Video
library

  • 771
Comments (0)
Login or Join to comment.
Transcript
Bogdan: Hello, everyone, and welcome to a new Novalis Circle Webinar. My name is Bogdan Valcu. I am the director of Novalis Circle. And today I have the pleasure to introduce topics pertaining to implementation, commissioning, and clinical versatility of ExacTrac Dynamic. It is my pleasure to introduce you to three speakers today representing different institutions that have implemented ExacTrac Dynamic over the last two years and are kind enough now to share their clinical experience. To begin, we have Dr. Philipp Freislederer from LMU Hospital in Munich, Germany. Philipp is a medical physicist and a scientific employee in the Department of Radiation Oncology. He is also one of our early adopters of ExacTrac Dynamic. In fact, LMU was the first Elekta installation for ExacTrac Dynamic. And together with LMU, we have various ongoing clinical projects that Philipp will address today.

The second speaker is Dr. Thomas Kole from Valley Mount Sinai Comprehensive Cancer Care in Paramus, New Jersey. Dr. Kohl is a clinical assistant professor in radiation oncology at the School of Medicine at Mount Sinai. He is also an attending at Valley Medical in New Jersey. In his lecture, Dr. Kole will review his rationale for adding ExacTrac Dynamic to his extracranial program, as well as review the versatility of the system for various anatomical applications. To conclude, we have Scott Jones from Highlands Oncology in Arkansas, who will address their implementation and clinical experience with ExacTrac Dynamic. Scott is a director of medical physics there and the president of the Arkansas Radiation Physics Consultants. And the Highlands Oncology Group actually represents the first installation in United States for ExacTrac Dynamic, so I'm sure Scott has some interesting details to share with us today.

As true with all our webinars, we are providing CE credits if you are in need of CAMPEP, MDCB, ASRT credits upon successful completion of this webinar. Please complete the test and e-mail us at info@novaliscircle.org should you have any questions. As always, please remember to utilize Google Chrome or Safari to access the webinar, and should you have any connectivity issues, simply refresh the page. Utilize our chat interface to send us questions. We will answer your questions upon completion of the three lectures. Check the polling session for questions that we might like to ask you. And if you'd like to follow us on social media, please do so by utilizing the hashtag, #novaliscircle. With that I'd like to turn it over to Dr. Freislederer for the first lecture.

Dr. Freislederer: Hello, everybody. My name is Philipp Freislederer from LMU University Hospital, the Department of Radiation Oncology. Thank you for inviting me to this talk. And, today, I would like to give you some insights about the clinical perspective for Brainlab's ExacTrac Dynamic and also for some of the commissioning and routine QA parts. A short introduction of our department here in Munich, we are treating roughly 2,400 patients a year and are equipped mostly with Elekta Linac. And onto Elekta Versa HD Linac, we installed two different Brainlab ExacTrac Dynamic systems starting last year in roughly May I would think.

So, what is ExacTrac Dynamic exactly? This is a combination of an optical surface scanner with optical structured light for patient positioning. In addition to that, you have a thermal camera which has one dedicated use, and that is to give the optical surface or the optical surface information some additional information for image registration. So it serves as sort of a fourth dimension for registration, which also makes multiple in-room optical cameras unnecessary. In addition to that, you have X-ray positioning and intrafractional monitoring system with stereoscopic X-ray sources and detectors. So the whole system consists of X-ray positioning and monitoring with an additional real-time tracking workflow using surface-guided radiation therapy.

Why would you need the thermal information? If you would look at this image, this is the image of a patient, or rather typical patient abdomen, and it is quite flat and, therefore, any registration algorithm would have some problems of getting accurate results especially in the longitudinal direction. Now, there is thermal data available due to the body heat. So, if you add this information to it, what the registration algorithm would see would be an image roughly like this. So, there are a lot of peaks and valleys onto which the algorithm can grab and really hold on to so you have a unique topography of your patient surface, which increases the accuracy potentially dramatically.

And so, a little guide through the workflow of ExacTrac Dynamic. The patient is roughly pre-positioned using the surface only. The thermal information is not added at this point. So, what you see here is the patient after it has been pre-positioned to a roughly good direction, which makes it easier for the X-ray registration algorithm to get accurate results. And here you can already see some breathing motion of the patient, and this was afterwards after the prepositioning, where you would select an area of interest, which you would like to track during treatment intrafractionally. After the selection of your area of interest, you're acquiring stereoscopic X-ray images, and this is actually the fusion in six dimensions in real time. You have some quite nice tools for registration or to check and verify if the 6D registration is done correctly between the X-rays and the DRRs from the treatment planning system. This registration result is then used as your reference. So, you always reference to the internal anatomy, and this is used also to create a surface reference later on for intrafractional monitoring.

After your X-rays, you would then go into an intrafractional monitoring mode. You see the area of interest, the region of interest to be tracked on the right part of the image. So, this would be the opening of an open face thermoplastic mask and how it looks like during treatment. And this is also acquired in real-time. You see the gantry rotating in the back and you see some values from some displacement for your surface. And during certain gantry positions, you shoot your stereoscopic X-ray images to have the highest possible surveillance during the fraction. And every time you shoot your stereoscopic X-rays, the surface is then again set to zero until a new reference is created for the surface and thermal imaging. And during intrafractional monitoring, it's not only the surface which is being tracked, but also the thermal information is being used.

