Transcript
Bogdan: Hello. Hello everyone. My name is Bogdan Valcu. I'm the director of Novalis Circle. And I would like to welcome everyone to a restart for our online webinar series. First, let me just say that I hope everybody is staying safe and we hope to soon be able to return to normal interactions with you. As you may have already heard from us, we have decided to migrate the rest of our calendar year 2020 scientific activities online. So, please feel free to go to our novaliscircle.org website and see the list of upcoming events. We have also decided to make free our entire video library of past events, so more than a decade of presentations are available for you should you have time and catch up on some of the talks. However, we strongly encourage you to sign up for a membership for the website as well, as we have quite a good number of new features coming up including our forum and pretty soon some online tumor boards as well.
Together with me today is Philipp Freislederer from Ludwig Maximilians University in Munich, and he will cover some new topics pertaining to ExacTrac Dynamic. LMU together with a group in Copenhagen have been the two early adopter sides for ExacTrac Dynamic. They've had the chance to work with the products as it was still under development, and they can share with you some unique experiences.
Before getting into the webinar, just a few topics. First as you remember, from the webinar interface the majority of you just have the opportunity to listen. So, feel free to send us via the chat line questions. We will put all those together, and at the end of the presentation, we will go through a question and answer session. I would also like to remind you that you have to login via Google Chrome or Safari. EasyWebinar isn't terribly friendly with Microsoft products, so Internet Explorer Edge will not let you see the webinar. We're also recording everything, so the information will be available afterwards on novaliscircle.org. We are providing continuing education credits, and if you'd like either a CAMPEP, MDCB or ASRT credits following the webinar, please feel free to send an email to info@novaliscircle.org, and we'll provide you with specific instructions on how you can obtain credits.
Lastly, we are happy to announce that ExacTrac Dynamic has been released in the European community. However, it is not yet available in all other markets. So, we intend this webinar to be for educational purposes only, and please bear in mind the disclaimer attached. Should you have questions in your local markets for product availability, please ask our commercial teams for updates on that end.
So, with that being said, I'd like to again thank you, Mr. Freislederer for his time today, and we'll look forward to his presentation.
Philipp: Well, thank you very much for the introduction. I'm gonna now turn my camera off for the moment. And if that's all right with you, I can also leave it on, but I will not see your comments at the moment. All of the questions will be gathered by Brainlab later on, and we can discuss it afterwards. I will now switch to...one second. This is technically now the hard part for it.
First of all, thank you very much Brainlab for letting me have this webinar with every one of you. One second. Okay, I'm guessing everyone can now see my slides. And I'm gonna just start off with the webinar. So, Brainlab has asked me to do a webinar on the ExacTrac Dynamic as Bogdan already said before. We're using...we're not using the system, because we just got the...just the CE mark has just been released. and...but we were able to test it a little bit to do some validations together with...not together with, but also Rigshospitalet in Denmark, in Copenhagen did some validations for the ExacTrac Dynamic. And obviously, the COVID-19, the corona crisis has struck everybody, and we are waiting to get the final installation at the moment to do some additional tests, but at least we did some validation testing before, so we had the chance to play around with the system a little bit.
So, this is going to be my topic, ExacTrac Dynamic solution for IGRT/SGRT for Linac-based radiosurgery. I have to disclose some things. My department has research grants from Brainlab. And I also am receiving a speaking fee.
What is the ExacTrac Dynamic? So, we have a combination of an optical surface scanner using structured light. And additionally, and that's sort of the main new addition to it, is a thermal camera. So, why are you using a thermal camera? There are various options...various reasons for it. One very important part is that you have additional registration information. The thermal camera, the thermal information serves as a sort of a fourth dimension in addition to the optical surface scanner. It also makes multiple in-room cameras not necessary anymore. You can sort of see it as a system where the optical surface scanner with a structured light provides you three dimensions, and using the thermal information, you have additional access to something, or the registration algorithm has additional access to something which it can grab on.
In addition, you have x-ray positioning and monitoring just as the "old ExacTrac." So, you have your usual...or x-ray positioning, x-ray monitoring. So, this serves now as a surface guided radiation therapy tool, because you have real-time tracking using the surface and the thermal information, and also using the x-rays. And as Bogdan said before, it has now been CE certified, so after everything is getting...it's gonna get a little bit more quiet with coronas. Hopefully some clinics are able to install this system and use it.
I'm gonna guide you a little bit to the workflow. Unfortunately, these are all some slides with phantoms, because we didn't have the chance to get...obviously didn't have the chance to have patients on it yet. So, I'm guide...I'm gonna guide you through the workflow, how it is proposed in the last version.
The system starts off with some prepositioning based on the surface. You see it here, there's a phantom being placed and you see some shifts that the system calculates. These shifts are afterwards sent to the couch. And after you move your couch with the shifts, obviously, you should get in a very good direction of prepositioning, and this is all considered prepositioning. So, the...I'm gonna come to that after on. The x-ray is still something that's quite important in all the cases in this...in this version. So, these are obviously phantom and obviously therefore very small. This is expected. But this considers...this is considered a rough prepositioning according to the Brainlab R&D department.
Afterwards, you see, this is the camera view with thermal information from the thermal and surface camera. You select your region of interest. In this case, what you see is a specially designed phantom by Brainlab, also where you have a heated wire inside. So, you see the red areas are so...are the warmer ones and the green areas in this case where there's no heated wire. So, you select your region [crosstalk 00:10:54]
Bogdan: Philipp, one second. If you don't mind the interruption, just making sure that you've entered the presentation mode on the PowerPoint. We were still on the introductory slide. Sorry.
Philipp: Oh, I'm very sorry. I'm very sorry. One second, one second, one second. I'm very sorry. I might have gotten the wrong button here. So...better now?
Bogdan: Yeah. If you don't mind going back some slides.
Philipp: Okay.
Bogdan: Nobody was actually able to see it. Thank you.
Philipp: I'm gonna go back again. So, this is actually the first actual real slide. As I said before, you have an optical surface scanner and you have a thermal camera, and then you sort of see the thermal camera as the fourth dimension. And using your x-rays, you have the additional positioning and inter-fractional monitoring. And this is the part where it actually gets interesting. So, you do your pre-position workflow. You get some values for your pre-position based on the surface only. These values are then sent to the couch to enable a 3D movement. And afterwards the...here in this case, the phantom or the patient afterwards is positioned quite all right as I said before, in a phantom...in the stationary phantom, obviously quite good values. And now you should see the heated phantom with the heated wire inside. So, the red and yellow spots on the surfaces is actually where the heated wires inside and you get a temperature profile. And in our case, you can select them...in any case, you can select the temperature profile for the session or for the couch angle, and you select your temperature profile, or in this case and so you select your region of interest for the specific patient, for the specific fraction.
After the selection of your region of interest, you acquire x-rays, and you acquire dual x-rays. You have the console. In this case the phantom also has some...has a spine inside, has some ribs, has some bones inside, so this...it also here has some heated wire inside, so this makes it a little easier for any registration algorithm, but I'm gonna come to that afterwards. So, you acquire your x-rays and you use the x-rays afterwards with the registration with a previously generated DRRs. You use same as the...same case as the old ExacTrac, you use the x-ray DRR registration for your initial positioning. These values you're getting from the x-rays are then afterwards sent to your six-dimensional table in our...for Elekta cassettes, this would be the HexaPOD table, and you can enable directly the couch movement also for any rotations.
