Hello everybody. Welcome to the webinar. My name is Neil Miller. I am a Director of Product Marketing here at Sun Nuclear. We are Miriam Medical Company. It is my pleasure to welcome Doctor Manuela Borgella, medical physicist at Institute Jules Bourdais in Brussels, Belgium. Her professional trajectory is marked by a unique fusion of expertise in physics and a fervent enthusiasm for interdisciplinary exploration. Having earned her doctoral degree from VIA University in Brussels, where she focused on pioneering treatment delivery approaches for radiotherapy, Dr. Borgella transitioned to a collaborative role at I GB I GB IGIJBS. Boy that was tough to say. Nuclear Medicine Department Doctor Borgella's current pursuits involve the clinical transition of the CT innovations in radiation oncology, exemplifying the modern medical physicist who thrives at the crossroads of scientific collaboration. Just a a few quick logistics before we begin. Please note that all attendees are muted. We do encourage questions though. If you have any, please enter those in the questions box that you should see on the left side of your screen. We will address those at the end. We are also recording us and we will send a link of the recording out to everyone who signed up. So if you miss anything you will be able to to go back and do that. I do have a couple of quick polls just to run before before we get going. I'm just interested to to get a little background from the folks that are online right now. So the first poll here is, are you using 40 CT in your treatment planning currently? So if you could just take a couple of minutes or a couple of seconds really and and answer that for me, I would really appreciate it. OK. Looks like we're getting quite a few people filling that out up to about half of you and give you just about 10 more seconds. All right, let's see what people had to say here. So wow so at 73% of you are doing 4D CT or using 4D CT that's awesome 12% planning to start few that are investigating and and about 10% of you who who are not currently as I mentioned to Manuela when we were getting ready to open up the the webinar this morning we had honestly surprising number of people sign up. I wasn't sure you know what the interest would be out there on 40 CT So I can see why I'm a lot of people are doing it and I'm sure you're really interested in what Manuela has to say. So got one more quick poll. If you are doing 40 CT, so it look like a bunch of you are, do you have AQA program in place again, if you just, you know, 30 seconds or so to to fill this one out. OK, looks like getting up to almost half of you now. Give me about 10 more seconds. OK, Let's see. All right. Yes. Oh, wow. Almost even across the board, people who have one, who are developing one and and don't have one right now. So for those of you who are developing one and don't have one right now, I think what you're going to hear over this next 30 to 40 minutes is going to be very, very interesting to you. And with that, I am going to pass it over to Manuela. You have the floor. Thank you very much. So today I'm going to start presenting some of the argue that we're doing in institutional Borde. So the presentation it's assessing the accuracy of current for digit imaging protocols and the multi institutional study. So how everything started in 2021? We had to move from old board to a brand new hospital that is literally called the new board where we have 6 new bankers equipped with Electa systems and one with the Electa Unity machine. Also we have two simulations room and MRI simulation room and the CD simulation room and I was in charge of the CD simulation from the public tender preparation till the 1st patient treatment simulation going into the market. We decided to go for some Matongo Oprahum Pro which is the latest simulator proposed by Siemens. And here below you can see the specific features that were really interesting for us. For example, Director Lasers is the lasers that are attached into the board of the city and they are directly coordinated with the interior, interior, the laser of the city itself. The mobile workload means that there are two tablets next to the inside of the city and you can take them out and you easily position your patient and you. This way you have more time to go to remain next to the patient. I might it's the artifact reduction workflow between spiral dual energy. It's the solution that it's you can have dual energy and algorithm. You just have one One scan with one KV and another scan in a second can be and you can have a reconstruction of dual energy direct density. You don't have to reduce, you don't have to do different KV images. You just to have to do one KV and there is an reconstruction algorithm that you can implement to complete that will be independent of the KV that you are using in in your clinic. And the last solution is the 4D CT solution, prepare the 4D solution, the direct I4D solution proposed by Simmons. We after the commissioning of the city, we continue to do research at the level of the 4D because we came from two different reasons. Firstly, we came from an old Toshiba. We really had very bad quality for the city images and we really wanted to come I directly for this the latest solution presented on the market. And secondly, we try to do more research at the level of this part because Director I4D was introduced maybe only two, three years ago, so we didn't find a lot of literature on this subject. So as you can see, this is a small content of my presentation around for the city. I will start with a short background on for the city. We'll continue with the quality assurance literature review and I will continue furthermore with the two articles that together in my team we published last year. The first article it's regarding a compressive quality assurance program for for the city in radiotherapy and we continued after the program was developed and implemented in our clinic. We applied it in a multi centric audit for quality assurance, what is for the city, So actually for the city was introduced more than 20 years