This is an image of what it would look like if the surface would detect any out of tolerance values. So this would be indicated in red, and your beam would be held automatically. And also you could either shoot X-rays automatically or move the gantry to a position where you can acquire stereoscopic X-rays, so 0, 90, 180, or 270 degrees, always plus-minus 10 degrees. So you can shoot stereoscopic X-rays and then reposition the patient using your six-dimensional couch during the fraction. Of course, couch rotations are non-coplanar. Treatment angles are quite of importance during stereotactic treatments in the head. This was some X-ray images of a non-coplanar couch angle. And this would be what the monitoring workflow would look like during a non-coplanar treatment beam.

So when is it always good to see when which type of information is actually used for the patient for setup and during treatment? So during setup only, the 3D surface is used for a rough prepositioning of the patient. And if I say rough, this is also in a...rough would be still...as far as we experienced this in the range of 1 to 2 millimeters for the isocenter. So this is done only with the 3D surface, and at this point, no thermal information is added. After you were to correct using the internal anatomy, your X-rays, and you define your area of interest to track with surface and thermal information. During treatment, this area of interest is used for intrafractional monitoring in four dimensions, 3D surface plus the thermal imaging. And every time a new X-ray is acquired, the surface is updated.

A short walkthrough our routine QA with the system, so just you would roughly doing, every day we perform, obviously, a daily check. This takes five minutes in total with a dedicated phantom from Brainlab. There are two things to be checked, one is the deviation between the surface camera or the surface thermal camera, and the X-ray positioning systems, so a so-called consistency check just in case there will be some movements of the camera or of the flat panels. And also you check if there's a deviation from your calibrated radiation isocenter later on. Once a month, we need to calibrate the thermal camera to the surface camera, which takes also about roughly five minutes. And this is done also with a dedicated phantom on the right bottom image. We recommend to calibrate the ExacTrac system to your radiation so isocenter or your Linac radiation isocenter center monthly. This is done by placing a ball bearing, any type of ball bearing, either a Winston-Lutz pointer or an anthropomorphic phantom in the radiation isocenter with the Linac beam and afterwards performing a Winston-Lutz test. This radiation isocenter is then checked. This is also called SRS check weekly because we are hopefully always on the day where we actually treat the SRS patients. And there's also an anthropomorphic phantom placed to the isocenter using ExacTrac Dynamic, or the X-rays in this case. And then, you check for your regular Winston-Lutz test for 2 by 2-centimeter fields to see the position of the ball bearing compared to your radiation isocenter.

At first, when we started treating with ExacTrac Dynamic, we shifted from cone-beam CT workflows to ExacTrac Dynamic-only workflows. So we wanted to see if there's a big deviation between those two different systems, are they both calibrated to the same center? Do we get the same results when positioning a phantom or then later a patient with the two modalities? So we took two anthropomorphic phantoms and had six randomized standard locations. And we set up the phantoms according to ExacTrac Dynamic and to cone-beam CT. And we verified the position then afterwards with both modalities. What we have seen is that the deviation between the two IGRT modalities is quite low for the initial correction, even lower for the post corrections of the verification image afterwards. So we were quite sure that we could position the patients with ExacTrac Dynamic quite good sub-millimetric metric without less than one degrees of rotation. And then, we chose to treat the first patient even with an ExacTrac Dynamic only workflow and no cone-beam CT and haven't changed this workflow ever since.

From a physics perspective, what was also interesting is, are there some deviations between cold phantoms? Do we need warm phantoms? And what is the influence of the thermal imaging at all? So we had a dedicated phantom from RTsafe filled with warm waters. It's a head phantom with bone equivalent material. We filled it with warm water and we compared the accuracy of the surface thermal imaging, this combination to the X-ray positioning from the ExacTrac system. And we did that at zero degrees and also at some non-coplanar couch angles and also compared to an empty or filled with cold water phantom. The results look quite promising as it is obviously a phantom and we didn't expect too much of movement. We saw that the deviations between the surface thermal imaging and the X-ray imaging for the phantom, everything below 0.1 millimeter. So, the system is quite good in itself. And also the surface is also very useful for the detection of any displacements later on.

Now, to the more clinical part of this, if we look at...we are not only treating multiple isocenter SRS patients. We're also using a lot of single isocenter SRS treatments, so multiple brain metastases with 1 isocenter up to 15 metastases and 1 treatment session. And if you treat those multiple brain metastases patients with one isocenter, you're always very afraid that you have some rotations that you cannot really account for because if you have too much of rotations and also translations rotations, they're quite the bigger deal. The farther those multiple GTVs are apart from each other, the further they are away from the radiation isocenter or your treatment isocenter. And also, the higher the rotation is, the more likely is a risk of compromised target coverage. So, we need to verify the position of the patient, or of the GTVs, or of the patient itself quite thoroughly.

