We're using the patient with one to four brain metastases as a control and compare the patient who has 5 to 10 brain metastases who receive radiosurgery and look at the impact on neurocognitive function. For patients with more than 10 brain metastases, we more routinely consider whole brain radiation treatment and in particular, hippocampus-spearing whole brain treatment if those patients are eligible. And if the patient progress after whole brain radiation treatment, then we will consider radiosurgery and those could be more than 10 metastases. And there are obviously a lot of technical challenges when we are moving from single targets or a few targets of brain metastases to multiple brain metastases.
There are multiple solutions currently being utilized in clinics including gamma-based systems, LINAC-based systems. But we can summarize this approach into two major categories, one is multi-isocenter techniques, basically, we treat each individual brain metastases individually and this will be a perfect example for gamma knife or CyberKnife approach as well as some of the LINAC approach. The drawback of this is very cumbersome delivery in the patient who has multiple brain metastases, particularly more than five, the treatment time could be long. And if we use the LINAC system, then the treatment planning complexity can be particularly challenging when we're treating multiple brain metastases with multiple isocenters.
Other solutions including mono-isocenter techniques, particularly using VMAT techniques, and it has been shown that we can achieve very similar plan quality when we compare to gamma knife plans, maybe a slightly more low dose wash in the dose less than 3 gray or so range. However, it's really highly complex inverse planning and heavily dependent on the user experience, and requires complex patient individual QA. So, this is the challenge we're facing in the clinic and the Brainlab Elements software is a perfect solution for this challenging clinical problem. And I will have Dr. Liu to take over and explain this...to introduce the software.
Dr. Liu: Hello, everyone. This is Haisong Liu and I'm going to introduce the Brainlab Elements for a multi-mets SRS planning. So, Brainlab multi-mets Elements, it's one isocenter non-coplanar dynamic arc technique, treating multiple metastases and with the current version, it can treat up to 15 mets in one session. So, this is the workflow of the multiple mets SRS and starting from import all the images into the software including CT MRI, and image fusion is basically a rigid fusion between CT and MRI and it's fully automatic, it takes seconds to fuse the CT and MRI, and the evaluation always showing very good results. And the next one is called anatomical mapping, basically, identify or automatic contour all the organs at risk including both eyes, both nerves, optic chiasm, and brainstem, and also the whole brain.
So, for the cases that we have done, all these automatic contouring really need no or very rare human intervention. And the next one is called a SmartBrush, it's their kind of tool to doing the contouring. And instead of in the traditional planning system, we have to contour slice by slice on each axial slice, the SmartBrush automatically picks the higher density in the T1 contrast MRI images and you can just draw the circles on one axial view and one of the sagittal view or coronal view, basically, two brushes that create a three-dimensional volume for you. And that is a good starting point and then you can just go through the slice and manually make revisions to make it perfect. And the next one, they call it "Object Manipulation," and it has several Boolean operations there. But most of the time, we're using this to generate margins for the smaller brain mets.
So, in our clinical practice, we usually generate a one-millimeter margin for most of the mets, but when the math is very small, for example, the one that we show here, the GTV is only 0.06 cc, so in this case, we generated 2-millimeter rounded. And then we're going into the multiple brain mets planning itself. And this is a planning interface and basically, you choose from two lists of protocols, one is the protocol determines the dose, the prescription dose. So, we have several levels, say for example here, all the mets treated to 15, 18, 21, or 24 gray, or three-level doses depend on the size or volume. But this is just a starting point, you can feel free to select any mets to change it to, say, even 17 gray or 16.5 gray. And then, the next column is the setup.
So, the setup, basically, the arc geometry to either use five table angles. So, the first one is coming with the default installation, it has five table angles starting with couch angle at 0 and 40-degree arc separation all the way to 160 table angle. So, here we listed a couple of different arc geometry. So, these is the different templates we make, we have a total of 10 of these templates and this is an example of the default template, 5 table angle, 40-degree arc separation, and this one has 6 table angles starting from 10 and ends at 170 and 32-degree arc separation. The most complicated, we have 9 arc, 9 table angles, and 20-degree arc separation. So, after we select the dose and the arc template, we just click "Calculation" and everything here happened in the background.
