Elekta AB (publ) (EKTAB) Earnings Call Transcript & Summary

Good morning, and good evening to you all wherever you are in the world. Welcome to our ElectiUnity webinar today. It is wonderful to have you join us. My name is Joanne, and I'm your host for this webinar entitled ElectiUnity with CMM at UMC Utrecht first patient experience. A little bit of housekeeping first. This webinar will run for 1 hour today. [Operator Instructions] Any questions we have not been out to answer during the webinar itself will be followed up and answered after the broadcast. So we are honored to have 3 fantastic guest speakers from UMC Utrecht. UMC Utrecht is the first ElectiUnity clinical site with true tracking and automatic gating, and will be sharing their clinical experience of treating the first patients with this new technology. Today, we will be joined by Dr. Martijn Intven, a consultant radiation oncologist at UMC Utrecht in the Netherlands; Sara Hackett, a clinical physicist [ Lyka Myers ] , radiation therapy technician specialized in MR and MR-Linac. Thank you all for joining us today. It's really great to have you here. You're going to be sharing the details of the first patient treatment with CMM, true tracking and automatic gating, including the preparation work and what the next steps are for Unity in your department. It is with great pride that we can now present true tracking and automatic gating, an important next step in the ElectiUnity Comprehensive Motion Management journey. Comprehensive Motion Management, also known as CMM will combine Unity's market-leading imaging capabilities with powerful algorithms and technology to quantify movement and correct for it automatically as it happens, diagnostic quality 3D imaging, combined with multi-planar cine-imaging is the foundation of CMM. It enables precise treatment and spares the tissue you want to preserve. Multiple Motion Management strategies will also give you the freedom to choose the approach that is best suited for your patient and indication to deliver true personalization and designed to give you more time delivering care to your patient, CMM is seamlessly integrated into the existing Unity Workflow. No additional setup time for the motion management and the workflow is intuitive and user-friendly. Now let's see how UMC Utrecht are using CMM true tracking and automatic gating. I'm looking forward to hearing about their first patient treatment. Over to you, Sara, Martijn and Lyka.

So thank you for the opportunity to talk about and give a presentation on the first clinical experience on the automatic gating system on the Unity. In this presentation, we would like to cover the topics of what motion we want to cover with the motion management system. We want to give some physics background, talk about the workflow and share our first clinical experience. And in this case, I'll start first with talking about what motion we want to compensate for with the system. So actually, in radiotherapy, we have to deal with 2 types of motion, inter-fraction motion and intrafraction motion. And actually, the inter-friction motion is the motion which appears between our -- between the preparation phase, the scans you make for treatment preparation and the treatment phase, so each treatment fraction. During each treatment fraction, the anatomy is, in most cases, a little bit different than during the preparation phase and from fraction to fraction also the anatomy is a little bit different in the intrafraction motion and also things happened during the fraction. So the anatomy also changes from the time you do position verification scan to your dose delivery. That's the intrafaction motion. First, the interfraction motion. Yes, this is an example to show you what the differences are in a patient with rectal cancer and the upper row images, sagittal MR images of patients with rectal tumor from every fraction, which was a daily treatment in 5 fractions dose. So these are sagittal MRIs from day to day. You see really that tumor is moving quite a lot from day to day, and it's caused by difference, for example, in bowel gas, or bowel or stool, which is different from day to day. In the lower row, you see the intrafraction motion and that's the motion which takes place, in this case, during an MR-guided treatment of rectal cancer, and that's a little bit less than the interfraction motion. But what do we do with intrafraction motion? Intrafraction motion is corrected using our MR-guided adaptive workflow. Using an adaptive workflow, you make a daily plan, each [ treatment ] fraction, and that's the way you correct for intrafraction motion. What is not corrected with the renewed treatment plan is motion that occurs during dose delivery. So the intrafraction motion. What types of interfraction motion are there, they are generally 3 types, drift, breathing motion, and I think I call sudden movement. Drift. First is -- some kind of slow movement. It's caused by a relaxation of a patient while lying on the table, for example [ bladder filling ] during the treatment. And that causes a slow movement of your target and most known is the movement of the prostate during the treatment. You see images of prostate moving during the MR-guided treatment. And actually, those are 3 images in some kind of movie loop. And in between, it's around 20 minutes. So you see the prostate moving down and moving to the backward. And that's a slow process; here it looks quite fast, but it is a slow process, which occurs during the treatment fraction. It's Drift. Next, breathing motion. That's thing I don't have to explain a lot. I think it's more in the upper abdomen in thoraces, more craniocaudal motion caused by breathing of the patient again if you see to cine MRs of the patient, you see clearly the liver moving up and down, caused by breathing of the patient. And there is, types of movements, which are called sudden movement, but it's -- that they are different movements causing that sudden movement, like patient repositioning and the patient is on the table and is moving a little bit like for an achy back. He's moving a little bit to the side. You see that in the image at the right is made during the 3D cine-MR made during treatment delivery, the patient is moving a little bit to the side, this is some kind of sudden movement. You also have peristalsis or passing of gas. It's also not really predictable movement like you see in a patient with a rectal tumor with gas bubble passing by, you see if this was a patient with prostate cancer and you want to treat the prostate, then prostate moves forward. This is actually a female patient. But -- you can imagine that passing of gas in the rectum, is [ at the interface ] of the position of tumors and organs in the pelvis. And those are motions you want to compensate for. So Intrafraction movements -- interfraction movements are compensated by doing your adaptive treatments -- and for intrafraction movements, we have a motion management system, and we were the first ones to test it on the Unity MR Linac and first Sara will tell you about the physics background of that motion management system.