ago aiming to reconstruct the dynamic volumetric imaging or respiratory correlation city. It adds A temporal dimension to the traditional 3D city and provides a visualization of the target motion. It had a lot of benefits at the level of the therapy to reduce planning target margin. However, the knowledge of the limitations are still present 20 years ago and also now. The major limitation remains the motion uncertainties and the image artifact resulted from the irregularity of the patient breeding. This is a basic and general representation of what the for the CT process is about. The external breeding signal of the patient is recorded during the CT acquisition after the images are sorted retrospectively into images in the imaging in the bins, which corresponds to different phases of the breeding cycle. The final image. The final image represents a collection of 3D images acquired over entire respiratory cycle and each corresponding to a particular breeding phase. There are various breathing recording systems on the market and here I just listed the the most used clinically. I will start with the the RGSC system from Varian which is the reflector block positions on the Torrex of the patient and that will now be detected by the infrared camera and the movement will be projected back to the to the city. The ANSI belt system it's basically a fixation belt that is a with a pressure traductor and to the expansion and contraction of the belt during greeting. The pressure transductor will deliver a digital signal. Gate City from Vision RT and Sentinel 4D from Syria are the surface guided solutions available on the market. And for example the serial scanners scans over the surface of the patient with a laser and the text direct the text and the camera detect the laser projection. So basically they are just following the surface or define specific surface at the level of the tracks of the patient. And this information it's afterwards you sent to the city. However, not everything is plug and play. Recently the NCS Test Group published a very nice article regarding motion management and one of their comments and the strong encouragement was that the development of open platform and standardization interface by the vendors which will enable different devices to operate and communicate safely. And I can give you an example what is at the level of our institution in our city room we have the Vision RT, the Gate City solution installed because we are a Vision RT representative site and it was already in our institution before we moved to the new board. And behind it we have the LPM, a new camera just because the I4D solution, the new solution for Zemos is not compatible with the Vision RT camera. So we had to introduce the LPM back into our clinic. We are not very happy but this is the way that we had to do it. Just a quick discussion about the sorting approaches in 4D. So I mentioned before after the acquisition is regarding about the the images so that it comes to have a sorting binning approaches and the binning approaches of the 4D can be phase based and amplitude based. Phase based solution means the reading cycle is divided in time equidistant beams and each image is assigned to a specific beam depending on which phase of the cycle was acquired on amplitude based each cycle is divided in amplitude equidistance beam and each image is assigned corresponding to bleeding amplitude. This is a small summary from the cooking book of Siemens and these are pros and contacts of different two solutions and you will see later on the two solutions are both clinically available and used in at the level of the city. It's very important to say that all the time, the most of the time the retrospect, the analysis and the binning solution are done retrospectively based on the breeding signal and the projection data, breeding signal of the external surrogate. Just a quick word about the I4D because as I said it was it's quite recently presented implemented on the market. However, I left here behind Werner and all he had very nice publications. He was involved in the development and and validation of these solutions and you can just Google and the on apartment the his publications are really specific on the quality of assurance that you can do at the level of I4. DI want to mention this because he's the first solute for the city clinically available solution that is using prospective data that is doing the binding. The idea of the 4D CT is to continuously acquire projection data in sequence mode for each couch position of the field of view until it's assured based on the real time analysis of the breeding signal that the acquired data covers the covers the entire representative cycle of the patient. I did hear a small screenshot of the cooking book that Zimmons is a it's offering for this solutions and it's really nicely presenting the difference between their conventional for the city and the I4D solution and how it's nicely improving the quality of the image. Of course I 4D it's an add on and comes with a pricey cost. So as I said before we started, once we did the commissioning of the city, we came up, we want to do more development at the level of the for the city quality assurance and we try to come up with the quality assurance program and we figure out some specific criteria meaning that we need the relative test that can clinically representative of breathing patterns. And this was done to literature review. We wanted to be sure that the city parameters and reconstruction should be sufficient and available in clinical routine, that our software should be able to identify artifacts and deviations and we can do reputability test and all steps of the evaluation should be relative and easy to follow. The entire workflow should be as automatic as possible with limited user interface and we also try to have a procedure that was easy to implement and user independent from. This is a short representation of the 4D CT literature review mainly what papers we already used in our part. So these are the bookmarks and Lambert the the in 2011 and 2018 are the