Transcript
So I'd like to cordially thank Dr. Vonn [SP] for introducing the topic and I would like to jump into the topic directly because as you heard, oligometastatic disease in the brain may be some really interesting indication in the future. And the special thing about Munich is that we are mainly an ELEKTA site, as you can see.
So just some short information on Munich LMU. So we've run some 2100 patients per year in 2 headquarters with 6 LINACS, 4 ELEKTA ones, two from Siemens. We have to replace them at some date. And we have an MRI-LINAC in preparation and MOSAIQ and the usual things that you have in the ELEKTA world, as you can see here. So the interesting part is now we're the first ones who are using type of ExacTrac ELEKTA combination.
I would like to present these first data from our department. We did this one preliminary study on our ELEKTA machine as you have cone-beam CT and the ExacTrac. And the big advantage of ExacTrac is, of course, you can verify if your non-coplanar arcs are also in the right position and that's something that you cannot do with the rigid information with 0 degrees with a cone-beam CT. And as you can see the end-to-end precision is very high. So this is the kind of prerequisite if you want to do these kinds of treatments.
So this is now the workflow that we currently do. So we have high-resolution MRI of the brain with a slice thickness of 1 millimeter in amperage. In reconstruction, we use a dual-layer mask system with a planning CT with contrast and use a dose calculation grid of 1 millimeter. Two millimeters gives a good result as well. In the end, you can speed up the whole process. The GTV is, of course, contoured in the contrast-enhanced MRI and the planning CT as well. And the fusion algorithm, I will present that later on. It is very precise, and we use a 1-millimeter GTV to PTV margin. And the benchmarks for the endpoint radiation necrosis we use usually V10 and V12. And, of course, we focus on a very steep gradient index and a low conformity index for these volumes.
So this is a parallel phase II trial that we have started to do because we didn't want to just do it because we can do it. We wanted to collect the data that we produced in the clinical setting and compare it with our form indication. Because we were very strict in the former days, we said everything beyond three metastases has to receive whole-brain radiotherapy. And so we did that prospective trial and at least started to do that, and the indication is that we have patients with 4 to 10 brain metastases from any disease and we excluded certain ones where you have at least up to today, of course, an indication for whole-brain or C2 radiation such as lymphoma or small cell lung cancer. We could discuss the latter, of course, but we try to exclude this from the beginning. And the prospective setting, of course, and semi-randomized comparison because of these elderly whole-brain patients that we will co-match and then match comparison in the end. And the isodose line is 80%, normally 80% to 90%, and the dose range is 15 Gy to 20 Gy depending on the size of the lesion and the clinical context. And we also fixed hypofx. I will come to that later.
So the co-PI of mine is sitting here, Steffi Corradini, and we are now trying to fill this study. And the primary endpoint and this comparison with whole-brain radiotherapy is overall survival. Secondary endpoints are intracranial control and toxicity endpoints. And, of course, patient-reported outcome and neurocognition.
So now we'll jump into the software itself. So Elements Multiple BrainMets SRS. Actually, this is a very, very short way because I'm from Munich. BrainMets is situated in Munich, so they presented the software to us, and we said, "Well, okay, let's try it. Let's give it a try." And you have this like app-based system. So Elements you have this automatic brain metastases contouring and planning in one hand and the workflow is as follows. You have like a demo patient. You can choose between different apps. For example, anatomical mapping for the co-registration, smart brush object manipulation, patient-specific QA.
And now we'll jump into the major tool itself. You start with the registration. I skipped this part. And then you will perform the delineation and the nice thing is that you can do that with a smart brush within the MRI and can do it in two layers and it reconstructs the lesion in a very fast way. So this is a very comfortable way of performing these settings.
And you have the organ-at-risk delineation which is done automatically. So it speeds up the whole process. And a very nice feature is that you can have an overview on all these metastases and this nice multi-slice view, and you can see it in the sagittal and the coronal, and the axial view. And you can click on single metastases and choose them and adapt these lesions. So, and you can do that on CT and MRI as well.
It is template-based. So, for each situation, like, for example, if a metastases would like to perform a plan, you can preselect from different templates. And it depends a bit on experience which template you use, but if you use the template it works in a very speed way. So you have your plan that is produced, [inaudible 00:06:08] which will talk a bit more about this topic. And you have this DVAs and you can click on the single metastases and for each metastasis, you have the information and conformity on gradient, on the volume, on the normal tissue exposure like V10/12. You can directly retrieve this from the DVA. So it's a very convenient way to control for your plans. And as you can see that is the registration that you do in the beginning. Like, you can compare it in this color view whether it's okay or whether you have to adapt it.