So, the software will automatically select the isocenter based on the center of mass of all selected PTVs and use the pre-configured table angles and arc lengths to calculate which arc to select. So, each arc may deliver a dose to a subgroup of metastasis, so if two mets lined up in the same direction of the leaf motion, then the software will add another arc in the reverse direction treating the other met in line. So, this concept leads to optimized conformal treatment combined with low dose spread to normal brain tissue. ANd MLC margins, the MLC will conform to each individual PTV with up to one-millimeter margin, and it can vary between zero to one and it also automatically select which PTV to treat for each arc. So, after the selection matched with MLC with each individual met, there's the weight optimization of each arc to achieve the best conformality index, it's based on the inverse Paddick calculation for conformity index, and then the final dose calculation based on pencil beam algorithm.
A planning time for a simple case like three or four mets, it should take about 1 minute and 10 seconds, and for the most complicated 11 or 12 mets with 7 arcs, this may take up to four minutes. So, we're going to show you several planning cases and how we evaluate the plan in the next...in the last slides, but I'm going to jump into the patient QA part. So, for patients-specific QA, we don't do patient-specific measurements, rather we do an independent calculation with independent software, which is Eclipse planning system. So, after the MME planning, DICOM export CT, structure, and plan to Eclipse, and then we will recalculate the plan with both AAA and the Acuros dose grid with one-millimeter dose grid. And then we compare the mean dose of each PTV, so for our first 56 patients and 233 targets, the percentage difference between the MME and AAA is 2.7% and between MME and Acuros is 3%.
And also, we observe that the smaller the PTV volume, the larger the percentage difference. And our Eclipse calculation or Eclipse planning system is not configured for small fields, so our smallest output factors PTV and dose profile end at the field size three centimeter by three centimeter. So, that's why when we go to the smaller fields, we have larger discrepancies and that's our assumption. So, during the summer, we also work with Mobius. We sent a couple of our cases to Mobius and they use their Mobius3D and/or Mobius Fraction Check with our Truebeam trajectory files to calculate the TPS calculated dose and Mobius calculated dose and also the real delivered dose coming from the projection calculation from the Truebeam trajectory file. And one of the good things I like about this software is other than the windows, it also provides us 90% coverage or even probably 95% coverage as you defined, so that is a more meaningful comparison.
So, for the other several tests that we did for initial commissioning, one is the hidden target test. We already do a hidden target test for the BB and isocenter, but for this one, we put the BB at the off-center location. In this case, it's almost nine centimeters in 3D, far, far away from the isocenter. And then we treat the Rando phantom with patient treatment protocol and IGRT tolerance at the same as we treat the patient, and then we acquire EPID images at each start and stop angle if available. So, in this example, this is arc number one, gantry at 270, table at 0. So, you can see this is the intended-to-treat, the BB here, and this is the EPID image we got, and really, it's at the center of this MLC aperture. And this is arc number two with a table at 70, gantry at 160, and you can see the other two mets are here and this is the BB here, and that's still at the center of this MLC aperture. And for quantitative analysis, we import all these...we acquire actually a total of 10 EPID portal in this test case, and we import them into the RIT analysis software.
And you can see the results, on the lateral direction, we have two parts or one part over one millimeter, and on the longitudinal, we have two parts a little bit over one millimeter, and the other part is within plus-minus one millimeter. And another test, we use our Delta4 Phantom which is a 3D diode array phantom. We intentionally put all 10 mets on the tiles and this is the...these tiles have 10 mets, range from 1 centimeter to 1.3-centimeter in diameter, 0.3 to 0.4 cc, and we use 6MV Flattening Filter Free mode to treat these 10 mets. And the middle picture is showing the dose profile along the blue lines here and you can see the red line is the planned dose and the green dots is a diode measurement, so it's matched with very good results.