So the new Elekta software offers 4 different strategies for motion management. And the first one is predominantly useful for kind of stationary objects. So things like the prostate, the bladder, perhaps head and neck; things that aren't going to be influenced by respiratory motion, but still are influenced by the kind of drifts from relaxation or the sudden changes that Martijn just described. The next 3 strategies are aimed at targets that are influenced by respiratory motion. And now we have the choice of where exactly in the respiratory phase you want to gate. So the [ Full XL ] strategy gates the delivery when the tumor is in the [ end XL ] position. The free-breathing gates the delivery when the target is kind of in a mid ventilation, mid position. And the last option is actually almost kind of stationary delivery. You assume that the target is going to be stationary when the patient is able to hold their breath for a reasonable amount of time. You can also see that each strategy is associated with particular recommended 3D MRs. I'm not going to go into detail about the MRs, but it is important to think about that the MR that you choose represents the tumor in a position that corresponds to your motion management strategy. This software also offers 3 new capabilities. And the first one, anatomic position monitoring is probably already quite familiar to those who are currently using motion monitoring strategy on the Unity system. This means that you can visualize the tumor, including the contour on the tumor in both the sagittal and the coronal projections. You have the choice of using a surrogate volume instead of a tumor to follow. But in this case, it's really important to choose something that is going to move rigidly with our target volume because that surrogate volume, the registration volume is going to be used to determine the movement of the target volume. The next strategy is -- or the next capability is the anatomic tolerance check. And this is actually used to gate being on and off. So here, you can determine what your tolerances on the tumor movement are. In this example, you can see a projection of the PTV, it's the kind of light orange volume. And here, the radiation will be turned off when the GTV moves outside the PTV. The first capability, the baseline shift [ plan ] is intended to correct for these kind of systematic drifts or systematic changes in the anatomy that Martijn described. So for example, you see a relaxation of patient or their breathing changes systematically, so your tumor is in kind of systematically different position to where it started. And this new feature enables you to make a very rapid shift of the position with the MOCs of the remaining segments from planned to the new target position. It means that you're able to change the position of all of your remaining segments and continue with your radiotherapy in under a minute. So I'm going to show some films of these different capabilities that we made using the 4D Modus Phantom. This is an MR-compatible phantom and it contains a cylinder with an object that shows a different contrast to the surrounding material. And you're able to program the movement of that cylinder using this separate MR-compatible drive with preset frequencies and amplitudes. So here you see the movement of our GTV, the green volume on both the sagittal and the coronal images. These images are acquired with the frequency of 6 hertz. And now you see the movement in both the left-right craniocaudal and anterior-posterior axes. But you'll also see on the uppermost line, when the target starts to move outside of the GTV. So that determines when the radiation is turned on and off. In the upper most line, the gray parts mean that the radiation is on and the orange parts means that the radiation is off. In this case, we chose our XL strategy because the tumor generally spend most of the time in the XL position. And so the delivery should be most efficient. And here, you see that we've programmed a kind of systematic change in the position of the target volume. And in order to kind of account for that systematic change, we're making a baseline shift plan. So you can either use the average change in the position, to time it over the last few breathing cycles where you can fill in your own values. And this will send the new values of where your kind of average target position should be through to Monaco. Monaco will update all of the remaining segments, it will recalculate the new dose distribution as no reoptimization, so you don't need a new 3D MR. And you can see here, the plan -- the new plan is already prepared. It's set to MOSAIQ. It's automatically imported with proprietary MOSAIQ, and the new relation is sent through to the treatment session manager. There is a slight pause in the cine-images, and this is because the location of the sagittal and coronal cine images is always chosen so that it precisely passes through the center of the target volume. So these positions have to be updated every time you make baseline shifts. And the reason that both the coronal and sagittal images are chosen is because you could also have some parts of plane movement. On the coronal images, you can determine the superior-inferior and the [ referenced ] information. On the sagittal images, you can determine superior-inferior and post information. And so by a combination with these, you get twice as much information about the sup-inf movement. And certainly for respiratory motion expected, that's going to the largest component of your movement. But you're also able to check if there is out-of-plane movement by looking at the projection of our target volume on these coronal and sagittal images. Of course, the latency of the process is crucial to ensuring that you're accurately able to deliver the treatment. So here, we looked at both the latency between the target volume moving within tolerance. So for example, moving inside PTV and the beam turning on and also the latency between the target volume moving outside of the tolerance and turning beam off. The system uses the prediction model. So it looks at the movement over the last 3 or 4 [ beam ] cycles. And it estimates where the tumor is going to move to next. So you can see that the average latency is very close to zero for both beam on and beam off and it's never more than about 200 milliseconds. We also made some [ field ] measurements, looking at tumor with 2.5 centimeters radius -- sorry, diameter, but with a very large motion. So 2 centimeters peak to peak. But we said that the gating tolerances to the PTV which is the tumor volume plus the margin of 3 millimeters. And without gating, you see a really smeared out dose distribution because you have such a large motion amplitude. But here, you see on the left-hand side, the planned dose distribution and on the right-hand side, the dose distribution as measured in Gafchromic film. And you see a really nice agreement between the calculated and the measured.

