Transcript
Thanks a lot for this kind introduction. Ladies and gentlemen, I would like to show you the results of our clinical implementation of the software. I am extremely honored to speak after Dr. Yamamoto because he has laid the ground lines to implement this and actually, I think there's still more, far more reluctance in Europe to replace WBRT by multiple mets, SRS, and I think we can discuss that later on. And today I will show you the technical framework that was implemented and we'll talk a bit about precision radiotherapy uncertainties, will show you the literature piece, and some of the things that we discussed, some talks before. And I have some clinical results and the plan comparison of VMAT versus DCA, and as you realize, most of the data that you've seen before are Gamma Knife data. So now we will switch to the Linac world a bit.
So department of radiation oncology at LMU serve two university hospitals in that city and a lot of other centers doing SRS, and we have around 2,150 patients per year and for ELEKTA machines. And actually we're preparing to install an MRI-Linac, but obviously at the moment we're one of the two centers worldwide that use this multiple brain mets elements software together with an ELEKTA accelerator because in former days it was used in connection with variant machines and now they're also introducing HyperArc. So I think we're at least cutting-edge in this kind of treatment and that's the difference compared to other MLC systems. It's a 160 MLC system. It is 40 times 40 maximum field size with 80 leaf pairs and it's a 5 millimeter project leaf width at the isocenter. So compared to other minimal MLCs, it's rather let's say rough, but actually works.
We'll show this data on the EXACTRAC sub-millimetric accuracy because this is used to perform the image guidance during treatment and yes, you can see compared to all these files that is shown in here, it is very clear that we have a sub-millimetric accuracy. This works quite well in that setting and we have our own data comparing that with the cone beam CT and as well it's up to 0.5 millimeters. And this represents the truly achievable targeting accuracy of the system. So it's ready to control with these XVI panels that you have also pre-size non-coplanar arc and that's a big difference compared to the classical cone beam CT. If your patient, you shift into zero position with your imaging and then you treat. But with this system you can also verify your non-coplanar arc. That's important as we have a 12% rate of non-coplanar arcs, but you have deviations showing the course of the disease of more than 1 millimeter. And if you come to this treatment of multiple metastases in just one isocenter, that's very important as every deviation from that causes a large error in the end. So that's the difference compared to if you do single shot Gamma Knife, for example.
And of course, some of the work has been presented by the master himself, Dr. Yamamoto. So it's very important that we have this multi-institutional prospective work and the most important information for me at least is that the cumulative volume was not more than 15 milliliter. That's not much if you have for example, one very large metastasis, many small ones, you have to take that into account. And some of these papers I will present to you have far larger metastasis volumes in the end. So 1 to 10. There were 1,194 patients and as you can see here, there is no difference compared to these groups, 2 to 4, to 5 to 10 metastasis. So this is actually the base to say, "Okay, we can use this data as kind of evidence to do that instead of whole brain RT as you have no the difference for this oligometastatic patients compared to those with more metastasis." And this work has been shown by Dr. Yamamoto. So we have minimal side effects in the end. So we'll skip these slides, but in the end you can say it's a safe treatment though it is just a prospective trial. It's the most important one on the topic and it shows you that this can be done in a safe way.
So this is one of the first benchmarking paper by Gevaert and colleagues. It has been published in Radiation Oncology in 2016 and they had these inverse with plants and optimized dynamic conformal arc algorithm compared to volume metric and modulated arc therapy with 10 patients included and they could show that they had a higher efficiency and better gradients. And these are some of these images. As you can see here, it's specifically on the low-dose spillage, you have advantages that you would expect from this inversely plant [inaudible 00:05:22.062] algorithm.
Actually this is very nice paper from the Green Journal two years ago from Bohoudi and the group of Frank Lagerwaard from UMC in Amsterdam. And they did a risk modeling of radiation necrosis based on single shot metastases with defined dose ranges and then they could calculate what you should prescribe for multiple metastases according to which with 10 to 12 parameter we choose. So that's a very nice formula that you can use to determine where to give your dose. And as you can see here between 24 and 18 gray, you have this range and if according to your size of the volume you could choose this isotoxic level. It was very easy calculation if you for example assume if a 5% risk, 1 metastasis then it's poor statistics that for 6 metastases, it's 25%. So you have to somehow multiply the 1 minus risk by the number of metastasis and that's the independent risk and probably it's higher. So to avoid this, you of course have to decrease your risk for lesion and that's one of the reasons why this isotoxicity approach is a nice one.