And for a 3% dose deviation, one millimeter DTA, we get a gamma passing rate of 99% and that's based on almost 400 diodes. So, we also did a 3D gel phantom and this whole process, we did with a Greece company called RTSafe. We sent anonymized patient CT to them and they build this skull and internal bony structures phantom using 3D printer, filled it with gel, and sent it back to us. So, we treat the gel phantom as if treating a patient and then use their provided protocol to do a post-irradiation MRI and they will provide all the analysis for us. So, this is the plan that we make, there's nine targets planned, the smallest volume is 0.1 cc and the largest is 0.5 cc, and 24 gray prescription, about 30 gray at each target center.
And this is the compare between the planning CT versus the post-irradiation MRI, and after the fusion, we can see that the colored circle here are the intended-to-treat targets and the black hole is what we irradiated. And RTSafe provide us the dose profile on one dimensional and the two-dimensional gamma index map with isodose overlay. And cognitively, we have about 0.2 to 1.1-millimeter three-dimensional targeting displacement for the nine targets, and D95 for the nine targets range from minus 2 to 6.5%, and the gamma passing rate is 99%. So, Dr. Shi is going to continue with our current clinical experience and plan evaluations.
Dr. Shi: Since the implementation of this Multi-Mets Elements in our institution probably a little bit over two years ago, this is quickly being embraced and we have a very, very steady flow of patients. And currently, in the past about two years or so, we have already treated over 100 patients using the software. For patients with single brain metastases, we just using the standard Brainlab iPlan system to plan. As you can see, based on our treatment guidelines, the majority of our patients are still patient with limited brain metastases, two to three brain metastases, comprise about 50% of our total volume, and we also have about 40% of patients who have between 4 to 10 brain metastases. The only very select patients that are getting radiosurgery are 10 or more brain metastases, usually for recurrence or patient who has a highly resistant tumor and without interruption for chemotherapy for other consideration.
This is just our initial evaluation for our overall outcome, local control, and intracranial control rate. As you see, we have close about 90% local control rate at six months, and we have not really seen any high-grade toxicity, and those numbers without digging into more detail about histology and GPA are in par with historical data. And I want to go through some of the planning philosophy we have here in our institution and then I will show a few cases, hopefully, to illustrate those points and make it more meaningful to you. So, first of all, currently, all the MME software came with a default template as Dr. Liu already mentioned, which is the 40-degree arc separation with 5 table angle, and this is as shown on the top right...top left-hand side.
And then we quickly learned that, you know, using one template is not sufficient and one template is not necessary the optimized template to use because the distribution, the number, and the volume of the brain metastases vary greatly from one patient to the other. As a result, we develop multiple custom templates using different table angle as well as arc separation, and we run all those different templates for each case and we select cases based on dosimetry, the ideal dosimetry. As you can see, here are the list of the 10 currently tested templates and there are a different number of arcs from 5 to 9 arcs and arc separation anywhere from 40-degree to 20-degree. And then after 100 or so cases we treated, as you can see, this is the history of distribution of which arc we would end up picking as results based on the idea of dosimetry.
So, it's relatively evenly distributed with most of the arcs...or template we picked are using around 6 table angle, between 28 to 35-degree separation as highlighted in the yellow. And the default template does being used as well but we will also see that if we using more table angles such as 7 to 9 table angle or even 10 table angle, we not necessarily getting better results, it's a diminishing return, and then when we running complex cases, more table angle doesn't necessarily help and it stresses out the planning system sometimes. And another key parameter we looking at the dose is V12. We know that the V12 is linearly related to the individual brain metastasis and the volume and we have a quick reference point to give us an idea if a brain metastasis...single brain metastasis of a certain volume, which roughly V12 range will it be to help us to evaluate a plan.
