Mike Tallhamer, DABR (00:02):

We have a PSA with Vision RT, as well as a COE with our Florida institutions on the technologies you’ll see in this presentation.

Mike Tallhamer, DABR (00:13):

Just a little bit about the Rocky Mountain region goes without saying, probably better than Florida, but we’ll skip that piece for right now. SGRT is our standard of care, so if you were at our institutions this morning, you see, we treat a hundred percent of our patient census across all indications. And when I say all indications, somebody always asks, well, what about electrons? What about HO hips? What about Osteoarthritis? What about this all means all? There is no exception to that rule. At our institution, we treat everything with SGRT. Those indications are treated across the entire treatment pathway. So we’re talking about CT simulation, planning, patient setup, motion management, both standard motion management for wild motion that will stop the beam. Also, respiratory motion management if you’re doing purposeful actual beam gating treatment delivery and dose visualization. So Rocky Mountain region, we have all of our patients running through a completely uniform SGRT process. And if you were there this morning, you saw that process. And if you’re coming tomorrow, if you’re one of the 90 to a hundred people who will be visiting with me tomorrow evening, you’ll see that process as well.

Mike Tallhamer, DABR (01:14):

That being our philosophy, I was asked to talk about pushing the boundaries of SGRT, and I always love these kind of talks, cause there’s gonna be no graphs. It’s just me pie in the sky. Let’s see what we can do with these things. But a lot of people get very uncomfortable with those types of ideas of pushing the boundaries, especially in medicine. There’s certain lines that we think we should not cross. We need lots of controlled studies and those types of things. And so I’m here to convince you that maybe it’s not going forward.

Mike Tallhamer, DABR (01:45):

Pushing the boundaries is not crossing a line, and so we have seen in history that SGRT is the standard of care. Many years ago, actually, 10 years ago we were here at an SGRT meeting in Denver. And the idea of the, it becoming the standard of care was considered one of those outdoor or out there type of ideas. And this is the AI definition of out there that’s very scientific. So if you wanna read the definition of what an out there idea is that was something that was not really well received. It was like, this is a DIBH tool. This is maybe an SRS tool. This is maybe this type of tool. And it was very pigeonholed into kind of a motion management type of type technology.

Mike Tallhamer, DABR (02:33):

Treating a hundred percent of your patient population with SGRT is still considered kind of wildly out there. I do a lot of consulting work, and this is something that is not always well received. The idea of tattooist treatments is still very uncomfortable for some people. Applying the principles of SGRT to simulation treatment planning and all that sometimes is considered, why is that necessary? Or is this an ill-advised, you know, know, adventure into something that we don’t need to be looking at? And you may fall, I don’t have any strange ideas that someone in this room is not in one of these categories, and that’s absolutely okay. I want you to know.

Mike Tallhamer, DABR (03:16):

I think it’s officially dead at this point. this is a safe space. The jokes go way better when it’s on time. . This is a safe space for you guys. I, I don’t want to push anyone into feeling like you have to do something or that there’s some proper or improper way of doing this. I don’t have a button, so we’re gonna wait. At least we’re ahead of time. man, you ruined my joke by clicking that button, man.

Mike Tallhamer, DABR (03:39):

So my philosophy tends to be a little bit different. So if you work at my institutions or you’ve been to my institutions as on a site visit, as some of you were this morning, I, I have a little bit of a different approach. I tend to like to cross those lines just a little bit and push kind of some of the boundaries much to the screen of some of my staff who happens to be here today.