two publications that we sue at the level of for that. They really verified the the audit on for the city quality assurance and different multi centric trials and for example Lambert it was really at the level of the redefined SBRT at the level of lung tumors and for the city was part of this quality assurance only in 2018 there was really really an interesting CPQR technical quality assurance guideline who really presented the first time they presented validation tests and tolerances that could be that could be easily implemented in clinical routine and the tolerances were available and easy to understand, nicely explained and easy to understand. And in 2022 we we found out, we found out that somebody really tried to implement that in their clinical routine through a poster presentation at the double APM. This really gave us a lot of idea that we are going on the right track. And also this is just a small demo that we found. One of our colleagues from Germany implemented a software frame solution for City based quality assurance that it's open source and we just wanted to do the same similar solutions that he proposed for quality of the city at the level of for the city quality assurance. So basically this was our our reference how do you want to do our QA motion software and the others were really helpful for the level of implementation of what we want to do. So quality assurance at the level of the acquisition and implementation, we used our Sears to rest dynamic Phantom. The Phantom was previously supported by Sears and now it's under the umbrella of Sun Nuclear and the Median Medical company. It's a phantom. I think everybody knows about this. So I'm not going to go into details. If you want details that Neil can give you without any any problem. It's a phantom that it allows comes up with a software and algorithm that you can do 3D motion solutions and also implement breeding, breeding patterns at the level of the of the software and it came with different inserts and different positionings. And this is in the image you don't see my screen but in the image you can see the positioning of the positioning of the phantom at the level of the CT and the camera that it's following. The in this case we are we are having RGS external, the block from the RGS system and the camera that is mounted on the table. At the level of reconstructions we use 3D CT reconstruction as a ground through to really determine the the image insert and the 4D CT 10 phases, phase based solution, phase based binning. And we do also the MAXI maximum intensity projection and average density projection reconstructions to be used because we use this in our clinical solutions. Again, not not everything is plug and play before we came up that the imaging insert it's it's what we need for our audit. We also tested different inserts and some of them by accident, some of them because we didn't have the imaging insert initially in our institution and not everything works. For example, the micro chamber insert has the as you can see here has the the line, the insert of the utilization chamber that gets like a line through the unit, through the through the target. So that's very difficult to try to automatize the computing, the auto computing of the target. The SBRT Rd. has the SPL has specific roads for the visible fiducials near the target. In those fiducials in for the city, we realized that they give even more artifacts than the target itself there, So that was also out and some colleagues by mistake also they did they use the imaging insert but with a different diameter of 3 centimeters. This is much easier to apply in our software because but you still have to do changes at the level of the software. So this after this threat we really started to set aside one day and wrote a procedure about how the QA should be done at the level of of the. For the city breeding patterns and evaluation metrics, we took as a reference the the president, the President, the article presented by Julianne where she developed a lot of breeding pattern solution to test the I4D solution in comparison with the old, the the older solution for instruments with the new I4D solution that they're preparing. And we came up together with her to this to have breeding patterns that are simulating as much as basically the clinical routine. So in the first columns you can see breeding patterns that are five regular solutions and three irregular breeding curves. The regular solutions are the first one, it's 3D that we try to have it more complex to be sure that we can is not only superior, inferior, but it has a lot of more movement at the level of of the tumor. The second one it's with amplitude 5 and period of five diseases. A normal breeding patient. Afterwards we have a patient that has a higher and higher amplitude and the higher higher altitude, higher. And a very small amplitude and very small. The the 4th one on the part, it was really to mimic the lower amplitude. It was definitely based on breeding curves of pancreatications to mimic the compression belt that you use during acquisition. The irregular breathing patterns we tried to mimic as much like a really agitated person who has. He has a very high amplitude like 2.5 amplitude that if you see we can also see that in our clinical settings double amplitude for example, if the patient has to come up with doing immerse rest during the during the simulation you can have double amplitude solutions or irregular frequency or purely irregular breathing patterns. As I mentioned before, the CPQR was the first one that really presented the first introduced guidelines once for the four DCD performance and mostly for regular patterns and for amplitudes smaller than 2cm. The Canadian partnership, the guidelines were introduced for regular sound, but however the tolerance that we wanted to to have, they're presented here, we wanted to go a little bit further to see also the limitation of our systems. So we go to went for the irregular patterns. We did electrical tests, multiple States and intuitionally came up also with tolerances that could be