Transcript
I'm here to talk about, in our initial experience in our institution, building the new elements cranial SRS 1.5. It's actually very similar to the previous presentation. First I want to start off by talking about some workflow challenges in our institution, the main one being the time period from the MRI to treatment. In our institution, we try to have treatment for metastases occur within seven days of the treatment planning MRI scan. Sometimes that MRI scan is done before the simulation, even before consult, so that puts a lot of pressure on our whole SRS team to get this done quickly, from the therapist who discovered these long treatments on the machine on short notice, to the physicist who has to turn around these plans quickly. Then the residents and the attendings need to draw the volumes and approve the plans quickly.
In addition in our institution, we have a limited number of authorized planners. It's our policy that only physicists are able to plan SRS and fSRS and iPlan due to the quality of plans in iPlan being very operator-dependent. And then after we come up with an awesome plan, the physician and the physicist have to stand at the machine for 40 minutes to 2 hours depending on the number of arcs, the number of isocenters being treated. While you go through a multitude of gantry angles, kV image checks, and while this is onerous for us, it's infinitely worse for patients who are immobilized on a hard table for an hour to two hours.
So, this brings me to what I perceive in our experience using Elements Cranial SRS for the last few months compared to iPlan. And the three major things we've noticed are there's a decreased planning time, there's an increased consistency in dosimetric quality of radiosurgery plans, as well as decreased treatment time. So, first, let's talk about decreased planning time and increased consistency in dosimetric quality with Elements Cranial SRS. They're collated, and they really come from three main factors. The template formation, automated inverse planning with a four-point algorithm, and a clever hybrid approach planning algorithm.
So, first, let's talk about the template formation. You can form templates. The first is a protocol template that can be created and saved, so if you treat small metastasis to 10 Gy regularly at the periphery, you can create a template with volumetric dose guidance and save that template so the planner can then use that in the future repeatedly. And after this is created, you would then pick a setup template. You can see this, I don't know if people can see the picture. But you can pick a setup template, which has the location of the tumor and the number of arcs, and after this is achieved the automated inverse planning occurs with the four Pi algorithm, and the four Pi algorithm yields ideal arc positions and lengths as well as table positions. The planner also has input with sliders for PTV vs OAR weighting, normal tissue sparring, modulation, and OAR dose weighting. And obviously, this makes this much less planner-dependent for the quality of the plan. And then the third factor, the hybrid planning algorithm, which is very clever, uses Pencil Beam for the majority of the early stages of optimization, and then the user or planner can switch and toggle onto Monte Carlo so the final calculations of your dosage to your OAR and your targets are more accurate.
Also, the new Elements has improved anatomical mapping over iPlan, as well as a better smart brush, as was mentioned previously in the day, which decreases physician volume contouring time. There's also MRI distortion correction, which we're looking at internally right now to see how much of a difference that makes in the dose around the target structure as well as the OARs. And here you can see a picture of the ability to toggle on Monte Carlo, and the ability of the planner to have some controlled automation with the sliders.
So, I'm just gonna go over a small sample of our patients that we use for internal QA with iPlan versus the new Elements. So, these patients were all treated in iPlan, and then we replanned them with Elements Cranial SRS. They consisted of seven vestibular schwannomas. Six of them were treated with 3d dynamic conformal arcs, one was treated with fixed-field IMRT. Three brain metastases cases we have treated. They were all treated with 3d dynamic conformal arcs as well. These were located, one tumor was located in the brain stem, and the two other ones were located abutting the motor strip.
This is a very small sample set, but we show that we found that the Elements had a significantly lower conformity index and a lower gradient index. After listening to Dr. [inaudible 00:05:42] talk, I'm not sure how clinically significant this is, but I do know that V10 and V12 are correlated with a range of necrosis. And this showed that V10 was reduced by more than 40% on average with these patients in Elements as well as V12 by around 50%. This is an example of one of the patients whose target was in the brainstem. He was treated the max point dose on iPlan that was 16 Gy as opposed to a less anxiety-provoking 12 Gy when re-planned in Elements. The brainstem mean dose also went down.
So, there's also an overall reduction to OARs for all the vestibular schwannoma cases, and significantly lower brainstem max doses as well as the cochlear doses for those that were drawn. We also have noticed faster treatment times. At Strong Memorial Hospital, my institution, we use ExacTrac for setup and verification at each couch for kV imaging. This usually takes about 40 minutes to one and one-half hours depending on how many arcs and isocenters, and most of that time is due to machine motion. Personel walking in on the machine, waiting for the gantry to move, kicking the couch, checking kV images. We've noticed with Elements Cranial SRS that these treatment times have been drastically decreased, probably by about half to less than 20 minutes with four to five arcs. And you can see most of that saving in time is due to decreased machine motion. There is some increased beam time due to the modulation of VMAT format, and there's some decreased imaging time.