Mike Tallhamer, DABR (03:59):

We like to explore some of the possibilities with some of the technologies that are put in there. So today we’re gonna review a couple of the quick historical cases of things that have been presented, what that look like, then what might be happening. Now we’re gonna look at some of the present opportunities with the SGRT technologies as they’re constituted in this sim plan treat dose paradigm that we’re, we’re talking about promoting. And then we might just peek over the fence just at the end and see what we can do with these technologies. If you don’t push the boundaries, you will never know where they truly are. This is often associated with TS Elliot. He didn’t actually say that. We don’t let the truth get in the way of like really good presentations. So we’re gonna say he said that the actual quote. Only those who will risk going too far can possibly find out how far one can go. That’s a little bit more of a scary quote. So we’re not gonna use that one. We’re gonna go with the boundaries one. And so for moving into our review, let’s review

Mike Tallhamer, DABR (05:01):

I’m gonna take you back to a time, a more innocent time, a more gentle time. We’re gonna go back to the SGRT meeting in 2019. This is so much easier when I’m doing it in, in Las Vegas. So maybe not a more gentler time, but we were in Las Vegas. We presented at that time the first 10 patients that we had done in, in what was called a massless head and neck setup. We had seven patients that were treated in a chin strap for simplicity for our therapist, but we also had three patients who we presented 30 fractions on each of those. So 90 fractions of patients being treated completely without a mask, just a posterior mobilization. We also presented a talk on the principles of IGRT augmentation, an idea that SGRT can be used in conjunction with IGRT, where we’re actually correcting the posture because we all know that we have six off couches, but we don’t have posturepedic couches, so we cannot correct posture with our couches. We move our patients as one unit. So SGRT is a principle of using multiple ROIs to correct the posture before doing IGRT and seeing how much better the IGRT can be lowering your residual shifts, eliminating the need to kind of split the difference as we all oftentimes here. This was presented at the meeting and I, like I said, I oftentimes get a lot of fan mail. So after I returned, I got lots of questions about this.

Mike Tallhamer, DABR (06:23):

This is an actual email. I saved this and I blocked the person’s name out because if they’re here, I really don’t want to offend anybody. But the it promotes an unsafe tech technique for treatment is, was one of the things that I got. It posterior mobilization with SGRT will never be a viable way of treating a patient. So we learned in 2019 that what’s happens in Vegas stays in Vegas. So we are just not gonna talk about that anymore. So let’s come back to 2025.

Mike Tallhamer, DABR (06:56):

Let’s see what’s happened since we were gone. So in 2019, as we said, Nick enters the conversation. We talked about a first cohort of patients that were treated either in a very minimally masked chin strap or in a completely masked regimen. We talked about postural correction and IGRT augmentation, and we also presented the first look at being able to pull SGRT data into the IGRT workflows for treatment, 2020 was a great year. This is what I consider one of the best QA tools that have ever entered the radiation oncology realm. And I know that’s saying probably something probably over the top, but postural video has been something that I believe has changed the way we set up patients completely across the board. If this was ever to leave our clinics, I feel we have done a disservice to our patients. And so 2020 was a great year in that we had this postural video. But we took some hits in SGRT in 2020. In 2020 we had done years of work trying to get masks head and neck off the ground in our institutions. And before you know it, we were masking everything. So we went from removing the mask to masking everything. And this is the year obviously the pandemic hit. Some vendors took advantage of it. Elect was promoting a portable version of body fix. 2021, AlignRT Advanced continued to expand the opportunities for us to do new things in SGRT. The future of SGRT I was asked to present on a similar talk to this about the future of SGRT where we saw it going. In 22 at the SGRT meeting. We talked about promoting SGRT to the level of IGRT within our OIS and looking at how does that work within the system and getting images into a review process that was more promoting IGRT and SGRT as equals. And then the new patient id, the facial recognition system was released as well, which took away the name, date of birth, what are we doing manual process of doing that and entering a kind of forcing function, Biometric facial identification, unlocking treatment plans. 23 we got our DoesRT system. So this entered clinical practice at our centers. And we’ve seen a lot about DoseRT. 2024 MapRT entered the realm. We got FDA clearance or not FDA clearance. We got MapRT into the clinic. We were using this in a nonclinical way, but we were able to use these collision mapping softwares to explore more of the solution space and look at what of our calculation models about collision modeling, if those were maybe far too conservative or maybe were they good enough? And so we were able to cross compare these things with MapRT.