acceptable or that can have a trigger at the level of a trigger alarm that you should investigate more certain certain solutions. The metrics you can they are presented here and they are divided in amplitude deviations in volume deviations and city number deviation. And something that we added also extra. It's because they use amplitude, the maximization, the projection solutions and the special integrity. Special integrity means that you draw a line, Special integrity, you draw a line through the. It's an evaluation independent on the contours that we are creating. It's a line through the for the city and based on the House field units have with the measurement you are determined the value of how big is your tumor and how big is your amplitude at the level of the movement. This is an example of a short overview of the QA motion, QA Motion. It's the name that we gave for our image analysis workflow. It was implemented in meme software as an as an workflow. So basically once the data the once the data is imported in MIM, the user selects for reconstructions the stat for reconstruction, the, the 3D, the static CT which will be used as a reference for the which will be used as the reference and the for the CT the maximization projection and the average IP projection. During the first step, we also are looking at the DACOM data and take out some relevant important values that are used to finalize at the at the in the final report. Second step, the static city is used as a reference and city number threshold based method was implemented to contour the target for D with a volume consistency smaller below 0.01 milliliter. So the auto controlling is done on the static city. It's determined have a threshold value in the static city and this threshold is afterwards applied to the exhale phase. Once the outer contour is done on the excel phase on the for the rest of the phases for the city different propagation. It's used to have the contours on all the 10 phases. Once the auto contouring is done in the third state, the for the city image analysis it begins. We start with the monitoring system. So actually the monitoring system data it's retrieved automatically from a specific folder and it's compared with the input data that we have from the, from the, from the, from the Sears. I didn't mention that actually once the data is in in the workflow, the software automatically detect what type, what type of breeding pattern was implemented, was used at the level of the acquisition. And this reference that I will be compared with the monitoring data solution because this is really important considering that this factor it's completely independent on the acquisition. Once the external BIMA system it's verified, we go for the image imaging data and that's like amplitude deviation, volume deviation, city number deviation and special integrity. These are all the all the data. It's automatically analyzed in the workflow and the results are on the display and the analysis is finalized with a standardized report and color labeled results. And also we made sure that for people who are more interested in going to details on the results, there is also an available and exporting a CSV format. If you want to go for the new details I will come now. You should see a video currently regarding a small demo of the workflow that we implemented in MIM. I see it a little bit, I hope you see it better than me. But basically this is like once the data it's loaded and we run the workflow, he's automatically starting doing the contouring. As you see it's not just plug and play, the outer contouring is done with the different steps and to come up to a nice target validation of up to 0.1 volume. This is a small visualization of the target that the movement of the target and how the contour it's following the moving target. He's still doing the as you can see he's still doing the contouring in the December moment on the 10 phases applying the deformable registration. And he also you can see here the contouring of the phases that was average phase in the the industry projection phase. Here the last step is creating the report, the PDF report that can be easily accessed and visualized and here are the example of it and also if you want to go in the workflow, each workflow has a different window. In each evaluation it's presented in a different window and you can go through the statistics and through the images to see where you have. For example, in this part you see that we have artifacts at different series and the 10% and you just go to the 10% and see if you have an artifact that corresponds to your results. Can you see now my slide regarding the validation? I hope because I'm still seeing the video. Would you like the video stats? Yes, I don't. I didn't stop it. Yes, thank you. So can you see my slide with the new? Yes. So currently these are the results of our single slide, the repeatability test. So I said once we developed the evaluation and the evaluation program, we tried to validate it in our institution by using our for the city solution I4D. And we did the reputability test. So we did all the eight tests available in three times and we saw that for regular curves I can see this is the result of the. For the regular curves we can have very good results with the into the that are into the tolerances of city, city, QR volume deviations below 5%, city number valid evaluation below 10 Honsfield units and diameter deviations below 2mm. This was really nice to see and evaluate. However, we saw that for irregular and high motions we can have artifacts and this is you can see them by seeing the volume, the outliers in the volume deviation and each year small outlier represents is the representing on a small artifact. And for example for the irregular frequency, we really have a huge artifact that you can have volume deviation up to 20% of of of difference. Once we had together and we sit together and we saw that everything goes well with the quality. With the for the QA motion, the evaluation program that we implemented for the city, we came around and the seed were