Bogdan: Welcome, everyone, to a new Novalis Circle webinar. The topic for today is cranial radiosurgery for treatment of vestibular schwannomas. Today's talk are a hybrid between a standard webinar and a case of the month in that we are reviewing three cases involving single fraction, three fractions, and five-fraction prescription regimens for patients with vestibular schwannomas. We are honored today to introduce, once again, the clinical team from Instituto Zunino in Córdoba, Argentina, and we will start with Dr. Sylvia Zunino, who will introduce the clinical presentation for the disease and review the certain strategies for deciding on a clinical treatment path. Dr. Zunino holds several academic appointments in Argentina. She is the head of the Instituto Zunino in Córdoba, and she serves as president for the Marie Curie Foundation. Following her lecture, we have Dr. Daniel Venencia, who is the head of physics at Instituto Zunino, and he will review the technical aspects of generating radiosurgery plans for the three cases. Dr. Venencia will also highlight practical aspects of utilizing Elements Cranial SRS and even touch on a few QA and commissioning topics.
As always, please remember to use Google Chrome or Safari to access the webinar. And should you have any connectivity issues, simply refresh the page. Utilize the chat line to send us questions. We will review and answer your questions upon completion of the two lectures. Follow the polling interface for questions that we would like to ask you. And should you choose to follow us on social media, please utilize #NovalisCircle. With that, I'd like to turn it over to Dra. Zunino for the first lecture.
Dr. Zunino: It is an honor for us to participate in this Novalis Circle meeting. We thank the organizers for the opportunity to show our institutional experience in the use of Elements Cranial SRS for vestibular schwannoma. Dr. Daniel Venencia will speak about the physics aspect, and I will deal with some medical concepts. As introduction, in this slide, we can see that one-fraction SRS Gamma Knife or LINAC with tumor marginal doses between 12 to 14 Gy revealed 5-year tumor control rate of 90%, 99%. The same results were obtained with SRT in 5 by 5, 3 by 10, 6 by 3. The results comparing SRS and SRT for tumor control, hearing preservation, facial nerve preservation, and trigeminal preservation were not significantly different.
In this trial, patients were selected for SRS or SRT techniques according to pretreatment hearing function, tumor size, and proximity to critical organs. