RaySearch Laboratories AB (publ) (RAYB) Earnings Call Transcript & Summary

Hello, everyone, and welcome to a webinar global rebooting highlighting the additions to our newest RayStation release this is Version 11B. My name is Dayna Bodensteiner, and I'm the director of Director of Product Management for RaySearch Americas. And I have the privilege of introducing our key speakers today. First, [indiscernible] who is an application physics specialist located in Sweden on our RaySearch Laboratory support team. He is a resident expert on many applications as he's been on our team for almost 7 years. Before that, he has extensive clinical experience doing brachytherapy for 6 years and also worked for 4 years at the Sweden Radiation Safety Authority. Next, we have Craig McKenzie, who's our senior clinical specialist from the RaySearch Americas team. Many of you have probably recognize Craig as a past AAMD President. He was in the leadership at the University of Florida Protons and then came to work for us. He went back to the clinic as the Director of Dosimetry at Miami Cancer Institute for 3 years, helping them get established with RayStation and then rejoined our team again in 2021. He has left the paradise of Miami and now resides in Texas where he is joining us today. So before I hand it over to Emil and Craig, I'll give a high-level glimpse of our Version 11B and some of the options and features that you can find. So excitingly, we have LET evaluation for ion. We have EQD2 reporting for brachy and photon dose. We have the ability to do a cone beam CT conversion, and we've implemented your enhancement requests, a few of which are improving the registration workflow and persistent visualization settings. All exciting features and these guys are going to dive deeper into them -- into all of them throughout the presentation. One last reminder is that our software is a subject to regulatory clearance in some markets and specifically in the U.S. The 510(k) for 11B is currently pending and under the review of the FDA. With that, if you have any questions throughout, feel free to use the Q&A panel. I'll be monitoring it. And if I see questions come in, I might interrupt these guys or I might wait until the end of the presentation. But with that, I'll go ahead and turn this over to Emil.