Mike Tallhamer, DABR (09:57):

So what does it look like today in 2025 early in 2025? The OIS module entered our clinic for testing. This allows us to take the IGRT or the SGRT imaging and place this into the IGRT workflows for our physicians. This allows you to basically enter this directly into the ARIA timeline. Your physicians can then review these just like they would review any other IGRT image. This means no more are the PDF printouts and then reimport them into ARIA and then have your physician remember to do their documentation on days when you’re not doing whether imaging. SGRT is actually a full-fledged IGRT equivalent now. This includes all of the approvals for RV use and all the things that administrators have to track. This makes that process extremely is easy.

Mike Tallhamer, DABR (10:48):

The Maskless H&N in 2025 we actually see publications now using the macromedic system and this dorsal shell, this was published earlier this year. That was promoted at ESTRO. If you’re interested in seeing the process with the dorsal shell, you can scan that QR code and there’s a video that you can actually watch for that process and see what that looks like. Massless H&N is being promoted on social media, either by vendors or as you see on the right there. Another study out of, I believe that’s Zurich, Switzerland, maybe I can’t remember right off the top of my head, but something looks extremely familiar to me. If you look at that dorsal shell and you think back, have I seen that somewhere else? It looks very familiar. Just something else I’ve seen, I can’t remember when, maybe it was 2019. I don’t remember.

Mike Tallhamer, DABR (11:52):

We also received a beam light system. So we have a Cherenkov imaging system. Cherenkov imaging has been talked about extensively. This beam light system went in about a month ago. This is a hybrid QA thing that we’re trying to develop with our arc checks. But if you look very closely, you can see our HDMLC leaves shaping the field individual leaves. What this means from an image quality standpoint, the beam lights are tuning the lighting for the delivery and the imaging of Cherenkov imaging in a completely separate way than what you’re doing in the room. So you can operate in a normal lit room, but the cameras are seeing a tuned version of the lighting for imaging. This allows us to get extremely nice imaging at this level of quality. I mean, if you see that, I mean, that to me is one of the coolest things I’ve ever seen. You can start looking at potentially using these images for cross correlations from day to day fraction to fraction. Even control point to control point if you really wanna get crazy. So if we’re looking at pushing the boundaries, there’s machine learning models, there’s AI models that can actually use these things and basically flag each approved fraction and then say, okay, fraction seven is an anomalous fraction compared to all of the other fractions. You can look at things like contralateral breast, which we just heard. You can look at different types of deliveries and say, okay, this position may be off because the projection of that dose on the patient is anomalous compared to the other ones that you’ve done because you’re getting such high quality image now you can do that. So this also entered our clinics just literally a month ago. So I know what everybody’s thinking in the room. What could possibly be next?

Mike Tallhamer, DABR (13:15):

I think we may have traveled too far into the future and hoverboards are the next thing we’re gonna go. I don’t know if anyone has seen back to the future, but that’s one of the greatest movies from my childhood. So I think this is the next thing we need to push for from Norman and the team.

Mike Tallhamer, DABR (13:30):