I would like to tell something about the workflow at the Unity and about simulation. I can still very well remember when we treated the very first patient at [indiscernible] few years ago. I was very excited that we could finally see what we were treating, just before treatment, during treatment and after treatment. But now with the implementation of comprehensive motion management, we are finally able to see within a few milliseconds how are target and the other organs are moving within that patient. So that is very exciting. And finally, we're also able to adapt for those movements. First, I want to talk about patient simulation. All patients in all departments who were selected for gating treatment, they underwent patient simulation. And the main purpose to get the simulation is to see if the patient is eligible for this treatment. First, we want to refine the 3D image contrast. There are different image contrasts available and also the radiation-oncologist decides which 3D scan he or she wants to use for delineation. So for all patients, we acquired a T2 3D free-breathing scan and a T2 3D navigator-triggered scan. For the T2 3D navigator-triggered, a navigator is used in the sequence, which is placed on the liver-lung interface and so the system knows the breathing pattern of the patient and the scan can be acquired in XL. The scan type of such a navigator triggered scan can be for a normal breathing pattern can be around 4 minutes. But when a patient has a very shallow breathing pattern, it can take up to 8 to 10 minutes. So that is what you have to take into account. There are also other contrasts available in the new CMM workflow, for example, T1 3D flare fat-suppressed images or balanced and most of them are accelerated by compressed sense. Also, we want to use the simulation to define if we want to treat the patient with or without abdominal compression by using a corset. Also, the simulation was performed to see if the tumor was visible at 2D cine images. We acquired a coronal and sagittal 2D cine image with a time resolution of 6 frames per second. And we wanted to see if the tumor is visible. Also, the tracking quality was defined. So we want to see if the software was able to track the tumor. You can see this in this little movie. This was one patient we simulated at the Unity. You see that the software is tracking the structure very well. To perform a patient simulation, you need to create a dummy patient with [indiscernible]. And within this dummy patient, you can perform the entire workflow until that treatment delivery. And another advantage of simulation is that the MRI can be used as a reference for the first fraction so that both 3D scans have the same contrast and you have a more effective contour propagation for the fraction. This is an overview of the standard unity workflow. So after patient is set up, a 3D MRI is acquired followed by contour propagation to the new daily MR scan. After approval of those contours, treatment plan is generated and parallel to that verification 3D images acquired to see if the patient is still within the same position. Then the radiotherapy is delivered, and we also have the availability to acquire 3D intrafraction scans and post scan to see if there's any interfraction motion. This is an overview of the new CMM Unity Workflow, the green and the red box, they illustrate the new steps of the workflow. So during treatment planning, a template is acquired, which means 3D registration to 2D images and later on during radiotherapy delivery, live cine images will be acquired for tumor tracking, and they are used for automatic beam gating and baseline shift planning. And also, we still have the availability to acquire a 3D post scan. This movie shows that the slice of the 2D cine images are placed automatically within the center of the region of interest, which is usually, like Sara said, the target or it can be a surrogate structure. Okay. This is an overview of how the algorithm works. First, there are 2 planes extracted from the daily 3D MR, sagittal and coronal plane. Later during treatment, then generation training images are acquired. There are 60 training images, 2D cine images, which are averaged and registered to the extracted images from the 3D image. That template has to be approved. And later on, the live cine images are registered to the template and the average and actual displacements are shown in the CMM software. So still, we will try to improve the workflow all the time in calibration with Elekta. This is one example from prostate patients. We still want to improve the image contrast of the 3D scans but also the 2D scans. And this is an example for prostate patients. On the left, you see a balanced TFE coronal sagittal image, 6 frames per second. So with a very high temporal resolution which you can wonder, is this really necessary for a prostate patients or is a lower temporal resolution also fine and I think because the prostate doesn't move because of breathing motion, you will also be fine with a lower temporal resolution, and that's why we have invented now new cine image with a T2-weighted image contrast, which has a temporal resolution of 1 frame a second. I want to end with the RTT challenges and key learnings challenges are most of our workflow is now at the Unity are driven by RTTs only, but now in the beginning with the implementation of the comprehensive motion management, I think the physicist and the radiation oncologist will be present again, at least in the beginning, also to implement some limits or guidelines during this treatment. Also another challenge is the focus during beam on, usually when beam on time, some kind of waiting time and we just keep an eye on the patient. But now you really have to focus on the CMM software and check if the tumor is tracked correctly. And if yes, what the efficiency is of the treatment so that the patient doesn't stay too long on the treatment table. Also, you have to make decisions very quickly because if you want to perform a baseline shift then, you have to, yes, handle very quickly. And what we discovered was that training takes a lot of time. There are very -- several different workflows. And yes, you want to get to know them all. And yes, that just takes a lot of time. So what we learned from RTTs' perspective, I think it's very important to involve your RTTs in the training because in the end, they have to perform the whole workflow and also never underestimate the importance of phantom and volunteer testing, including -- and testing with the entire treatment team I think is very important.