Thank you, Dayna. Let's get started with the general improvements. And all of you very welcome to this webinar, of course. As Dayna said, I hope we will find it interesting and insightful. So let's start. The latest release is RayStation 11B. And what we have done is for the physics part, now the [indiscernible] commissioning process that you can compute all the dose curves at once before you needed to do it energy by energy, which could be quite cumbersome and also time inefficient since the Monte Carlo curves take quite a bit of time to calculate one by one and then commission the machine now you can just do it with one click. So this will greatly improve the commissioning process, especially if you're using our Monte Carlo dose engine. And also one improvement, before when you just calculated 1 curve, for example, 1 profile -- 1 profile or 1 depth dose for the Monte Carlo, the system automatically computed all the fields for that profile. So if you had a 10x10 profile, the system would calculate all the best profiles, but only show the one you have selected. But now we are showing them all at the same time because when we are calculating just 1 depth in the Monte Carlo dose engine, it will automatically calculate for all depths. So that is information for free. So if you're using 11B, you should not be confused that you see more curves than you perhaps expect when running the Monte Carlo because we are choosing to give you all the information that actually is in the system. We also made it able to have to be used a couple of materials for different patients -- many patients. And the full element of composition of the different materials can now be edited and use to different patients. And the left of this slide is that we have improved our [indiscernible] template. So if you buy RayStation for the first time, you have an amount of 10 big machines in your database, where we have machine specifications according to how we think they should be. And over the years, these have been changed, and we have now improved our templates quite a lot. And we also have the support for different affluence models with on energy. We can give example for the 6 MB we can have both a flatting filter, you can have the flattening filter free, and you can also have a fresh new quality in the same machine and energy. So now no need to use different machines for the same physical machine. Another thing that we have improved is the ROI visualization settings. So now it is a 3-click toggle, you can say. So if you click the header of 1 ROI type, the first click, you will remove them all. So no other ROIs will be visible. If you click again, you will make them all visible. And if you click it third time, you will go back to the state where you began. So say, for example, you choose to show the P2P together with, let's say, the bladder and over those structures, if you click once, then they will all be unhighlighted you will see nothing of the ROI, if you click again you will hold the ROIs and if you click a third time, you're going to go back for you for your [indiscernible] bladders if you pick up. And also the visualization settings to deeper systems now before when you call for patient and open it, you sometimes had to do this all over it. Say, for example, let's go back to the bladder and B2B example. So if you select those and you close the patient and then open again, those settings would not persist between the different opening of the application, but now this will persist between after you save the patient and open it again, it will persist. You can see it in all types. Also for the clinical goal, if you have a template, for example for clinical data goals and these will now be mapped automatically in the clinical goals. And if you, for example, import a patient, say, for example, that you have at a different nomenclature for the naming of the ROIs doing these patients, you can now associate that ROI OPI to the clinical goals without creating a new clinical goals for the name of the structure you have in your plan. And one small -- yes, that's important. Before you could remove approved plan without signing it, you didn't -- you got the wording message, of course, but now you need also to sign it with your passwords. So you cannot accidentally remove a plan as easy as you could before. And one last thing is that if you go to the main menu, it's a ball to the top left in RayStation 6. You can go to the shortcut dialogue, and you can now search for the different shortcuts because we implement more and more shortcuts all the time. So that list was static before, but now is searchable for easy access. So you can quickly find in the short that you need to use. Nowadays, it's also possible to use the fee [indiscernible] for ROIs of the site bolus port situation in the dose computation. So there's no need to do a [indiscernible] material anymore. You can use the CT data instead, and the material is also visualized in that way. And now also, if you go to the view, you can also choose to toggle between the image data or the material data. You have a drop down; you can see it in the bottom picture in this slide here. Then we did some improvement in the beam specification support and fixation ROIs, as I mentioned, in last slide. And this is prior to 11B and now we have better visualization of the material and the CT raw data, so to speak, enhance big unit. And other general improvements in 11B is that you can now compute the total dose with the rotation -- the patient rotation. And another thing that we implemented was that the nominal contribution will now be rounded to the closest centigrade or 0.01 grade, if you choose to use grades as the units because sometimes before you can get some strange decimals that could lead into problems when you export the plan. But now it's rounded to the closest full centigrade. And you can also export with either the both beam dose of clinical point or the number of dose distribution. And this can be selected in the export windows that you see here in the image on the bottom right. And by that, I leave the arena to Craig, who will speak a little bit of the patient modeling improvements that we're implemented in RayStation 11B. So take it away, Craig.

Thank you, Emil. So today, I'd like to talk about in-patient model and some of the improvements and in our approved image registration workflow. We now support multiple rigid image registrations using the same data set. And typically, this is used for the contouring. At the same time, we also can rename them and add comments. As you can see, demonstrated here, we can rename and add a description of what the registration was done on, and we now can approve the registrations as well. A new feature also is to be able to create POI registrations with less than 4 POIs, even down to 1, which I'll try to demonstrate soon. There is a pan and zoom, which stays the same even after you change directions between actual current own and sagittal. And then we have a new triangulation algorithm that's been updated and is now faster. Okay. So with the new improved registration workflow, this is what it would look like on the tool bar. You'll have this new button here, which creates a new frame of reference. And with that frame of reference, that actually allows us to create more image registrations or multiple using the same data sets. There will be only 1 frame of reference per pair of image sets. We'll use this frame of reference to compute dose on additional data sets as well as creating the formable registrations. Then when we click on new registrations, and you can see here there's a frame of reference and here's a head and a lung that was created as used for contouring. So it's going off of the same frame of reference, but yet we can actually go back and view each of these different registrations, even in contouring mode. Okay. And just kind of a little bit of a reemphasis here is the 2 buttons are here. And as we create a frame of reference, you can see it here. In this particular case, we did a for a mouth registration. You see is kind of centered on the mandible. However, the see spinus off a little bit. Therefore, we created a second image registration using the same data sets and now we've recentered it so we can use this for contouring. Okay. And just a short video here to kind of demonstrate some of the features that are in the [indiscernible]. So the 2 image sets are there -- excuse me, the 2 buttons are there, but we are centered on the floor of the mouth, you can see the mandible. Again, the is spinus off just a bit. Therefore, we'll create a new image registration. With this, we can rename it as well as we can write a description of what the registration or how the registration was done. Okay. And now using the new triangulation fast algorithm, we're able to quickly register these 2 recentered on the C-spine now. And you can see how it recentered the new registration. Now we have 2. If we go back up to the tool bar, we can actually flip between the views of the frame of reference and any other registrations we have. We also now can approve them. So as you can see, we and found them and do approval, which would lock the registration. And if we switch over to structure definition, now we can actually change that view within this module so that we can see the different registrations as we decide to contour. That includes with each of the views that we decide to look at with the layouts, whether it's side by side or overlapping. And like I mentioned before, now we have where you can do less than 4 POIs for registration. In this particular case, we have 1 point that's located on 2 different image sets. It's placed over a bone spur next to the vertebral bodies. And if we just use that 1 POI, we can do a registration. And you can see if we click on it, we have 1 POI with different locations for that. Okay. Now I'll pass it back to Emil. Thank you, Emil.

