Name: Novel Multi-Isocenter Accuracy Testing and Patient-Specific QA for Gyroscopic Radiosurgery System
Start: 2024-10-10T12:00:00.000-07:00
End: 2024-10-10T12:33:00.000-07:00

Hello everyone and thank you for joining us today for a continuation of our Best of QADS webinar series presented by Sun Nuclear. Today's presentation novel multi isocentric accuracy testing and patient Specific QA for gyroscopic radiosurgery system will be presented by Carlos Rodriguez, Director of Clinical Transformation Services for Mercius Health in Portugal. Before we get started today, I do want to just cover a couple quick housekeeping items. All attendees are able to enter questions to the box on the left of the presentation screen. We will address these at the conclusion of today's presentation as time allows. In addition, a recording of this presentation will be sent out to everyone following the broadcast. Today's presentation, novel multi ISO center accuracy testing and patient specific QA for gyroscopic radiosurgery system will be delivered by Carlos Rodriguez from Mercius Health in Portugal. Carlos has always been attracted by the physics fabric of biological processes and his technology physics degree sparked a genuine interest for clinical application of radiation therapy. Carlos has an inquisitive personality and is passionate about collaborative work. His current interests are efficient production and project management practice, transformation, regulation, compliance and advanced artificial computing. During QADS 14 earlier this year in Portugal, this presentation was delivered as part of the novel QA methods for SRSSPRT and Single ISO multiple target panel. In today's webinar, Carlos will discuss his clinical experience with the novel gyroscopic system and highlight the latest clinical uses. Carlos, I'm going to turn the presentation over to you. Thank you, Andy. Hi, I'm Carlos Rodriguez. I'm. I will be presenting on the behalf of the team of IRCA, which is a a center that has a ZAP in Madrid. The work that we did to test the accuracy of the ZAP tyroscopic radiosurgery system, but also the QA program that we put in place for patient QA, I'm here with Boja Aguilar, which is the the medical physicist that is managing the clinical aspects here at ERCA. So the purpose and the objectives of the work that we developed was to evaluate what was the beam delivery accuracy of a system that was essentially frameless radiosurgery. So the the first aspect was to look daily about single ISO center accuracy doing with Winston lose anthropomorphic phantom. Second aspect was to establish weekly a verification for multi ISO center targeting using also a phantom for that. You can see that on the lower left part of the of the screen where we used the sterile fan with the multi met with some Lutz cube. And the third aspect was to implement a patient specific QA and review what's what we could achieve with this novel equipment. And you can see part of of that on the lower right hand side where you have the rest map check and the dose distribution of one of the of the patients. So before we dive into the testing, we'll give an overview from the material aspect. And first one is obviously the machine. So I would say we're live in in Madrid. So the the picture that you have on the screen is the machine that I have behind myself. It's a three MV linac, which is Volt free self shielded machine. So as you can see, there's a glass ceiling. There's also windows for that inside the machine. There's also a stereoscopic IGRT system and the bore of the machine is like a big bore city. It's 83 centimeters and the inside of the machine is 1 meter 65, which you might have an idea that could be claustrophobic being inside the machine, but there's quite some room once you're with your head inside. The beam orientation of this machine is done by a combination of rotation of two gimbals. So this is represented on the on the image that's have on the right hand side where you have an actual rotation of one of the gimbal, which is represented by the green arrow and you have an oblique rotation which you can see with the orange arrow over there. So this leads to, I would say one of the the the challenge when you first see this machine and, and and manipulate it is as medical physicists, we are used to some coordinate systems where if we say for example, again 345° and table at 90, we have a mental representation this machine that that changes due to this combination of rotation. So to give you an example, if I take the middle left hand side picture, which would be a beam coming from the front, so anterior part to the patients that would be gantry 0 Table 0 on the normal lean act. But for this machine is actual 270 and gantry at 90°. So I'm mentioning this just that one of the things when we want to set up QA protocols are are manipulating the machine. It's one of the things that we have to put our head around about how do we generate the beams and the orientations we're looking for to treat the tests that we are doing before also going on the rest of the material. I would like to go through some consideration about ISO, ISO center and the how the treatments are done with these machines. So the the first aspect is at planning. So at planning what you do is having a selection of ISO center is each ISO center as a collimator associated to it and you also