So with a clinical example, one of the first patients that we treated was a patient with a metastasized malignant melanoma. So initial date of diagnosis was in October 2016, pT3b and an M1c status with lung, brain, and lymph node metastases. And this patient was treated in November, so one of the first ones we had. And he had BRAF and MEK inhibitor in combination afterward a systemic therapy. In the first follow-up in March now and is in complete remission with prescription of 17 Gy and two, three mets and one was to 19 Gy and one to 20 Gy. And all together with 9 Arcs, 6 table angles that we used and 5417 monitor units, 12 minutes treatment time for all these 5 metastases. So if you add up this with a single isocenter approach for each one you will end up with a far longer treatment time. And this patient get a fixed steroid treatment and the tapering, a very fast one. So we really reduced the acute side effects that could occur without tapering.
And here the OAR constraints. One issue is that in the next version of the software you will have more control on your OAR constraints. For now, it's just optimizing on optimal conformity, optimal grading index. So if you have some violation of a predefined constraint you have to adapt your dose for now. So next step will be, or your volume of course, into the direction of this critical organ. And this is probably something where there is still room of improvement.
So this is the case with these five metastases. One, for example, was close to the optic system. That's the plan. And as you can see, this is the follow-up imaging or just a short demo of how you can take control of these annuals. It's important that you look at these template annuals because it's like an r-forth and off-back and you can see how many of these arc positions are actually used. The more the better normally. And you can see from each view which metastasis is actually treated. And this is the imaging. So initial metastases, as you can see here, the small one over here near to the optic system, first follow-up imaging so nothing left. So nice case. Of course, I will show you the best case, but I will also show you one case that was not optimal.
So like this was a patient 50-year-old with an uterine sarcoma, had presented in January with five brain metastases, was in favorable clinical score, had no deficit, so we performed SRS within our trial. And this patient had a multifocal progression, as you can see here. So a lot of metastases outside the treated ones. So we have different options now and several studies are very important in this setting.
So one option would be like Yamamoto and colleagues did, so perform salvage radiosurgery and all these other ones. We did something a bit different in this case. And that's a nice feature. You can completely import those data sets to your planning system, which is in our case, Monaco from ELEKTA and then you can reduce the dose in this whole-brain treatment in these pretreated regions. And, for example, you can take a 10-grade isodose and prescribe a certain under dosage in this region.
And this is a short comparison. We're currently working on comparing these data for all our patients between VMAT and DCA, like the DAM conform arc treatment. As you can see here if you do VMAT on the first approach, you have this rather large low-dose bath with 5 Gy isodose line. And if you do the same, it is a much steeper gradient in the DCA. Of course, you would argue different planning systems could make a difference, of course, and that's something we have to correct for in future.
Just to show you the numbers that were produced by the systems you have a far lower cumulative V10 and a far lower V5. So that's already an issue that's important if you just try to use VMAT. You have the thing that the total dose exposure might be rather high and that's especially important if you perform re-irradiations with whole brain.
So this is the cumulative experience up to today. So we started with the first patient in January of last year. So 15 patients in, that were treated with the software mainly melanoma, lung, and breast. The median number of metastases per patient was 4 and ranged from 2 to 16. As I told you, 4 to 10 was the inclusion criteria for the stereobrain trial. And we had some patients who also had a [inaudible 00:11:36] for about two and three metastases. So these were put into the registry. So overall we treated 75 metastases and that was what we discussed before, and the cumulative volume was up to 25 milliliters. So this is really rather a large number and with acute toxicity in one patient with a very large metastases.
And these are actually my concluding remarks. So to skip into my timeframe. So this is actually a very effective treatment with few side effects and up to now at least with a high local control, but of course, we are at the beginning. So we started in November. So we can just report in the short-term outcome and it's a very fast treatment option because it's from 12 to 20 minutes. If you think of 60 metastases treated simultaneously, it's really impressive in my eyes.
And of course, you have to carefully select these patients and that's the reason why we do these prospective trials because you can miss some patients and they have multiple progression, and you have to retreat them in very early time points. And of course, we inform these patients on the alternatives like they could have standard whole brain through the standard of care or could be included in the HIPPORAD trial run by Professor Grosu in Freiburg which, for example, spares the hippocampus. So these are the alternatives that we'll discuss on individual base. And, of course, we have to be aware of side effects, so careful dosing is necessary. In my eyes, you cannot take the RTOG dose prescription and put it to 12 metastases. You have to adapt this depending on the size and location.
And in the end, I would strongly advise to be aware of combining each kind of hypoFX with the single shot technique because you might end up with a high disparity between the low single-dose compared to the single high-dose volume. So this is not actually something the software can produce an easy way and you could end up with, let's say, acute toxicity from the single session if you perform further fraction with a hypoFx. So this is something, at least, you will have to think about.
And then, I would like to thank you all, it's a small of you, to Munich, whoever wants to come it's really nice between the Alps. And I'd like to thank colleagues from the physics department for providing some of the content, Brainlab, and some of my colleagues like Steffi Corradini, Raphael Bodensohn for providing some of the images. So let me thank you for your attention and I would like to hand over to neural radiology.