Thank you, Craig. Another thing that we have implemented in the RayStation 11B is that it's now possible to generate synthetic CTs from the CBCT. And we have 2 ways of doing this, the CBCT correction and the virtual CT. And this functionality is part of the license that's called [indiscernible] it's part of our dose tracking functionality. So the first one, the corrective CT is working in the iterative manner with 2 main parts: find a conversion from the CBCT intensity scale it to the planning CT Hounsfield unit intensity and then find a correction map that can remove low frequency artifacts. For example -- this is -- for example, if gas in one of the CT system up in another, we can take that. We also have a second method that is called the Virtual CT, which is a hybrid between a default CT and the CBCT correction method I described above. And this can be used if the first one isn't giving the results that you wish for. So you can -- you have the other options to use both, one or the other, to get the perfect result. It depends a little bit on body side and other changes in the patient that has to be from treatment and planning CT, of course. And both these methods would also populate the CBCT if you have a limited seal of view of your original CBCT. So let's say, you have a smaller part of the patient that's to come in CT when treating or just before treating, this will fill in the remaining information taken from the planning CT. So you have the whole image set representative from that synthetic CT that you generate. So it will take the data from the planning CT to fill in the gaps that might be clinically relevant, of course, but it might be of interest to still have it there. So it's good need. We're also working more on our AI functionality or deep learning segmentation. So we hope to release new models. We have released some and there are some in the pipeline that is both new and also some are improved of our older models. We will see a list shortly about the different models. We're also learning the supporting deep learning segmentation templates that can be used to use specific needs. So you don't have to -- you can save some clicks there. And also the segmentation of lymph nodes levels of the patient. So I will put this slide here for your reference for a while. I will not read through it. But as you can see, the models in white are already released models. And the models in green, a models that we plan to release this spring in 2022. So this list is really starting to get longer and longer, which is very nice. And of course, our aim is to be able to provide models for all relevant treatment practice that is common and uncommon for that matter. But we want, of course, to have a comprehensive model library as possible. And also here because you see a model here, it is why that has been -- that is -- it means that it's released, but it also means that it's continually improved upon as soon as we get access to more data. And as you know, the model themselves does not contain any patient-specific information. So these models can be shared by us to you without jeopardizing any privacy loss that meet feature of our model. No patient data is in the model file. Yes. As I said, our goal is to have a research standard will available for infrastructure relevant to radiotherapy. Another feature that has been limited in RayStation 11B is that you can do a machine learning planning model on the beam set level. And also the licensing has been a little bit changed. So now there are only 2 specific treatment technique licenses. It's the RAY B planning photons and RAY B planning protons. Some of the photon planning improvements here. We have improved the CyberKnife planning excessively. We now have the ROI protection option. And if you're a user of RayStation, you might have used the protection when using for example the Tomo functionality or VMAT functionality. It basically means that it would block -- it will -- the end that it will block the structure, the ROI that you set to protect, so you won't get any openings when going around that ROI. And we also entered some beam entry filtering during automation. And the reason for that is to have a greater chance that the plan will be delivered -- deliverable because that it today, we could sometimes get a plan that looks pretty very nice, but then it violated some of the constraints of the machine itself, which is, of course, not good. And it would show at a later stage, whether you calculate the final dose, but now it's also taking into account optimization. So you can have these violations reported during the optimization in an effort to save your time on planning. We introduced the support for the jaw setback for various machines so that the jaws will follow the leads. Another thing, which might not be very obvious in the [indiscernible] is that we try to -- or we did do a lot of member improvements in the segmented doses because if you had a large patient with a lot of data in it, sometimes the memory could run out and leading to a crash. But by tweaking and optimizing the algorithms, we are now much more easy on the memory of the computer. So these crashes will be very seldom to occur. It's also possible now to reverse or make a copy of an arc and reverse it if the copy become the reversal of the original arc. And lastly on this slide, we did some improvements for the segment based VMAT in the multicriteria optimization. For example under additional motion of the MCD is not under strictly enforced, which leads to more details of reading and planning, also improves the operation plan quality and maybe most important all it's less likely to violate the constraints we entered. Another thing that we added in 11B is that we now can -- as you can see in the bottom left picture is the new icon in the optimization dialogue where you can press fine-tune. So let's say, you have a plan that looks quite good, but you're not really happy. And instead of trying to reiterate or reevaluate objective constraints for that plan, you can now pinpoint the ROIs that you're interested in by using this tool bar clinical goal fine-tuning. So what you do is that you press this fine-tune icon, as you can presented with the dialogue to the right. And you can select the specific clinical goals that we really want to fine-tune. And what we are going to do is as we try to fulfill these clinical goals and prefer preserving the VH as well as the overall station dose resolution, but still try to work locally on the specific fine-tuning ROIs that you have selected in these dialogues. So it is also a time saver we hope. So instead of doing a reiteration of your objectives and constraints in the optimization process, you can pinpoint the exact ROIs that you want to treat, and you will save some time. And by that, I give the word over to Craig.