And the software also gives you pretty user-friendly features based on QVH, you can look at either each individual V12, which here as example in the highlighted area, you can see the individual V12 is 2.6 cc, and we evaluate each individual V12 and in a cluster metastasis, we'll also make sure the V12 isodose line does not converge or the converged isodose lines still have a limited number...a limited volume of V12 total volume. And this is one other example, we use individual V12 to help us with dose prescription. This is a patient with four metastases with two rather large brain metastases, 5 cc and 10 cc. And if we give 21 gray and 18 gray, both of them have a V12 number in the range of 10 cc, and given that there are four brain metastases in total, we feel that it is too high a V12 volume, so we reduce the prescription dose to 18 gray and 15 gray, that bring down the V12 to 6.7 and 7 cc, which is more clinically appropriate.
And this software also gives us a very well-detailed parameter of each lesion. For example, here, as you can see nine brain metastases, we will have the volume of the metastasis, maximum dose, minimum dose, and as well as the mean dose, and the conformity index. And they also help us to further evaluate if something that's outlined that we can potentially individually evaluate if something needs to be, you know, adjusted. And based on our current experience, we look at all our brain metastases plan with Elements software as listed in the first column, you can see the conformity index is quite much in bar as the VMAT plan or gamma knife plan based on our own institutional experience published in the prior paper. And the gradient index is actually in favor compared to VMAT planning and very comparable to the gamma knife plan. The isodose line is very similar to what we usually do with a single target, around 80% isodose line.
One other important feature or issue we need to recognize is the off-access effects. And this is a paper that's published in 2015 by Dr. Roper, they look at the impact of the geographic location of the metastasis based on the distance from the isocenter. As you can see here, the further away from the isocenter, the larger the displacement if there's a rotational or translational error. And if we do some simple calculation at the table here, if there's a 5 degree rotation error...0.5 degree rotation error, sorry, then the displacement at 7 cm from isocenter will be about 0.6 millimeter, and this number can increase beyond 1 to even 2-millimeter if the rotation error is more than 1 degree. So, currently, based on this calculation, we try to restrict the lesion treated more than seven centimeters away from the isocenter because the five-millimeter rotation error can lead to more than 0.5 millimeter transition disagreements.
And so, if we have multiple brain metastases and some of the lesion is more than seven centimeters, we'd prefer to use more than one isocenter to keep everything less than seven centimeters away from the isocenter. And so, this is our current planning procedure and we will consider using two isocenters if there's lesions more than seven centimeters away from the isocenters. And we usually do a one-millimeter margin for regular targets, but something that is very small, less than five-millimeter or four-millimeter in diameter, we may consider adding about a two-millimeter margin. And we generate treatment plans with different customized templates and based on the arc geometry, we will select the best plan based on the V12, dose to OAR, and the conformity index as well as the mean dose. So, right now, we run all the 10 templates for each individual plan and we compare all the parameters and we'll treat...we will choose the plan that gives us the best dosimetry.
And regarding the treatment delivery, in our institution, we use ExacTrac as IGRT solution, so the residual error for the IGRT setup is 0.5-millimeter, that should be a degree in rotation, and 0.5 millimeter in translational error. And we will reimage with ExacTrac each arc angle and if the error is exceeding our cutoff, which is 0.5 degree and 0.5 millimeter, we will apply the shifts and reimaging with ExacTrac to verify. And in the next couple of minutes, I will show you a few clinical cases and really want to just illustrate some of the points I just discussed. And first of all, this is a patient who has renal cell carcinoma, unfortunately, develops multiple brain metastases, and he has more than 10 metastases both supratentorial and infratentorial and involving both hemispheres of the brain.
Given the heavy disease burden and a high number of brain metastases, we recommend whole brain radiation treatment with hippocampus sparing techniques. So, the patient received the whole brain 30 gray in 10 fractions. Unfortunately, when she returned for follow-up after finishing the whole brain, quite a few of the lesions already progressed, so the whole brain, unfortunately, does not have a meaningful benefit for this patient who has renal cell carcinoma back, which is not surprising because the resistance nature of that. So, we get the patient for reduced surgery salvage, and as you can see, there's further 12 brain metastases and the volume is anywhere between 0.2 to 4.2 cc, and the total volume is around 15 cc.