You can choose afterwards to verify the position, and obviously for the phantom, we can easily verify everything, and you see that the registration now gives you obviously very good results. I was...from a physics point of view, I was quite happy to see that at least we have a vertical deviation of 0.1 millimeter. If it would have been 0 millimeter in all directions, that would have made me think a little bit if I really could trust everything I will see, but at least, obviously it's a static phantom and no dynamic movement, and this should work out quite fine.
After your x-ray verification, you go into monitoring or tracking mode. So, on the right upper side, now you see the patient, or in this case the phantom surface, and your region of interest which you have chosen before. And you'll also see the life online values for the full surface and thermal information provided by the system, and you see deviations calculated to the isocenter of the plan. You can then choose afterwards to enable automatic x-ray mode or you can take an x-ray in the beginning. This is still quite open and you can...just you can start treatment afterwards.
So, you have an initial x-ray, and every time you take an X-ray, obviously this...or not obviously, but the x-ray is always chosen to be the ground truth in this case and in any case. So, you have your x-rays. You really monitor internal or bony registration, and the surface just in this case provides...it gives you a hint of how the patient is behaving and gives you some monitoring. So, during monitoring, you monitor the surface and you can decide if you want to...what do you do afterwards. There are several options in this case, one option would be is if the surface goes out of tolerance obviously it is easily monitored without any additional dose. If the surface goes out of tolerance, the beam can stop or the beam can go on. You can have automatic x-rays to be taken, dual or mono depending on the gantry position at the moment, or you can continue and just monitor, or you can stop the beam and take x-rays. That's pretty much for any patient or any...depending on how tight you wanna set...you want to set your margins for monitoring, you can actually decide what you want to do patient specific. So, in this case, we've set some tolerance level for the patients on the surface, and there are some x-ray verifications taken automatically at certain predefined gantry angles, so that you not only have inter-fractional monitoring using the surface, but in some cases where you need higher accuracy, then you can also choose to have an automated inter-factional monitoring based on the internal structures using the x-rays.
What you always see is the deviations calculated from the surface and the deviations calculated from the x-rays. Anytime there's a certain shift seen on the x-rays, the surface is set to the position of the x-rays. So, the x-rays are always your sort of ground truth. Now, in this special case, three verification images would have been taken. Those are the two blue dots on the bottom where you see below, patient movement. And so these are predefined gantry angles, where you...wherever you want to shoot x-rays. For VMAT, this would be the case. If you have a 3D plan, you can decide to shoot to or to take some x-rays after some monitor units have already been applied to the patient.
So, what did we do before? We did some very basic preclinical validation tests, as we are a clinic, where we already do a lot of surface guidance, and we have quite good routines for our surface guidance. One of the interesting parts for us is or has been, how does the surface look when the...when the x-rays look different, or how can you compare...how's the surface comparable to the x-rays? So we used exactly this warm phantom with a heated wire inside to have a temperature profile and to get the whole...to get everything we can out of the new ExacTrac. So, we have this modified phantom designed by Brainlab. And we just did just a couple...just 10 random isocentric couch deviations, some random translations, some random rotations to simulate isocentric uncertainties at couch zero, and we did a couple of deviations at two different couch angles to simulate inter-fractional movement. So, for each of these couch deviations, we had a value for the surface registration and we had the value for the x-ray registration, and we wanted to calculate the deviation between those two.
So, couch zero degrees, everything stayed below one millimeter. We had the largest deviations between surface and x-rays in the longitudinal direction. This is something where surface guidance in general has some problems with sort of a tube-like shaped object. It's always harder for any registration algorithm to grab on. And for inter-fractional monitoring, we had quite lower deviations. In this case also, the thermal component has been used more than the surface only. So, this is an addition of the thermal...sort of can be seen as an addition of the thermal information. In general, we can say that the differences between surface and x-ray positioning were always below one millimeter for a...obviously for a static phantom and for not...not for a patient.
As the main feature for the moment, and in my opinion, or in the opinion of many other people also is the...for the old ExacTrac is SRS or cranial treatments, I would like to go into the topic of cranial and SRS, and what is interesting in this case is a little bit...and I'm gonna give later on an outlook on what the future of the ExacTrac Dynamic will look like. But the first applications in our clinic and in most clinics which are installing the ExacTrac in the moment or the ExacTrac Dynamic will be cranial or SRS types of applications. So, just in general, why would you need special care in SRS or SBRT? This is something not new. We have smaller margins for error, and we have a higher dose, and we have one, single fraction only. And this has already affected some patients in the United States, some patients in Europe. So, everything...a lot of things can happen. [inaudible 00:22:46] cranial localizations accessories or some failures to properly set up some backup charge has actually led to the death of some patients. So, extra special care has to be taken when it comes to SRS.
How do we do it at LMU University Hospital in Munich? For SRS for multiple brain mets, we use a diagnostic high resolution contrast MRI of the brain with a slice thickness of 1 millimeter for target definition. We use a double-layered mask for patient immobilization. And we plan on a 2 millimeter planning CT with contrast. We use a 1 millimeter margin in addition to our GT read to get our PTV. So, 1 millimeter margin is not really a lot of room for errors. There are indeed multiple error sources in cranial SRS. Some are related to the symmetry as the output factors for small fields or the MSE calibration, which is also affecting the output factors. And there are imaging-related errors, some geometric distortions in the MR images, the accuracy of the image registration of the two systems, and obviously also the calibration of the imaging and the radiation isocenter. And in this edition, you also have mechanically related errors, and mechanically related errors can...also movement in the mask can...is considered a mechanically related error. And we will see later on that movement in...that there is a lot of movement in immobilization masks. So, monitoring of the internal structures is something that is quite crucial to SRS treatments.
Also you do have run-out of the gantry, you have run-outs of your collimator, and also you have run-outs of the couch. As we did a lot of...we have some experience with the as they call it, old ExacTrac over the general ExacTrac, not Dynamic, here in Munich, and we did some...and we're also doing multiple SRS treatments of multiple metastases, SRS treatments with one isocenter. I just want to share some brief experience of what happens if you go from isocentric to monocentric SRS. What you see here is, you have a 1 millimeter PTV...1 millimeter margin on your GTV, and in green that's the 80% isodose line. If you have one isocenter per metastasis and you have small translation, this might affect the coverage slightly. If you have small rotations for pretty much round metastasis, there's not a lot of...there's not that much what can happen in this case, but this is always "not that much" can also be quite high.
When it comes to a monocentric SRS, we have an isocenter outside of your metastasis, of your target volume. Some translations will also affect the position of the isocenter...or not the position of the isocenter, but the position of the metastasis, but if you have any rotations, you will get some very high errors. And in general, you can say, the smaller your target volume is and the higher your rotational error is and the higher the distance between your target is, it is also an increased risk of compromised target coverage. So, that's...these are reasons why we need special care. For the ExacTrac, we did a couch run-out test in our clinic and compared it with actual inter-fractional motion data from patients. And this can be a little complicated at the beginning, but what it says we measure the actual isocenter couch run-out using a anthropomorphic head phantom, so we did really basically our workflow, which we are doing for the patients, like positioning the head phantom on the couch, doing x-ray images, and then continuing with a normal workflow after verification and so on.