Thank you, Emil. Just to an overview of some of the proton planning improvements. We now can compute dose average LET as part of the final computation using our superfast Monte Carlo dose engine. The water equivalent depth is computed and displayed in the beam specific -- beam dose specific points alongside their physical depth. We also have range modulation name displayed for the ocular gaze plans. We support single scatter delivery systems, and we support nonuniform fluence for all proton broad beam techniques that's in patient modeling, and we support set of beams for the boron neutron capture therapies, which would include DICOM export. We also -- just to point something out, we also can also do background dose. We're doing our 40 optimization. So we're using the background dose as one of the functions. But we do limit this functionality of robust to be just with beam set dose. And with LET, I'm proud and excited to say that we're the first commercial TPS, which supports clinical evaluation of LET. Computation of dose average LET for protons and light ion plans. This is fully integrated into the plan evaluation workspace where you can see it in 2D, you can see the LET volume histograms, you can see line dose, those statistics. And with that, we'll demonstrate in a video showing how it looks like in RayStation. So we're kind of -- we're in the plan optimization now. I just wanted to demonstrate under the beam -- those specification points that we do have water equivalent depths that are displayed as well as physical depth. And of course, we change to a different DSP, we'll be able to have those updated as well. And in order to get to LET, it's as simple as checking the box, it says compute LET and hitting final dose. It's going to be part of that fast Monte Carlo algorithm for the dose engine. I'll switch over now to the planned evaluation, and I will compute it. And over on the left side column, where we have our evaluation doses, you can see it computes it that fast. Now in our display, we actually can put comparison between the actual dose and the LET. Within the LET, we actually have a color table, which we can adjust. We can actually set a threshold -- minimal threshold for that dose in order to filter out some of the LETs that are displayed. We add line dose. In this case, going straight down to the access of one of our beams. Okay. We do display our LVH? You can see on the right side at the bottom, and we do that next to our DVH for the dose on the left. And if we switch over to line dose each one of our scale, we have dose on the left column, and we have LAT on the right column. We can make those fit for max. And we can go down the graph and look at each one of them and see where the 2, the dose as well as LET overlap. Now I'll go back, and I can adjust my line dose a little to kind of see what location within the anatomy of the CT that where that ends or were they 2 intersect. Okay. So now I will just show that we do this for each beam sets with the LET. And we also can compare LETs at different plans if we choose to do that. And the difference is on the bottom right. And we also will show those statistics. So here I'll turn on the difference between the two. And we also can change that color table as well to be able to show the differences between the two. We just open it up. We'll change our values. Now we'll be able to show the difference between the two. All right. So it's exciting with LET, a lot to do. I know a lot of people have been waiting on this. So it is in the 11B now. And I know Emil had mentioned this earlier, but it's also a very important feature, particularly in protons as one of the reemphasized that we now do patient rotation within the compute perturbed dose function. So now we have the ability to add that as part of the evaluation. And again, we do include the statistics for LET as well as RP -- excuse me. Okay. I'll pass it back to Emil. Thank you, Emil.