And there's some unfavorable feature of that, one of the patient...one of the lesion is close to the optic chiasm and overall, these brain metastases have an average separation, means they're not heavily clustered but some of them are relatively close to each other. And we use a uniform prescription of 18 gray to all those lesions and we're using a 6MV Flattening Filter Free delivery mode. As you can see, I've just shown an example of the six templates, we've added actually...you know, as you can see, based on the conformality index, V12, and the chiasm max dose, we ultimately choose template number four, which give us the best conformity index, the lowest V12, and keep the chiasm maximum dose within our tolerance, which is at 10 gray in our institution. And this is the ultimate plan the patient was treated with.
And this is a summary of dosimetry using, you know, 3D illustration of the arc separation and the DBH graph as well as the isodose line. And as you can see, we have a conformal plan and there are very tight V12 clouds. And we also look at individual V12 as I mentioned earlier to see if any individual lesion or clustered brain metastasis dose need to be adjusted, and our largest brain metastasis V12 for the individual lesion is only 7.3 cc, which is reasonably well within our clinical tolerance for the patient and we receive the treatment without further plan adjustment, 18 gray to all 12 lesions in single isocenter.
And this is the post-treatment follow-up MRI scan. As you can see, two months later when he returned, we can see very marked improvements of all the brain metastasis treated, some achieve even ACR as shown here, and some of the large lesion is significantly reduced in size. So, it's a very, very favorable clinical outcome with no toxicity. And the second case is a patient with melanoma with multiple brain metastases. This patient has 11 brain metastases, they are relatively well-distributed, which means there were good separation, there's no clustered lesion. And this patient's number of lesions is slightly above our usual cut-off of 10 brain metastases, however, his melanoma histology was highly radioresistance, and also resistant to systemic. Given the burden that requires the initiation of systemic therapy quickly, we ultimately agreed to give him radiosurgery and we also treat every single one with 18 gray with 6MV Flattening Filter Free mode.
And the mets are pretty small, the total volume of the brain metastases is only about five cc or so. And this is the dosimetry illustration here, as you can see, very nice DVH, very conformal dose distribution, all the mets were getting the prescription dose and the average of the conformity index is very good at 1.3. The total V12 is only 12 cc with 11 mets, and the mean brain dose is 1.6 gray also is favorable. And when we look at the individual V12 as we do at the early, the largest individual V12 is only 2.2 cc, so, again, no further dose plan adjustment is made. The next case I want to show you is a case with two large brain metastases, the patient was with stage 4 non-small cell lung cancer and he was found to have two large brain metastases. One is in the brainstem, another one is resected, so it was irradiated by radiosurgery. For resected lesions, we usually give two-millimeter margin based on a Stanford experience to reduce the local failure rate.
And when we're doing these two large volume brain metastases, you can see the total volume of the lesion is quite high, about 17 cc. And even though they have good separation, there is also one lesion very, very close to the brainstem, so as a result, we decided to use a fractional regiment. So, the software also allow you to plan fractional radiosurgery, so you don't have to do everything in one delivery, you can do three, four, five, or whatever fractionation size you prefer. So, we end up using the 8 gray times 3 to the left temporal one and 7 gray times 3 to the brainstem one to observe the brainstem tolerance. And this is a four-arc plan as illustrated here. And this is the DVH, as you can see, we have an excellent conformity index of 1.2, and the brainstem 0.5 cc dose is very, very favorable at 21.5 gray and the patient was delivered with three fractions over three different days.