So just some short information on Munich LMU. So we've run some 2100 patients per year in 2 headquarters with 6 LINACS, 4 ELEKTA ones, two from Siemens. We have to replace them at some date. And we have an MRI-LINAC in preparation and MOSAIQ and the usual things that you have in the ELEKTA world, as you can see here. So the interesting part is now we're the first ones who are using type of ExacTrac ELEKTA combination.
I would like to present these first data from our department. We did this one preliminary study on our ELEKTA machine as you have cone-beam CT and the ExacTrac. And the big advantage of ExacTrac is, of course, you can verify if your non-coplanar arcs are also in the right position and that's something that you cannot do with the rigid information with 0 degrees with a cone-beam CT. And as you can see the end-to-end precision is very high. So this is the kind of prerequisite if you want to do these kinds of treatments.
So this is now the workflow that we currently do. So we have high-resolution MRI of the brain with a slice thickness of 1 millimeter in amperage. In reconstruction, we use a dual-layer mask system with a planning CT with contrast and use a dose calculation grid of 1 millimeter. Two millimeters gives a good result as well. In the end, you can speed up the whole process. The GTV is, of course, contoured in the contrast-enhanced MRI and the planning CT as well. And the fusion algorithm, I will present that later on. It is very precise, and we use a 1-millimeter GTV to PTV margin. And the benchmarks for the endpoint radiation necrosis we use usually V10 and V12. And, of course, we focus on a very steep gradient index and a low conformity index for these volumes.
So this is a parallel phase II trial that we have started to do because we didn't want to just do it because we can do it. We wanted to collect the data that we produced in the clinical setting and compare it with our form indication. Because we were very strict in the former days, we said everything beyond three metastases has to receive whole-brain radiotherapy. And so we did that prospective trial and at least started to do that, and the indication is that we have patients with 4 to 10 brain metastases from any disease and we excluded certain ones where you have at least up to today, of course, an indication for whole-brain or C2 radiation such as lymphoma or small cell lung cancer. We could discuss the latter, of course, but we try to exclude this from the beginning. And the prospective setting, of course, and semi-randomized comparison because of these elderly whole-brain patients that we will co-match and then match comparison in the end. And the isodose line is 80%, normally 80% to 90%, and the dose range is 15 Gy to 20 Gy depending on the size of the lesion and the clinical context. And we also fixed hypofx. I will come to that later.
So the co-PI of mine is sitting here, Steffi Corradini, and we are now trying to fill this study. And the primary endpoint and this comparison with whole-brain radiotherapy is overall survival. Secondary endpoints are intracranial control and toxicity endpoints. And, of course, patient-reported outcome and neurocognition.
So now we'll jump into the software itself. So Elements Multiple BrainMets SRS. Actually, this is a very, very short way because I'm from Munich. BrainMets is situated in Munich, so they presented the software to us, and we said, "Well, okay, let's try it. Let's give it a try." And you have this like app-based system. So Elements you have this automatic brain metastases contouring and planning in one hand and the workflow is as follows. You have like a demo patient. You can choose between different apps. For example, anatomical mapping for the co-registration, smart brush object manipulation, patient-specific QA.
And now we'll jump into the major tool itself. You start with the registration. I skipped this part. And then you will perform the delineation and the nice thing is that you can do that with a smart brush within the MRI and can do it in two layers and it reconstructs the lesion in a very fast way. So this is a very comfortable way of performing these settings.
And you have the organ-at-risk delineation which is done automatically. So it speeds up the whole process. And a very nice feature is that you can have an overview on all these metastases and this nice multi-slice view, and you can see it in the sagittal and the coronal, and the axial view. And you can click on single metastases and choose them and adapt these lesions. So, and you can do that on CT and MRI as well.
It is template-based. So, for each situation, like, for example, if a metastases would like to perform a plan, you can preselect from different templates. And it depends a bit on experience which template you use, but if you use the template it works in a very speed way. So you have your plan that is produced, [inaudible 00:06:08] which will talk a bit more about this topic. And you have this DVAs and you can click on the single metastases and for each metastasis, you have the information and conformity on gradient, on the volume, on the normal tissue exposure like V10/12. You can directly retrieve this from the DVA. So it's a very convenient way to control for your plans. And as you can see that is the registration that you do in the beginning. Like, you can compare it in this color view whether it's okay or whether you have to adapt it.