Thank you, Craig. So as you know, we have released [indiscernible] factuality a few releases back. And a lot of things has happened. It's now possible, for example, to look specific well positioned in the plan. To say, for example, if you look at the picture to the right, let's say, that we in the ring and the intrauterine applicator, we want to fixed dual positions. Now we can enter them and lock them and then use the interstitial needle that is coming with this particular applicator. We can only use our optimization function on them, and they will take the dose given from the locked dual position in the ring and intrauterine into account doing the optimization and adhering to the eco straight there. Another thing that we have put in the 11B is that you now have direct access to the fusion functionality that you might have seen in patient modeling before. So you don't have to switch between the different workspaces if you want to do something that involves fusion between different image sets. So it's a workflow improvement. It's the same functionality as you might be used to if you use the fusion in the patient for before, the exact same [indiscernible] on, but it's more easily to use. Now the minor tweak is that you can now switch the visualization on and off for both channels -- channel candidates as well as the applicator template. And if you played around with 11A, you might know that we introduced functionality that's called smart draw. And the smart draw is an algorithm that tries to find the channels in the patient. So before 11A, we only could automatically detect solid markers. But with this smart draw improvement, it's now also possible to have a dotted markers inside and it will -- the system or the algorithm will find the correct source path even if you have a dotted marker. And I think, maybe the last -- the least, it's now possible to define clinical goals in the [ DTG2 ] in the brachy module based on the [indiscernible] model. And as we mentioned before, it's also possible to calculate [ DTG2 ] dose for photon plans as well. I wanted to carry out here is that the clinical goals [ DTG2 ] only works for brachy at this time. But if you have the planned evaluation license, you will be able to calculate [ DTG2 ] those for all photon plans, including that of course. So here is a dialogue. You cannot, of course before the dose, I will show this in the video very shortly, computer form and accumulate the [ DTG2 ] for different treatment times of fractions. And you can also deploy in some of them. So you get the big picture of what we have really given the patients how much dose. And another minor improvement, which is very good, I believe, is it's possible to rename the different doses now. So you can have a -- you have a free text that you can actually name the dose. So it will lead to less confusion. And this is not only for [ DTG2 ], it's for all version and some doses. So let's have a look how this works. So first, I need to compute [ DTG ] dose. And the priority here is if all right overlapping the one with the highest priority will take priority. And then we entered the beat the value for all this. So B2B gets then as with the B2B and overall it gets a 3. And we choose which doses we want to calculate on, and we press compute. It's quite fast. And if we go to the arrow mix of revolution doses and press it, we will be able to see what we did. So now we have 2 B2B two doses, 1-1 is the first fraction and 1-2 is for the 3 and 4 fraction. And if you hover over the good affair, we can see the exact properties of the different settings that we use when calculating this dose. So now we want to sum the doses. And this call fraction 1-2, this is 2 fractions giving as fractions 3 and 4. So it's 2 implants, same patient, but 2 implants. The 1-2 is the last -- 2 last fractions with the saving plan and the B2B 1-1 is for this first implant and the first and second traction. So we have doubled this as we did with the last one. And this is, as I said, fraction 1 and 2. Now we have sum the dose for the 2 first fractions and the 3 or 4 fraction. Now we have a deformable registration in the background. So we deform this dose to the original image set. Now we get the new dose at this fraction 3 and 4 that is mapped to the first image set, the MR5. Now we sum those doses, and we get the total dose in ECD 2 for both tractions in -- on the same image set. What we've done with this, we can just to evaluate our DVHs. We can drag and drop the doses in the plan evaluation. We can look at the difference, and we can also evaluate the DVHs. The dash lines is ECD 2 the solid lines is the physical dose. Of course, we can also look at dose statistics and the clinical goals to make sure that they are fulfilled or if they are not how we perhaps could take this in consideration where we give fraction 5 and 6. So that is about what I had to say today. Thank you very much for listening. And now I give the stage to Dayna again.

