Dr. Nzhde: Okay. So my presentation will be mainly focused on margin assignments for treating multiple metastases with single isocenter. More specifically focusing on our recent decision in our institution to reduce margins used for these patients. These are my disclosures, and credits to my colleagues who have helped me with this presentation.
So we have done a lot of work in commissioning and validation of this methodology. And if you're interested to learn more about the commissioning and validation methods, there is a SAM [SP] session right after this symposium at 1:45 where I will present more about commissioning and validations. But just in summary, everything that you see in this one slide has been utilized for commissioning purposes. And they all are viable solutions for using at your institutions.
Most notably, we have used this polymer gel to do commissioning of the system where this bigger cap that you see there sort of holds the gel, and the smaller cap that you see outside, that's where the water goes. And the smaller vials that you see here, those are for calibration purposes. And not all gels require this, but this particular one used here requires vials to be irradiated to known doses. And the results that we get from that kind of measurements are just fantastic. Just to give you one example over the line 1-D profile two, three targets, you have fantastic results in terms of measurements versus calculation.
In terms of patient-specific QA, again, at that session, I will cover the details of this. But just in summary, many of the methods used in commissioning are also good candidates for using for patient-specific QA. One note is that since this is an enhanced dynamic conformal arc and not a volumetric modulated therapy, you do not really, at least at our institution, we do not have to make a measurement for every patient. Yes, we did make a measurement for the first 10 or 15 patients, but currently all we do is we utilize another planning system to do a forward calculation as a secondary calculation, and we are ready to go with that patient. So it's a very, really quick process of validating the plan prior to treating the patients.
I do have to step back and mention that implementation of this treatment metadata, our institution has been a great success. And I really mean that. These patients, instead of coming...like Dr. Rahimian mentioned, they have patients with 30 or 40 metastases, instead of coming multiple days and getting treatments for multiple hours, they just come once and the treatment lasts less than 30 minutes and they're done. This is great benefit to the patient. It also has provider benefits. And I'll show you some data. It requires less physicist's time at the machine, less physician time at the machine, it increases your throughput on the machine, more patients are treated. And as from the Kaiser experience, actually also increases the eligibility of the patients that are eligible for radiosurgical intervention.
So here, one of my colleague physicists tracked his time in terms of how much time does it take to plan these cases using iPlan method, multiple isocenters versus MME with a single isocenter. And in one extreme example that I'm showing you in this corner here, he spent more than six hours of planning that, and it's real time tracked by him, versus 20 minutes and you're done with the plan. So even on a preparation part for treating this patient, it takes much less time.
This is one of the MD residents, physician resident tracked the time that they spent at the machine treating patients using multiple isocenters versus single isocenter. So significantly less time and it doesn't scale the number of isocenters necessarily. And it takes less than 30 minutes to treat them. How much time do physicists spend at the machine treating these patients? At least in our institution, we have to show up at the machine, the physicians have to show up at the machine. And it turns out that using this method significantly reduces that time as well. And it's shown here. And by the way, this is not in one shot. In other words, if you have a patient with 16 metastases, at least at our institution, it means I will be paged to the machine 16 times. I have to leave my office 16 times, go to the machine, do an isocenter, and back. Just imagine that scenario, okay? It's a very disruptive process and it's not good for the patient either.
So we did talk about margin assignments at the previous symposiums and I did mention that rotations in terms of having multiple isocenters, one isocenter per target are not that detrimental for treating these patients. Whereas when you have a single isocenter, then rotations do become detrimental. And we decided that we will be utilizing large margins. And I'm here to share with you that we have, sort of, backed down on that decision of using large margins and we are going to reduce our margins. So our decision was specifically for those targets that are further away from the isocenter and they're smaller, you may even completely miss it if it's not aligned correctly. The problem is they're aligned correctly, so we will have to rely on our IGRT system to align those targets.
So we have communicated with you that we utilize 0.5 millimeters and 0.5 degrees for using ExacTrac. And we said the margins will vary from 0 to 2 millimeters depending at the location of the target. But the truth is, we uniformly use 2-millimeter margin for all of the targets at our institution, which now we are reconsidering to change.