So with a clinical example, one of the first patients that we treated was a patient with a metastasized malignant melanoma. So initial date of diagnosis was in October 2016, pT3b and an M1c status with lung, brain, and lymph node metastases. And this patient was treated in November, so one of the first ones we had. And he had BRAF and MEK inhibitor in combination afterward a systemic therapy. In the first follow-up in March now and is in complete remission with prescription of 17 Gy and two, three mets and one was to 19 Gy and one to 20 Gy. And all together with 9 Arcs, 6 table angles that we used and 5417 monitor units, 12 minutes treatment time for all these 5 metastases. So if you add up this with a single isocenter approach for each one you will end up with a far longer treatment time. And this patient get a fixed steroid treatment and the tapering, a very fast one. So we really reduced the acute side effects that could occur without tapering.
And here the OAR constraints. One issue is that in the next version of the software you will have more control on your OAR constraints. For now, it's just optimizing on optimal conformity, optimal grading index. So if you have some violation of a predefined constraint you have to adapt your dose for now. So next step will be, or your volume of course, into the direction of this critical organ. And this is probably something where there is still room of improvement.
So this is the case with these five metastases. One, for example, was close to the optic system. That's the plan. And as you can see, this is the follow-up imaging or just a short demo of how you can take control of these annuals. It's important that you look at these template annuals because it's like an r-forth and off-back and you can see how many of these arc positions are actually used. The more the better normally. And you can see from each view which metastasis is actually treated. And this is the imaging. So initial metastases, as you can see here, the small one over here near to the optic system, first follow-up imaging so nothing left. So nice case. Of course, I will show you the best case, but I will also show you one case that was not optimal.
So like this was a patient 50-year-old with an uterine sarcoma, had presented in January with five brain metastases, was in favorable clinical score, had no deficit, so we performed SRS within our trial. And this patient had a multifocal progression, as you can see here. So a lot of metastases outside the treated ones. So we have different options now and several studies are very important in this setting.
So one option would be like Yamamoto and colleagues did, so perform salvage radiosurgery and all these other ones. We did something a bit different in this case. And that's a nice feature. You can completely import those data sets to your planning system, which is in our case, Monaco from ELEKTA and then you can reduce the dose in this whole-brain treatment in these pretreated regions. And, for example, you can take a 10-grade isodose and prescribe a certain under dosage in this region.
And this is a short comparison. We're currently working on comparing these data for all our patients between VMAT and DCA, like the DAM conform arc treatment. As you can see here if you do VMAT on the first approach, you have this rather large low-dose bath with 5 Gy isodose line. And if you do the same, it is a much steeper gradient in the DCA. Of course, you would argue different planning systems could make a difference, of course, and that's something we have to correct for in future.
Just to show you the numbers that were produced by the systems you have a far lower cumulative V10 and a far lower V5. So that's already an issue that's important if you just try to use VMAT. You have the thing that the total dose exposure might be rather high and that's especially important if you perform re-irradiations with whole brain.
So this is the cumulative experience up to today. So we started with the first patient in January of last year. So 15 patients in, that were treated with the software mainly melanoma, lung, and breast. The median number of metastases per patient was 4 and ranged from 2 to 16. As I told you, 4 to 10 was the inclusion criteria for the stereobrain trial. And we had some patients who also had a [inaudible 00:11:36] for about two and three metastases. So these were put into the registry. So overall we treated 75 metastases and that was what we discussed before, and the cumulative volume was up to 25 milliliters. So this is really rather a large number and with acute toxicity in one patient with a very large metastases.
And these are actually my concluding remarks. So to skip into my timeframe. So this is actually a very effective treatment with few side effects and up to now at least with a high local control, but of course, we are at the beginning. So we started in November. So we can just report in the short-term outcome and it's a very fast treatment option because it's from 12 to 20 minutes. If you think of 60 metastases treated simultaneously, it's really impressive in my eyes.
And of course, you have to carefully select these patients and that's the reason why we do these prospective trials because you can miss some patients and they have multiple progression, and you have to retreat them in very early time points. And of course, we inform these patients on the alternatives like they could have standard whole brain through the standard of care or could be included in the HIPPORAD trial run by Professor Grosu in Freiburg which, for example, spares the hippocampus. So these are the alternatives that we'll discuss on individual base. And, of course, we have to be aware of side effects, so careful dosing is necessary. In my eyes, you cannot take the RTOG dose prescription and put it to 12 metastases. You have to adapt this depending on the size and location.
And in the end, I would strongly advise to be aware of combining each kind of hypoFX with the single shot technique because you might end up with a high disparity between the low single-dose compared to the single high-dose volume. So this is not actually something the software can produce an easy way and you could end up with, let's say, acute toxicity from the single session if you perform further fraction with a hypoFx. So this is something, at least, you will have to think about.
And then, I would like to thank you all, it's a small of you, to Munich, whoever wants to come it's really nice between the Alps. And I'd like to thank colleagues from the physics department for providing some of the content, Brainlab, and some of my colleagues like Steffi Corradini, Raphael Bodensohn for providing some of the images. So let me thank you for your attention and I would like to hand over to neural radiology.