What would happen if we would not do this? This is a single patient study we did with the old ExacTrac, where we wanted to look at the difference, how would the patient be treated if we would not correct for non-coplanar couch angles. And we took the values which we need for correction, and we did some dose calculations for the single patients. And we saw that we had those coverage difference of up to 5 gray in this one metastasis. If you look at the DVHs, also especially for the coverage, you have quite difference if you would not correct for non-coplanar couch angles. And we have quite a lot of treatment beams which are not coplanar, especially for this multiple brain metastases cases. We also have seen that according to our institutional standards, we had some tolerances. I think it's now 0.5 millimeters and 0.5 degrees as upon 0.5 millimeters translation, 0.5 degrees rotation. If something is...oh, increases that threshold, we would need a couch adjustment. And we saw for a number of 20 patients and median number of 6 couch angles per patient, 23% of those couch angles needed couch adjustments for them after our non-coplanar image guidance workflow.

So there is definitely a need to adjust the couch if you're treating with non-coplanar treatment angles. This is what a typical plan would look like for a multiple brain metastases patients. We also, with ExacTrac Dynamic, are able to show some preliminary results for the difference between the surface of those patients inside a mask and the X-rays. And so, we acquired for 10 patients 76 fractions. We acquired the difference between the surface in thermal or combined surface thermal imaging and the X-ray imaging. And what we saw for this relatively small patient cohort that there was very good correlation between X-ray and surface thermal imaging, and the difference between those modalities was below 0.5 degrees rotations and then a sub-millimeter range for translations. But obviously, we had some very small deviations to be actually corrected for or monitored. So this study will continue with more patient data hopefully.

The other case where we're using ExacTrac Dynamic quite freely is spine SRS or spine SBRT, however you would like to call it. And as you can see here is some example image of a treatment plan of a dose distribution. We have a very steep dose gradient to the spinal cord. This is a workflow of what we're doing, how we're treating these patients with ExacTrac Dynamic with quite high margins for the patient surface as there is a lot of breathing motion. But we still want to have the surface to account for a larger translation during the beam. And additionally, we have very tight margins for X-ray surveillance, 0.7 millimeter, 0.5 degree. And we are using the X-ray surveillance with the highest frame weight, frame rate possible, so for dual images per arc. And the patients are placed on the couch without any vacuum mattress, without any fixation, and usually with the arms up. If you look at our patient cohort, we also analyzed 9 patients, 34 fractions of spine SBRT with those tolerance settings. And we've seen that roughly 14% showed at least one of those tolerances or threshold levels out of tolerance, and the patient needs to be repositioned during treatment with also quite high maximum deviations calculated from ExacTrac Dynamic. So we are quite confident to treat these patients with a high surveillance and account for those rotations and translations.

To conclude this, so far, we roughly treated 160, 170 patients with ExacTrac Dynamic. We saw that for phantoms and also for the first patients, the system offers a patient setup with sub-millimetric range and sub-degree accuracy and precision, especially for non-coplanar treatment angles and also continuously during those delivery with a comparable small amount of imaging dose compared to multiple cone beam CTs I would say. We have very good correlation between those internal two modalities, so X-rays and the combination of surface thermal imaging, and there's the potential to reduce the X-ray frequency for certain indications. And this is also what we will do in the future to trigger an X-ray when the surface exceeds a certain threshold and not to trigger an X-ray at every dedicated 90-degree angle. As I said before, we had 20% of the misalignments and brain SRS and 13% of the spine SRS were over our institutional tolerance levels and the patient repositioning. So, in total, we can say that we can reduce the risk of compromised PTV coverage or [inaudible 00:23:01.623] coverage with ExacTrac Dynamic and is a very safe treatment application for targets close to sensible organs at risk.

In the future, there will be some novel workflows. One, a very exciting workflow with the next software version is a surface-only workflow, and secondly, deep inspiration breath-hold will be also, from what I heard, will be released quite soon. We also will have a patient study plan for this workflow here at LMU Munich. And to sum it up again, the ExacTrac Dynamic offers patient prepositioning with 3D structured light only. This improves X-ray registration, and it also accounts for local deviations such as the arms, the chins, rotations, etc. Secondly, you have X-ray positioning and verification during the beam. And the more important part I think is before each new couch angle. So this accounts for your treatment couch runout and also for the potential intrafractional motion, which also happens inside any thermoplastic mask. The third advantage I would say is that we will soon have a workflow change from the strict intrafractional X-ray monitoring with dual X-rays to a more SGRT-only workflow with less imaging dose, so you only trigger X-rays when your surface is out of tolerance. This would exclude some special treatments such as those spine SRS I said before and also some high-dose rate treatments with flattening filter-free. So thank you very much for your attention. And I am very happy to answer all of your questions.

Bogdan: Thank you, Philipp, for the great review of your ExacTrac Dynamic program. And we'll switch now to Dr. Kole for the subsequent lecture.

