Transcript
Thank you for the introduction and I would like first to thank Reinhardt and company for inviting me to talk. And my collaborators for this study were M. Ayadi and P. Dupuis, who are physicists in my department and M.P. Sunyach is the physician. So, first, as an introduction, I would like to present to you our activity in brain stereotactic treatments. So, we started in 2010 with the acquisition of the CyberKnife, and from 2010 to 2017 we've been treating 4,000 patients. And from them, 56% of these patients were treated for brain metastases.
So, in 2017, we treated more than 700 patients and 444 patients were treated for brain metastases. Here are the main protocol we use. So, we use...for tumors less than 2cc, we use 1Gy, and 1 fraction of 20Gy when the isodose is 78% to 80%. And for larger tumors we used two different methods, either a fraction of 6 or 7 Gys, depending on the size of the lesions and its positions. So, in routine, we treat patients with multiple targets sequentially mainly because of the treatment plan with the Cyberknife, it would be...the treatment session, we want to treat all the treatments in the same time. So, consequently, the last target is treated remotely from the fusion MRI, which is problematic from a medical point of view.
In 2016, we acquired [inaudible 00:01:58] with the ExacTrac device, so if we were to perform making different table angles. So, we decided to examine the possibility to treat with multiplanar arcs and to perform multiple [inaudible 00:02:15] radiations. So, why we could do this instead of say, Linac? Because the treatment session time would be reduced, and because we could treat all the targets in the same time so we could also reduce the whole treatment duration, and because also, it could be a good alternative in case of breakdown of the Cyber, or in case of import delay. So, the objectives of the study were first to evaluate the performances of the dedicated treatment planning system, which is Elements, with a view to treating the different targets, and to compare it to Multiplan, and at the same time, we also needed to look at the operational comparisons between...with the VMAT in terms of efficiency.
So, for the first study, we took seven patients with two to five brain metastases. GTV-to-PTV margins were two millimeters, and PTV sizes were less than two ccs, so we used the one fraction protocol, and we tried to cover 96% to 99% of the PTVs. For the dosimetric studies, we used the Elements MultiMets version 1.5, and for the delivery, with the collimator Agility having MLCs of five millimeters at the hydro center. And we used the Multiplan TPS, version 5.2.1, and for the delivery, the CyberKnife equipped with fixed-sized cones.
So, here are the parameters for the planning with the Elements, so the Elements image is the dynamic conformal arc therapy, so we used 6 MV arcs, non-coplanar arcs of 160 degrees. We used five to six tables angles, with a single isocenter. The optimization is automatic and the algorithm was the Pencil Beam. But the Monte Carlo was also available on this. And planning time to get one solution was about five minutes. For Multiplan, we used multiple fixed beams, so there were over 200 beams in that plan, and we used two to three fixed-sized cones for each plan. So, again, algorithm was the Ray Tracing, and Monte Carlo was still available but we did not choose it clinically. And planning time was about two hours to get one solution.
A few words about Elements, but you have heard from many talkers about this CPS. So prescription protocol and the actual CT ballistic are in planning templates, which are apart from the TPS interface, so you've got an example here. The arcs are automatically generated, and the targets are automatically also assigned to specific arcs. And there's also automatic avoidance when two targets are too close, which is problematic with the [inaudible 00:05:33] treatment because you sometimes, with other TPS, you've got to kind of [inaudible 00:05:39]. So, it's really a good solution. So, as you see here, all these are automatic, and you have very low inter-operator variability with this treatment planning. Just very fast, just a screenshot of the interface, which is a protocol which we will set up to calculate [inaudible 00:06:01] solutions. So, that means, when the physician assists [inaudible 00:06:05] his controls, he can perform [inaudible 00:06:06]
So, for the treatment plan comparison, the dosimetric criteria we used the PTV Paddick Conformity Index and we compared the mean dose, and also, we compared the mean dose to the healthy brain which I believe can work well. And for the VMAT, we also compared delivery efficiency in terms of number of MUs, treatment time, and we automatically [inaudible 00:06:38] log files and in particular, gantry speed, mean segment area and dose rate.
Here are the results on the seven patients for the [inaudible 00:06:52] for the CyberKnife. So, as you can see the mean conformity index inferred for the two techniques. But the mean dose is higher for the CyberKnife compared to the DCAT, and also the variability is less important with the CyberKnife.
I tried to see if there was an impact of the tumor size or the number of tumor of patients, per patient, but it seems that tumor size has no impact on the results, and it seems also that when the number of lesions is less than five, you have a better conformity index with Elements, but the setting is too low so we have to make more patients to confirm that result.
On the actual brain, the actual brain is better spared with Multiplan. Now, I'm going to present you two patients. So, there's one who had four brain metastases. We used for planning the CAT, six arcs with five table angles. Here are the results in a healthy brain and our patient and at least, as you can see, the dose profile are more or less equivalent with the Cyberknife and not exceeding the maximum acceptable that is 25Gys, 20Gys divided by the isodose prescription was reached on each PTV. And there is more autonomy for dose profiles for the Elements, and the maximum dose is less important. Here are the 1.5Gy, the Elements with the spare [inaudible 00:08:34]
In detail, you have the conformity index. So half of the time it will be [inaudible 00:08:44] to have a conformity index than we find a better conformity index than the Multiplan. And the dose to the actual brain is slightly better with Elements.
Patient two had five brain metastases. We used five arcs and five table angulations. Here are the results on the image. So, here are those profiles are marked with the DCAT. But, here as you can see, the healthy brain is really...we have really better solutions with the DCAT, compared to CyberKnife.
Here are three out of five combination [inaudible 00:09:30] conformity index with Multiplan, but to the actual brain you can see the results are better with the dose, in particular, [inaudible 00:09:38] is twice lower in the DCAT compared to CyberKnife. About the delivery time, it will take 92 minutes with this plan, with treatment with the Cyber, while it would take 18 minutes for the Elements. But here, this time comprised the images here, it does not comprise images on the table rotation [inaudible 00:10:07] a few minutes.
Now, a small comparison with VMAT. In that case, we took eight patients, we used the Monaco treatment analysis to do this imagery. We used four to six arcs equally spaced. This TPS is using the Monte Carlo algorithm for [inaudible 00:10:32] sequencing parameters we used for the arcs. Here are the dosimetric results, so in terms of conformity index, the conformity is not significantly different between the two treatments. And again, if we look at the dose of a healthy brain, we see that the dose is considerable with the VMAT treatment. And also, the number of Mus.
Now, if we look at the delivery efficiency and the behavior of the VMAT, when delivering VMAT plans you see that the [inaudible 00:11:15] This is the statistic of patient two [inaudible 00:11:19] six arcs delivered for patient two. So, you see that the gantry is moving slowly from VMAT delivery, why those rates are higher compared to the DCAT. You see also that there's more fluctuations in the speed and dose, right? But you see that the segment area is more important in VMAT than in DCAT. About delivery time, the delivery time was more important in VMAT compared to DCAT.
So, as a conclusion, what we can say is that Elements MultiMets is a robust automatic planning system for treating multi metastases. The planning time is reduced compared to CyberKnife and VMAT, and this is a system that is nearly operator-independent. Also the delivery time is reduced, there are less MUs than with the two other techniques. We obtained a similar conformity index and we observed that the healthy brain was better spared, in particular if we look at P5. It could be interesting to investigate the impact of tumor location and tumor size on the number of PTVs, so we need more statistics to be investigated for this factor, to see if it impacts the results. However, with this system we have more heterogeneities in dose distribution profiles to the PTV versus the other techniques. In the other techniques we [inaudible 00:13:07] PTVs, so we cannot say if it has a clinical impact or not, but it seems that when looking at the other presentation of yesterday, it could have more low-impact. So, in that case, this is, Elements is the performing tool, [inaudible 00:13:27] tool in most of the cases. But Cyber and VMAT will be kept in cases of...if we need highly modulated beams, and when a PTV is very close to the critical organ [00:13:47]. And I thank you for your attention.
So, in 2017, we treated more than 700 patients and 444 patients were treated for brain metastases. Here are the main protocol we use. So, we use...for tumors less than 2cc, we use 1Gy, and 1 fraction of 20Gy when the isodose is 78% to 80%. And for larger tumors we used two different methods, either a fraction of 6 or 7 Gys, depending on the size of the lesions and its positions. So, in routine, we treat patients with multiple targets sequentially mainly because of the treatment plan with the Cyberknife, it would be...the treatment session, we want to treat all the treatments in the same time. So, consequently, the last target is treated remotely from the fusion MRI, which is problematic from a medical point of view.
In 2016, we acquired [inaudible 00:01:58] with the ExacTrac device, so if we were to perform making different table angles. So, we decided to examine the possibility to treat with multiplanar arcs and to perform multiple [inaudible 00:02:15] radiations. So, why we could do this instead of say, Linac? Because the treatment session time would be reduced, and because we could treat all the targets in the same time so we could also reduce the whole treatment duration, and because also, it could be a good alternative in case of breakdown of the Cyber, or in case of import delay. So, the objectives of the study were first to evaluate the performances of the dedicated treatment planning system, which is Elements, with a view to treating the different targets, and to compare it to Multiplan, and at the same time, we also needed to look at the operational comparisons between...with the VMAT in terms of efficiency.
So, for the first study, we took seven patients with two to five brain metastases. GTV-to-PTV margins were two millimeters, and PTV sizes were less than two ccs, so we used the one fraction protocol, and we tried to cover 96% to 99% of the PTVs. For the dosimetric studies, we used the Elements MultiMets version 1.5, and for the delivery, with the collimator Agility having MLCs of five millimeters at the hydro center. And we used the Multiplan TPS, version 5.2.1, and for the delivery, the CyberKnife equipped with fixed-sized cones.
So, here are the parameters for the planning with the Elements, so the Elements image is the dynamic conformal arc therapy, so we used 6 MV arcs, non-coplanar arcs of 160 degrees. We used five to six tables angles, with a single isocenter. The optimization is automatic and the algorithm was the Pencil Beam. But the Monte Carlo was also available on this. And planning time to get one solution was about five minutes. For Multiplan, we used multiple fixed beams, so there were over 200 beams in that plan, and we used two to three fixed-sized cones for each plan. So, again, algorithm was the Ray Tracing, and Monte Carlo was still available but we did not choose it clinically. And planning time was about two hours to get one solution.
A few words about Elements, but you have heard from many talkers about this CPS. So prescription protocol and the actual CT ballistic are in planning templates, which are apart from the TPS interface, so you've got an example here. The arcs are automatically generated, and the targets are automatically also assigned to specific arcs. And there's also automatic avoidance when two targets are too close, which is problematic with the [inaudible 00:05:33] treatment because you sometimes, with other TPS, you've got to kind of [inaudible 00:05:39]. So, it's really a good solution. So, as you see here, all these are automatic, and you have very low inter-operator variability with this treatment planning. Just very fast, just a screenshot of the interface, which is a protocol which we will set up to calculate [inaudible 00:06:01] solutions. So, that means, when the physician assists [inaudible 00:06:05] his controls, he can perform [inaudible 00:06:06]
So, for the treatment plan comparison, the dosimetric criteria we used the PTV Paddick Conformity Index and we compared the mean dose, and also, we compared the mean dose to the healthy brain which I believe can work well. And for the VMAT, we also compared delivery efficiency in terms of number of MUs, treatment time, and we automatically [inaudible 00:06:38] log files and in particular, gantry speed, mean segment area and dose rate.