So data on HyperArc. That's the option I mentioned from Ruggieri and colleagues. It's very late publication. Twenty patients included in this one and they compared HyperArc to the RapidArc treatment and they could show that HyperArc outperformed actually RapidArc with the very short time slots. So 10 metastases in 20 minutes really fast treatment option, but the conformity and great in index and lower B12 in this case as a metric for radiation necrosis. So just some images and as you can see here, you have far better conformity and here with VMAT you have more low-dose spillage.
So another paper showing actually the same. So many of these papers are still at planning stage and showing as well better grading index for the HyperArc treatment. So this is our own experience. So we're using [inaudible 00:07:32.713] mapping system, very stable one with planning-CT that's given with contrast with the GTV that is contoured and amputated MR sequence and use a 1 millimeter PTV margin in the end, and uses a benchmarking endpoint for radiation necrosis V10 and 12 and try to keep it below 10 and 8 millimeter, but of course, depending on the size. If its getting larger, these metrics fail to work and that's actually prospective trial because as I said in the beginning, still lot of reluctance in Germany on replacing whole brain radiotherapy by radiosurgery. So we're doing that in a let's say, same or randomized way, we're doing a prospective trial and we'll do a propensity score comparison with all our patients that we treated earlier with whole brain radiotherapy and of course, are expecting that we will outperform these old data.
So as you can see here, those ranges between 15 and 20 gray with dexamethasone taping for this single shot, high dose delivery treatment part for overall survival. And I will show you how to use this software from Brainlab to delineate these lesions and to generate a plan. So you have a software tool that is really easy to use because as you can see here, every metastases in a formal versions was delineated automatically. I think FDA said that's not possible so you have to draw it into slices. It's very easy that it interpolates your remaining volume and then you can actually delineate all these metastases. The organs at risk are automatically contoured and in the end you can start your planning algorithm. That's relatively fast. So actually you cannot go away and then come back hours later. It's relatively clear that you get these results in actually no time. If you have some experience on the templates, it's very easy and you have a very good way of interpreting these deviators as you can click on each metastases, and then view for example, your radiation necrosis mark or the gradient index, and the conformity index. And there's a nice tool because Dr. Yamamoto mentioned difficult if you have two lesions very close together, but it can correct for some part of the volume that is treated by the other close besides that. So there's possibility included.
So co-registration also with the form of the registration possible and that's one clinical example, metastasized malignant melanoma patient who was... He had five metastases. He was treated in November 17 at the [inaudible 00:10:21.581] Afterwards it is considered prescription on as I said, 1 millimeter PTV margin, 80% isodose. So plus minus 5% and with 9 Arcs and 6 table angles, 5,417 monitor units. This took us 12 minutes treatment time. So really fast. As you can see here, our doses are very good and this patient is now... The plan of course, the patient is for now tumor-free. As you can see here, the metastasis before have gone completely, all lesions work on, off three months. And this patient is still tumor-free and this was one of the first ones we treated with that. But of course, the other patients in this case for example, at five metastases from endometrioid uterine sarcoma and sometimes later with Milroy disease and in this case, there is a good option you can actually transfer the complete those that you have calculated with this element software tool, your for example monarchy system and then you can for example, reduce the dose to these usually treated areas, but you can of course, do another SRS session. But in this case as it was Milroy disease, we voted against it.
So this is the experience up to now. It's since almost a year, it's 37 patients which is rather small number, but most of them lung cancer patients, median number of the treated metastases was four. And as I mentioned, prescription as it was around 80% in our institution. So we're rather conservative in this way with 20 gray maximum, but had almost no acute toxicity and all irradiated metastases were controlled. And we have some of course, intracranial failures, but they were treated a salvage situation. And some of these patients die due to systemic progressions. So we have to find a way in between.
So this is an example again on this VMAT versus DCA comparison and I will really quickly go to important dose. If you go down to the low-dose spillage for example, for this case, you really have a clear difference between this dynamic conformal arc algorithm if you compare these values. And this is something I really like because we started to evaluate the metastasis of 7 patients and these 23 lesions of course, will increase that total number was really clearly correlated. You have this diagram. So this is the difference of your necrosis parameter and this is the PTV volume, then for the smaller volumes, you have clear advantage for these small lesions with a DCA algorithm. And as they become larger, the advantage is smaller. So VMAT is not an option. My eyes is for the smaller lesions, DCA will or might outperform this in that case.
So let me complete. So this is very effective treatment with very few side effects. It's a very fast treatment. So between 5 to 20 minutes, that's really great. If you have 10 metastases, think of former day, single radiosurgery you would probably take hours to treat them. So it's a very quick treatment, patient has tolerated in a way, good way, but of course, you have to carefully select the stations as you could have those who fail or the low necrosis. So you have to find a good way to have this isotoxicity approach and in the end, I would at least make one statement to be aware of. If you have hypofractionated treatments for larger ones to go and this was with SRS, this might cause some problems depending on the size. So total volume of the lesions is still a matter of debate and in this regard, I would like to thank you for your time and I am hoping for good questions and discussions afterwards. Thank you.