who wants to join or who wants to test their for the city. Currently I was in the same. I was in an extra workshop, the city City Innovations and that implementation at the level of city and it was really really nice and they really helped us and give us a lot of feedback at the level of implementation of the quality assurance. So we asked around if everybody from the workshop would like to join and also we liked around as you see Belgium is a small country. So we asked around who wants to join our study and luckily a lot of people wanted to join. So we invited the 14 institutions, 14 institutions were invited over a period of four months, 11 in Belgium between France and one in Germany. Each institutions can from 14 institution 13 were able to join. One institution postponed the audit because the machine wasn't break was in breakdown during the when we were planning to visit for the audit. And another institution postponed the audit because they wanted to install their new external monitoring system. So we postponed it for for six months afterwards each center was allowed to use their four DCT acquisition protocol for SBRT and use their in house monitoring system. We didn't give them any any limitations and we were able to have images data from 12 institutions in this side. We did nine side visits the the Belgium sites where we did the visited and did the audit with our Phantom and the three remote support. That means that the three institution had their own Sears Phantom and they were able to do the audit just following our procedure. And maybe I think we had one sort Tyson online call to if they had the questions or or discussions about the phantom usage or the other procedure. Of course again nothing is plug and play. What we saw it's that the QA motion that we implemented on in our department was not always compatible with all the data that we were receiving. So in order to raise the data from various imaging and treatment planning system, the initial workflow was slightly expanded. For example, the average IP and the maximum IP reconstructions were not only uploaded and because they were not mandatory, the monetary data was 3D and for the phases while every JP were optional. So when we saw that there were the available were not available in the data, we decided to create them in automatically in MEME. So we implemented this in the workflow anonymized template was done Meme meme software requires a specific discussion of the series to be sure that is creating the monoclotor of for the city naming. So then if this was not presented in the protocol. Therefore we try to internize an anonymization template to systematically rename the series to be sure that we don't have problems at the level of creation of the for the city movie. The last one was also because we didn't know how to implement the external. Most of the people had different external monitoring systems, not only NGS system as us. Firstly, we tried to we had to adapt our algorithm analysis to be a way to be available also for different signal than the NGS system in our part. And secondly, we did a drop off list to be sure that we can, the user can select the external monitoring system that is using their Clinique and this information can be applied or introduced at the level of the results and the report. This is an overview of the scanner and the models that we had over the 12 institutions that we investigated. We had three different city vendors including seven different cities scanner types. We had three different breeding solutions, three different breeding systems, the LGS, the Sentinel and the Gate CT. This is also a small overview about the that was recently published by the NCS group and they did a questionnaire regarding motion management and what devices are recorded in recording surrogate breeding signal. And 67% of the people are using market block and this is also in agreement with the double APM Group 324 who also did a similar motion management inventory And from they had quite a large number of responses, I think more than 500 responses. And for them was also more than 67% of the people are using the real time positioning management and the secondary is followed by the surface tracking or the the anzide belt other people are using. The other people are using the device list for example. There was also some questions, some of them are using the device list from GE or non commercial solutions. The external monitor system gave us a really good results as you see at the level of measurements. This is an example what we implemented in our software. So basically we were importing the signal of the 40s of the external monitor that were the fixed of the motion of the breeding motion and we were comparing the motion, the breeding the pixel from the external motoring with the motion that we gave as a reference in the Sears Phantom. And as you can see in the graphic, the Sergei deviations were in between were in very comparable with the reference data and they were below 2mm at the level of amplitude deviation at the level of the reconstruction. We we we used all only phase based city data because this was that we use in the clinical routine and most of the people are using I think from 12 institutions. 