So let’s talk about what the future could look like. We’re gonna talk about MapRT and the APIs. This is something new. We’ve heard a lot about MapRT. You’re gonna see two simulation processes with MapRT. MapRT in this case, on the left is a simulation workflow. We’ve heard about the simulation workflow. We can get the patient on the couch, we can take a reference. We can then essentially do a dry run sim place isocenter. And if you look at the way this is done, they’re moving around an axis orientation three axis. On the right is an API application that we’ve written that has had feature creep. But this is something that we engage our therapist with and say, okay, how can we change this process and make it faster for you? One of the things that they recommend is, I’m used to seeing lasers on a patient. I’m not used to seeing this three axis thing. It’s easier to see where lasers are. And so we’ve added lasers to the system. The ability to look at multiple collision maps that are cached is another thing. Cause It’s a, it’s a slow process if you have to go in and out and select different types of machines. And so we’ve actually added the ability to add the different rooms. On the left, you can see they’re adding the imaging setup. So you can see the arms that come out. This is a new orientation, which will give you a new collision map. On the right, you can see that the collision maps have already been cached. And so in a second you’ll be able to see that the therapist, after they’ve done that, they can flip back and forth between the collision maps and compare the imaging collision map based on the treatment collision map. So now you can optimize very quickly look at your isocenter based on what you’re used to seeing with lasers and move forward with that on the right. The reason why the cache is nice is because you can see how long some of the collision maps can take if you have a very complicated geometry.

Mike Tallhamer, DABR (15:04):

Other minor tweaks that we can do. We can change the way information is presented for our dosimetrist. So in on the left you can see a number of additional pieces of information. You have two different things. You have imaging fields and you have additional imaging fields. So the in red, you see what is basically the port films from the plan. They show up as duplication of the treatment fields. And so this is sometimes confusing for our dosimetrist. They have imaging fields and they have duplication of, of port films that are actually treatment fields. On the right is the application that we’ve written that basically eliminates those things by properly reading out the DICOM and being able to isolate the things that the dosimetrist is used to looking at. So they can have a profile. And if they don’t wanna see the port films and those types of things, they can isolate that information and look at that very quickly. The other thing is, and this, I should have blown this up, we have the collision map on the left. We all the fields are yellow. If you were to look at those things immediately, you could not tell which were imaging fields. Which ones are treatment fields on the right, the imaging fields and the treatment fields are different colors. One is green, one is teal. And you can isolate these things and make them very quickly for the dose of interest as they’re optimizing a plan. Not worrying about imaging fields if that imaging field is not something mission critical to the, the optimization of the plan at the time.

Mike Tallhamer, DABR (16:22):

We’ve already talked about non coplanar treatments. One of the nice things about MapRT. It has lots of different features. We’ve talked about all the features today. This is actually workflow choreography. Essentially we can use MapRT to actually optimize the delivery of the plans. Cause we look at the 3D model and we can move from start to stop on each one of the angles and make sure that optimization looks appropriate. This allows us to treat SS with five couch kicks in about 11 minutes at our institution because our dosimetrist is doing all of the optimization for the delivery for the therapist. So they are optimizing not only the plan itself with the non coplanar beams, it’s they’re also optimizing the delivery for them by actually working through to make sure that it makes sense to kick the couch in one direction. If they have to come back the other direction, they can switch those gantries around. So the interface works extremely well for optimizing those things.

Mike Tallhamer, DABR (17:12):

The reason for that we talk about SRS, one of the nice things about MapRT is the collision risk for SRS because we have non coplanar beams is very high especially on some types of delivery. Certain vendors have made this easy for us. And I referred to it today during a site visit as SRS for dummies. And I, I think I insulted some people very hyper is great. It makes things very, very easy. The thing is, is they solve the problem for you by limiting the solution space. They say, we are going to take care of everything. You’re gonna use our mobilization, you’re gonna use this type of beam geometry. And then we know you’re not gonna hurt anybody, which is great if you’re just looking to do SRS and get through the SRS. But there’s lots of other solutions that are much more refined, especially when you’re doing things like single isocenter, multi target treatments.

Mike Tallhamer, DABR (17:55):

MapRT allows us to explore that entire solution space. You would say, well why do I care? I can look at things like this SIMT plan. It’s a single isocenter. I’m treating maybe four lesions. Not a big deal. This same patient has been a frequent flyer at our institution. This is what their last plan looked like. It is four grouped SIMT plans all optimized around previous deliveries and I think it has like, I think 15 or 16 beams. They’re all synchronously optimized. So if any one of those plans presents a collision risk at a collision check, that means you’re retiming the whole plan. That is not cool and not fun. So we want to be able to quickly look at these plans and make sure that we don’t have any type of optimization problems.

Mike Tallhamer, DABR (18:37):

So if you have something like this currently in the MapRT workflow, you would export each one of these for plans. You would go through that workflow that you saw where you select the plan, you select the patient surface, you select the room, you then look at the collision map, you then close outof that, select a new plan, select the same surface, select the same room, look at a new collision map. And you would do that four times. That’s a little bit cumbersome, especially when you’re under the gun and you’re doing seven or eight of these in a week. So for us, we wrote some ESAPI scripting. We have the API, but we also have variance API.

Mike Tallhamer, DABR (19:09):

And with that we can now integrate MapRT into Eclipse. So we saw the integration of MapRT with RayStation and it’s right in the interface. This is a script that we’ve written that optimizes the workflow for our dosimetrist. You can now select one isocenter. You can find another isocenter that may have a collision beam. And instead of finding this out after you’ve done the entire plan, you can explore the solution space directly inside of Eclipse. Find the appropriate beam angle, find the start and stop angles. Put that right back in your plan. Reoptimize your’re on your way. You don’t have to worry about this thing going to the machine with other problems. This is a very, very quick workflow, very easy cause the dosimetrist never has to leave the interface for eclipse.

Mike Tallhamer, DABR (19:51):

When you’re talking about advocating to do things like this where you’re replacing collision checks with some more automated process or a new process, you have to be very careful that you’re not eliminating safety checks that you need. So we’re gonna just kind of talk about both of these. We need to verify that the collision maps we’re using are appropriate. And so in this case the collision map on the top, we’ve pushed our plan. We have an SRS and we’ve pushed it to the treatment machine. It looks like everything passes. But if you look, the isocenter is outside of the head. That is not the appropriate isocenter. The reason for that is because the DICOM is the isocenter that we’re we’re localizing. The MapRT surface is very sensitive to workflow errors. So if you’ve zeroed your CT in the wrong location, if you took a capture of a patient, they had to get up and use the restroom and come back and you captured in the wrong location. All of those things can lead to a collision map that is not appropriate for this treatment plan. If you look at the lower one, this is the exact same patient. So this patient had an SRS, but he also had an SBRT lung. And when we were looking at the SBRT lung, can you tell me if that isocenter, can you find it one? ’cause The isocenter marker is so small inside of a very large man, you can’t tell if it’s off by a centimeter. You can’t tell if it’s off by five centimeters. You can’t tell if it’s lateral shift. Where’s the isocenter and is it in the right location? Is it off by a lot? Is it off by a little? Is it off at all? We don’t really know. And so one of the things that we said was this is something that we are using calculation models. If we’re going to use this as a tool we need to be very careful. Next slide because here be dragons. This is where we make our mistakes. We don’t want to take an institution that is heavily involved in four PI type deliveries, non coplanar deliveries, not having collision problems, and now using a new tool and introducing collision problems because we’re trusting data blindly without the proper QA. So this is where we can use again, tools that we develop.

Mike Tallhamer, DABR (21:37):

This is using the API in an external interface. So we have our DICOM surface. we have our MapRT surface. We can merge these in a software very easily and then we can check to see if they match. An SRS patient who was an inpatient. MapRT surface. You can see there’s extensive amounts of copious stuff on top of him. The merge surface still looks good. So both of these collision maps can be trusted.

Mike Tallhamer, DABR (22:05):

We have our SRS patient, our very large SRS patient that we saw where the s the isocenter was outside of his head. We have his MapRT surface. We have the merged surface. And this is why the isocenter is outside of his head because the DICOM surface and the MapRT surface are not married in the same coordinate system. So we can look at correcting these things. There’s multiple pathways to correcting this so that cannot be trusted.