So, we turned the couch and we measured the clockwise couch run-out, and we saw that for this specific couch, it's not too bad. We had a couch error of 0.25 millimeter in the XY direction. And what you also see, the square...or the circles are the actual average non-corrected ExacTrac positioning results. So, for roughly, I think around 50 patients, we actually determined some thresholds on where we want to correct using ExacTrac. In our case it was 0.7 millimeters and 0.5 degrees of rotation. Everything which was below these cases or these values, we did not correct for couch rotations. And so we got some values, some mean average non-corrected ExacTrac positioning results. And we wanted to compare these results with the actual run out of the isocenter. And what we saw...and for those...for the status, we have no correlation...we could not observe any correlation between couch run-out and the patient setup. And we sort of thought that the positioning uncertainty mainly is therefore caused by inter-factional motion and we can detect the couch run-out by ExacTrac.
So, in general, for cranial SRS, small margins and a high single dose require a lot of special attention. One [inaudible 00:28:58] has multiple isocenters, the gantry, collimator, couch imaging system and could be also the cone-beam CT and your Brainlab ExacTrac system, or your surface guidance system. If you use KV imaging throughout the whole fraction for every culture angle and for every gantry angle, you can account for some of these isocentric errors, and you can carry out frameless radiosurgery with an accuracy of 1 millimeter. The increased potential of the dynamics workflow is...will be in the future. So, we have...obviously we have now...we already have x-ray tracking and x-ray inter-factional monitoring, but now in addition we have a surface guiding system, where you can actually monitor some parts of the...some regions of interest is...and once these once these regions of interest drift out of your desired position or out of your tolerances, you can in an integrated workflow, shoot or take some x-rays and correct for anything which you might not see in the beginning.
So, what you see here now is a video of the motion monitoring. So, on the right topside you see a patient, now the patient is moving out of some tolerances. And now, there's...some miracle happens, some magic happens, the patient has moved back by himself. What happens now, you're gonna shoot some x-rays, the system calculates difusion, the radiation is still going on, because you don't have any deviations higher than your tolerances. And if the patient would move out of the...would move out of some tolerances for the surface, you could shoot x-rays again, or stop the treatment or go inside and reposition the patient if you cannot reposition the patient using your HexaPOD couch in one workflow.
As I said before, and this is something I've said a lot already, you have surface motion monitoring during the entire fraction. So, this can be considered as an...or is an SGRT workflow. And an SGRT workflow automatically means you have additional patient safety. This is not the case for SRS only. This is the case for any other type of treatment. Every SGRT system at the moment available, as far as I know, has an automatic beam hold when the patient is out of tolerance. This is something which is a main safety factor...safety feature and will become more important in the future also.
In the case of the ExacTrac Dynamic, you now have automatic or manual extra triggering, if you have deviations between your planned and your actual surface. And you can directly reposition the patient. So, you have an integrated workflow, which has the potential to save a lot of time when it comes to not only SRS treatments, and all the timing is not that vital for SRS, but it can save you a lot of time if you use a system like that in a regular workflow. That's the potential we see.
In general, I'm gonna talk a little bit about imaging dose, and these are quite...the values are quite old. So, I know there have been some improvements and doses on cone-beam CT and on MV imaging, maybe also a little bit, but mostly on cone-beam CT or on KV portal imaging. But in general, the case is still all the same. You have quite a...you obviously have a high effective dose if you use MV portal images for positioning. Also, your...also your cone-bream CT doses, obviously, these are old, non-low-dose protocols. There have been novel low-dose protocols already used by a lot of people, but you still have quite a high dose and that's something quite natural for a cone-beam CT. And then you have for two orthogonal KV images. So, in a regular workflow, we could guess dependent obviously on the side and on the tolerance level you're using for the patients if it's SRS or non-SRS. You have 2 positioning images, 2 verification images and roughly 6 to 12 inter-factional images per fraction. So, you have a quite a low entrance dose, at least what you can sort of roughly accumulate compared to cone-beam CT positioning and obviously compared to MV positioning. And the goal...this is also something quite general, in any case, image dose reduction for specific treatment sides is one of the main goals and can be...this also has to be carried out with...regarding safety and regarding your tolerance levels, which are going down and down and down. And so, systems like the ExacTrac or systems that have surface imaging, in general have the potential to use...to safely use...reduce the imaging dose you have to take.
I have to also say, obviously, sometimes you need soft tissue contrast for specific sites and you have to evaluate your imaging protocols always for certain locations to have a possible dose reduction. And In this case, the rationale of identify those image guidance steps that can potentially be accomplished without the use of ionizing radiation and planning the image technique to be consistent with the image quality for the treatment decision, and these are also...this is the perfect summarization of a system which uses SGRT in a sort of a safe way, because you have some imaging dose obviously left, but you can reduce it for specific sites. But this has to be still evaluated. Which sites are the perfect ones for it, obviously, everyone sort of has an idea.
Are some cone-beam CT protocols replaceable?
[00:35:52]
[silence]
[00:36:07]
Phillip: ...which is now considered still the gold standard of positioning, you might say, so we want to compare the cone-beam CT positioning with the ExacTrac positioning to see if we can reduce some dose by positioning some patients with the ExacTrac. So, we compare the detected errors by ExacTrac and the cone-beam CT positioning errors for 32 patients, and we saw that those 2 systems are in quite close agreement, at least here for the Elekta site and we have an...we had neglectable cross model discrepancies in the position detection for 32 patients. And this patient data, obviously the accuracy and precision for phantoms was very, very low.
So now the future, one major part, when we first got contacted by Brainlab to start off with the validation of the system, they contacted us because one part is quite essential of a surface guided system and it's deep inspiration breath hold. I know Brainlab has put a lot of work into making the system the way it works right now, and it's...and we think it works and the workflow is quite sophisticated now. And therefore, deep inspiration breath hold will not be or is not available in the first version. I know that Brainlab is working on a workflow, and it's expected sometime this year where you have the deep inspiration or a gating workflow. And this is something quite interesting, because that could be a real reduction in imaging dose, and we in our clinic, we're positioning the patients in deep inspiration breath hold or any breath patients with MV imaging. And three or four MV images a week, obviously accumulating quite a high dose, the patients are sometimes also quite young, so you really want to reduce the dose in this case, and you want to increase the time efficiency. And this has to...potential we have to look into it what the final version will be, but this has the potential to have a very nice deep inspiration workflow, because you have patient positioning and you have patient monitoring, not using...not looking at the surface only, but the idea is you monitor the chest wall or the spine also at first and the chest wall, and you also monitor the heart, the most essential part of a deep inspiration breath hold workflow.
This is a short video obviously by Brainlab, but I did not do it myself for the workflow of deep inspiration breath hold. So, you also select a region of interest, you have your thermal tracking, and then, you monitor one spot or one region on the patient surface. And as soon as the patient is breathing into her breath hold surface, you are taking some x-rays, now that was good timing. And so, you have x-ray verification, and then, there are a couple of things you look at. You look at the frontal ribs. And afterwards you can also look at the position of the spine and the DRRs you get. And now the interesting part, you see the position of the heart as your most vital structure for deep inspiration breath hold treatments. And after your verification, you go into treatment and you can independently...of the gantry position, you can always monitor the position. And you have a surface triggered beam hold just as any regular inter-fractional motion monitoring.
So, this one is for the slides. I'm quite happy to get a lot of questions from you. I shortened it a little bit, so we have more time for questions. And I want to thank also my team who are working on the project. And what you see here, this is actually the thermal-to-surface calibration image, and it was quite a warm day when we took it. So, thank you very much for your attention. I'm very happy to get [inaudible 00:40:48] questions now.
Bogdan: Hi Philipp, it's Bogdan. I think we already had some technical difficulties broadcasting with too many participants. So, I haven't seen too many questions. So, maybe I'll ask you some. But for anyone that's still on the chat line, if you have questions, feel free to send it to us.
So, as you plan on going clinical with the system, you know, maybe...I don't know if you put any thought into it, but what kind of validation testings have you thought about doing, and also, what do you think will be some of the first patient treatments do you think you'll undergo?
Philipp: So, the first patient treatment... So our plan is to get clinical. Obviously, the whole corona crisis have...has stopped a lot of the processes we wanted to have and we wanted accelerated at this moment, but this is one of the process that has been stopped around the world. But as soon as we get clinical, the first patients we will have I think which is already scheduled will be a multiple brain mets patient with an open mask, and one single isocenter. And we will use it with an open mask on the system, so the workflow will be pretty much the same as on a regular ExacTrac, and...but we have a motion monitoring of the open mask of the patient movement inside the mask, and then, triggered x-ray imaging to control...to verify the position and obviously to control the position inside the mask during the whole treatment fraction.
Bogdan: It shows there's a few questions from the participants. How do you calibrate the thermal system and is there even a need to calibrate the thermal system?
Philipp: A calibration on the thermal system is not something we have looked into at the moment. There are some calibrations phantoms provided by Brainlab. And I think we have a thermal-to-surface calibration. But this is done with a regular phantom, and we are not 100% sure how this workflow will look like when we get it.
Bogdan: Here's a more technical question. I'm not sure if we can answer this. What are the requirements for installations?
Philipp: I know that one of the requirements is a HexaPODic table if you want to use the whole workflow. You can always...the system has two parts in it. So, you can either choose a HexaPODic or a full ExacTrac only workflow, or you can use a cone-beam CT workflow. So, at the moment, for at least I can say for Elekta installations, you can only send...you can...you have your cone-beam CT and you position according to the cone-beam CT, and after that you monitor using the ExacTrac. So this is one of the possibilities. The requirements for installation, we will install it on an...on a brain...on an Elekta Versa HD with a HexaPODic table.
Bogdan: Right. So, Elekta configurations, Versa HDs with HexaPOD are supported for Varian TrueBeams with PerfectPitch are supported as well. Besides that, you know the same x-ray requirements apply to ExacTrac Dynamic as the previous versions of ExacTrac. If most of you remember, we had started shipping ExacTrac 6.5s with the same x-ray technology as is now present in ExacTrac Dynamic, and we have a new installation data package for the new camera mount which goes on the caudal end of the couch. So, there's just one camera system that attaches to the ceiling and we provide electrical guidance on how to install that in the treatment room. Okay, we have more questions. What is the average time...treatment time with the ExacTrac Dynamic?
Philipp: Hard to say. The average treatment time, I know I had some figures on average treatment time with a regular ExacTrac for multiple brain mets, or obviously, depending on how many metastases there are. But in general, what you could...what I would think is that, if you're not using cone-bream CT positioning, if you can get sort of rid of your cone-bream CT positioning and use a full ExacTrac workflow, you have the potential to reduce your treatment time quite largely. Obviously, the cone-bream CT takes some time. The registration, not so much anymore. But where you can save time is actually running inside repositioning the patients. This is also something quite difficult to ask, because if you have...now you have additional information, and that's a general problem with everything where you monitor the surfaces all the time. The first question is, can you reduce treatment time with surface guidance? Yes, you can, because you have...you might have a workflow where you can reduce it, but you also see more. And that's...that might be the main issue. Now, you see the patient moving all the time. You see the patient moving, and now you have to compensate for this movement. You don't see that if you don't have a system like that, where you see it...see it a little later. The...what is...there have been some studies, I don't have it on me right now, but the general...what I've heard from other clinics is, a lot of movement is happening after the patient is positioned by the RTTs. And the RTTs turn their back around and leave the Linac room, that's the time when the patient starts stretching a little bit, starts breathing in a little bit. So, all systems which monitor the patient all the time, not only inter-fractionally, but also before the fraction even starts after the initial positioning. All of these systems help you in gaining accuracy, gaining safety, and if you use a fully integrated workflow, also getting a little bit faster.
Bogdan: Here's another question that I'm sure you don't have yet to hit out to answer, but has the system been used in pediatric patients and what are recommendations? So, obviously, you're not yet clinical with the system. But if you are planning on doing pediatric cases, what kind of measures would you have in place?
Philipp: There are possibilities of doing pediatric patients. At the moment, we're using the old ExacTrac on pediatric patients. Obviously, pediatric patients if they are not under...narcotized fully, they will move, and a surface guidance system or a system like that...in general, a surface-guided system is very helpful for pediatric patients, although there's not really a lot of data on this. And in...and for pediatric patients you want to be extra precise and extra accurate. So, it has the potential for pediatric patients, yes, but there are no recommendations on it yet, and...but I think we will look into it as we install the system. We have some...but we see some potential.
Bogdan: Okay, and last question. Do you have any recommendations for future customers on necessary steps to commission ExacTrac Dynamic?
Philipp: So...it...that's quite a tough question. Commissioning is always a little tough. It sort of depends on what kind of phantoms you have. And the commissioning will also be dependent on your workflow. So, if you have a pure cranial workflow, you might not really need a heated phantom or you might not really need anything with thermal information inside, because a head or an open mask in the head has a lot for a registration algorithm to grab on compared to a tube-like surface like the abdomen. So, you could use a regular phantom, you could use a regular end-to-end test which is already...have been done in commissioning. This is a system in the beginning. So, a clear recommendation on commissioning is still tough, but we are working on it at the moment. After we get the final version of it, we wanted to learn something in the last few weeks, but obviously we did not really have time for playing around so much at the moment. But in general, I would say, an end-to-end test is possible for the first cranial versions. The commissioning using a standard head phantom with some bony information inside is also feasible, also with films. When it comes to any deep inspiration breath hold workflows, there's obviously a lot of more commissioning to be done regarding latency. Regarding also now the thermal information, you might need a thermal phantom or a heated phantom to use, and that's going to be quite of the tough part in the future.
Bogdan: Okay. Philipp, thank you very much for your talk, and we look forward to hearing more from you as both you and Rigshospitalet are going clinical with the system. So, again thank you, and thank you everybody that managed to get in, and we will have the recording available for everyone else.
Philipp: All right. So, thank you very much. And as soon as we...we will start...as I said before, we will start hopefully in a month with everything clinical, we're getting clinical and as soon as we have some more data on commissioning, as soon as we have sort of recommendation on commissioning, we will also start publishing our information as fast as we can, and we might think of another webinar where we can actually present our first clinical data. But this is something we have to look at after sort of we figure out all of the current drawbacks and the crisis.