10 institutions was amplitude phase based reconstruction and everybody and two only amplitude based in the construction. Because we wanted to be sure that we have correct analysis of the data and we don't have this as another parameter to be tested. We just from for our analysis we only used amplitude phase based binning. However, the workflow is able to do also amplitude based binning. This was in agreement with them recent publications. As you see the NCS group published at 60% of the of their responder were using face based binning and the other 180% almost 80% of the of the 500 respondents were also using face based binning. In our case the reconstruction is done in our base. The you can say this in our in our institution the the reconstruction image reconstruction used for treatment planning system was most of the time average intensity projection and this is in agreement with what the other institution also are using at the level of double APM responders. Out of 500 responders 61% were in agreement that they use average data set. And we are also coming up with the similar results for our test coming to results basically the iterative approach implemented in our software. We realized that actually it's the auto controlling solution works very well but also gives a very large threshold differences ranging from -500 hours unit to -280 house unit variations. These are large variations to determine the thresholds to be sure that your the threshold is done. But also this is also because as you see you saw before different cities are using also different filters for the for their acquisition. This was not an limitation of of our study, but however it was really important to be sure that we have the auto controlling thresholder well implemented as as a golden standard for the regular patterns. The mean the mean volume deviation was below 2% and the mean city number deviation below 5 house units while the amplitude deviation across institutions was within 2mm. So this the overall the the the institutions, the results were very well we had here. You can see the outliers there are small, There are small volume deviations, there are small outliers everywhere. But these are really small influences and that are not really influencing the mean value of the deviations. And are still in the CPQR recommendations. However, for irregular patterns, the result irregular patterns were much more variable. The the city number, the deviations were up to 150 house units per registered. For the the up to 130 house units of variation were registered for the high velocity pattern for more than one institution. The high velocity pattern that is the the first one gave a lot of variable results between institutions. Large amplitude deviation also observed for this irregular patterns specifically for double amplitude for irregular frequency where over estimation up to 6mm were detected. This is also interesting to follow because, and not very concerning because the irregular batteries we really wanted to push the systems as much as possible to their to their limitations. So double amplitude is not easy to to follow, not even for the, not even for the CS phantom. But at the level of the city, we really wanted to see how we can introduce some artifacts. If the artifacts are there, if we can quantify them and repeatedly see them in different images. The line profile analysis of the four DCD space, the special integrity was very good at the level of at the level of the regular patterns. But you can pay attention that the regular patterns in the in the first diagrams have a different scaling compared with the 2nd diaphragm and the the. The regular patterns had the amplitude deviation of maximum 2mm which is in agreement with the CPQR. However in the irregular in the irregular breeding patterns we had the deviation up to 6mm and also the up to. This was in agreement with the amplitude deviation that we see in the in the previous slide. This is also a small example of how you can see in the line profile the artifacts you can see initially in the first part that you have on on double amplitude solutions and also an underestimation at the level of in the middle of the of the target. That's creating that you have a double structure inside of near or inside of your of your target. And that was really interesting to see and to quantify to where is your artifact and and to see it. This is also an example of irregular frequency patterns, artifacts that were were seen on different scanners and you can see they are really in general completely different. It's double structure or irregular patterns. And this is also in agreement with the regularities that were represented by Werner and are in 2019 where they were nicely presenting types of artifacts that are can be implement, can be introduced at the level of the for the city. So in summary for the city quality assurance was one of the major limitations of the of the of this audit was the dynamic phantom of ability. So we saw in Belgium for example from the 12 institutions, only two institutions had their own phantom and everybody else participated in the phantom. But we had to go or share our phantom or go to the Audi together with them because they were not the the the the phantom was not available. The external monitoring system was very good independent on the external that the system that we used it was the the deviations were below 2mm. The the average IP was the dominant cities used for longest BRT treatment delivery. This was available in our study. As you can see also the other most recent presentations also had the same reply regular breeding patterns they the results of the for DCT systems were within the CPQR provided thresholds and that was really nice to see her and to hear about it. And one nice thing that we also notice is that the direct I4D solution had more robust results against breeding variations compared with the compared with the other systems. The largest variations with I4D solutions were still in the CPQR recommendations. So they they're they were not perfect, they still had some small artifacts, but the artifacts, the latest of the active set was still limited compared with the other systems. Some other specific questions to finalize. How do we transfer knowledge to other clinics? This is a small summary, so the four DQA procedure it's implemented as supplementary file. You have to just be sure to search it. It's in the in in our published article as supplementary data, the 4D CDQA. It can be as a program flyer, as a commercial solution by shared at national conferences or as a different users as this webinar. The to improve the data. The foreign city evaluation workflow MEME is working to to have an online solution through MEME anywhere and improve the access to Dynamic Phantom. Or maybe this is a bit. This is quite possible in the Belgium because it's a small country, so to have a renting solution this is also up to discussion. How do we clinically cope with artifact artefact? Read for the city images. As we saw this is also from the NCS group. The quality of the image, it's at the level of evaluation the breeding curve or stack artifacts or interpretation artifacts and really done at the level of the patient when the patient is on the table. For example in our case we see that this is not only the case because in we are using I4D module with RGS and the in the Siemens Ingovia console we need to do 5 clicks to be sure that we are we can visually see the 4D CD movie. The RTDS are not doing that only and at the end we come up with the 4D CD. When we do under geometry, maybe one hour later when the patient is not there anymore and even we see artifacts, it's a little bit quite late. So the biggest advice trainers RTT responsible for simulation to detect and and discuss about the foreign city that they just acquired. This is a slide from Evelyn what they implemented in their institution for quality assurance of for the city. Basically it's a work chart that depending on the artifact and the the size of the artifact, the type of the artifact and the place of the artifact, you can have different solutions called the called the physicist redo Retest guide the patient. This is very nice and it takes I think to well discuss it. It takes a little more time to implement it that was implemented. It's much smoother and standardized in your clinic breathing guidance. There are also this is was part of of the presentation there. Apparently there are 74% of the people who are doing patient coaching at the level of the city. But this is doesn't mean that it's like audio coaching. Not a lot of them are doing outdoor coaching or visual coaching. And the last part of the literature are really saying that it's not easy to coach at the level of the 4D because it will be really difficult to implement it at the level of treatment delivery, what the future holds for for DC team. I also want to highlight a very nice parallel opposite editorial that was last year in the medical physics journal for DCT. It's a long overdue for improvement and this is a debate for should we call for vendors attention to the devote resources in innovating for DCT technology since it's long overdue for commercial development or should we have be aware of the increments and the fundamental limitation of the 4D CD are very hard to overcome. These are really nice. I really suggest that you should go and and have the editorial. I just took some two small slides for that. So they really printed that. There is one publication that presented that for the city artifacts affect the local control and is BRT for lung and liver metastasis. Please look at the presentation the the article is very interesting and when we saw this we really tried to mimic this part because as I said before we had a very bad experience for years with our older city. So we had this type of a lot of artifacts and for the city images, however, it was very, very difficult to do these correlations because the LGS data we were not accessed, it was never recorded or saved in a specific or standardized ways. We were not having a proper follow up of the SBR team. Most of the time it's the ONCO log and the feedback was never arrived at the radiotherapy with so it was difficult to find local control data and also difficult to determine if the doctor used for the four DCD for contouring. Sometimes the images were so bad the doctor was completely ignoring the 4D, so we were not able to see for the space shifting. Even if we had local control data, if the 4D was used or not for the contouring, something else, the 4D CT simulation. If you have a perfect art for the CT, that doesn't mean that your motion management and level of the machine will be perfect. So the 4D CT will not impact its ability to predict how the patient is going to breathe during treatment. And this is a nice paper from Jennifer quite a long time published where she presented that they were following lung and liver patients and the baseline was the CBCT amplitude deviation in cranocaudal and they saw that you can have at the level of treatment delivery still a lot of deviations. For three patients they had deviations higher than 5mm and even for one lesion, the lesion the lesion 31, the 4D CT overestimates the motion amplitude during treatment with 2.8mm and and this was really interesting to see and don't believe everything what your for the CT is telling because at the level of treatment things can change drastically. Something else that also mentioning but this list briefly to say they are really the people from UCLA are going for five DCT respiratory motion modeling. This is a quite a new solution and yeah I'm shortly presenting the floor chart describing the technique. If you go into details please see their literature. I think they have already 3 nice publications on the subject and this seems very interesting and very nice. Currently it's only available at at their institution and is not commercially available on the market. Just one final word, I would like to thank all the people from the clinical translation of first city innovation in Radiology group. It was really helpful and interesting to have them on support and discussions at the level of our research. And of course, thank you very much to my physics team at institutional board. Thank you so much Manuela. That was a really amazing presentation. There was a lot of great information there and and obviously it shows in the questions. We actually got quite a few. There's there's sometimes with these webinars, nobody asks a single one and we've got a whole bunch. We don't have a whole lot of time left. So I'm going to try and pick a few here that I think are important and anyone who we don't answer your question well we'll get in touch with you afterwards. Well at least you know we'll send out a you know an answer kind of key to what was asked and we couldn't get to. So first one is the is the mem workflow custom made or is it a standard workflow available for everyone? Currently it's it can be available for people who have