Mike Tallhamer, DABR (22:30):

We have two different workflows. On the left you are going to see a workflow that corrected this in CT sim. The therapist caught the fact that they took a surface in the wrong location. This is the software. You can see obviously the body is in the wrong location. It’s up on top of their head. They actually captured the surfaces in two positions out of order which resulted in this. But you can see if you call MapRT, you’re going to get a collision map. So there’s the collision map and there’s the plan and everything passes and everything’s fine except for the fact that the person’s abdomen is on on their face. The therapists were close enough and actually smart enough to know, Hey, I did this wrong. Let’s, let’s correct the workflow. But we send this on so you can grab the second surface and you can see by blending that this surface is actually appropriate. On the left hand side, we could do a new collision map. And if you look at the new collision map, that new collision map looks exactly like what we would expect and where plan is still good. But now the collision maps are very, very different in that case. So if you were looking at something, you would want to do it. On the right, we have a collision map that is not appropriate because we have again, a misalignment. This is our SRS patient, but we didn’t get a second surface. So how do we fix this? Well now we have the API, we can actually move the CT couch. So now we can correct the MapRT surface to the DICOM coordinates. We can move that DICOM surface in three dimensions to make sure that it fuses. We can verify that with some blending of the things. Look at the laser, say okay, this matches up to our plan. And the interesting thing about this is if you see the collision map that we have now, it says that there was a collision on that lower portion of the collision map. You could shorten that arc and eliminate that collision. But after you’ve corrected the collision map, you will notice that the arc is actually going to collide on the other side of the patient because he’s a very large man. So if you look as the oscillate, you can see there’s a collision at the top of the map as well as the bottom of the map. So had we corrected that, sent that to the machine, we’d be lik we’ve used MapRT to fix this and MapRT didn’t actually fix it. And so we would’ve had to have corrected two different start and stop angles in that case. So different ways of doing that. We’re gonna go to the next use case.

Mike Tallhamer, DABR (24:30):

This is something, this is actually a throwback to 2015, the last time we were in Denver. So we’re gonna start this. This is a maskless head and neck. Again, one of those things that don’t work with posterior mobilization. I’m not bitter at all or anything like that. You’re gonna see blending of the surfaces. This is again a massless head and neck. He’s just sitting with a posterior mobilization mold. We are actually going to take this, move this into a export utility that we’ve designed, allows us to trim the surface. So we’re gonna take some of the sheets away. We’re gonna trim out some of the table and then when we’re done, we’re gonna set some vox actualization parameters, some smoothing parameters and some volume parameters and say okay. What that does is creates a synthetic CT for us. We can now place the synthetic CT wherever we want. There’s different ways of creating synthetic CT volumes. There’s a bunch of papers out there on how to do it. You can see two examples down here. So we have a synthetic ct and volume pixelation and we also have a synthetic ct, what’s called a fill down method. So both of those can actually be pushed to your treatment planning system. So now I could CT this much of my patient and I can get that much of my patient and they’re all fused in DICOM coordinates, which is fantastic. So we can push these things into the planning system. Everyone’s like, well, why the heck would I wanna do that for an SRS? Well, if I wanna edit out the sheets, if I don’t like what I’m seeing, I can actually edit out a surface. So if I’ve pushed the mesh to the planning system, I can edit the mesh, I can take the sheets out, I can take something out that’s temporary and then I can put it into a secondary collision check. This is rad formation collision check. This surface is now in rad formation. I can do a collision check and it’s check MapRT’s collision map. I can put the entire volume. So this is my massless head and neck because he’s very, very elevated. You can see I’m not gonna probably be able to get arcs underneath of him. So now I can put that in a collision check and check out MapRT’s accuracy as well. I can use this for commissioning. And then the magic thing that everybody wants to do. So we can take that surface and put it right inside of AlignRT. So now when I CT my breast to here and the elbows are way up here, I can get the entire thing for postural video. I can take the MapRT surface, convert it into a structure, put it into my planning system, and now I can set up the entire patient eyes to thighs and I don’t have to CT them that far. I can just do the CT that I need. So very, very interesting things that can happen. This was first done in 2015 when we did this with the at the time Gate CT surface. If you’re old school enough to remember what that is. Congratulations. You’re old like me.