Bogdan: Great. Well, thank you very much, Philipp. And stay safe everyone. Have a good day.
Philipp: Thank you. Goodbye.
Bogdan: Bye.
Together with me today is Philipp Freislederer from Ludwig Maximilians University in Munich, and he will cover some new topics pertaining to ExacTrac Dynamic. LMU together with a group in Copenhagen have been the two early adopter sides for ExacTrac Dynamic. They've had the chance to work with the products as it was still under development, and they can share with you some unique experiences.
Before getting into the webinar, just a few topics. First as you remember, from the webinar interface the majority of you just have the opportunity to listen. So, feel free to send us via the chat line questions. We will put all those together, and at the end of the presentation, we will go through a question and answer session. I would also like to remind you that you have to login via Google Chrome or Safari. EasyWebinar isn't terribly friendly with Microsoft products, so Internet Explorer Edge will not let you see the webinar. We're also recording everything, so the information will be available afterwards on novaliscircle.org. We are providing continuing education credits, and if you'd like either a CAMPEP, MDCB or ASRT credits following the webinar, please feel free to send an email to info@novaliscircle.org, and we'll provide you with specific instructions on how you can obtain credits.
Lastly, we are happy to announce that ExacTrac Dynamic has been released in the European community. However, it is not yet available in all other markets. So, we intend this webinar to be for educational purposes only, and please bear in mind the disclaimer attached. Should you have questions in your local markets for product availability, please ask our commercial teams for updates on that end.
So, with that being said, I'd like to again thank you, Mr. Freislederer for his time today, and we'll look forward to his presentation.
Philipp: Well, thank you very much for the introduction. I'm gonna now turn my camera off for the moment. And if that's all right with you, I can also leave it on, but I will not see your comments at the moment. All of the questions will be gathered by Brainlab later on, and we can discuss it afterwards. I will now switch to...one second. This is technically now the hard part for it.
First of all, thank you very much Brainlab for letting me have this webinar with every one of you. One second. Okay, I'm guessing everyone can now see my slides. And I'm gonna just start off with the webinar. So, Brainlab has asked me to do a webinar on the ExacTrac Dynamic as Bogdan already said before. We're using...we're not using the system, because we just got the...just the CE mark has just been released. and...but we were able to test it a little bit to do some validations together with...not together with, but also Rigshospitalet in Denmark, in Copenhagen did some validations for the ExacTrac Dynamic. And obviously, the COVID-19, the corona crisis has struck everybody, and we are waiting to get the final installation at the moment to do some additional tests, but at least we did some validation testing before, so we had the chance to play around with the system a little bit.
So, this is going to be my topic, ExacTrac Dynamic solution for IGRT/SGRT for Linac-based radiosurgery. I have to disclose some things. My department has research grants from Brainlab. And I also am receiving a speaking fee.
What is the ExacTrac Dynamic? So, we have a combination of an optical surface scanner using structured light. And additionally, and that's sort of the main new addition to it, is a thermal camera. So, why are you using a thermal camera? There are various options...various reasons for it. One very important part is that you have additional registration information. The thermal camera, the thermal information serves as a sort of a fourth dimension in addition to the optical surface scanner. It also makes multiple in-room cameras not necessary anymore. You can sort of see it as a system where the optical surface scanner with a structured light provides you three dimensions, and using the thermal information, you have additional access to something, or the registration algorithm has additional access to something which it can grab on.
In addition, you have x-ray positioning and monitoring just as the "old ExacTrac." So, you have your usual...or x-ray positioning, x-ray monitoring. So, this serves now as a surface guided radiation therapy tool, because you have real-time tracking using the surface and the thermal information, and also using the x-rays. And as Bogdan said before, it has now been CE certified, so after everything is getting...it's gonna get a little bit more quiet with coronas. Hopefully some clinics are able to install this system and use it.
I'm gonna guide you a little bit to the workflow. Unfortunately, these are all some slides with phantoms, because we didn't have the chance to get...obviously didn't have the chance to have patients on it yet. So, I'm guide...I'm gonna guide you through the workflow, how it is proposed in the last version.
The system starts off with some prepositioning based on the surface. You see it here, there's a phantom being placed and you see some shifts that the system calculates. These shifts are afterwards sent to the couch. And after you move your couch with the shifts, obviously, you should get in a very good direction of prepositioning, and this is all considered prepositioning. So, the...I'm gonna come to that after on. The x-ray is still something that's quite important in all the cases in this...in this version. So, these are obviously phantom and obviously therefore very small. This is expected. But this considers...this is considered a rough prepositioning according to the Brainlab R&D department.
Afterwards, you see, this is the camera view with thermal information from the thermal and surface camera. You select your region of interest. In this case, what you see is a specially designed phantom by Brainlab, also where you have a heated wire inside. So, you see the red areas are so...are the warmer ones and the green areas in this case where there's no heated wire. So, you select your region [crosstalk 00:10:54]
Bogdan: Philipp, one second. If you don't mind the interruption, just making sure that you've entered the presentation mode on the PowerPoint. We were still on the introductory slide. Sorry.
Philipp: Oh, I'm very sorry. I'm very sorry. One second, one second, one second. I'm very sorry. I might have gotten the wrong button here. So...better now?
Bogdan: Yeah. If you don't mind going back some slides.
Philipp: Okay.
Bogdan: Nobody was actually able to see it. Thank you.
Philipp: I'm gonna go back again. So, this is actually the first actual real slide. As I said before, you have an optical surface scanner and you have a thermal camera, and then you sort of see the thermal camera as the fourth dimension. And using your x-rays, you have the additional positioning and inter-fractional monitoring. And this is the part where it actually gets interesting. So, you do your pre-position workflow. You get some values for your pre-position based on the surface only. These values are then sent to the couch to enable a 3D movement. And afterwards the...here in this case, the phantom or the patient afterwards is positioned quite all right as I said before, in a phantom...in the stationary phantom, obviously quite good values. And now you should see the heated phantom with the heated wire inside. So, the red and yellow spots on the surfaces is actually where the heated wires inside and you get a temperature profile. And in our case, you can select them...in any case, you can select the temperature profile for the session or for the couch angle, and you select your temperature profile, or in this case and so you select your region of interest for the specific patient, for the specific fraction.
After the selection of your region of interest, you acquire x-rays, and you acquire dual x-rays. You have the console. In this case the phantom also has some...has a spine inside, has some ribs, has some bones inside, so this...it also here has some heated wire inside, so this makes it a little easier for any registration algorithm, but I'm gonna come to that afterwards. So, you acquire your x-rays and you use the x-rays afterwards with the registration with a previously generated DRRs. You use same as the...same case as the old ExacTrac, you use the x-ray DRR registration for your initial positioning. These values you're getting from the x-rays are then afterwards sent to your six-dimensional table in our...for Elekta cassettes, this would be the HexaPOD table, and you can enable directly the couch movement also for any rotations.
