Transcript
Thank you for the invitation to Brainlab. And thank you for the nice introduction, and I really have to say thank you to Maximillian because he gave an excellent introduction to my talk. And I will just pick your last beware as more or less the idea behind what I would like to talk about. So, yeah, here is my little disclosure. So the title of the talk is more or less Complex Multifocal Planning. And so, therefore, first of all, we need to know or need to make up our minds what is considered a complex planning situation or what do you want to talk about? So, first of all, we could think about a large number of lesions, so the software from Brainlab itself, it's right now possible to treat 15 metastases in a row, if you talk about the multiple brain metastases element. So a large number of lesions could be something more than 15 maybe, or even 10 or so. And then we should talk about if it's clustered. So if all the lesions or most of the lesions are in the cerebellum, for example, and some are somewhere else. So, to make a long story short, the idea behind large numbers of lesion would be to treat them maybe with two plans, with two isocenters or so, and take one of the clusters within one plan and the other one in another plan or so.
Another good idea would be, at that point, maybe just go for something like, not use just five arcs. So within your template, you heard already that it's template-based. So just enter another one. Do six arcs or so and then get the system another degree of freedom at that point. So, that's all I want to say about that.
So because I want more or less focus on different prescription regimes within one case, for example. So what if you got 10 lesions or so, and you wanted to go for a fractionated regime for one of these lesions and you want to treat all the other ones in a single shot? At that point, you get into trouble when you want to use it within one plan because the software itself only allows you to use the same fractionation regime for all lesions you want to treat at that point. Same situations, maybe if you jump over to different indications. Like you got multiple brain metastases and you got a cavity that might be there just because of surgery of larger metastases or whatever, or you got a primary tumor as well there and you want to do something for that. So this will more or less end up in the same situation that you get a fractionated regime where you want to do some hypofractionation and maybe some single-shot situations like Maximilian already said. So we need to be aware what to do normally. So and this is more or less what I want to talk about.
One of our case reports, so we did a couple of patients in this way. I think we had roughly 120, 130 patients treated with multiple brain metastases, 30 to 40 treated with cranial SRS, and now we have already a couple of patients treated with both systems together. So, where we got a fractionation and we got also some single-shot situation, and this will be my case report more or less. So we got this patient, 56-year-old, when he came to our department, he had a lung tumor at that point. And in July 2017, they found six brain metastases at follow-up or at an MRI because he had had problems. And they figured out that there was one large lesion, roughly 3 by 3 by 3 centimeters. So really, really large lesion. And the neurosurgeons in our department then decided to take that out. So they did a resection of the large lesion. You can see it down here. So this is the cavity that's been left, and here are some of the other metastases. So you see, one is pretty near to the brainstem, and there are some others here, just some of these.
And then the patient came to the Interdisciplinary Tumor Board in our university medical center and we start to discuss what to do with this patient. And at that point, we thought, "Okay, let's go for something like, we would like to do single-shot stereotactic radiosurgery for the smaller lesions, so for five of the lesions. And we would like to do something like a hypofractionation, 7 times 5 gray for the cavity itself." So and then at that point, we figured out, "Okay, how should we do it?" Normally we do multiple brain metastases using the Brainlab element, but this situation is a little bit tricky so we cannot do it at once. We need to mix some kind of software at that point. So we decided we will do multiple brain metastases for five of the lesions, and then we will do hypofractionation of the cavity afterwards.
So, we then decided to do something like that, but at that point, you need to think about what Maximillian had said already. How can you really be aware of the dose you would give to the other lesions? For example, to the cavity by just using dynamic conformal arc? So you will end up with some dose that you would like to give to the other lesions that will also will go to the cavity at that point and you will also get some scattered dose there. And you have nothing that's really, really defined. So you find a solution where you got some dose there, but not exactly what you want. And how can you take that into account?
And at that point, we figured out, okay, what would be a good approach, and what I would like to talk about is we treated the five lesions, we wanted to give single-shot SRS with multiple brain metastases and we included at that situation already the cavity within the treatment plan. So the first session of the hypofractionation will be part of the single-shot as the rest. So we planned already 5 gray to the resection cavity in this situation, so we only had to do a follow-up or a new plan using cranial for six fractions because that was all what was left at that point. We were able to give a really nicely-defined dose to the cavity, so that we have a good situation where we already have a dose there. We do not give something additionally or so we might get in trouble with side effects or so. So we had something really nicely predefined already. So you can see here the dose given to the other lesions. At that point, we already figured out, "Oh my God, one of the lesions is pretty close to the brainstem, so we can't give the 20 gray we would like to." We just give 20 gray at that point just to be sure that we do not give too much dose to the brainstem.