Mike Tallhamer, DABR (26:50):

So what’s next? What can we do with all of these cool little things? Is this just toys to have toys or is this actually something that we can use clinically? So let’s look at a sim plan treat that we are envisioning actually implementing at our institution. What could a future sim plan treat dose look like? Well, we could do a CT sim but not with a ct. We can do an optical sim. Now we can take something like osteoarthritis, patients who are getting very low dose radiation. We can take a MapRT surface. Instead of being simulated asking someone, taking someone who has osteoarthritis and asking them to lay prone on a table with their hands above their head or hands above their head. This way you can just simply have them stand next to your CT couch. Take a MapRT, walk to the other side of your CT couch, take another MapRT, and then you can export those surfaces. Trim it to whatever you like. And so in this case, we’re trimming this down to the right hand. We’re exporting that ct. That CT is then evaluated sent to the treatment planning system. We can now put that CT in the planning system. We can do a plan on this hand. It’s 50 monitor units. We can actually use wedges to account for the fact that we want uniform dose through a wrist and uniform dose through the tips of the fingers. So we don’t just do open fields and get hotspots in in the nail beds. We can actually use an EDW. We can use that surface and now push it onto AlignRT and set this patient up. We can just simply put their hand on the table, line them up with AlignRT. We can then treat it on the true beam and actually take an image and align that image to that synthetic ct. And then we can watch that patient get treated. So in the course of this entire sim plan tree dose, this patient received zero radiation until they were actually treated, which is really, really cool in my opinion. Cause these patients are getting low dose radiation for osteoarthritis. This is not a malignant condition and this is something that we don’t need to be doing additional imaging because you think about even a simple port phone with two monitor units. If I’m delivering 50 monitor units and two monitor units every fraction for six fractions, I’m delivering a substantial amount of additional dose that’s not needed in that case. And it doesn’t even have to be sim plan, treat dose. We can then actually after delivery, this is all sitting inside of REL, like we did this with any other standard CT, took any other standard images and our physicians check this every single day, get their RVs and everybody’s happy and everybody goes home and the patient didn’t get any additional radiation. So that is something that we think is pushing the boundaries of SGRT in the right direction. Not crossing a line, but maybe using the technology in a more efficient way.

Mike Tallhamer, DABR (29:27):

If you wanna know where the future of SGRT is, do what everybody else does. The millennials ask ai. I like perplexity. AI will tell you exactly what the future of SGRT is. This is a table if you wanna see this search. It’s actually really good if you use it on Grok or anywhere else. You get really funky answers. Perplexes is a pretty decent research AI that will read papers and give you citations. So that link is the search of what is the future of SGRT. These are just some of the experts. The things in yellow are things that we’re actively looking at right now in our centers. And I don’t have anything hard and fast to present to anybody right now, but there are some unbelievably cool things that you can do with SGRT in the future. Looking at portability of the technology, looking at expanding it beyond radiation therapy to other disciplines. Enhancing the accuracy of adaptive therapy, which we heard about some of that as well. There’s so many different things in room. Scene mapping, augmented reality. I mean, we’re all buying couches on Amazon now and sticking ’em in our own living room before they even show up. We can like, why not map out our stools and not collide with our stools? Not that the therapist ever collides with stool in my center, by the way. That never happens. But you could do interesting augmented reality that makes things and deliveries safer. So I’ve taken up way too much time but I don’t have any conclusions. I just think that everybody should think about what these technologies can do. Get a little creative. Don’t get nuts. Don’t cross that line. But you can definitely do some very interesting things with SGRT.