have a selections of beam per ISO center. So the computer supports the selection of this. But basically the the irradiation is intrinsically multi isocentric treatment that you do. The second aspect is related with the treatment where the table is moved to, to place each of the plan ISO center sequentially at the machine ISO center. It does it through three movements. You have a table movement which is the translation and then you have a beach and you have a yo. So that's also the machines you have registration for positioning down in six degree of freedom, but this is translated in a 3D movement for positioning. The last aspect is just definitely indicating that the dose disposition is highly sensitive to the accuracy of the IGRT registration and table movement because this is this combination that leads us from 1:00 isocenter to the other to treat the, the dose distribution. As an example, I put here an optic nerve meningioma treatment to show the distribution that we can have wrapping around the optic nerve, the dose distribution and to to show what is the obviously the the type of accuracy we're we're achieving with this type of of machine. So in terms of the machine itself, the alignment of the of the patients is made using imaging SO2D stereoscopic imaging and using that to do then some calculation for the table movements. So you'll see a video that will be showing this, this part where you have one irradiation cava second one where you will then create DRRS and do a registration based on the matching of these images. Out of that, the system will extract what is then the movements you have and you have here a depiction of the movements that are done here. They're exaggerated because to to see obviously translation yo and and tilt. A second important aspect of this machine is the collimation. So there's an automatic selection of the collimator and also the alignment of the collimator with the beam, the beam axis. So you can see a revolver kind of paratus in yellow that was indicated here and you see all the different type of collimators that can be selected. And once the selection is done, there's a mechanical lock to have the alignments for for treatment. The third aspect is the beam delivery. So the the ZAP X machine is a machine obviously that can rotate around the patient. This sets the delivery is a sequential delivery beams at at this at this moment. So one of the same thing. The video here will provides a display of that. And I think this is what I hope you're seeing something which I don't seem to to be seeing. But what you have is that the beam is delivered, you move to the next position, then again a delivery, then again a delivery. And this is done then for all the different types of centers in terms of the apparatus that we use to verify the the different aspects we wanted to look. So the first one was to verify single ISO center accuracy. So what we did in this case is use Whatsappix as a software in their console that allows automated with some loots steel bowl analysis and the algorithm in terms of positioning optimized for training and positioning. So one of the reflection we had was to try to use a phantom that has human anatomy features to try to minimize potential positioning errors that could be of using something that was features. So we custom designed ball bearing holder with the tungsten ball in it for for this this part. The second aspect was verifying the multi ISO center accuracy since the system moves from one ISO center to the other one to track if there would be any issues with that. So we selected the stereo fan system since it's an industry standard and reliable system and also demonstrated and several publications in radio surgery. So nuclei are also at specifically designed base plates for that phantom. The ZAP requires different distance to be be able to position the phantom at the ISO center. The multi pet met Winston Lutz inserts were used in to verify how the conversion of the 6th degree of freedom was done to the three degree freedom movements of the table and we got stuck with the 11. Issue is that the Zapic software doesn't expert in the images in DICOM RT. So some nuclear as a software that allows to do this type of analysis automatically. But we had to develop our own routines to be able to do this. The third aspect is the patient specific QA. Same thing here, we went to use the map check due to his wide usage in radiosurgery context. Also, the, the system had a cap AP parameters QA called Cyber knife routine. And for us it was an opportunity also to see if we could develop part of AQA of, of the beam, not specifically the, the patient QA. But this is not the, the, the aim of this presentation, but it, it's a little bit directional as well. Why we went this route in terms of the design of this matrix, it creates a a spatial sampling resolution that goes between 175 and 247mm for that. And also you and Al have already used this type of system for the the usage of these patients for in the ZAP context as well. So in terms of method, what we did was to daily use the prime time RT 4mm diameter think symbol in in the phantom to simulate the patients. After the auto alignment using the KV with the DRR, the steel bowl would be placed at the center of the KV imaging. At this moment we would do then the tests using different gantry angles, patient left, right, what