The last case I want to show you is one with complex distribution and this is a patient also who has brain metastases from melanoma and required a radiosurgery to all those lesions. And as you can see, he has one dominant lesion, which is quite sizable, it's about 13 cc in size, and other lesions are pretty small, usually in subcentimeter, and then some lesions are relatively close to each other but overall has reasonable separations. And this is a couple of arcs we evaluated, as you can see that when we have a very uneven distribution of the size of the lesion as well as the geographic variable in the location, the templates make a huge difference. As you can see just for the six temporary that we show you here with different arcs separation as well as the number or arcs, you look at the conformity index vary from 1.4, which is very, very favorable, to 4.2, and the total V12 vary from 77 cc to as high as 177 cc by the default template.
So, ultimately, the patient, because of the large V12 and the high number of brain metastases, we using a fractionation regimen and as everybody can see, we use Template 3 which has a good conformity index of 1.4 and the V12 cc is 77.7. And this is the illustration showing you that the distribution as well as the DVH. Because the large lesion of the brain metastasis, the V12 is 39 cc, however this patient is fractionated, so that's a very reasonable and accepted V2. So, as a summary, so we feel that radiosurgery alone is a very appropriate option for patients with multiple brain metastases and that's really being adopted globally and the treatment for patients who have multiple brain metastases with radiosurgery, we expect it to continue to rise. However, further prospective study is really needed to better understand the neurocognitive impact from radiosurgery for patients who has more than four metastases, and prospective trials will address this.
And the Multi-Mets Elements software is highly efficient for treatment planning for patients who has multiple brain metastases, it really minimized inter-operator variation. And the plan quality is highly comparable to multi-isocentric plans, as well as gamma knife plans. The delivery time is very short and that's really helped with the patient experience as well as clinical flow. And lastly, I will just show an acknowledgment for the people who help this practice available and feasible, including our neurosurgeons, our radiation oncologists, our dedicated dosimetrist and many physicists, administration support, as well as Brainlab collaboration with us. Okay. All right. So, we'll save the last few minutes to answer some of the questions.
Dr. Shi: So, we got some questions and the first one is, "Do you use fine-tuned AAA or Acuros model for this reason, and that's patient-specific QA independent calculation with Eclipse?" And the answer is the only tuned part for AAA or AXB is the source...what they call it? I forgot the term but we're using variant-recommended source size, which is for both AAA and AXB, variant published a technical bulletin, they have a study showing that what type of...either 1 millimeter or 0.5 millimeter or even 1.5 millimeter, they have a couple of choices, and they have recommendations to the source size to use in those AAA and AXB model and that is what we used. And as I mentioned in the talk that the beam model is not dedicated for small field dosimetry, it's a general radiotherapy planning system. And the next question is how long did it take us to commission the multi-mets software? And I'd say it's between two to three months and we did the in-house measurement with [inaudible 00:46:59] for measurement, that takes about one month, and then the RTSafe gel phantom is about two months because of the shipping between U.S. and Greece does take some time. And what's the next question?
Dr. Liu: So, there's a question regarding what criteria do we use to determine the plan is clinically acceptable aside from conformity and gradient indices like V12 volume, etc. So, first of all, I think each institution should have their established guideline regarding their radiosurgery practice and that's including, you know, they take into consideration their equipment and commission data. But generally speaking, I think we'll cover some of the key parameters we evaluate for the radiosurgery plan. And obviously, the majority of data to support the evaluation of the plan for clinical consequences is based on single isocenter or a very limited number of brain metastases, such as V12 and association with renal necrosis.
Generally, we will, based on the size of the individual brain metastases, recommend a clinical targets radiosurgery dose and that's also mainly based on the old RTOG study and we usually consider a 15 gray or 18 gray or 24 gray based on the size less than one-centimeter in diameter, one to two or more than three centimeter in diameter but a patient who has multiple brain metastases, we frequently slightly reduce the dose. And when we have the plan, we will run multiple templates and we'll compare the parameters of the multiple templates, we will select the plan that will give us the most ideal conformity index as well as gradient index and the plan with the lowest V12. And based on individual V12 and the volume relationship as a graph as I showed you, we have a general estimation based on individual V12, what's the combined V12 would be like if every single isocenter is optimized in the plan and we would compare that number to make sure that is in agreement.