Okay. Thank you. Now I'll continue with our real first clinical experience. So with the patients we've treated on using the motion management system. We actually now treated 3 patients with upper abdominal tumors; the images of those tumors you see in the screen, actually 3 male patients with either a renal cell tumor metastasis in the upper abdomen or neuroendocrine tumor of the pancreatic head. We [indiscernible] deliver those tumors at this moment for the grade 5 fractions. So we did this also in these treatments. And we used a workflow we regularly use in to treat these patients with abdominal compression. We used lumbar support corset for that and with the trimming of the PTV margin. And we did that in a free breathing treatment. So the treatment starts with the same procedure as a normal treatment on Unity. There's a contour propagation and you adapt your contours to the current anatomy. And as [ Lyka ] told some slides ago, at that moment, there's also cine MR made to simulate up to give the system the opportunity to track the tumor. And there's a template generated and you have to check the template if the structure is defined well by the system, in all patients in all fractions very well. So if that's okay, your plan is okay, then you're going to start with the treatment. The motion management system looks like this. We've already seen it in some slides of Sara and also from [ Lyka ]. This is from the patient with the tumor in the pancreatic head. In the purple structure is your GTV and the green structure is a structure called the gating envelope, and that's a structure you want to keep the GTV into. You see 4 lines in the low part of the screen, the 3 lowest lines are the movement of the structure in the 3 directions. So there's the left-right direction, the superior-inferior direction and anterior-posterior direction, you see the actual displacement of the tumor in the -- in the left column next to the lines. And you see in the right column there, the average displacements of the tumor compared to the original location. And that's actually really interesting because that if that numbers are going up or down, that's actually the drift of the tumor during the treatment fraction. The line above is a line and percentage which is there is the line which shows the overlap between the gating envelope and the GTV. And you can set the percentage you want to accept for overlap to be when the beam is on. So in this case, we selected 95%. So if the overlap is higher than 95%, the beam is on; if the overlap is less than 95%, the beam is off. And I'll show you, yes. So you see the image of the -- movie of the patient treatment and you see -- so the -- if you see the upper lines, you will see the percentage changing with inspiration when the GTV is outside -- a little bit outside of the gating envelope and the percentage goes down and overlap percentage tend to be missed off. So at the orange bars of the upper line, the beam is off. And on the other sides, the beam is on. So that works really nice. And actually also in all patients we treated and simulated right now, the tumors were quite well visible on the cine MR. So next, the system has its own way of tracking the tumor. You can nicely also visibly track the tumor and check if the tracking is going okay. Then there's, I think, a really nice feature in this system, and that's the drift correction. Sara already showed it for the phantom. So if you see that right column of numbers is increasing. For example, if you see in this case, a drift of almost open 5 centimeter or few millimeters to the superior side and you can correct for that, you can do this correction. And we actually started now with this correction. It's the same procedure as Sara showed in the movie and in the slides. It's recalculated, the plan is based on MOC movement. So the MOC is moved to compensate for that drift that new plan is calculated really quick, and the new plan is sent to MOSAIQ and it's approved actually within 1.5 minutes, you are up and running again. So it is a really quick correction of drifts you see during your treatment. And this is actually also very nice for the tumors which aren't affected with breathing movement for example, like that drift shown in prostate cancer, you can really nicely use this system for those drift corrections. So even when there's no breathing