So the situation then came out with 4 lesions getting 23, 1 lesion, 20 gray, and 5 gray to the resection cavity already within this. We used a template that came up with five table angles in total at that point, just to give you an idea about that. And so in this situation, you could see the 5-gray isodose line, now is the purple line that you see. It goes nicely around the cavity. So we end up with a nice dose of 5 gray in the cavity already, and we could do a good summation at that point and add the next plan, to be honest, and get a good result.
So, here are the DVHs just to give you an idea. So the purple line down here or the orange line here is the cavity. So you see we end up with something like 5 gray for the cavity. Here are the other lesions, the one that I only gave 20...we only gave 20 gray to. So you see nice dose conformity here at this point. We end up with an overall conformity index of 1.4 in this situation, but the cavity was excluded because that wouldn't make sense at all because that's not the typical lesion would be treated in this way. Just to give you an idea about the 12 gray, and the low dose, and this situation here.
And then we added the next plan. So we added up the cranial plan, cranial SRS in this case. This is the VMAT approach. That was not really part of this, of the session, so it's a completely VMAT approach in this case. So there will be no thing like dose-escalation like in multiple brain metastases. So it's more or less the ICLU-conformable dose prescription. And you see the DVH up here, you see the nice 5-gray dose path here given to the cavity. And what we as well did in this situation, so on the first side, we wanted not to give a not defined dose to the cavity during the single-shot SRS treatment. And on the second part of this so-called hybrid plan, so on the cranial plan, we do not want to add additionally dose to the already-treated SRS lesions. So we introduced them as organs at risk to make it easy and simple, and we just said, "Okay, maximum dose will be 1 gray that we would allow in this situation." And then we did the plan. We did the VMAT approach. We used a single arc in this situation, just table angle 0, just a full arc, just to be sure that we do not use the same arcs that we used before more or less not to get there into trouble or so. And to give it nice and clean. And you see that we ended up here with...here's the DVH for one of the lesions, firstly treated lesions, and we ended up with 0.74 gray in this situation.
So this all took place in August 2017 up to September 2017 with the SRS, with the SRT. And the patient was for follow-up already. All our patients got a follow-up after every three months using MR images at that point. And the last follow-up of this patient was in March, just one month ago now. And as you can see, so the disease is controlled. Here is also a little image of the cavity and the disease control. Patient is doing well.
So, but if we start to treat something like that, and this is already important for situations where you only use one single isocenter, we need to make up our minds what is about misalignment. So in the old days, one isocenter per lesion was always...in the isocenter, that was a good idea, that was easy. Misalignment wouldn't make too much at that point if we had a little margin, maybe, for instance. But now we are dealing with multiple lesions. None of these lesions is covered by the isocenter, normally, because the isocenter is positioned in the center of mass of all lesions. So we need to make up our minds, what happens to a misalignment.
So I fooled around a little bit with the DICOM data of this patient, here in this case. And we introduced some rotational and also translational errors. I just wanted to show you the rotational errors just to give you an idea, one of those lesions here. And you see, if we introduce 2 degrees of rotation and pitch and roll, or even just 1 degree of pitch and roll error, we will already start to do a big under-dosage to the lesion and also an overdosage to the healthy tissues surrounding this.
So we need to think about how can we deal with this and how can we get rid of these problems. And so we have to introduce some imaging solutions, IGRS. So here are the three that came up to my mind. Maybe you could also add Calypso, but Calypso, it's not really an imaging device. So I just kicked it out of my list. It was in there before when I thought about image guidance solution. So first of all, we got the ExacTrac image system, we all know that. So 2D image pair. We have it here. So orthogonal images, room-based system, so it's completely independent of gantry and couch position. So you can do images wherever you want at every thinkable angle. We got Cone Beam CT, which is more or less the gold standard for imaging at that point. This is really 3D image set, so this is perfect. But the worst part of this, it's Linac-based.
So we got the kV source and the imager also on this machine if you look here, fixed to the gantry. And if we change the table angle, we get into trouble because we can only do Cone Beam CTs for coplanar angles. So the system for SRS is completely out if you really want to compensate for translational errors and so on, and you have to compensate for that.
So another option at that point would be the Optical Surface Monitoring System, OSMS. So this really monitors the whole surface of the patient. So it will be monitored, will be projected as speckled pattern onto the patient. And then you got three-camera pods here in the room, so you can always see what happens. You can then compensate for problems there. You have to use an open mask. So what is monitored is not really the head, so it's more or less the part of the open mask. Otherwise, you will only look on the mask and you won't see any movement at that point. System is also room-based like the ExacTrac system, so it also can do images at every given couch angle. But as you have to use an open mask, you will introduce already some degrees of freedom for the patient just by opening the mask. So it will not be that strict like the Brainlab mask, and so on.