And usually, we were looking for the conformity index and the gradient index less than 1.5 in V12, again, as we look and pay more attention to the individual V12 than the total V12 because the number of the brain metastases is heavily variable. Other parameters we'll use when a patient who would surely have a lot of brain metastases is we're looking at the mean brain dose, that is something that based on Dr. Yamamoto's experience, he recommended looking at the total joules of the radiosurgery to the brain less than 10 joules, but that is basically mean brain dose multiply the volume of the brain. So, we'll try to get a mean brain dose less than three gray...no, less than two gray and that will be considered a safe range for patient with multiple brain metastases. Okay.
Dr. Shi: So, the next question is, "The current version focuses on conformity, but the gradient between close lesions may not be optimized. So, when you see those bridges between lesions, you just ignore it or you just look at V12 gray?" So, I would say we don't ignore it, we look at V12 gray, and if the two lesions are very close , so sometimes we have to generate helping structures to encompass the whole region and calculate V12 gray by that means. And if V12 gray over our acceptable range, sometimes we reduce the prescription dose, or sometimes we have varies cases that we have to send the whole plan to Eclipse and using VMAT technology to solve it.
Dr. Shi: I just add to that, sometimes with patient who has multiple clusters of brain metastases, all the V12 isodose converged and overall V12 isodose high, we may even consider a fraction of 8 gray times three as a solution for that as well.
Dr. Liu: The next one is, "Was there a reduction in planning time or treatment delivery time using MME versus previous treatment planning system?" The answer is yes, definitely. We used Brainlab iPlan before when we are using the multi-isocenter for multiple mets as basically a serial planning. So, after 4 or 5 mets or 20 arcs, the workstation becomes really slow and that will also eventually crash after close to 30 arcs. And with this MME, as I show in the slides, that the average planning time for a case is about two to three minutes, so the planning time is really saved. And also treatment delivery time, our treatment delivery time is about 20 minutes per isocenter. So, like before, we have to separate the treatment...we have to limit the session to one hour, in other words, three mets for a patient, to be treated in one session, because otherwise, the patient may get very anxious after an hour of treatment. That is three mets. But now no matter how many mets, the treatment time will be 20 minutes or less.
Dr. Shi: So, our institution also has 4C gamma knife, and so that's also clearly very, very favorable when we're treating multiple brain metastases, it dramatically reduced treatment delivery time. You know, as you know, the gamma knife has its own planning system and so, each isocenter requires at least one shot, and if your source decay over time, the treatment time becomes very, very long and potentially non-tolerable for patients. And the planning time is also very, very equivalent compared to the gamma knife plan, relatively straightforward, but it's clearly a significant improvement in clinical efficiency and patient experience.
Dr. Liu: And also regarding the planning time, not also the calculation planning but also the contouring. For contouring like 10 or 15 mets, if we do like the old way, you have to contour slice by slice, and that really, you know, time-consuming. But with the SmartBrush function, you just give SmartBrush two views, axial or coronal or sagittal, and it automatically constructed the 3D volume for you and that's really saved a lot of time also.
Host: Great, thank you very much to all of you for your participation and for submitting questions, and a very special thank you to Drs. Shi and Liu for their lectures and for sharing their clinical experience utilizing the Brainlab Multiple Brain Mets SRS Elements with us today. As always, the webinars are recorded and they are available on the Novalis Circle websites if you'd like to rewatch them or watch them in the future. And for those of you who participated, credits are available provided that you have a successful completion of the survey following up the webinar.
The next Novalis Circle webinar will be a technical webinar. This actually will be on Friday, November 17th, and we will address some planning considerations that will help you improve the dosimetric output of the software when you have more complex cases, in particular answering questions to, "What can you do to reduce the bridging dose between lesions that are very close to each other?" If you join the webinar on Friday, we will teach you how to do this within the Brainlab SRS Element. Thank you to all for your participation again and for any follow-up questions, please feel free to use the form on the Novalis Circle, and we'll see you next on the next webinar.