So, here you can see, for the 10 patients that I was talking about previously, that when you plan in Elements versus iPlan there are fewer couch kicks. On average there's four couch kicks for the Elements Cranial SRS plan, and on average there are nine for the iPlan. So, finally, I want to talk a little bit about Elements Cranial SRS Monte Carlo implementation, which I think is important and cleverly done. We all know the value of Monte Carlo for cranial radiosurgery planning due to its improved accuracy for small fields and inhomogeneous media, especially at tissue interfaces, and complete Monte Carlo simulations are expected to result in the highest dose calculation accuracy. But due to the long time of calculations, this can be cumbersome and clinical implementation is not always feasible. But Elements gets around this by doing most of the calculations in Pencil Beam, and your final calculation can be done with Monte Carlo.
And so, here what we did was we took the seven vestibular schwannoma cases and we planned them using the Monte Carlo algorithm versus the Pencil Beam algorithm, and you can see here that the dose covering 99% of the target volume is quite different in some of these patients, probably due to the ones where it's at the surface interface, like patient number three. You can see here that the Monte Carlo shows 19.34 Gy covering 99% of the volume whereas the Pencil Beam shows 18 Gy covering 99% of the volume, which is an 8.3% difference. You can also see this difference in the OAR, the brainstem. You can see that the Monte Carlo brainstem max dose is 11.5, whereas the brainstem max dose with Pencil Beam is shown to be 9.38. Pencil Beam is 82% of the Monte Carlo calculated max dose.
So, in conclusion, in our institution, we feel that Elements Cranial SRS has improved consistent dosimetry over iPlan and it's less planner-dependent and decreases planning time. We theorize that the increased conformality, sharper dose fall-off in normal tissue, and improved avoidance of OARs will likely translate into a lower risk of late toxicity including necrosis and damage to adjacent structures for patients. Elements Cranial SRS incorporates Monte Carlo in a non-cumbersome, efficient manner. The biggest demand, I think, for a patient is probably the faster treatment time, which is probably almost half of iPlan.
I just want to thank BrainLab for inviting me to present, as well as Rebecca for helping me put the slides together, our entire SRS team at Unversity of Rochester, and our patients who bravely fight every day and who are the why to what we do. Thank you.
In addition in our institution, we have a limited number of authorized planners. It's our policy that only physicists are able to plan SRS and fSRS and iPlan due to the quality of plans in iPlan being very operator-dependent. And then after we come up with an awesome plan, the physician and the physicist have to stand at the machine for 40 minutes to 2 hours depending on the number of arcs, the number of isocenters being treated. While you go through a multitude of gantry angles, kV image checks, and while this is onerous for us, it's infinitely worse for patients who are immobilized on a hard table for an hour to two hours.
So, this brings me to what I perceive in our experience using Elements Cranial SRS for the last few months compared to iPlan. And the three major things we've noticed are there's a decreased planning time, there's an increased consistency in dosimetric quality of radiosurgery plans, as well as decreased treatment time. So, first, let's talk about decreased planning time and increased consistency in dosimetric quality with Elements Cranial SRS. They're collated, and they really come from three main factors. The template formation, automated inverse planning with a four-point algorithm, and a clever hybrid approach planning algorithm.
So, first, let's talk about the template formation. You can form templates. The first is a protocol template that can be created and saved, so if you treat small metastasis to 10 Gy regularly at the periphery, you can create a template with volumetric dose guidance and save that template so the planner can then use that in the future repeatedly. And after this is created, you would then pick a setup template. You can see this, I don't know if people can see the picture. But you can pick a setup template, which has the location of the tumor and the number of arcs, and after this is achieved the automated inverse planning occurs with the four Pi algorithm, and the four Pi algorithm yields ideal arc positions and lengths as well as table positions. The planner also has input with sliders for PTV vs OAR weighting, normal tissue sparring, modulation, and OAR dose weighting. And obviously, this makes this much less planner-dependent for the quality of the plan. And then the third factor, the hybrid planning algorithm, which is very clever, uses Pencil Beam for the majority of the early stages of optimization, and then the user or planner can switch and toggle onto Monte Carlo so the final calculations of your dosage to your OAR and your targets are more accurate.