So, what I found out is that it might not be as accurate as the ExacTrac system by looking at literature. Nevertheless, it's also very fast, but I mean, you have to open up the mask at that point. So, what I then looked at, is the ExacTrac system, at this point, possible to check or to correct for those problems. And the good point is there's this publication out there. Yes, it is, compared to the gold standard, to the Cone Beam CT. You see after ExacTrac correction when you introduce 1 degree of roll and pitch error where you end up with 2.3 millimeters discrepancy, you will end up with 0.6 millimeters already.
So the last couple of slides will be on... So we have complex planning situations. We have IGRS requirements, so we have to think about statistic evaluation and maybe the risk of radionecrosis we had already. And there's this paper out there published in 2010 and says there is. Yeah, you have to consider the volume given, 10 gray, of the healthy brain, of the whole brain when you want to be sure that you do not get radionecrosis. And there is this paper and it says, 10.5 cc should be below V10 gray for the whole brain to be sure that you won't get any problems. But the problem with this is that it's only for single-shot or single lesions. Nowadays we're dealing with multiple lesions, so we have no real number at that point. And so we came up in our department and just took data of our first 50 patients. So we ended up with adding the total PTV volume here versus the volume of the brain receiving 10 gray, and we ended up with this nice linear thing here. And nowadays, whenever we treat a patient, we just plot in the volume we would have for the total brain, or for the total PTV volume. And we saw at which volume we would end up because we already treated, as I said, 120 patients. We do not see any radionecrosis more than we had before. So, and then we just have also a nice tool to see if our applying quality is quite nice because everything is done automatically.
And if you're not so familiar with this, the system came up with a plan and you have no idea, is it a good plan or not? So at this point, we came up with this. And then we plot in the volume and we see, okay, we got some...we had only 9 cc or 6 cc for the PTV and we end up with 50 cc here for the brain. Then we see, okay, we can do better at that point, then we start to use another template, and so on.
The same we did for the gradient index here. So we also see how the gradient index looks in our department just to have some kind of quality tools. So you should have something like that, otherwise, you have no idea if you're doing really good, or if... I mean, you will do good always with the software, but you won't be able to see if you can do better. So, and that's what we want to do, we get better results. And the result of that, what happened in our department is you can see our gradient index is 4.36 nowadays. And we got in the mean, 1.39 conformity index. Compared to literature published just one year ago, you see we're doing a lot better just by looking at these numbers and then just thinking about how to deal with that. So this is just my conclusion. And yeah, thank you for your attention.
Another good idea would be, at that point, maybe just go for something like, not use just five arcs. So within your template, you heard already that it's template-based. So just enter another one. Do six arcs or so and then get the system another degree of freedom at that point. So, that's all I want to say about that.
So because I want more or less focus on different prescription regimes within one case, for example. So what if you got 10 lesions or so, and you wanted to go for a fractionated regime for one of these lesions and you want to treat all the other ones in a single shot? At that point, you get into trouble when you want to use it within one plan because the software itself only allows you to use the same fractionation regime for all lesions you want to treat at that point. Same situations, maybe if you jump over to different indications. Like you got multiple brain metastases and you got a cavity that might be there just because of surgery of larger metastases or whatever, or you got a primary tumor as well there and you want to do something for that. So this will more or less end up in the same situation that you get a fractionated regime where you want to do some hypofractionation and maybe some single-shot situations like Maximilian already said. So we need to be aware what to do normally. So and this is more or less what I want to talk about.
One of our case reports, so we did a couple of patients in this way. I think we had roughly 120, 130 patients treated with multiple brain metastases, 30 to 40 treated with cranial SRS, and now we have already a couple of patients treated with both systems together. So, where we got a fractionation and we got also some single-shot situation, and this will be my case report more or less. So we got this patient, 56-year-old, when he came to our department, he had a lung tumor at that point. And in July 2017, they found six brain metastases at follow-up or at an MRI because he had had problems. And they figured out that there was one large lesion, roughly 3 by 3 by 3 centimeters. So really, really large lesion. And the neurosurgeons in our department then decided to take that out. So they did a resection of the large lesion. You can see it down here. So this is the cavity that's been left, and here are some of the other metastases. So you see, one is pretty near to the brainstem, and there are some others here, just some of these.