Dr. Kole: Thank you for giving me the opportunity to speak today. I'm going to speak to you about the clinical versatility of ExacTrac Dynamic. These are my disclosures. So at Valley-Mount Sinai Radiation Oncology, we have 2 Varian TrueBeam linear accelerators, both embedded with the 6 degrees of freedom couches, Variant Realtime Imaging, and Variant RPM gating. We also have a tomotherapy unit, and a Gamma Knife Icon system. So we use our Icon system for the majority of our SRS. So we are looking to ExacTrac Dynamic not to enhance our SRS capabilities, but we're really in the market at the time for a surface guidance system. After doing a bunch of research, we felt that the ExacTrac Dynamic system was going to be the most versatile for our department. And also, outside of just the surface guidance, was going to be able to enhance almost everything we were doing on our linear accelerators. So, we ultimately decided to upgrade both of our machines with the ExacTrac Dynamic system in March and April of this year. And Brainlab was really great to minimize our machine downtime. We've performed the upgrade sequentially, with any machine only being down for a week at a time.

Once we were up and running, we very quickly were trained and then began treating prostate, breast, brain, and spine cases using the new system. Very quickly, after seeing the versatility of the system, we expanded to other disease sites, which is head, neck, lung, upper and lower GI, extremity sarcoma, and GYN. And as of today, about 70% of our treatments are currently using some form of ExacTrac Dynamic for all of our linear accelerator treatments. Once we got the system, our physics really did the heavy lifting of putting together workflows for every disease site based upon the primary ExacTrac Dynamic workflows. So the primary workflows within ExacTrac Dynamic includes a standard where you have a surface-based repositioning followed by anatomic alignment using the ExacTrac Dynamic stereoscopic X-ray alignment to the planning DRR. Then for more soft tissue aligned sites, we have the CBCT workflow where we're doing, again, a surface-based prepositioning, followed by a soft tissue CBCT alignment. Once the CBCT iso is set, ExacTrac Dynamic takes over, along with the stereoscopic X-ray reference, and then subsequent verifications performed to that reference. Then, we have the implanted marker workflow. We're using that for the majority of our prostate cases. There, again, we have a surface-based repositioning, and then we go into an ExacTrac Dynamic stereoscopic X-ray fiducial match to the planning DRR.

What we did is we went through for every disease site, and we used different combinations of these workflows, along with different imaging parameters to create separate workflows for every site. And so what you're looking at here is sort of the work that our physics team put together, along with our clinicians. And we have the prostate and pelvis workflows here, including our SBRT and conventional prostate treatments. We have cranial workflows that include the SBRT brain treatments performed on our linear accelerators, along with our conventional brain and head, neck treatments. We then also have our breast, lung, spine and sort of miscellaneous workflow, which we use for the remaining disease sites.

Once we put all these workflows together, we then started building tables of tolerances for our ExacTrac Dynamic templates. And here, you see, this is sort of a living document that we have in our department. It contains all of the six-dimensional tolerances for X-ray shifts, as well as surface tracking for each of our disease-specific workflows. This document is always changing, and we'd be happy to share it with any site that is interested. So I'm going to take you through some of the setups that we've been using now for different disease sites and sort of what our experience has been like with the ExacTrac Dynamic. For all of our patients, we're using 6-degree of freedom couch. All patients receive non-coplanar X-ray imaging for all the ExacTrac treatments. For the majority of our volumetric arc therapy patients, they're going to be having periodic stereoscopic X-ray verifications and the function of cardinal angles. And then for our static fields, we have MU-triggered X-ray imaging.

First, so we actually were doing with our conventional brain cases. So this is a 74-year-old gentleman with a glioblastoma. He had a gross total resection, which has been followed by adjuvant chemo radiation with Temozolomide. He was prescribed 60 gray in 30 fractions, which was planned using RapidArc IMRT with 2 arcs and standard thermoplastic immobilization. Two-millimeter PTV margins were used. Given the cranial nature of the disease, we are using the standard workflow to align to cranial bony anatomy, so the patient is first repositioned to the surface. We perform a CBCT on the first day of these triggers just for volume verification, especially with glioblastoma. We can find times where there's been some change in volume from the time of simulation to the time of treatments, so we always verify with a CBCT on the first day. The ExacTrac Dynamic then takes over, where we perform stereoscopic X-ray imaging. You could see that the red where the therapists have excluded portions of the cranial anatomy from alignment. The images are fused to the treatment plan in DRRs, and then a subsequent stereoscopic X-ray verification is performed. During the course of the arc, we are monitoring to surface for motion, and then a stereoscopic X-ray verification is performed prior to the next arc. We also repeat a CBCT every fifth treatment, again just for volume verification.