Also, the new Elements has improved anatomical mapping over iPlan, as well as a better smart brush, as was mentioned previously in the day, which decreases physician volume contouring time. There's also MRI distortion correction, which we're looking at internally right now to see how much of a difference that makes in the dose around the target structure as well as the OARs. And here you can see a picture of the ability to toggle on Monte Carlo, and the ability of the planner to have some controlled automation with the sliders.
So, I'm just gonna go over a small sample of our patients that we use for internal QA with iPlan versus the new Elements. So, these patients were all treated in iPlan, and then we replanned them with Elements Cranial SRS. They consisted of seven vestibular schwannomas. Six of them were treated with 3d dynamic conformal arcs, one was treated with fixed-field IMRT. Three brain metastases cases we have treated. They were all treated with 3d dynamic conformal arcs as well. These were located, one tumor was located in the brain stem, and the two other ones were located abutting the motor strip.
This is a very small sample set, but we show that we found that the Elements had a significantly lower conformity index and a lower gradient index. After listening to Dr. [inaudible 00:05:42] talk, I'm not sure how clinically significant this is, but I do know that V10 and V12 are correlated with a range of necrosis. And this showed that V10 was reduced by more than 40% on average with these patients in Elements as well as V12 by around 50%. This is an example of one of the patients whose target was in the brainstem. He was treated the max point dose on iPlan that was 16 Gy as opposed to a less anxiety-provoking 12 Gy when re-planned in Elements. The brainstem mean dose also went down.
So, there's also an overall reduction to OARs for all the vestibular schwannoma cases, and significantly lower brainstem max doses as well as the cochlear doses for those that were drawn. We also have noticed faster treatment times. At Strong Memorial Hospital, my institution, we use ExacTrac for setup and verification at each couch for kV imaging. This usually takes about 40 minutes to one and one-half hours depending on how many arcs and isocenters, and most of that time is due to machine motion. Personel walking in on the machine, waiting for the gantry to move, kicking the couch, checking kV images. We've noticed with Elements Cranial SRS that these treatment times have been drastically decreased, probably by about half to less than 20 minutes with four to five arcs. And you can see most of that saving in time is due to decreased machine motion. There is some increased beam time due to the modulation of VMAT format, and there's some decreased imaging time.
So, here you can see, for the 10 patients that I was talking about previously, that when you plan in Elements versus iPlan there are fewer couch kicks. On average there's four couch kicks for the Elements Cranial SRS plan, and on average there are nine for the iPlan. So, finally, I want to talk a little bit about Elements Cranial SRS Monte Carlo implementation, which I think is important and cleverly done. We all know the value of Monte Carlo for cranial radiosurgery planning due to its improved accuracy for small fields and inhomogeneous media, especially at tissue interfaces, and complete Monte Carlo simulations are expected to result in the highest dose calculation accuracy. But due to the long time of calculations, this can be cumbersome and clinical implementation is not always feasible. But Elements gets around this by doing most of the calculations in Pencil Beam, and your final calculation can be done with Monte Carlo.
And so, here what we did was we took the seven vestibular schwannoma cases and we planned them using the Monte Carlo algorithm versus the Pencil Beam algorithm, and you can see here that the dose covering 99% of the target volume is quite different in some of these patients, probably due to the ones where it's at the surface interface, like patient number three. You can see here that the Monte Carlo shows 19.34 Gy covering 99% of the volume whereas the Pencil Beam shows 18 Gy covering 99% of the volume, which is an 8.3% difference. You can also see this difference in the OAR, the brainstem. You can see that the Monte Carlo brainstem max dose is 11.5, whereas the brainstem max dose with Pencil Beam is shown to be 9.38. Pencil Beam is 82% of the Monte Carlo calculated max dose.
So, in conclusion, in our institution, we feel that Elements Cranial SRS has improved consistent dosimetry over iPlan and it's less planner-dependent and decreases planning time. We theorize that the increased conformality, sharper dose fall-off in normal tissue, and improved avoidance of OARs will likely translate into a lower risk of late toxicity including necrosis and damage to adjacent structures for patients. Elements Cranial SRS incorporates Monte Carlo in a non-cumbersome, efficient manner. The biggest demand, I think, for a patient is probably the faster treatment time, which is probably almost half of iPlan.
I just want to thank BrainLab for inviting me to present, as well as Rebecca for helping me put the slides together, our entire SRS team at Unversity of Rochester, and our patients who bravely fight every day and who are the why to what we do. Thank you.