motion. So -- and actually, now we are up and running again. So it's a really quick correction of that drift. What are the actual duration of the treatments, these [indiscernible] fractions. Actually, the last 2 fractions of the last patients -- of the last -- of the 2 patients we were treating weren't done when this webinar is going on. So this is data from 20 patients. We had a median on table time of 42 minutes. Yes, almost 40% of the time we spent on plan adaptation, contour propagation, that is manual adaptation of the contours and thus the recalculation of the new plan. Just 5% of the time you do those delivery and rest is like data, time spent on like data transfer or acquisition of MR. If you compare that to the normal treatment times of patients receiving a fully online elective procedure for upper abdominal tumors. And it's quite comparable. And if you really compare those diagrams, and you see that [indiscernible] is little bit faster but that's probably also because we have more -- a little bit more experience than the data sets of the nongated delivery. Those delivery takes a little bit longer, and the time spent on other procedures is a little bit shorter, that is most of the times due to that you don't need, for example, a position verification scanning in your scan. So the delivery time is a little bit equal, the things you do during the delivery time or during the treatment fraction time is a little bit different then in the nongated [ nerve block ]. So to illustrate that this is the normal workflow we use and when you have gating, then the position verification step in our center is replaced by the motion monitoring during treatment [indiscernible] that makes that the time is in step longer even if motion monitoring or gating makes treatments a little bit less effective -- more accurate but less effective because that was only on usage when the tumor is in the exact position you want to access. So what are the potential advantages, the treatment is more accurate. That means that you can reduce PTV margins potentially. This means less healthy tissues in the field and a reduced toxicity, probably more hypofractionated treatments. And for example, if you can do really a drift correction for prostate cancer during your treatment, you really can have small PTV margins that's already also shown by Thomas [indiscernible]from our center, it is published this year. If you really can correct for that drift, you can go to really small PTV margin [ percent ] for prostate cancer. And if you're able to do that, then you also really can hide fractionated and get really accurate treatment for those patients. And yes, what to do with the abdominal compression at this moment, we used to use abdominal compression, and we're going to see in the next patient we treat if probably also we can skip the abdominal compression and treat patients in free-breathing with motion management system without abdominal compression. And the future [indiscernible], I think, really, this is the beginning, the beginning of real online tracking of your tumors. Do you have, as Sara explained, there is a prediction in how tumors move in this system. You can use that, you have MOC movement with the drift correction, you can combine all those things in really MOC tracking. And we did a proof of principle of [indiscernible] proof-of-principle experiment for MOC tracking in an indication of cardiac ablation on your Unity and this is really cardiac motion, which is compensated for and also respiratory motion, and it was really feasible to deliver such treatments on Unity and hopefully, it because becomes available in some time on Unity to treat patients with really tracking of tumors. So the comprehensive motion management strategies in Unity, I think it's really positive that you can see what you treat, while you treat and really can compensate for what you see and that goes quite automatically. It's feasible, treatment times weren't that longer with gated delivery [indiscernible] treatments. It was well tolerated by patients. We do simulation before to see a patient tolerated treatment well and we haven't had a patient which wasn't able to be treated with the motion management system. In my opinion, it's the first step to watch real track dosing effort. So I thank you also on behalf of [ Lyka ] and Sara for your attention. Thank you.