For our head and neck patients, we're looking more for soft tissue alignment. So they're using a CBCT based workflow. This patient happened to be a 95-year-old gentleman with a locally advanced squamous cell carcinoma of the right tonsil. He was not a candidate for combined modality therapy given his age and medical comorbidities, so he had to be treated with RT alone. He was getting some mild hypofractionation given the fact that he was unable to get chemotherapy. She was prescribed 66 gray in 30 fractions to sites of gross disease while simultaneously delivering 54 gray in 30 fractions to some critical sites, again, planned using RapidArc IMRT with two arcs, standard thermoplastic immobilization. We also used a vaclok up to the shoulders in patients to help reproduce the cell daily. Again, 2-millimeter PTV margins. So in this patient, after the surface repositioning, the CBCT is performed, the therapist then perform a soft tissue alignment. Once the CBCT isocenter is set, we then enable ExacTrac Dynamic for an X-ray reference. Then, we go into surface monitoring during the course of the arc. Prior to the next arc, a verification set of stereoscopic X-ray is performed within a given surface tracking through the remainder of the treatment.

For our breast treatments, we've been using ExacTrac Dynamic now for right whole breast. This is a patient who's a 58-year-old female with the right stage IIA breast cancer. She had a partial mastectomy with negative margins and was then recommended to proceed with adjuvant full breast RT using a moderately hypofractionated technique, opposed tangents with field in field. So for this case, our target is the whole breast and, therefore, lining up to the chest wall and using the standard-based workflow. So the patient is prepositioned using the surface. Currently, we are still using AP and lateral simulation tattoos, however, we anticipate in the future to be discontinuing that as we spend more time using the system. We then indicate on the patient surface the area for which you'd like to monitor the surface during the course of the treatment. And then ExacTrac Dynamic uses the stereoscopic X-ray system to align the patient using six degrees of freedom to the treatment planning DRR. Once that alignment takes place, we perform an MV portal verification on the first day and then every fifth treatment. During the course of treatment, the patient is monitoring with surface tracking, and then periodic stereoscopic X-rays as a functional monitor unit, as well as prior to each new field.

For our lung treatments, again, we're now moving to again soft tissue alignments. We're using a CBCT based workflow. This patient here is an 83-year-old male with a medically unresectable stage IIB non-small cell lung cancer by virtue of a satellite nodule adjacent to his primary. The patient has very, very poor pulmonary function. Given the location of disease, the volume that needed to be encompassed, it was recommended that he have a modified stereotactic technique. He was simulating using a 4DCT, with various stereotactic fixation and abdominal compression and was ultimately prescribed 60 gray in 10 fractions again using RapidArc IMRT with two arcs. So this patient, again, after surface prepositioning, we enacted CBCT workflow. After a CBCT alignment was performed, a CBCT isocenter was set. ExacTrac Dynamic then performed a stereoscopic X-ray reference within surface monitoring throughout the course of the arc. The patient then had a stereoscopic X-ray verification performed prior to the next arc with again surface monitoring throughout the course of treatment. By having this stereoscopic X-ray referenced before each arc, as well as surface monitoring, it has given us a lot of confidence to decrease our margins from ITV to PTV.

For upper GI cancers, we're seeing also a tremendous utility with ExacTrac Dynamic. This happens to be a patient with an intrahepatic cholangiocarcinoma. She had a resection performed and found to have positive lymph nodes. She completed adjuvant chemotherapy and then went on to receive adjuvant chemoradiation. Then, because of the location, she had a 4DCT simulation through respiratory motion and was ultimately prescribed 50.4 gray in 28 fractions, which was delivered using RapidArc IMRT, and this time four arcs. Given the soft tissue alignment, we went with a CBCT workflow. Also, a CBCT isocenter was set. Stereoscopic X-ray referencing was performed, and then surface monitoring throughout the course of the arc. X-ray verification was performed prior to the next arc, again, with surface monitoring throughout the arc. What we find is that we've become very confident now in our setups for these patients because not only are we monitoring during the course of their treatment, of the surface, but we're able to verify their bony alignment in between each arc and make any small adjustments that we need to. So that really helps with these multi arc treatments, with keeping these patients aligned and not significantly increasing the treatment time.

Now, we get to the best part, which I saved for last, which is prostate. As you are all well aware, conventional treatment for prostate are becoming increasingly rare these days. 1.8 to 2 gray refraction treatments probably come close less than 5% of what we do for our prostate population here at Valley-Mount Sinai. And so as moderately hypofractionated stereotactic body radiotherapy treatments become increasingly prevalent, we really need to be taking into account intrafraction across the motion as the dose refraction is increasing. In the past, we were using the varian real-time imaging system, where we were doing fiducial monitoring using real-time 2D kV imaging. And what you're seeing here is some data from a manuscript that's in preparation, where we're looking at the fiducial excursions during our prostate SBRT treatments. And what you see is about 20% of our treatments have excursions of 3 millimeters of greater throughout the fraction of SBRT that's been delivered. So it's imperative that when we're delivering these high-dose refraction treatments, we account for that positive intrafraction motion.

So what we're doing now with the ExacTrac Dynamic system is using the implanted marker workflow. And then we'll use additional workflows as necessary when we need to incorporate things like, of the fiducial coverage. We always perform a CBCT prior to the start of treatment, and this is mostly for anatomic verification. We want to verify [inaudible 00:38:54.024], but also we do an approximate CBCT fiducial alignment, which helps with the automatic marker detection for the ExacTrac Dynamic. The other thing the CBCT does for us is it tells us whether or not we should receive treatment the way it is. We've come up with some guidance for our therapists during our prostate treatments to determine whether or not the patient needs to be really taken off the table, have their rectum emptied or their bladder filled more, and sometimes they just need to be reset up. And we use this by looking at the pitch required to approximately get the fiducial aligned and CBCT. And if we find that we're having problems over consecutive treatments, then maybe that's a patient that needs to be re-simulated.