And then the patient came to the Interdisciplinary Tumor Board in our university medical center and we start to discuss what to do with this patient. And at that point, we thought, "Okay, let's go for something like, we would like to do single-shot stereotactic radiosurgery for the smaller lesions, so for five of the lesions. And we would like to do something like a hypofractionation, 7 times 5 gray for the cavity itself." So and then at that point, we figured out, "Okay, how should we do it?" Normally we do multiple brain metastases using the Brainlab element, but this situation is a little bit tricky so we cannot do it at once. We need to mix some kind of software at that point. So we decided we will do multiple brain metastases for five of the lesions, and then we will do hypofractionation of the cavity afterwards.
So, we then decided to do something like that, but at that point, you need to think about what Maximillian had said already. How can you really be aware of the dose you would give to the other lesions? For example, to the cavity by just using dynamic conformal arc? So you will end up with some dose that you would like to give to the other lesions that will also will go to the cavity at that point and you will also get some scattered dose there. And you have nothing that's really, really defined. So you find a solution where you got some dose there, but not exactly what you want. And how can you take that into account?
And at that point, we figured out, okay, what would be a good approach, and what I would like to talk about is we treated the five lesions, we wanted to give single-shot SRS with multiple brain metastases and we included at that situation already the cavity within the treatment plan. So the first session of the hypofractionation will be part of the single-shot as the rest. So we planned already 5 gray to the resection cavity in this situation, so we only had to do a follow-up or a new plan using cranial for six fractions because that was all what was left at that point. We were able to give a really nicely-defined dose to the cavity, so that we have a good situation where we already have a dose there. We do not give something additionally or so we might get in trouble with side effects or so. So we had something really nicely predefined already. So you can see here the dose given to the other lesions. At that point, we already figured out, "Oh my God, one of the lesions is pretty close to the brainstem, so we can't give the 20 gray we would like to." We just give 20 gray at that point just to be sure that we do not give too much dose to the brainstem.
So the situation then came out with 4 lesions getting 23, 1 lesion, 20 gray, and 5 gray to the resection cavity already within this. We used a template that came up with five table angles in total at that point, just to give you an idea about that. And so in this situation, you could see the 5-gray isodose line, now is the purple line that you see. It goes nicely around the cavity. So we end up with a nice dose of 5 gray in the cavity already, and we could do a good summation at that point and add the next plan, to be honest, and get a good result.
So, here are the DVHs just to give you an idea. So the purple line down here or the orange line here is the cavity. So you see we end up with something like 5 gray for the cavity. Here are the other lesions, the one that I only gave 20...we only gave 20 gray to. So you see nice dose conformity here at this point. We end up with an overall conformity index of 1.4 in this situation, but the cavity was excluded because that wouldn't make sense at all because that's not the typical lesion would be treated in this way. Just to give you an idea about the 12 gray, and the low dose, and this situation here.
And then we added the next plan. So we added up the cranial plan, cranial SRS in this case. This is the VMAT approach. That was not really part of this, of the session, so it's a completely VMAT approach in this case. So there will be no thing like dose-escalation like in multiple brain metastases. So it's more or less the ICLU-conformable dose prescription. And you see the DVH up here, you see the nice 5-gray dose path here given to the cavity. And what we as well did in this situation, so on the first side, we wanted not to give a not defined dose to the cavity during the single-shot SRS treatment. And on the second part of this so-called hybrid plan, so on the cranial plan, we do not want to add additionally dose to the already-treated SRS lesions. So we introduced them as organs at risk to make it easy and simple, and we just said, "Okay, maximum dose will be 1 gray that we would allow in this situation." And then we did the plan. We did the VMAT approach. We used a single arc in this situation, just table angle 0, just a full arc, just to be sure that we do not use the same arcs that we used before more or less not to get there into trouble or so. And to give it nice and clean. And you see that we ended up here with...here's the DVH for one of the lesions, firstly treated lesions, and we ended up with 0.74 gray in this situation.
So this all took place in August 2017 up to September 2017 with the SRS, with the SRT. And the patient was for follow-up already. All our patients got a follow-up after every three months using MR images at that point. And the last follow-up of this patient was in March, just one month ago now. And as you can see, so the disease is controlled. Here is also a little image of the cavity and the disease control. Patient is doing well.
So, but if we start to treat something like that, and this is already important for situations where you only use one single isocenter, we need to make up our minds what is about misalignment. So in the old days, one isocenter per lesion was always...in the isocenter, that was a good idea, that was easy. Misalignment wouldn't make too much at that point if we had a little margin, maybe, for instance. But now we are dealing with multiple lesions. None of these lesions is covered by the isocenter, normally, because the isocenter is positioned in the center of mass of all lesions. So we need to make up our minds, what happens to a misalignment.