I should mention that what you see on this graph, on the X-axis, is the threshold for the ExacTrac utilized. So in our case, we're utilizing a 0.5-millimeter threshold, and the Y-axis is the frequency of the therapist realigning the patients. So 50% of the time, the patients actually are being realigned by the therapist, which means several things, but one of them means that utilization of ExacTrac for each arc is actually necessary because at the end of each arc, when the therapists are taking an image, there is a 50% chance that they have to realign the patient. So it is a necessity to have that sort of an imaging capability when you're doing single isocenter treatment. The experience is really not that different from our previous experience with the trigeminal patients, where, also with the 0.5 threshold, we were utilizing 50% of repositioning of these patients.
Now, okay, we showed this graph and we said, well, if you have a 0.5-degree rotation, which is that orange line here, and you're 6 cm away from the isocenter, that basically translates to just because of the rotation, you have to add 0.5-millimeter margin. Well, that's true. And one thing I would like to share with you is that depending on the number of targets that you have and the sizes of target you have, the consequence of adding margins is different. So when Dr. Rahimian is showing you that they have many small targets, adding a margin to many small targets has a different consequence than adding a margin to a few larger targets. And that's what I'm demonstrating here.
So if you have one large target of 10 cc, and you're adding margins, the X-axis's margins, that's the volume, the increase is not that steep. But if you have 10 targets of 1 cc and you're adding a 2-millimeter margin, like we do at our institution, the increase is much steep. So there's a big consequence of adding larger margins to smaller targets that are sort of spread around the brain.
This is one patient example whereby adding margins, 0, 1 millimeter, and 2 millimeter, we're doubling or tripling treatment volume as well as V5, V8, V10, and V12. And just to convince ourselves, we just basically went through an exercise of planning eight patients using six different margin schemes. Okay? So these eight patients, and what I'm showing you is the average data, that have been planned with zero millimeter margin, with the one millimeter margin, and the two millimeter margin. And the graph that you see here is the average for those eight patients.
And let's look at the target volume, for example, or the V12, or V10, or V8, or V5. Comparing 0 to 2-millimeter margin, we're basically tripling any of those endpoints that I'm showing you there going from 0 to 2 millimeter. And at least in our institution, when we're prescribing the multiple metastasis treatment, V12 is a strong consideration for the prescription. If you don't meet the V12 criteria, then we drop the dose. In other words, what I'm saying is by adding a margin and not meeting the V12 criteria has a consequence in terms of lowering the prescription for that patient.
The other three columns that you see here is the same patients except their planned with the variable margin. Any target that is within 6 cm distance from the isocenter, in this case, for example, is planned with a 0.5-millimeter margin. And anything that's more than 6 cm away is planned with a 1-millimeter margin. And you can see that even a technique or approach like that has a sort of big improvement in terms of V12 all the way down to V5.
So just to reiterate, when we use variable margin, and I'm gonna go back here for a second, for these schemas that I have here, using variable margins, only about 10% or 15% of the targets get the larger margin. The 90% of them actually get the smaller margin. So when I'm telling you we're using 0.5 or 1-millimeter margin, in this case, about 90% of the targets get 0.5 millimeter margin. The rest is ones that are really far away will get a 1-millimeter margin. And that's shown in this histogram, basically. Only small portion of them are more than 6 cm away.
Okay. Just to demonstrate the same idea one more time, these are different patients with different number of metastases. So basically each of those lines is a different patient. And the yellow one is with two metastases. And for example, the red one is with the 12 metastases. And when we go from 0 margin to 1, to 2-millimeter margin, the increase in terms of the V5 is much steeper for the ones that have more metastases. And the same is true for V8. And these are now actual patient data that I'm showing you here. And this is V10, V12, and the treatment volume.
So basically, the increase in terms of number of metastases is much more significant. While the approach of using single isocenter for multiple metastases is more beneficial for having more sort of targets. Adding a margin sort of contradicts that. It hurts us. So we really need to be mindful of the margins that we are utilizing for these patients. So in general, we're using wider areas, we are irradiating wider areas if we use larger margins compared to the old method of utilizing multiple isocenters.
There is another change that happened going from iPlan to multiple metastases. The dose distribution within the target is more inhomogeneous. If you look at this one example here, although the prescription is 20 gray, within the target, we have about 30% or 40% hotspot, which you may argue to be a good thing, I'm just pointing out that it's different than what we used to have. So if you're adding too much margin, then we're irradiating wider areas, the V5 and V10 and V12 are wider, and you're delivering more dose to the target. So the patients are not really getting the same treatment. That's the message I'm delivering to you. And we have been sort of revisiting this idea and trying to see if there is any room for reducing the margins.
What I'm showing you on this graph is the RTOG prescription criteria. On the X-axis, what you see is the size of the target, the Y-axis is the prescription. I understand that some institutions don't exactly follow the RTOG criteria, including ours. So in parenthesis, I'm showing you the doses that we use, instead of 24, we use 20, instead of 18, 18, instead of 15, 16. But the phenomenon is there basically. As you're adding a margin, your targets are becoming larger, and your prescriptions may change for these targets that are sort of at the line of the threshold going from one prescription level to another.
So if you have a target here, potentially that was going to get 20 gray, now you're adding a margin and it drops to 18 gray. And moreover, as I mentioned, it could also happen because of another thing. It could also happen because your V12 is much larger now and you're not meeting the V12 less than 10 cc criteria, and because of that you drop the prescription. So both of those can happen. So there is some consequence in using larger margins. And we need to be mindful in terms of what margins we're utilizing.
This is data showing that when we acquire ExacTrac images after each of the arcs delivered, this is the residual motion that we measured for those patients. And really, there is really not that much residual motion that we see. In all directions together, if you look at the X, Y, and Z directions, the mean residual motion is 0.2 millimeters with overall 90th percentile or 0.7 millimeters. And if you look at the translations, same thing. The mean is about 0.2 degrees and the 90th percentile is 0.6 degrees.
So because of those considerations, we just had a meeting with our lead physician who is leading our radiosurgery service. And we've decided that instead of using margins, this is our current assignment margin here, we are going to transition to one of these criterias, most likely to this one here, where you're utilizing 0.5-millimeter margin, and 1-millimeter margin for the ones that are further than 6 cm away. And keep in mind that these are only a small fraction of the targets that are further than 6 cm away.
Ultimately, you probably will need a clinical trial to determine what margins are effective for these patients. But at least in terms of analyzing the technology and the results that we have, we have made a decision that the margins that we have been utilizing probably have been too big. And we have decided to reduce the margins. Okay. That was my message for today. Thank you for your attention. So at this point, we will be taking questions. We'll answer your questions, Dr. Rahimian and I. And then I'll introduce you to the next speaker. Dr. Sharma here.
Dr. Sharma: Excellent presentation [inaudible 00:16:26]. My question is about the [inaudible 00:16:32] criteria that you use. At what stage instead of reducing the dose would you go about multiple isocenters rather than single isocenter to keep the same dose, which is prescribed, rather than reducing it? And, you know, just [inaudible 00:16:51] single isocenter to multiple isocenters?
Dr. Nzhde: Let me see if I understand the question. When you look at the V12 criteria, at what point you will make a decision that you will not be using single isocenter and you will use multiple isocenter? So we haven't had a scenario where we did that. We had a scenario where with the V12 starts bleeding between different targets. So what we do in that case, we consider that as one contiguous V12. Let's say if you have two targets that are nearby and the dose bleeds between those two, we consider that as one target. We look at the V12 as one target and we reduce the prescription for that. But I didn't answer your question directly, which is, is there a limit where you will abandon the single to the single iso? I'm not sure if you will get that much benefit by using multiple isocenters. I'm just not convinced that you will get that.
Dr. Sharma: [inaudible 00:17:54] you mentioned that when they move from single to multiple isocenters, the amount of dose between the 12 gray dose reduces inside?