You can choose afterwards to verify the position, and obviously for the phantom, we can easily verify everything, and you see that the registration now gives you obviously very good results. I was...from a physics point of view, I was quite happy to see that at least we have a vertical deviation of 0.1 millimeter. If it would have been 0 millimeter in all directions, that would have made me think a little bit if I really could trust everything I will see, but at least, obviously it's a static phantom and no dynamic movement, and this should work out quite fine.
After your x-ray verification, you go into monitoring or tracking mode. So, on the right upper side, now you see the patient, or in this case the phantom surface, and your region of interest which you have chosen before. And you'll also see the life online values for the full surface and thermal information provided by the system, and you see deviations calculated to the isocenter of the plan. You can then choose afterwards to enable automatic x-ray mode or you can take an x-ray in the beginning. This is still quite open and you can...just you can start treatment afterwards.
So, you have an initial x-ray, and every time you take an X-ray, obviously this...or not obviously, but the x-ray is always chosen to be the ground truth in this case and in any case. So, you have your x-rays. You really monitor internal or bony registration, and the surface just in this case provides...it gives you a hint of how the patient is behaving and gives you some monitoring. So, during monitoring, you monitor the surface and you can decide if you want to...what do you do afterwards. There are several options in this case, one option would be is if the surface goes out of tolerance obviously it is easily monitored without any additional dose. If the surface goes out of tolerance, the beam can stop or the beam can go on. You can have automatic x-rays to be taken, dual or mono depending on the gantry position at the moment, or you can continue and just monitor, or you can stop the beam and take x-rays. That's pretty much for any patient or any...depending on how tight you wanna set...you want to set your margins for monitoring, you can actually decide what you want to do patient specific. So, in this case, we've set some tolerance level for the patients on the surface, and there are some x-ray verifications taken automatically at certain predefined gantry angles, so that you not only have inter-fractional monitoring using the surface, but in some cases where you need higher accuracy, then you can also choose to have an automated inter-factional monitoring based on the internal structures using the x-rays.
What you always see is the deviations calculated from the surface and the deviations calculated from the x-rays. Anytime there's a certain shift seen on the x-rays, the surface is set to the position of the x-rays. So, the x-rays are always your sort of ground truth. Now, in this special case, three verification images would have been taken. Those are the two blue dots on the bottom where you see below, patient movement. And so these are predefined gantry angles, where you...wherever you want to shoot x-rays. For VMAT, this would be the case. If you have a 3D plan, you can decide to shoot to or to take some x-rays after some monitor units have already been applied to the patient.
So, what did we do before? We did some very basic preclinical validation tests, as we are a clinic, where we already do a lot of surface guidance, and we have quite good routines for our surface guidance. One of the interesting parts for us is or has been, how does the surface look when the...when the x-rays look different, or how can you compare...how's the surface comparable to the x-rays? So we used exactly this warm phantom with a heated wire inside to have a temperature profile and to get the whole...to get everything we can out of the new ExacTrac. So, we have this modified phantom designed by Brainlab. And we just did just a couple...just 10 random isocentric couch deviations, some random translations, some random rotations to simulate isocentric uncertainties at couch zero, and we did a couple of deviations at two different couch angles to simulate inter-fractional movement. So, for each of these couch deviations, we had a value for the surface registration and we had the value for the x-ray registration, and we wanted to calculate the deviation between those two.
So, couch zero degrees, everything stayed below one millimeter. We had the largest deviations between surface and x-rays in the longitudinal direction. This is something where surface guidance in general has some problems with sort of a tube-like shaped object. It's always harder for any registration algorithm to grab on. And for inter-fractional monitoring, we had quite lower deviations. In this case also, the thermal component has been used more than the surface only. So, this is an addition of the thermal...sort of can be seen as an addition of the thermal information. In general, we can say that the differences between surface and x-ray positioning were always below one millimeter for a...obviously for a static phantom and for not...not for a patient.
As the main feature for the moment, and in my opinion, or in the opinion of many other people also is the...for the old ExacTrac is SRS or cranial treatments, I would like to go into the topic of cranial and SRS, and what is interesting in this case is a little bit...and I'm gonna give later on an outlook on what the future of the ExacTrac Dynamic will look like. But the first applications in our clinic and in most clinics which are installing the ExacTrac in the moment or the ExacTrac Dynamic will be cranial or SRS types of applications. So, just in general, why would you need special care in SRS or SBRT? This is something not new. We have smaller margins for error, and we have a higher dose, and we have one, single fraction only. And this has already affected some patients in the United States, some patients in Europe. So, everything...a lot of things can happen. [inaudible 00:22:46] cranial localizations accessories or some failures to properly set up some backup charge has actually led to the death of some patients. So, extra special care has to be taken when it comes to SRS.
How do we do it at LMU University Hospital in Munich? For SRS for multiple brain mets, we use a diagnostic high resolution contrast MRI of the brain with a slice thickness of 1 millimeter for target definition. We use a double-layered mask for patient immobilization. And we plan on a 2 millimeter planning CT with contrast. We use a 1 millimeter margin in addition to our GT read to get our PTV. So, 1 millimeter margin is not really a lot of room for errors. There are indeed multiple error sources in cranial SRS. Some are related to the symmetry as the output factors for small fields or the MSE calibration, which is also affecting the output factors. And there are imaging-related errors, some geometric distortions in the MR images, the accuracy of the image registration of the two systems, and obviously also the calibration of the imaging and the radiation isocenter. And in this edition, you also have mechanically related errors, and mechanically related errors can...also movement in the mask can...is considered a mechanically related error. And we will see later on that movement in...that there is a lot of movement in immobilization masks. So, monitoring of the internal structures is something that is quite crucial to SRS treatments.
Also you do have run-out of the gantry, you have run-outs of your collimator, and also you have run-outs of the couch. As we did a lot of...we have some experience with the as they call it, old ExacTrac over the general ExacTrac, not Dynamic, here in Munich, and we did some...and we're also doing multiple SRS treatments of multiple metastases, SRS treatments with one isocenter. I just want to share some brief experience of what happens if you go from isocentric to monocentric SRS. What you see here is, you have a 1 millimeter PTV...1 millimeter margin on your GTV, and in green that's the 80% isodose line. If you have one isocenter per metastasis and you have small translation, this might affect the coverage slightly. If you have small rotations for pretty much round metastasis, there's not a lot of...there's not that much what can happen in this case, but this is always "not that much" can also be quite high.
When it comes to a monocentric SRS, we have an isocenter outside of your metastasis, of your target volume. Some translations will also affect the position of the isocenter...or not the position of the isocenter, but the position of the metastasis, but if you have any rotations, you will get some very high errors. And in general, you can say, the smaller your target volume is and the higher your rotational error is and the higher the distance between your target is, it is also an increased risk of compromised target coverage. So, that's...these are reasons why we need special care. For the ExacTrac, we did a couch run-out test in our clinic and compared it with actual inter-fractional motion data from patients. And this can be a little complicated at the beginning, but what it says we measure the actual isocenter couch run-out using a anthropomorphic head phantom, so we did really basically our workflow, which we are doing for the patients, like positioning the head phantom on the couch, doing x-ray images, and then continuing with a normal workflow after verification and so on.