Okay. Thank you, Emil. Great stuff both of you. And we have had some great questions coming in. I have been typing answers to a few of you, but -- there are some that will go ahead and ask live, if I'm going to put you guys and ask you. The first one is coming in with regards to the material override. I know that there's some excitement and interest in how we've changed showing that in the software. And so the question is, can material -- these material override CTD exported? And also, can it be used for CRR generation? So either of you want to tackle that one. Emil?

I don't have a straight answer. So I will refrain from guessing.

Okay. Well, I'll jump in here and say that per se material overread in CT cannot be exported. But when we do ask for a CT, the material overrides of the ROI are included with that export of the CT, so the CT got material export material overrides are. That is included in that export. So it's not per se the CT uniquely with the material overrides. And as far as our generation, do we know if the material overrides are taken into account there? I believe that they are. That's how some of the MRI we have -- can use in MRI as a primary and we can use the material overrides as that in the DRR, but we can follow up on that to make sure. So I'm going to jump some topics here and go to the deformable image registration. So we have a customer asking that when we do a deformable image registration and we choose the CT as a reference and the cone beam CT as the target, they want to brought back and they want to do it the other way around choosing the cone beam CT as a reference and CT as the target. Can either one of you comment as to how that process works for them? I said now I'm just going to put you on the spot. Craig, either one of you can answer that, correct, right?

Okay. So we want to switch the two in the registration. That is something we should be able to create a new registration and swap those two, yes.

Yes. So it's as simple or is just checking, which is the reference and which is the target, and you can reverse it...

Correct.

In any order in that window, if possible, like Craig back dating new registration and then reverse those. And if you have questions on that, of course, as a customer, you can always call our support line. This is a question about the generation of the synthetic CTs and generating them from the cone beam CT. What are the requirements? Or can you maybe reiterate and then when we have a limited build of view between the two images? I know that's an important feature. Sometimes we don't have all the data there in front of us.

See the view wise when we are in the field of use to speak from the cone beam CT, we will do either of these two -- either the synthetic or the virtual. And once the deformable registration between those are made and we have the mapping structure, what is not in the field of view, we'll just be added from the planning CT still in the blanks, so to speak. So you have the whole CT set. Even based on the cone beam CT, you will see still -- in the synthetic CT, you will see the whole patient as it looked on the planning CT. And the data without a field of view will be copy and paste from the planning CT because we don't have any information in those areas.

Perfect. Okay. This is back to you also, Emil. And how about ECD 2 calculation? Is that possible for ion?