Dr. Rahimian: Yeah. I think there is definitely when you open the jaws because when you have a small lesion and the jaws are tracking it, the leaf leakage is obviously less but the issue is when you have multiple lesions, on the other hand, you're giving like nine times or seven times more MU, you have the head of the [inaudible 00:18:34] leakage. So it's a compromise between the two. The thing is that today in cancer management, especially with immunotherapy and targeted therapy, the patient surviving 5, 10, 11, 12 years with brain metastases, from breast cancer, and even lung cancer.
So what I would see I think the managing of these lesions is significantly improved using this technology compared with, you know, whole-brain radiation. And these patients, you will be surprised they're so functional, they're young and functional. But I think this is a good solution in combination with overall management of the patient. So I don't see a significant...I've never done, just going by based on the 12 gray volume. We never switched to multi isocenter because, unfortunately, or fortunately, our radiation oncologists, when they get used to a system, which is so efficient, nobody wants to go back including the physicists, spending eight hours reading.
So yeah, I think that's something that we haven't done it. Well, I think one thing is that the new version of this software, which in our institution we just installed, has the option of areas they can [inaudible 00:20:07] and the system software takes care of that. So it's actually up to now, you would get a plan and you had to deliver it. But with the new version, I'm hoping that if you...like whole-brain, call it a PT, as an OAR, and some criteria may lower those V12 and things like that. I'm actually looking forward to use that.
Dr. Nzhde: So Dr. Sharma, I'm not convinced that if you switch from single iso to multiple iso, you will get that much benefit in terms of the V12 volume. We already have a wiggle room, at least in our institution, because when we calculate V12, we include the PTV itself. And that's a conservative approach. So we could sort of be a bit forgiving and let it be with that and keep the prescription the same here. But it was an excellent question. And thank you for that. Yes, please?
Woman: Do you have any [inaudible 00:21:07]?
Dr. Nzhde: I couldn't hear the question very well, but it was related to the MR registration uncertainty?
Woman: Yeah. [inaudible 00:21:23] margins all times in addition to [inaudible 00:21:27]?
Dr. Nzhde: Yes. So that uncertainty, to be fair, existed when we were treating these patients with multiple isocenters to start with, right? So this question comes up all the time. And we don't want to do something extra in terms of margins just because we were treating with a single isocenter. So at least in our institution, if you were not utilizing margins, those uncertainties were still there when we were doing multiple ones. Why add that margin now for the single isocenter treatment? That's a bit unfair. But to answer your question a bit more scientifically, we have looked at the accuracy of the ExacTrac in vivo, in patients, by looking at the thalamotomy patients where we radiate and look at the MR enhancement, and we look at the isocenter, and we're convinced that the accuracy of ExacTrac, including registration of MR, basically, the overall pattern of the accuracy for these patients is close to 0.5, 0.7 millimeters. So if you want to add that margin, you could, but that also applies when you have multiple isocenters, not just single isocenter. So if you're not doing that prior to this, why would you do it now? Yeah.
Woman: So it should be for a single isocenter possible target and for multiple isocenters, so 45 to 47?
Dr. Nzhde: Yeah. I didn't hear the question very well, but yeah. So in our case, we're debating whether to do 0.5 and 1 versus 1 uniformly for our old targets. Yes.
Man: [inaudible 00:23:14]. I'm curious, in your variable margin study, how was the single iso selected with respect to different targets? And do you think the isocenter can be selected better to reduce the margin of the targets?
Dr. Nzhde: Well, that's a fantastic question. So I did have a cases example that was excluded from this target that will be included in SAM. But yes, even the decision which targets to include, makes a difference as to where the isocenter is. At the moment, the MME element itself decides where the isocenter will be. And you have no choice over that. What you have choice over is which targets to include. So in our case, we had two or three examples of patients where we divided the treatments to clusters. There were a few clusters that were superior, and then a few clusters that were inferior, and we separated them out. That way the isocenters were not too far from each of the targets. If you include all of them together, then the distances to the isocenter increase. But going back to the question, you don't have a direct control over where the isocenter will be other than inclusion and exclusion of the targets.
Man: So this question here can be addressed to either of you. Do you have any institutional rules as to when you're going to be using a single isocenter technique versus treating the target separately? So I imagine, for instance, we have three mets at opposite ends of the brain, you're not going to stick an isocenter in the center and the middle and treat all three at the same time when there's, you know, more than 12 centimeters per target, right?
Dr. Rahimian: We do one or two lesions with iPlan, which is two isocenters. But if it's more than three, we definitely do multiple.
Dr. Nzhde: Yes. So thanks for the question. To be fair, we don't have a strict rule, but we do follow sort of a soft rule that if you have targets that are further away from the clusters, then we separate the isocenters. And you have multiple choices there, either you can use multiple metastases elements with two isocenters, you have that option, or you can do MME for a group, and then the other one can be treated separately. You have that choice. And we have done that a few times. But I couldn't tell you that we have a strict rule. Yeah, maybe it's time to think about that.
Dr. Rahimian: I think sometimes we treat like 15 tumors in a single iso. Then if one lesion is very close to the orbit, for whatever reason, the [inaudible 00:25:51] doesn't do a good job. And we pick twice, so one for that tumor next to the orbit, which the dose follow-up is really sharp, and then the other 14 will be a single iso. It's kind of a thing that, as a physicist, you have to evaluate your deviations. Personally, I want my maximum dose not to be more than 26 gray even though we give like 22.5 and so we always...Because that's why we have a much lower necrosis in these lesions compared with even previously with iPlan because we're conservative. Even the neurosurgeons, they say, like Gamm Knife [inaudible 00:26:40], you can have 200% at isocenter. But I think the incidence of necrosis significantly goes higher. And management of that is actually very tough. You have to do surgery and things like that.
Dr. Nzhde: But very valid question. We need to think about it. We have a rule, we sort of try to exclude targets that are further away but it's not a very specific rule that we follow.
Dr. Rahimian: I think it would be nice in the next version if we could pick where the isocenter should be. And since we don't have that option currently in the current versions, you can have clusters and multiple like two isocenters. It's definitely 15 lesions, 2 isocenters is better than 15 isocenters. So still is really good.
Dr. Nzhde: The developers are smiling. Okay. Go. Please [inaudible 00:27:40].
Woman: Do have a [inaudible 00:27:42] for recurrence, you know, versus a margin? Or maybe [inaudible 00:27:49]?
Dr. Nzhde: Yeah. I knew this question was coming. No, we don't. But it's a very good question. I did make a disclaimer at the very end and said that basically clinical outcome has to dictate this ultimately. But currently, we don't have that data. We did have one scenario where there was a recurrence and we went back and we investigated, it was actually related to a mistake that was done by a physicist in terms of fusing the images. One image, MR image that was primarily utilized to determine the contours for the targets was not very faithfully fused to all others. And there was sort of a shift but that was because of the fusion, not really anything else. That was a mistake that we made.
Woman: The next question is the V12. And, you know, you evaluated V12, do you subtract PTV or do you look at everything?
Dr. Nzhde: So no, we take the conservative approach and we don't subtract the PTV. And in the previous question, I sort of addressed this. That gives us a bit of wiggle room. If we are, sort of, marginally close to where we need to change the prescription, we don't have to because we already include the PTV and it's a conservative approach. But I know that many other institutions exclude the PTV. And that's probably a better, sort of, definition of the V12. I don't know what you do, Dr. Rahimiam.
Dr. Rahimian: First of all, we have cases that a patient had whole brain radiation because they have 50 or some lesions and then they come for [inaudible 00:29:27]. For those cases, we subtract 10% of those from the grid that we have to make sure they don't get necrosis. In terms of your question, basically, if V12 is more than 10 gray or something, we adjust the dose, we lower it. It depends on radiation. Oncologists, they try not to get into trouble with problems of overtreating. So they adjust it basically. Yeah. I'm hoping that, you know, the new version, we can actually lower that V12. And that's something right now we don't have that option. But we can actually set up our OARs and lower those values.
Dr. Nzhde: But is the PTV included in the V12 calculation?
Dr. Rahimian: No, we don't.
Dr. Nzhde: We don't. Okay. Nicole.
Nicole: Yes. Thank you for those excellent talks and the following V12, the session further, I guess from our own experience, it's not uncommon to have a patient come in with a couple of large mets plus maybe 10 small ones. Do SBRT larger. They may have 10 mets [inaudible 00:30:45] maybe do 2 iso with the Elements. And then we kind of struggle, or at least I struggle with how do I report that V12? We might do the SBRT over five fractions, one or [inaudible 00:30:56] Monday and one on Wednesday. So I was just curious if you had any...
Dr. Nzhde: It's interesting. At the last symposium, I had an example. I didn't want to be repetitive and show the same example where there was large targets. So in fact, we had a patient with three kinds of things, right? There was one large target that the physician wanted to do an SBRT. There were multiple small ones that he wanted to do multiple metastases. And then there was a very tiny one in a brain stem that he wanted to do conical [inaudible 00:31:25] treatment. And we did all that.
So coming back to the original question, the SBRT one, you could include that in your initial plan of MME. So let's say you have 5 targets, they all get 18 gray, and the SBRT target gets a 6 gray, and then another 4 fractions separately for that SBRT target. That's one thing you can do. So there are many approaches you can have. Still doesn't answer the sort of the question of V12, if the delivery is over multiple days, but that also applied when we were using multiple isocenters per target, right? So their view, in fact, actually I would claim it's better now because you had 16 targets and you were treating them in the span of a week on different days. Now, actually, you're treating them at the same time. So biologically, we weren't sure what the V12 meant then, now it's actually one day, right? But it's an excellent question.
Dr. Rahimian: That's a great question. About this issue of V12, the alternative, let's say if the patient has 16 lesions or 15, if you give whole-brain radiation like 10 times through 50, your V12 is active in the brain. So basically, this is the greatest option that you have. You can have neurocognitive function preservation compared with that one. So, you know, you have to take things in perspective.
Dr. Nzhde: But you could make a counterargument to what I just said and say that the experience of the V12 was based on the old approach where patients were being treated on multiple days. That also is valid. Yeah. Go ahead.
Man: My question actually sort of is along those lines of V12 being the old approach. We're talking about a paradigm shift in how we treat. And we do a lot of gymnastics to shoehorn our current approach.
Dr. Nzhde: Exactly.
Man: That this old rule that, sort of, had questionable value to begin with. So I'm wondering...I mean, you guys have a great wealth of data, what do you see as our roles in, sort of, helping the community to think about it differently? I find it very frustrating as a single person on an island trying to wrap my head around how could it possibly...you know, one little spot down here, plus one little spot down there, doing the same thing?
Dr. Nzhde: Until we have something better to follow, that's what we have. Ultimately, to answer your question, what we can do is clinical trials, right? We do a lot of stuff in radiation oncology that's not based on true evidence and knowledge. And some of this is part of that. Even the margin assignments that I'm claiming that we need to reduce, ultimately, you need sort of a unified approach clinical trial to determine that, right? So until we have that, this is what we have. But I agree with you.
Dr. Rahimian: Then the first criteria is you shall not harm the patient. Be conservative and safe, that's the first thing in medicine, right? So I think, in that sense, obviously [inaudible 00:34:52] study and [inaudible 00:34:54] studies are very helpful, they look [inaudible 00:34:59]. But as a physicist, we make sure that everything is done precisely and accurately and also conservatively. That's our approach.
Dr. Nzhde: In fact, just to continue on the answer, by reducing margins, we're trying to basically do treatments that were similar to our clinical experience, and not vastly different treatments for these patients. It may be simplistic to say, "Oh, we're using one iso and we're treating these patients and they're getting the same treatment." They're not. So your question is very, sort of, just invalid, and there needs to be clinical trials to answer that. Any more questions? So I'm going to introduce Bogdan who is the director of clinical affairs as well as director of Novalis Circle. Okay.