So for our prostate SBRT treatments, again, we're using an implanted marker workflow. This happened to be a 60-year-old gentleman with intermediate-risk prostate cancer, who wish to proceed with prostate SBRT. He had stranded gold fiducials placed, along with SpaceOAR hydrogel prior to simulation with CT and MRI. And he was then prescribed 36.25 gray in 5 fractions, which is usually delivered with two arcs of RapidArc IMRT and 6x flattening filter free mode. There's always stereotactic body fixation and 3-millimeter PTV margins. So we perform an initial CBCT if one wants check for anatomical variations in the bladder and the rectum. We also then perform rough alignments of fiducials. From there, ExacTrac Dynamic takes over, where we perform stereoscopic X-rays and auto fiducial matches performed and then aligned to the treatment planning deal. Throughout the course of the arc, we are monitoring the surface, as well as performing stereoscopic verification X-rays at each cardinal angle of the arc, with subsequent adjustments made depending on tolerances.

Some of the more complicated treatments that we've been doing since getting ExacTrac Dynamic is actually incorporating pelvic nodal coverage into some of our moderately hypofractionated prostate treatments. So in moderate hypofractionation at our institution, we routinely prescribe 70 gray and fractions to the prostate and certain patients with high-risk disease where we want to cover prophylactically the pelvic lymph nodes. We also want to deliver 50.4 gray in 28 fractions to the pelvic lymph nodes. The problem with this is that we're treating to different volumes that really have two different references. If we line up to the prostate, there's a good chance that we'll be off our pelvic lymph nodes. So one solution is, okay, just increase the margin around the pelvic lymph nodes. The problem with that is that then you're increasing the volume of normal tissue that you're treating. So the solution that we came up with was to deliver a daily composite plan, one where we're treating a volume that aligns to the pelvic bony anatomy and encompasses all of the tissue including the pelvic lymph nodes, the prostate and the seminal vesicles with a margin that would allow for a positive motion, up to a total dose of 180 centigray. We then deliver the prostate boost of 70 centigray, which is aligned to the prostate fiducial. So this allows us to sort of cover everything we want with minimizing dose to the adjacent normal structures like the rectum and bladder.

So this is a case of a patient with Gleason 9 high-risk prostate cancer. He was actually treated on NRG protocol GU009. And we treated him using the exact same approach that I just described. What we do with these patients is actually treat the prostate and seminal vesicle in boost first. We perform a CBCT, again, for the anatomic verification and then rough fiducial alignment. We then go into the ExacTrac Dynamic implanted marker workflow and deliver the first 70 centigray of treatment. We're monitoring the surface during the course of treatment and then an X-ray verification is performed every 180 degrees. Once that plan has been delivered, we then move into the remaining portion of the treatment which encompasses all sites of disease, the pelvic lymph nodes, the remaining distal prostate and seminal vesicles with larger margin. In this view, then we use the standard workflow. So the ExacTrac Dynamic stereoscopic X-rays are now aligning to the treatment planning DRR pelvic bones. We perform X-ray verification at the start of each arc, and we monitor the surface throughout the course of treatment.

So, in summary, we have found ExacTrac Dynamic to be extremely versatile, and we have been applying it to nearly all of our EBRT treatments at this point. We're very confident with the accuracy and precision now that we have been provided by ExacTrac Dynamic. We're able to decrease PTV margins, very confidently in certain situations. And like I said, we're currently using ExacTrac Dynamic in some shape or form with 70% of our EBRT treatments at this point. One of the important things and what we found is that overall treatment times have not significantly increased with ExacTrac Dynamic, and surprisingly, we actually...and really not so surprisingly, our prostate SBRT treatment times have actually decreased. And this is really due to the fact that we're now, in real-time, able to verify the fiducial position and make corrections without having to reset the gantry or perform additional CBCT. So this has been really beneficial for our patients. That's all I have for today. Thank you.

Bogdan: Thank you for the insights into your clinical program, Dr. Kole. I will now switch to Scott Jones for the last talk.

Dr. Jones: Good afternoon, everyone. My name is Scott Jones. I am the director of medical physics at Highlands Oncology Group in Fayetteville, Arkansas. I'll give you a little background about myself. I'm a 1991 graduate from MD Anderson Cancer Center. My research there was based on using second generation portal imaging images to develop automated routines for positioning a patient. I have been the director of medical physics at Highlands Oncology Groups since 2004 and in clinical practice since 1991. The only disclosure that I have is that I have been provided a speaker's fee from Brainlab. I'll give you a little background of our clinic, our group. Highlands Oncology Group is a physician-owned, freestanding, multidisciplinary cancer center located in Northwest Arkansas, who provide multiple services including medical oncology and hematology, radiation oncology, surgical oncology, as well as imaging services and multiple ancillary patients support service. We have three clinic locations based in Fayetteville, Springdale, and Rogers, Arkansas. A little bit about the size of our clinic, we average approximately 1,400 new patients starts per year, with an average daily external beam load of 85 patients. And that is spread between three Varian accelerators: a Clinac iX at our Rogers facility, and a Trilogy and TrueBeam at our Fayetteville facility. And approximately 30% of our patient population consists of SRS, SRT, or SBRT type treatments.