So, we turned the couch and we measured the clockwise couch run-out, and we saw that for this specific couch, it's not too bad. We had a couch error of 0.25 millimeter in the XY direction. And what you also see, the square...or the circles are the actual average non-corrected ExacTrac positioning results. So, for roughly, I think around 50 patients, we actually determined some thresholds on where we want to correct using ExacTrac. In our case it was 0.7 millimeters and 0.5 degrees of rotation. Everything which was below these cases or these values, we did not correct for couch rotations. And so we got some values, some mean average non-corrected ExacTrac positioning results. And we wanted to compare these results with the actual run out of the isocenter. And what we saw...and for those...for the status, we have no correlation...we could not observe any correlation between couch run-out and the patient setup. And we sort of thought that the positioning uncertainty mainly is therefore caused by inter-factional motion and we can detect the couch run-out by ExacTrac.
So, in general, for cranial SRS, small margins and a high single dose require a lot of special attention. One [inaudible 00:28:58] has multiple isocenters, the gantry, collimator, couch imaging system and could be also the cone-beam CT and your Brainlab ExacTrac system, or your surface guidance system. If you use KV imaging throughout the whole fraction for every culture angle and for every gantry angle, you can account for some of these isocentric errors, and you can carry out frameless radiosurgery with an accuracy of 1 millimeter. The increased potential of the dynamics workflow is...will be in the future. So, we have...obviously we have now...we already have x-ray tracking and x-ray inter-factional monitoring, but now in addition we have a surface guiding system, where you can actually monitor some parts of the...some regions of interest is...and once these once these regions of interest drift out of your desired position or out of your tolerances, you can in an integrated workflow, shoot or take some x-rays and correct for anything which you might not see in the beginning.
So, what you see here now is a video of the motion monitoring. So, on the right topside you see a patient, now the patient is moving out of some tolerances. And now, there's...some miracle happens, some magic happens, the patient has moved back by himself. What happens now, you're gonna shoot some x-rays, the system calculates difusion, the radiation is still going on, because you don't have any deviations higher than your tolerances. And if the patient would move out of the...would move out of some tolerances for the surface, you could shoot x-rays again, or stop the treatment or go inside and reposition the patient if you cannot reposition the patient using your HexaPOD couch in one workflow.
As I said before, and this is something I've said a lot already, you have surface motion monitoring during the entire fraction. So, this can be considered as an...or is an SGRT workflow. And an SGRT workflow automatically means you have additional patient safety. This is not the case for SRS only. This is the case for any other type of treatment. Every SGRT system at the moment available, as far as I know, has an automatic beam hold when the patient is out of tolerance. This is something which is a main safety factor...safety feature and will become more important in the future also.
In the case of the ExacTrac Dynamic, you now have automatic or manual extra triggering, if you have deviations between your planned and your actual surface. And you can directly reposition the patient. So, you have an integrated workflow, which has the potential to save a lot of time when it comes to not only SRS treatments, and all the timing is not that vital for SRS, but it can save you a lot of time if you use a system like that in a regular workflow. That's the potential we see.
In general, I'm gonna talk a little bit about imaging dose, and these are quite...the values are quite old. So, I know there have been some improvements and doses on cone-beam CT and on MV imaging, maybe also a little bit, but mostly on cone-beam CT or on KV portal imaging. But in general, the case is still all the same. You have quite a...you obviously have a high effective dose if you use MV portal images for positioning. Also, your...also your cone-bream CT doses, obviously, these are old, non-low-dose protocols. There have been novel low-dose protocols already used by a lot of people, but you still have quite a high dose and that's something quite natural for a cone-beam CT. And then you have for two orthogonal KV images. So, in a regular workflow, we could guess dependent obviously on the side and on the tolerance level you're using for the patients if it's SRS or non-SRS. You have 2 positioning images, 2 verification images and roughly 6 to 12 inter-factional images per fraction. So, you have a quite a low entrance dose, at least what you can sort of roughly accumulate compared to cone-beam CT positioning and obviously compared to MV positioning. And the goal...this is also something quite general, in any case, image dose reduction for specific treatment sides is one of the main goals and can be...this also has to be carried out with...regarding safety and regarding your tolerance levels, which are going down and down and down. And so, systems like the ExacTrac or systems that have surface imaging, in general have the potential to use...to safely use...reduce the imaging dose you have to take.
I have to also say, obviously, sometimes you need soft tissue contrast for specific sites and you have to evaluate your imaging protocols always for certain locations to have a possible dose reduction. And In this case, the rationale of identify those image guidance steps that can potentially be accomplished without the use of ionizing radiation and planning the image technique to be consistent with the image quality for the treatment decision, and these are also...this is the perfect summarization of a system which uses SGRT in a sort of a safe way, because you have some imaging dose obviously left, but you can reduce it for specific sites. But this has to be still evaluated. Which sites are the perfect ones for it, obviously, everyone sort of has an idea.
Are some cone-beam CT protocols replaceable?
[00:35:52]
[silence]
[00:36:07]
Phillip: ...which is now considered still the gold standard of positioning, you might say, so we want to compare the cone-beam CT positioning with the ExacTrac positioning to see if we can reduce some dose by positioning some patients with the ExacTrac. So, we compare the detected errors by ExacTrac and the cone-beam CT positioning errors for 32 patients, and we saw that those 2 systems are in quite close agreement, at least here for the Elekta site and we have an...we had neglectable cross model discrepancies in the position detection for 32 patients. And this patient data, obviously the accuracy and precision for phantoms was very, very low.
So now the future, one major part, when we first got contacted by Brainlab to start off with the validation of the system, they contacted us because one part is quite essential of a surface guided system and it's deep inspiration breath hold. I know Brainlab has put a lot of work into making the system the way it works right now, and it's...and we think it works and the workflow is quite sophisticated now. And therefore, deep inspiration breath hold will not be or is not available in the first version. I know that Brainlab is working on a workflow, and it's expected sometime this year where you have the deep inspiration or a gating workflow. And this is something quite interesting, because that could be a real reduction in imaging dose, and we in our clinic, we're positioning the patients in deep inspiration breath hold or any breath patients with MV imaging. And three or four MV images a week, obviously accumulating quite a high dose, the patients are sometimes also quite young, so you really want to reduce the dose in this case, and you want to increase the time efficiency. And this has to...potential we have to look into it what the final version will be, but this has the potential to have a very nice deep inspiration workflow, because you have patient positioning and you have patient monitoring, not using...not looking at the surface only, but the idea is you monitor the chest wall or the spine also at first and the chest wall, and you also monitor the heart, the most essential part of a deep inspiration breath hold workflow.
This is a short video obviously by Brainlab, but I did not do it myself for the workflow of deep inspiration breath hold. So, you also select a region of interest, you have your thermal tracking, and then, you monitor one spot or one region on the patient surface. And as soon as the patient is breathing into her breath hold surface, you are taking some x-rays, now that was good timing. And so, you have x-ray verification, and then, there are a couple of things you look at. You look at the frontal ribs. And afterwards you can also look at the position of the spine and the DRRs you get. And now the interesting part, you see the position of the heart as your most vital structure for deep inspiration breath hold treatments. And after your verification, you go into treatment and you can independently...of the gantry position, you can always monitor the position. And you have a surface triggered beam hold just as any regular inter-fractional motion monitoring.