So I fooled around a little bit with the DICOM data of this patient, here in this case. And we introduced some rotational and also translational errors. I just wanted to show you the rotational errors just to give you an idea, one of those lesions here. And you see, if we introduce 2 degrees of rotation and pitch and roll, or even just 1 degree of pitch and roll error, we will already start to do a big under-dosage to the lesion and also an overdosage to the healthy tissues surrounding this.
So we need to think about how can we deal with this and how can we get rid of these problems. And so we have to introduce some imaging solutions, IGRS. So here are the three that came up to my mind. Maybe you could also add Calypso, but Calypso, it's not really an imaging device. So I just kicked it out of my list. It was in there before when I thought about image guidance solution. So first of all, we got the ExacTrac image system, we all know that. So 2D image pair. We have it here. So orthogonal images, room-based system, so it's completely independent of gantry and couch position. So you can do images wherever you want at every thinkable angle. We got Cone Beam CT, which is more or less the gold standard for imaging at that point. This is really 3D image set, so this is perfect. But the worst part of this, it's Linac-based.
So we got the kV source and the imager also on this machine if you look here, fixed to the gantry. And if we change the table angle, we get into trouble because we can only do Cone Beam CTs for coplanar angles. So the system for SRS is completely out if you really want to compensate for translational errors and so on, and you have to compensate for that.
So another option at that point would be the Optical Surface Monitoring System, OSMS. So this really monitors the whole surface of the patient. So it will be monitored, will be projected as speckled pattern onto the patient. And then you got three-camera pods here in the room, so you can always see what happens. You can then compensate for problems there. You have to use an open mask. So what is monitored is not really the head, so it's more or less the part of the open mask. Otherwise, you will only look on the mask and you won't see any movement at that point. System is also room-based like the ExacTrac system, so it also can do images at every given couch angle. But as you have to use an open mask, you will introduce already some degrees of freedom for the patient just by opening the mask. So it will not be that strict like the Brainlab mask, and so on.
So, what I found out is that it might not be as accurate as the ExacTrac system by looking at literature. Nevertheless, it's also very fast, but I mean, you have to open up the mask at that point. So, what I then looked at, is the ExacTrac system, at this point, possible to check or to correct for those problems. And the good point is there's this publication out there. Yes, it is, compared to the gold standard, to the Cone Beam CT. You see after ExacTrac correction when you introduce 1 degree of roll and pitch error where you end up with 2.3 millimeters discrepancy, you will end up with 0.6 millimeters already.
So the last couple of slides will be on... So we have complex planning situations. We have IGRS requirements, so we have to think about statistic evaluation and maybe the risk of radionecrosis we had already. And there's this paper out there published in 2010 and says there is. Yeah, you have to consider the volume given, 10 gray, of the healthy brain, of the whole brain when you want to be sure that you do not get radionecrosis. And there is this paper and it says, 10.5 cc should be below V10 gray for the whole brain to be sure that you won't get any problems. But the problem with this is that it's only for single-shot or single lesions. Nowadays we're dealing with multiple lesions, so we have no real number at that point. And so we came up in our department and just took data of our first 50 patients. So we ended up with adding the total PTV volume here versus the volume of the brain receiving 10 gray, and we ended up with this nice linear thing here. And nowadays, whenever we treat a patient, we just plot in the volume we would have for the total brain, or for the total PTV volume. And we saw at which volume we would end up because we already treated, as I said, 120 patients. We do not see any radionecrosis more than we had before. So, and then we just have also a nice tool to see if our applying quality is quite nice because everything is done automatically.
And if you're not so familiar with this, the system came up with a plan and you have no idea, is it a good plan or not? So at this point, we came up with this. And then we plot in the volume and we see, okay, we got some...we had only 9 cc or 6 cc for the PTV and we end up with 50 cc here for the brain. Then we see, okay, we can do better at that point, then we start to use another template, and so on.
The same we did for the gradient index here. So we also see how the gradient index looks in our department just to have some kind of quality tools. So you should have something like that, otherwise, you have no idea if you're doing really good, or if... I mean, you will do good always with the software, but you won't be able to see if you can do better. So, and that's what we want to do, we get better results. And the result of that, what happened in our department is you can see our gradient index is 4.36 nowadays. And we got in the mean, 1.39 conformity index. Compared to literature published just one year ago, you see we're doing a lot better just by looking at these numbers and then just thinking about how to deal with that. So this is just my conclusion. And yeah, thank you for your attention.