So department of radiation oncology at LMU serve two university hospitals in that city and a lot of other centers doing SRS, and we have around 2,150 patients per year and for ELEKTA machines. And actually we're preparing to install an MRI-Linac, but obviously at the moment we're one of the two centers worldwide that use this multiple brain mets elements software together with an ELEKTA accelerator because in former days it was used in connection with variant machines and now they're also introducing HyperArc. So I think we're at least cutting-edge in this kind of treatment and that's the difference compared to other MLC systems. It's a 160 MLC system. It is 40 times 40 maximum field size with 80 leaf pairs and it's a 5 millimeter project leaf width at the isocenter. So compared to other minimal MLCs, it's rather let's say rough, but actually works.
We'll show this data on the EXACTRAC sub-millimetric accuracy because this is used to perform the image guidance during treatment and yes, you can see compared to all these files that is shown in here, it is very clear that we have a sub-millimetric accuracy. This works quite well in that setting and we have our own data comparing that with the cone beam CT and as well it's up to 0.5 millimeters. And this represents the truly achievable targeting accuracy of the system. So it's ready to control with these XVI panels that you have also pre-size non-coplanar arc and that's a big difference compared to the classical cone beam CT. If your patient, you shift into zero position with your imaging and then you treat. But with this system you can also verify your non-coplanar arc. That's important as we have a 12% rate of non-coplanar arcs, but you have deviations showing the course of the disease of more than 1 millimeter. And if you come to this treatment of multiple metastases in just one isocenter, that's very important as every deviation from that causes a large error in the end. So that's the difference compared to if you do single shot Gamma Knife, for example.
And of course, some of the work has been presented by the master himself, Dr. Yamamoto. So it's very important that we have this multi-institutional prospective work and the most important information for me at least is that the cumulative volume was not more than 15 milliliter. That's not much if you have for example, one very large metastasis, many small ones, you have to take that into account. And some of these papers I will present to you have far larger metastasis volumes in the end. So 1 to 10. There were 1,194 patients and as you can see here, there is no difference compared to these groups, 2 to 4, to 5 to 10 metastasis. So this is actually the base to say, "Okay, we can use this data as kind of evidence to do that instead of whole brain RT as you have no the difference for this oligometastatic patients compared to those with more metastasis." And this work has been shown by Dr. Yamamoto. So we have minimal side effects in the end. So we'll skip these slides, but in the end you can say it's a safe treatment though it is just a prospective trial. It's the most important one on the topic and it shows you that this can be done in a safe way.
So this is one of the first benchmarking paper by Gevaert and colleagues. It has been published in Radiation Oncology in 2016 and they had these inverse with plants and optimized dynamic conformal arc algorithm compared to volume metric and modulated arc therapy with 10 patients included and they could show that they had a higher efficiency and better gradients. And these are some of these images. As you can see here, it's specifically on the low-dose spillage, you have advantages that you would expect from this inversely plant [inaudible 00:05:22.062] algorithm.
Actually this is very nice paper from the Green Journal two years ago from Bohoudi and the group of Frank Lagerwaard from UMC in Amsterdam. And they did a risk modeling of radiation necrosis based on single shot metastases with defined dose ranges and then they could calculate what you should prescribe for multiple metastases according to which with 10 to 12 parameter we choose. So that's a very nice formula that you can use to determine where to give your dose. And as you can see here between 24 and 18 gray, you have this range and if according to your size of the volume you could choose this isotoxic level. It was very easy calculation if you for example assume if a 5% risk, 1 metastasis then it's poor statistics that for 6 metastases, it's 25%. So you have to somehow multiply the 1 minus risk by the number of metastasis and that's the independent risk and probably it's higher. So to avoid this, you of course have to decrease your risk for lesion and that's one of the reasons why this isotoxicity approach is a nice one.
So data on HyperArc. That's the option I mentioned from Ruggieri and colleagues. It's very late publication. Twenty patients included in this one and they compared HyperArc to the RapidArc treatment and they could show that HyperArc outperformed actually RapidArc with the very short time slots. So 10 metastases in 20 minutes really fast treatment option, but the conformity and great in index and lower B12 in this case as a metric for radiation necrosis. So just some images and as you can see here, you have far better conformity and here with VMAT you have more low-dose spillage.