Not at this moment, but that's something that we're definitely looking into, but not in 11B.

Perfect. And I've been answering some of these questions. I think you guys can all see questions and answers that have been coming in. This next one, I can answer. It says we currently have 11A. Do we need to move to 11B to implement the new deep learning model? So 11A, we do have the release models that you currently have. The models actually are released independent of the software releases, but they will -- the new models will coincide probably with the FDA approval with 11B. So 11A, you can have segmentation models and then more will be coming out and how that timeline really lays out is yet to be determined. So please call us because we can help you with that registration to get that flip. So please call us and we can do so.

Absolutely. yes.

I see -- any other questions to ask or maybe some more coming in. Okay. So I think that's all that we have in coming in. Also I'll go ahead and just do a couple more slides here. I really -- as part of the product management team wanted to note that it's really important to us that we do listen to our customers. And I think you just watched through this presentation that many of these features are in there directly because you asked for them. This is a small sampling. Some of these were already spoken about, maybe I'll reiterate a few, but that ability to do an approval before you delete plans so that we don't accidentally delete some -- some of our ROI visibility setting, that [ ECD 2 ] dose -- remaining those some doses that was really asked for by you, et cetera. So we're listening. There are many, many enhance requests in each of our releases, and we're really proud of that. And lastly, we've really thrown a lot of information. These presentations so fast. I know it's hard to really bring your RayStation user absorb all of it. And so I just wanted to reiterate a couple of resources to you. We have this 11B new marketing [indiscernible] file is available. Anyone can send it to you, or you can download it from our website. And it really has the highlight is here what's in this release so that you can go looking for them. But when you really want to dive in, go ahead and find the release notes because we've meticulously gone through and really typed out so many of the features in there. And I think that our status moment is when we have a user that pay more to use -- monster print some doses and we're like, hey, it's already there, we added it. So just make sure you at least note release notes and find all that great information. I do have one more question that came in. With automatic optimization of geometry and table angles for VMAT be possible in 11B? Automatic optimization of geometry and table angle. Either of you one tackle that one?

This -- that is like geometry and table moved during delivery or you have to optimize?

I think this optimized of those angles, and that is capable possible already in that, right?

Yes.

Okay. Agree. Agree. All right. So with that, I'll do a reminder that this webinar is recorded. It will be available on the RaySearch Laboratories website, and it also will be available in the community where -- we want all of our RayStation users to be going. They are all of our past and present information about our webinars are found there. So if you don't have access community, please, please make sure that you contact us that you ask about that and then you will find that access today. And we do have another upcoming -- this is for RayStation users only. We have a series called the Ask the Experts, where we put our users in contact with the developers themselves. This is a full webinar on the synthetic CT in RayStation and its use in the formable and adaptive workflow. So that is upcoming February 17. And if you're a RayStation user, you can certainly sign up for that. Again, it will be recorded, and all of the Ask the Experts are found in that community. So I don't know that I have any -- there are more questions coming. Okay. So Emil, which HDR planning commercial systems do we plan for today? I think that you should note which system.

Out of the box, we support the [indiscernible] ScandiNova. But we're working really hard and it's a big priority for 12A to be able to deliver plans on other vendors machines.

Perfect answer. Question is here about printing the doses. So I said it was now available. Jared, he wants to know if that is scriptable? Do you guys know if it's possible to script the printing of the sums dose? My gut feeling is that, that is a scriptable option.

I agree. I cannot give answer set in stone, but it should be because we have almost all functionality that is not patient safety related is first of all.

I would agree with that. I would agree with that. And I think that we've answered all of the questions. I really love to thank Emil and Craig, for your hard work for being the face of this 11B release. Thank you for stepping up and presenting for us today. And with that, I will remind everyone that 11B was released in December. It is still going through FDA approval within the U.S. and some of the other markets also require it. We thank you for your time watching today. And with that, we'll go ahead and conclude this webinar. Thanks, everyone.

Thank you.

Thank you.