Oh, we were the first clinical install in the U.S. And I've had several people ask me why did we decide to be the guinea pigs. And really, this is the relationship based on trust that we have built with Brainlab over the years. We have a 14-year history of experience with Brainlab based on initially our fixed frame-based SRS treatment system that was incorporating Brainscan and iPlan. We moved to a frameless based treatment platform in conjunction with iPlan and the installation of an ExacTrac system on our trilogy in 2015. We also added Elements Treatment Planning in mid 2015, and then eventually, to expand our SRS program, installed ExacTrac Dynamic on TrueBeam in November of 2020. So, as far as the addition of the ExacTrac, as I mentioned, it was really added to expand our SRS program and to also allow for motion management, which we do not currently have available on our Trilogy unit. So the installation of the ExacTrac Dynamic began in November of 2020 and completed early December. The installation was performed over the course of a few weekends. And that was really important because it minimized our downtime. The training began immediately upon acceptance. And our first use for clinical treatment was December of 2020, and that was an intracranial SRS.

So here are two systems. Here's the Trilogy with ExacTrack 6.0. I think the main things to notice here are the in-floor installation of the X-ray tubes, the Brainlab 6D couch, as well as the infrared positioning camera. When contrast with ExacTrac Dynamic on our TrueBeam system, I think the things to notice of this image are the above floor installation of the X-ray tubes, the larger X-ray detectors which allow for a larger field of view, as well as the perfect pitch out from Varian, which integrates seamlessly with ExacTrac Dynamic. Here is a closer-up picture of the 3D/thermal imaging camera system for ExacTrac Dynamic versus the infrared system for ExacTrac 6.0.

What are the differences? Really the main differences are based on the differences in the camera system. So the ExacTrac 6.0 utilizes an infrared camera system. I think most people here are probably very familiar with this system. ExacTrac dynamic utilizes a 3D camera to detect patient surface. And that's coupled with a thermal camera system and improved X-ray system. As I mentioned, it's a beefier tube and a larger field of view X-ray image. And this allows for patient positioning and monitoring during treatment. That's the main thing that surface detection and thermal imaging systems add is the ability to track patient external surface during. Both systems allow for positioning in 6D and imaging prior to and during non-coplanar treatments, which I think is one of the big advantages of ExacTrac is just the fact that we can image the patient at a non-coplanar angles, which allows us to not have to move the couch back to the initial treatment position to verify.

What does the integrated surface monitoring for ExacTrac dynamic add? Really, one of the most useful tools that we find is the initial positioning utilizing the 3D camera. This allows us to position, say, like the hips need to shift to the right or the shoulders need to shift to the left. We're able to see that and align the patient's external contour to the predicted external contour based on the original treatment planning CT. Once the patient has been positioned based on X-ray imaging, we outline a thermal image to monitor during treatment. So we define a region of interest that we want to track during treatment and verify that the patient has not moved from the original setup. So how did we configure ExacTrac Dynamic for us and how do we verify its positioning early? Basically, we determine radiation isocenter using a Winston-Lutz test. We then enter ExacTrac Dynamic calibration mode, which is a very nice system. It provides pre-configured routines for calibration for both...well, for all of the systems, for the thermal 3D camera, also for X-ray correction images. That would be your dark fields and your flood fields. It also allows for image isocenter calibration, as well as radiation isocenter.

So this is just showing the calibration of the radiation isocenter once the radiopaque marker has been aligned with the true treatment isocenter from the radiation bank. We can also then take orthogonal OBI-based AV images. We can see a real quick check, just a visual check of the alignment of the pointer with the OBI system. Obviously, you can use MTXD to determine actual room. Here's a CBCT of the same setup. So we're able to verify the congruence of all of these imaging systems with the radiation isocenter. This is the phantom that is used for calibrating the ExacTrac Dynamic imager. For those of you who have used ExacTrac before, it's very similar system. It has predetermined markers. And there is a pre-configured software routine that will recognize these markers and calibrate isocenter based on this.

As far as calibrating the thermal image camera, this system is a thermal mat that we plug in and allow to warm up to our uniform key pattern across the entire phantom. There are pre-configured metal discs that allow us to reference the thermal camera back to the 3D camera. Another nice feature to be able to just quickly determine once you've done your Winston-Lutz for the day, if I want to determine if my ExacTrac calibration relative to that isocenter is still good, then I can use this handy little tool, the sphere that will give us an offset from isocenter. We also performed end-to-end testing, the scanned head, thorax and pelvis phantoms, the right plans for each site, the phantoms were placed purposely misaligned by known amounts, and we used ExacTrac Dynamic to line the phantoms, and then alignment was verified utilizing CBCT.