So, this one is for the slides. I'm quite happy to get a lot of questions from you. I shortened it a little bit, so we have more time for questions. And I want to thank also my team who are working on the project. And what you see here, this is actually the thermal-to-surface calibration image, and it was quite a warm day when we took it. So, thank you very much for your attention. I'm very happy to get [inaudible 00:40:48] questions now.
Bogdan: Hi Philipp, it's Bogdan. I think we already had some technical difficulties broadcasting with too many participants. So, I haven't seen too many questions. So, maybe I'll ask you some. But for anyone that's still on the chat line, if you have questions, feel free to send it to us.
So, as you plan on going clinical with the system, you know, maybe...I don't know if you put any thought into it, but what kind of validation testings have you thought about doing, and also, what do you think will be some of the first patient treatments do you think you'll undergo?
Philipp: So, the first patient treatment... So our plan is to get clinical. Obviously, the whole corona crisis have...has stopped a lot of the processes we wanted to have and we wanted accelerated at this moment, but this is one of the process that has been stopped around the world. But as soon as we get clinical, the first patients we will have I think which is already scheduled will be a multiple brain mets patient with an open mask, and one single isocenter. And we will use it with an open mask on the system, so the workflow will be pretty much the same as on a regular ExacTrac, and...but we have a motion monitoring of the open mask of the patient movement inside the mask, and then, triggered x-ray imaging to control...to verify the position and obviously to control the position inside the mask during the whole treatment fraction.
Bogdan: It shows there's a few questions from the participants. How do you calibrate the thermal system and is there even a need to calibrate the thermal system?
Philipp: A calibration on the thermal system is not something we have looked into at the moment. There are some calibrations phantoms provided by Brainlab. And I think we have a thermal-to-surface calibration. But this is done with a regular phantom, and we are not 100% sure how this workflow will look like when we get it.
Bogdan: Here's a more technical question. I'm not sure if we can answer this. What are the requirements for installations?
Philipp: I know that one of the requirements is a HexaPODic table if you want to use the whole workflow. You can always...the system has two parts in it. So, you can either choose a HexaPODic or a full ExacTrac only workflow, or you can use a cone-beam CT workflow. So, at the moment, for at least I can say for Elekta installations, you can only send...you can...you have your cone-beam CT and you position according to the cone-beam CT, and after that you monitor using the ExacTrac. So this is one of the possibilities. The requirements for installation, we will install it on an...on a brain...on an Elekta Versa HD with a HexaPODic table.
Bogdan: Right. So, Elekta configurations, Versa HDs with HexaPOD are supported for Varian TrueBeams with PerfectPitch are supported as well. Besides that, you know the same x-ray requirements apply to ExacTrac Dynamic as the previous versions of ExacTrac. If most of you remember, we had started shipping ExacTrac 6.5s with the same x-ray technology as is now present in ExacTrac Dynamic, and we have a new installation data package for the new camera mount which goes on the caudal end of the couch. So, there's just one camera system that attaches to the ceiling and we provide electrical guidance on how to install that in the treatment room. Okay, we have more questions. What is the average time...treatment time with the ExacTrac Dynamic?
Philipp: Hard to say. The average treatment time, I know I had some figures on average treatment time with a regular ExacTrac for multiple brain mets, or obviously, depending on how many metastases there are. But in general, what you could...what I would think is that, if you're not using cone-bream CT positioning, if you can get sort of rid of your cone-bream CT positioning and use a full ExacTrac workflow, you have the potential to reduce your treatment time quite largely. Obviously, the cone-bream CT takes some time. The registration, not so much anymore. But where you can save time is actually running inside repositioning the patients. This is also something quite difficult to ask, because if you have...now you have additional information, and that's a general problem with everything where you monitor the surfaces all the time. The first question is, can you reduce treatment time with surface guidance? Yes, you can, because you have...you might have a workflow where you can reduce it, but you also see more. And that's...that might be the main issue. Now, you see the patient moving all the time. You see the patient moving, and now you have to compensate for this movement. You don't see that if you don't have a system like that, where you see it...see it a little later. The...what is...there have been some studies, I don't have it on me right now, but the general...what I've heard from other clinics is, a lot of movement is happening after the patient is positioned by the RTTs. And the RTTs turn their back around and leave the Linac room, that's the time when the patient starts stretching a little bit, starts breathing in a little bit. So, all systems which monitor the patient all the time, not only inter-fractionally, but also before the fraction even starts after the initial positioning. All of these systems help you in gaining accuracy, gaining safety, and if you use a fully integrated workflow, also getting a little bit faster.
Bogdan: Here's another question that I'm sure you don't have yet to hit out to answer, but has the system been used in pediatric patients and what are recommendations? So, obviously, you're not yet clinical with the system. But if you are planning on doing pediatric cases, what kind of measures would you have in place?
Philipp: There are possibilities of doing pediatric patients. At the moment, we're using the old ExacTrac on pediatric patients. Obviously, pediatric patients if they are not under...narcotized fully, they will move, and a surface guidance system or a system like that...in general, a surface-guided system is very helpful for pediatric patients, although there's not really a lot of data on this. And in...and for pediatric patients you want to be extra precise and extra accurate. So, it has the potential for pediatric patients, yes, but there are no recommendations on it yet, and...but I think we will look into it as we install the system. We have some...but we see some potential.
Bogdan: Okay, and last question. Do you have any recommendations for future customers on necessary steps to commission ExacTrac Dynamic?
Philipp: So...it...that's quite a tough question. Commissioning is always a little tough. It sort of depends on what kind of phantoms you have. And the commissioning will also be dependent on your workflow. So, if you have a pure cranial workflow, you might not really need a heated phantom or you might not really need anything with thermal information inside, because a head or an open mask in the head has a lot for a registration algorithm to grab on compared to a tube-like surface like the abdomen. So, you could use a regular phantom, you could use a regular end-to-end test which is already...have been done in commissioning. This is a system in the beginning. So, a clear recommendation on commissioning is still tough, but we are working on it at the moment. After we get the final version of it, we wanted to learn something in the last few weeks, but obviously we did not really have time for playing around so much at the moment. But in general, I would say, an end-to-end test is possible for the first cranial versions. The commissioning using a standard head phantom with some bony information inside is also feasible, also with films. When it comes to any deep inspiration breath hold workflows, there's obviously a lot of more commissioning to be done regarding latency. Regarding also now the thermal information, you might need a thermal phantom or a heated phantom to use, and that's going to be quite of the tough part in the future.
Bogdan: Okay. Philipp, thank you very much for your talk, and we look forward to hearing more from you as both you and Rigshospitalet are going clinical with the system. So, again thank you, and thank you everybody that managed to get in, and we will have the recording available for everyone else.
Philipp: All right. So, thank you very much. And as soon as we...we will start...as I said before, we will start hopefully in a month with everything clinical, we're getting clinical and as soon as we have some more data on commissioning, as soon as we have sort of recommendation on commissioning, we will also start publishing our information as fast as we can, and we might think of another webinar where we can actually present our first clinical data. But this is something we have to look at after sort of we figure out all of the current drawbacks and the crisis.
Bogdan: Great. Well, thank you very much, Philipp. And stay safe everyone. Have a good day.
Philipp: Thank you. Goodbye.
Bogdan: Bye.