Here are the results on the seven patients for the [inaudible 00:06:52] for the CyberKnife. So, as you can see the mean conformity index inferred for the two techniques. But the mean dose is higher for the CyberKnife compared to the DCAT, and also the variability is less important with the CyberKnife.
I tried to see if there was an impact of the tumor size or the number of tumor of patients, per patient, but it seems that tumor size has no impact on the results, and it seems also that when the number of lesions is less than five, you have a better conformity index with Elements, but the setting is too low so we have to make more patients to confirm that result.
On the actual brain, the actual brain is better spared with Multiplan. Now, I'm going to present you two patients. So, there's one who had four brain metastases. We used for planning the CAT, six arcs with five table angles. Here are the results in a healthy brain and our patient and at least, as you can see, the dose profile are more or less equivalent with the Cyberknife and not exceeding the maximum acceptable that is 25Gys, 20Gys divided by the isodose prescription was reached on each PTV. And there is more autonomy for dose profiles for the Elements, and the maximum dose is less important. Here are the 1.5Gy, the Elements with the spare [inaudible 00:08:34]
In detail, you have the conformity index. So half of the time it will be [inaudible 00:08:44] to have a conformity index than we find a better conformity index than the Multiplan. And the dose to the actual brain is slightly better with Elements.
Patient two had five brain metastases. We used five arcs and five table angulations. Here are the results on the image. So, here are those profiles are marked with the DCAT. But, here as you can see, the healthy brain is really...we have really better solutions with the DCAT, compared to CyberKnife.
Here are three out of five combination [inaudible 00:09:30] conformity index with Multiplan, but to the actual brain you can see the results are better with the dose, in particular, [inaudible 00:09:38] is twice lower in the DCAT compared to CyberKnife. About the delivery time, it will take 92 minutes with this plan, with treatment with the Cyber, while it would take 18 minutes for the Elements. But here, this time comprised the images here, it does not comprise images on the table rotation [inaudible 00:10:07] a few minutes.
Now, a small comparison with VMAT. In that case, we took eight patients, we used the Monaco treatment analysis to do this imagery. We used four to six arcs equally spaced. This TPS is using the Monte Carlo algorithm for [inaudible 00:10:32] sequencing parameters we used for the arcs. Here are the dosimetric results, so in terms of conformity index, the conformity is not significantly different between the two treatments. And again, if we look at the dose of a healthy brain, we see that the dose is considerable with the VMAT treatment. And also, the number of Mus.
Now, if we look at the delivery efficiency and the behavior of the VMAT, when delivering VMAT plans you see that the [inaudible 00:11:15] This is the statistic of patient two [inaudible 00:11:19] six arcs delivered for patient two. So, you see that the gantry is moving slowly from VMAT delivery, why those rates are higher compared to the DCAT. You see also that there's more fluctuations in the speed and dose, right? But you see that the segment area is more important in VMAT than in DCAT. About delivery time, the delivery time was more important in VMAT compared to DCAT.
So, as a conclusion, what we can say is that Elements MultiMets is a robust automatic planning system for treating multi metastases. The planning time is reduced compared to CyberKnife and VMAT, and this is a system that is nearly operator-independent. Also the delivery time is reduced, there are less MUs than with the two other techniques. We obtained a similar conformity index and we observed that the healthy brain was better spared, in particular if we look at P5. It could be interesting to investigate the impact of tumor location and tumor size on the number of PTVs, so we need more statistics to be investigated for this factor, to see if it impacts the results. However, with this system we have more heterogeneities in dose distribution profiles to the PTV versus the other techniques. In the other techniques we [inaudible 00:13:07] PTVs, so we cannot say if it has a clinical impact or not, but it seems that when looking at the other presentation of yesterday, it could have more low-impact. So, in that case, this is, Elements is the performing tool, [inaudible 00:13:27] tool in most of the cases. But Cyber and VMAT will be kept in cases of...if we need highly modulated beams, and when a PTV is very close to the critical organ [00:13:47]. And I thank you for your attention.