meme. So people who have meme software, Meme software, it's a it's a big software that we're using clinical routine for computing and image data analysis. But the workflow to do it, it's like an independent workflow but can be just for example, I can export it from meme and import in another meme, but it's not it's still user dependent in the sense that you have to be a meme meme user to be sure that you have access to it. And that was a big debate that we had at the level of how to give access to other people to this workflow without having the meme software. But it's it's just too difficult because you need to have a meme installation actually. And the part is the the nice part is that they are using that they currently are doing a better version of meme Everywhere, which is an online MEAN platform of this solutions of their solution. And if this comes available you can just send a link, you open the workflow and you do the test. But this is still under discussion with me. OK, cool. Thank you. People who don't have me in their clinic. Gotcha. Among the CT scanners that you included in your audit, which type produced the most pronounced motion artifacts? Don't know if you know the answer to that. Or not. Irregular frequency and the double the Max, the high amplitude, the high amplitude it was really. You can see that the the cities were not just handling the the huge amplitude as good as possible and they were always giving artifacts, so that big amplitude, that's also a question that was really like 2.5 based on literature review, which was like a one the maximum that we also can have in your clinic. And maybe for example there's also like on message. If you see that you have high amplitudes higher than 2cm up to 2.5, you should try to talk with your patient, to relax and to coach, because this is really quite high and you will have problems at the level of the system. Thank you. I think that was a great answer and I think it was very much appreciated probably by the audience. But I think this particular person was also asking just across the CT scanners, the different types of CT scanners. Was there anyone in particular? Not really, but this one the maximum maximum amplitude and also double amplitude and double double amplitude. It was really more a regular curve that we really wanted to test the limitations of the scanner. So we were not sure that this can happen in real life, because in real life means that your patient has to really, really like consistently breed two times high, two times low, two times high, two times low. So that's not really specific for the patient, but high amplitude of for 2.5 it's more realistic at the level of the clinic. Gotcha. I think there were a couple questions that kind of go along this line. The question here is were there any findings regarding frequency of 40 CTQAI think another person asked. You know, are there any recommendations that came out of this as to how often you should do 40C TQA? The recommendations we are not giving recommendations, but for example, you've got a lot of feedback that a lot of people don't do. The CPQR recommends to do for the city each time when you have a new implementation, each time you have a new calibration at your city and twice per year, so two times per year, you should just take your end to end quality assurance of the 4D. But we when we did the audit we realized that a lot of people don't do that just because first of all they don't have the Phantom or they have an in house phantom that they do 1D and that is not is not really enough to test your your entire workflow. Or I think the the vendors are providing a small phantom. The variant is providing a phantom for their LPM and testing and they use that small phantom to test if there if the pharmacy is working, but without doing any measurements. So the CPQR is really recommended to do that twice per year and definitely at the level of the commissioning of your city. When you have a new city and the new 4D solution, you have to properly Commission it. Thank you. Looks like maybe one more question here. Do you ever do a 4D CT with compression? Yes, we do 4D CT compression for all the pancreatic cases because we have an SBRT pancreatic clinical trial ongoing. So all the pancreas cases are done with compression belt and used for the city and it's that's why we also tried to mimic this solution because we still have small problems. The problem is not at the level of the 4D is that the problem at the air PM is difficult where to position at the air PM because the compression belt is quite big. So we don't really have a place. If you put it on the compression belt you don't have signal and if you put it around it you still have it doesn't stay because it's still the the, it's still a a box, a small box that you have to have a small block that we have to add. So that's why we try to do to mimic 2.5 millimeter, 2.5cm of 2.5mm of amplitude that is very, very low and this was and with a small. That was purely to mimic our pancreatic cases and that value was taken by looking at the last 1010 patients of over 7 patients of pancreas that we did. And So what was the signal at the level of the RGS? And we try to see if we can mimic next with the phantom. Thank you. I mean again thank you so much for putting this together, for sharing all this information. We are at the hour. There are a few more questions here. I apologize we couldn't get to them all. As I said, we will follow up with you with some answers and thank you everyone for joining us today. As I said, we recorded this. We'll be sending out a link to the recording and we hope to see you all again online or at one of the upcoming conferences sometime soon. Again, thank you everybody and have a great rest of your day. Thank you. _1709633628895
Dr. Manuela Burghelea of Institute Jules Bordet, Brussels, will discuss the use of a dynamic phantom to benchmark different 4DCT protocols across multiple institutions. The 4DCT QA program (QAMotion) that was developed proved easy to use and implement, while highlighting the need for a standardized 4DCT QA program.