is called the ohm, anterior, posterior etcetera and image would be acquired with the MV imager which is a scintillator based detector that normally is used to measure the beam attenuation through the patient during the treatments. So for each beam it generates a 1000 by 1000 pixel image as a TIFF image and the pixel size is 0.04 ish pixel millimeter per pixel at ISO center. All image were used using the built in ZAP X software to to do the calculations for the multi pisocenter. The procedure was quite similar weekly. What we did was use the cerofan with the multi met with some loot insert inside which has different diameter, different steel bolts of 5mm, placing non position on the axis but also off axis. We would do the alignment in the same way as we did with the single ISO center and using the registration algorithms of the system. And we created 4 images, patient left, right anterior, posterior and posterior anterior. And we used well detector was exactly the same. And we use in house developed software using Python to get the offsets between the the steel bowl and the beam center. Finally, for what was the the patient QA, we use the SRS map check to which we had copied the patient plan and we positionize the center manually matched with the center of the metrics follow the same process in terms of what is alignment of this of this system did those measurements were collected and pro set with the SRS map check software benchmark against the patient plan. Also for benchmarking purposes, we use the Symocam Monte Carlo Dosengine from Scientific Care RT to assess the accuracy of the calculated plans. So the results that we got for single ISO center measurements of we've been now in activity to close two years. So the results that that you have here, we have now a little bit more than these ones. But initially and what we showed during the the workshop was that over the first year that we we were running, we analyzed around 2200 with some tests for singleized center in terms of mean and maximum deviation. What we got in all the different directions was submitted accuracy which is what we would expect obviously, but it was good that the results were pointing in very good direction. Also I put here you can see on the right hand side a plot which is interesting. Also we're in February 2024, we have done a table upgrade which is to allow treatment down to the C7 and as you can see it led to some noticeable changes. When you look at each of the graphic you have in the darker colour, which is the the black is what was the the new positioning and what you have in light colour was was the old ones. So by doing these upgrades, we do also recalibrations of, you know, what's the zero position of the table. And basically what we can observe is that there's some variation in what we could assume a change in accuracy and precision, but that shows also what are. I would say the the the limits of the the calibration accuracy that we can have with such a system. So was reassuring to see that we can have consistency along the time of what we are doing, but also that we are able to monitor what is happening when we have some change that might be subtle to detect. Also during the the course of of initially 6 months, we did this multi target analysis, same thing as with the single ISO center, we observed submillimetric accuracy, which was also, I would say reassuring. The machine normally moving between ISO centers as I would say not big movements, but when we're doing metastasis, several metastasis treatments, obviously movement can be more important. So being able to verify accuracy through quite long distance is also one thing that was reassuring on the capacities of the system. Finally, in terms of the results for a patient QA, we looked at the the first patients that we treated where we did 22D dose verification that we kept also be doing for the rest of the patients. But what we, we did was to observe the the different gamma index that we could have with, with the approach that we we took and also look at what was the best correlation that we would end up to have. So things was interesting to see that the the best correlations were obtained with two percent 1mm passing rates and also the dependency on the volume, which is something that obviously we we can assume due to the also heterogeneity of some of the the plants that we we treat. We use Simoka as well with a 2 percent 1mm mini gamma passing rate ended up to be 98%. And when we compared to the measurements that we we were doing, it was quite reassuring and consistent of what we were observing. We also had developed a separate way of verifying the monetary calculations that the TPS was doing. And also these results were reassuring as we were plus -0 point 3% on what we were observing. So in conclusions of what we have been doing initially, but also what we've been doing since we have observed millimeter accuracy possible for single colimeter targets. We don't mean deviation that was less than 0.38mm and the maximum punctual deviation of 0.7. The multiple ISO center QA was also reassuring with similar numbers. We had maximum punctual deviation a little bit bigger than what we had in single shots, but what it reassured also us. This type of tests is that it created integral validation of image based alignments and also the treatment couch find new movements to to