Thank you to all our speakers. We have received a number of questions from customers. So let's dive straight in with the first one.

So now Unity has CMM. What indications will you look to treat on Unity that you wouldn't have been able to treat or would have found it hard to treat with radiation therapy before?

Yes, I will answer that question. Yes, honestly, I think already the introduction of MR-guided radiotherapy opens a lot of opportunities for different indications, for example, in the upper abdominal to deliver quite high -- safely quite high doses. I personally don't think that the motion management system will add up a little bit more indications but with the indications we have right now, we can really more accurate treat those tumors. And by more accurately treating consumer, probably more going to hypofractionate, really reduce PTV margins, so reduce the side effects of our treatments. So actually, I don't think that there will be too many extra indications only with introduction of the motion management system. I think it will make the indications we already have by the introduction of the Unity will -- those indications can be treated better with the motion management system.

Do you see treating patient with true tracking an automatic gating as an opportunity to escalate dose and/or reduce toxicity?

Yes. That's actually, I think, the same answer I gave before with a more accurate treatment of the patient over the tumor using the motion management system. Potentially, we can reduce PTV margins, we can reduce toxicity by reducing the volume of healthy tissue in the field. And we can also -- if we have really small -- small treatment fields and accurate dose delivery then we could increase our treatment -- our total radiation doses. And so I think the motion management system allows us to dose escalate and to hypofractionate, certainly for example, in prostate cancer, as I showed in presentation but also in abdominal tumors.

Are there any indications where this new technology could potentially lead to fewer fractions?

I'd like to tackle this. So one of the indications that we're hoping to start treating with comprehensive motion management is peripheral lung tumors. And at the moment, [indiscernible] are currently treatment either 3 or 4 fractions depending on the proximity to the ribcage. With comprehensive motion management, we would dare to treat them with a single fraction. At the moment, that's not possible because you still see these kind of systematic changes. The patient relaxes, patient starts breathing different. With comprehensive motion management system, we could do that in a single fraction. And that makes a big difference, not just in terms of the toxicity of the treatment but for the patients themselves. One of the patients that Martijn showed that was treated with gating came from a city that is 2.5 hours' drive from us. So you can imagine for someone, there's a huge difference in driving here and back 3x or 4x versus just coming once.

Can you tell us the learning curve to get the team ready for the first gated treatment?

I can answer that question. I think when implementing such a new strategy as comprehensive motion management, yes, training is key. It's very important to train with the entire treatment team, also to test on volunteers with phantoms and also perform end-to-end testing. And besides, I think it's also very important to stimulate your patients you are going to treat. Since yes, you really want to see if the tumor can be tracked correctly. And so it can be, for example, more difficult with areas around lot of air or inflow from vessels. So I think that's also a very important thing we've learned so far.

Have there been any surprises with CMM, that is, any benefits that you didn't see coming?

Yes, I can answer the last question as well. Yes, we were very surprised how fast baseline shift correction works. It's very quick and effective. I think it's also very easy for RTTs to use, especially if there has been set some limits, for example, if there's a stable shift of 2 millimeters, you always form a baseline shift correction. So we were very surprised about that.

I think for me, certainly, it was how many baseline shift corrections were necessary -- as we mentioned, we've kind of set these limits of 2 millimeters. And based on what we previously estimated, we thought maybe 1 perhaps at most 2. But now that you're actually able to see what happens during the treatment delivery, you start to realize just how many of these little changes occur. So yes, it's both how quick and easy it is to do, but also how much benefit that delivers.

Yes. In my opinion. I too got one of my surprises first that these drift corrections are really fast and really accurate. So that is just engaging, sounds a little bit like this is only good for breathing motion, but even in tumors where breathing motion isn't that important for intrafraction movement, that drift correction is really important for prostate cancer -- so yes, I was really surprised to see the drift in the upper abdomen tumors, we now treated quite -- occurring quite often -- that we could really largely compensate for that, but I'm really positive on using the system also for locations where only -- the tumors are almost only affected by drifts. It's really nice to see that you can correct so fast for that.

We are really enthusiastic about the software. I think we've all seen what it could mean for our patients. And so we're really looking forward to that it can critical rollout.

Well, that's all we have time for today. Thanks to our speakers, Dr. Intven, Sara Hackett and [ Lyka Myers]. Thank you for watching. And I hope we've been able to show you how Elekta and Unity can enable you to see the difference. Thanks again, and have a really great day.