So another paper showing actually the same. So many of these papers are still at planning stage and showing as well better grading index for the HyperArc treatment. So this is our own experience. So we're using [inaudible 00:07:32.713] mapping system, very stable one with planning-CT that's given with contrast with the GTV that is contoured and amputated MR sequence and use a 1 millimeter PTV margin in the end, and uses a benchmarking endpoint for radiation necrosis V10 and 12 and try to keep it below 10 and 8 millimeter, but of course, depending on the size. If its getting larger, these metrics fail to work and that's actually prospective trial because as I said in the beginning, still lot of reluctance in Germany on replacing whole brain radiotherapy by radiosurgery. So we're doing that in a let's say, same or randomized way, we're doing a prospective trial and we'll do a propensity score comparison with all our patients that we treated earlier with whole brain radiotherapy and of course, are expecting that we will outperform these old data.
So as you can see here, those ranges between 15 and 20 gray with dexamethasone taping for this single shot, high dose delivery treatment part for overall survival. And I will show you how to use this software from Brainlab to delineate these lesions and to generate a plan. So you have a software tool that is really easy to use because as you can see here, every metastases in a formal versions was delineated automatically. I think FDA said that's not possible so you have to draw it into slices. It's very easy that it interpolates your remaining volume and then you can actually delineate all these metastases. The organs at risk are automatically contoured and in the end you can start your planning algorithm. That's relatively fast. So actually you cannot go away and then come back hours later. It's relatively clear that you get these results in actually no time. If you have some experience on the templates, it's very easy and you have a very good way of interpreting these deviators as you can click on each metastases, and then view for example, your radiation necrosis mark or the gradient index, and the conformity index. And there's a nice tool because Dr. Yamamoto mentioned difficult if you have two lesions very close together, but it can correct for some part of the volume that is treated by the other close besides that. So there's possibility included.
So co-registration also with the form of the registration possible and that's one clinical example, metastasized malignant melanoma patient who was... He had five metastases. He was treated in November 17 at the [inaudible 00:10:21.581] Afterwards it is considered prescription on as I said, 1 millimeter PTV margin, 80% isodose. So plus minus 5% and with 9 Arcs and 6 table angles, 5,417 monitor units. This took us 12 minutes treatment time. So really fast. As you can see here, our doses are very good and this patient is now... The plan of course, the patient is for now tumor-free. As you can see here, the metastasis before have gone completely, all lesions work on, off three months. And this patient is still tumor-free and this was one of the first ones we treated with that. But of course, the other patients in this case for example, at five metastases from endometrioid uterine sarcoma and sometimes later with Milroy disease and in this case, there is a good option you can actually transfer the complete those that you have calculated with this element software tool, your for example monarchy system and then you can for example, reduce the dose to these usually treated areas, but you can of course, do another SRS session. But in this case as it was Milroy disease, we voted against it.
So this is the experience up to now. It's since almost a year, it's 37 patients which is rather small number, but most of them lung cancer patients, median number of the treated metastases was four. And as I mentioned, prescription as it was around 80% in our institution. So we're rather conservative in this way with 20 gray maximum, but had almost no acute toxicity and all irradiated metastases were controlled. And we have some of course, intracranial failures, but they were treated a salvage situation. And some of these patients die due to systemic progressions. So we have to find a way in between.
So this is an example again on this VMAT versus DCA comparison and I will really quickly go to important dose. If you go down to the low-dose spillage for example, for this case, you really have a clear difference between this dynamic conformal arc algorithm if you compare these values. And this is something I really like because we started to evaluate the metastasis of 7 patients and these 23 lesions of course, will increase that total number was really clearly correlated. You have this diagram. So this is the difference of your necrosis parameter and this is the PTV volume, then for the smaller volumes, you have clear advantage for these small lesions with a DCA algorithm. And as they become larger, the advantage is smaller. So VMAT is not an option. My eyes is for the smaller lesions, DCA will or might outperform this in that case.
So let me complete. So this is very effective treatment with very few side effects. It's a very fast treatment. So between 5 to 20 minutes, that's really great. If you have 10 metastases, think of former day, single radiosurgery you would probably take hours to treat them. So it's a very quick treatment, patient has tolerated in a way, good way, but of course, you have to carefully select the stations as you could have those who fail or the low necrosis. So you have to find a good way to have this isotoxicity approach and in the end, I would at least make one statement to be aware of. If you have hypofractionated treatments for larger ones to go and this was with SRS, this might cause some problems depending on the size. So total volume of the lesions is still a matter of debate and in this regard, I would like to thank you for your time and I am hoping for good questions and discussions afterwards. Thank you.