reach the ISO center positions and for the patient QA verification. And in place we were able to demonstrate very good agreements with the TPS calculation, which when this DPS has pencil beam algorithm based. So as physicists we get often directly itchy with this type of algorithm and want to have some benchmark. So having benchmarked this with also independent Monte Carlo system reassured us and also the confrontation with the measurement that we did practically. So the superior results in the planes that we've verified were for smaller lesions and also when we meters as we would have I would say expected as as a result. So I would like to thank the team that you have listed here. I am with Baja that presented the waving at the beginning and that we will be taking any question that you would have for us and also acknowledging both Vanya Diaz and Katarina Soto that have been helping as well to start the activity here. Thank much. Thank you, Carlos. Appreciate the presentation. Wonderful presentation on a developing and and new technology to the marketplace. Relatively new, I should say. Always good to hear from an early adopter. I I do want to let everyone know while you go ahead and if you do have any questions, we encourage you to type those in the chat window. But our next and final Best of QADS presentation will be on Thursday, November 7th. That will be the clinical application of theranostics and how to approach patient QA. So again, another developing technology in use in radiotherapy spokesman theranostics and radiopharmaceutical therapies presented by Roy Hobbs. So again that can be registered on sunnuclear.com. Carlos, while we're waiting for some questions to come in, I know you touched most specifically on some of the patient specific QA that you've done. Maybe you can speak a little bit to the the clinical types of indications that you're treating with the ZAP X machine and and progress you've made in in in using the machine within your clinic. I will just do a like a small introduction to for for an answer to the question you had. I'll let Bokha also lenient with the system. We have been treating quite a a big variety of of pathologies, both oncological but also functional treatments. We have done very challenging, I would say challenging both for the machine in itself, but also in terms of what we needed to achieve clinically, which was for example having multiple metastasis treatments that would be more than 20. For example. We had cases where like for example, the case I was showing, which is a very optic meningioma where you're, you're basically in contact with the with with the organ at risk and having treatments also of Avms, which is also non oncological. But I'll let Boha also comment on the. Yes, Carlos said. We have treated a patient with up to 26 metastasis. We have treated very irregular volumes with very regular shapes and very big volumes up to 6250 CCS cubic centimeters, which is a challenge for this kind of machine. We've done also functional. We have treated trigeminals and facial spasms, yeah. Wonderful. Thank you. I will say it's a first for us to actually have someone presenting directly from the treatment vault. So thanks. Thanks so much for doing that with the the machine right in the background there. There is another question that that's come in speaking specifically about patient QA. Was most of your patient QA done with single ISO centers or did you actually do analysis on the different multi multi sites of those tumors? So I, I, I think, well, the question is or at least the way I understand it is that we do the machine QA on a daily basis where we, we, we do the, the single shot of telling it's, it's the ball bearing with one ISO center and doing it weekly. We do a verification for for the multiple position. In terms of the patient QA itself, we do it for every patient. Wonderful. Yeah. For patient for a specific patient QA, we use the SRS map check and we check all the ISO centers that are in including the planning. Thank you both. There's a question about some of the dose thresholds that you've chosen to use. I think on your slide maybe it said 20% dose threshold and there are some other publications out there that talk about using a higher dose threshold. Can you comment on on why you chose the 20%? Yeah, we use, usually we use threshold of 20%. That's the threshold that I've used in my clinical practice. That's my experience. If you use a higher threshold, normally you are going to get a better result. So we try to take into account low lower doses and to take into account those beams that and those doses that are close to, to the not to the tumor volume, but beyond to, to, to larger distance beyond the target volume. So if you have a lot of inhomogeneities or you have beams through your laminar tribosoi through your nose, you have to take into account those inhomogeneities and the way you take into account that is put in a low threshold. Thank you very much for the the explanation. With that, I do not see any additional questions and I want to thank both of you for the presentation. Both. At. QADS in February and then reprising that today as part of the Best of QADS presentation. Thank you again for joining us to everyone on the line, and we look forward to seeing you on our next webinar. Thank you again. Carlos, thank you. Thank you everyone. _1733887398797