So that brings us to the clinical aspect of the use of this device. The clinical workflows that we have utilized so far are intercranial SRS and SRT using the standard ExacTrac Dynamic positioning and monitoring workflow. We've also developed a left breast DIBH workflow that is very, very similar to other surface monitoring systems. We also have performed a few cases of marker-based ExacTrac Dynamic positioning and monitoring with CBCT verification. And we also utilize the CBCT workflow to monitor SBRT treatments for long in treatment. So the standard positioning and monitoring workflow is that we acquire 3D camera images. We position the patient based on those images. We acquired X-ray images. Once we've aligned the patient based on those X-ray images, we define a region of interest for monitoring on a thermal image. We send the shifts and position the patient, verify those shifts with post adjustment images, and then start the treatment and evaluate the surface motion during delivery.

So how does the system work clinically? Here's just a quick example of the pre-positioning utilizing the 3D camera. So this is all alignment based on just external contour. And this is aligning to the predicted patient surface based on the treatment planning CT. Once we've aligned the patient here with the external surface and we acquire a thermal image, we outline that image with a region of interest that we would like to track during treatment. And the things to note in this image are the graph at the bottom of the page. The green graph represents the motion of the patient surface utilizing the thermal camera. The 100% is the threshold that is established for this particular treatment technique. And these are pre-configurable per case. The two blue dots in the middle of this screen with the line running down represent where in the arc X-rays will be acquired to verify internal anatomy positioning as well. So you can see over on the far left bottom corner, there is an arc represented for this treatment. And at the principle angle of 270 degrees, this verification set of X-rays is taken and we can verify the patient position based on that as well.

This is just an example of acquiring and analyzing the fusion between the predicted BRR and the acquired X-ray. So, once we establish this thermal ROI, and we've taken our initial positioning X-rays, we verified the positioning, then we use the thermal system to monitor the patient treatment. And you can see that, again, in this same slide that I showed earlier, thermal position surface is monitored throughout the entire course of treatment. And I'll demonstrate that a little more later. Non-coplanar imaging is accomplished in the same method. There's literally no change. The only change there is that we will define a new thermal ROI per couch kit. But other than that the whole entire imaging process...I mean, that's the great advantage of ExacTrac because it's a room-based system. We're not dependent upon couch position to be able to image.

So, that brings us to our DIBH workflow. And as I mentioned previously, at the moment, I know Brainlab has a dedicated workflow that hopefully will be released in the next few months. But the workflow that we're utilizing at the moment is very similar to other surface monitoring systems. The process that we go through is just to obtain a free-breathing and deep breath-hold CTs. We do plan based on the DIBH CT. We then export that plan to ExacTrac Dynamic. We position the patient as closely as possible using the position the patient surface positioning in-room prior to imaging. We do use a deep breath-hold during this positioning process, mainly so that we can identify pitch errors if we need to move the patient down on the breast forward or up on the breast board to minimize the pitch adjustment. We then define a thermal ROI for surface tracking. We acquire X-rays for positioning. We take single exposure MV images at our treatment angles. And if any positioning adjustments need to be made, then we will require an ExacTrac image for baseline.

So here's a quick example of DIBH workflow. This is for a single field. At the moment, you can hear the...you might be able to hear the room door is closing. So at this point, the patient is free breathing. And so you see that their position tolerance, their surface position tolerances will come in and out of tolerance. Right now, what we're waiting for is for the beam to be ready. We'll then ask the patient to take in a deep breath. And you will see the tolerances all come within acceptable limits. The display of the thermal movement is displayed at the bottom screen. And there's a complete treatment for one field of the DIVH. So fairly simple process, as I said, very similar to other systems that are surface monitoring systems only.

We also have developed a marker math workflow...and I shouldn't say we developed, we've utilized the marker match workflow that positions the patient based on internal markers. In this case, this is a prostate case with three markers. You can see the image on the left. The markers are not aligned. Then once we perform the fusion and the shifts move out, reimaged. The image on the right represents that. And then we verify that with CBCT. You can see that this is a simple and straightforward process. The final process or workflow that I'll discuss is the CBCT workflow. And really, that is a very simple system. It is utilizing the OBI platform to position the patient. And then we acquire baseline X-ray images with ExacTrac. It opens on some stable internal anatomy for use and tracking movement during treatment. And so in other words, we're not looking for things that are moving with breathe, breathing. We're looking for some kind of stable anatomy to track to make sure that the patient has not moved. We use ITVs here. So this is not gated treatment. This is just to ensure that the patient has not moved during the treatment. Again, we established a thermal ROI for tracking the skin surface. And we begin treatment and monitor the patient during the treatment process. And we also obtain X-rays during these because most of these are beam-based.

So, that's our experience so far with ExacTrac Dynamic. I appreciate your attention today. And I want to thank Brainlab for inviting me here to share our experience. And I really want to send out a special thank you to my co-workers for giving me the time to be able to present. Thank you.

Bogdan: Great summary, Scott. Thank you all our speakers, and we can now proceed to a live question and answer session.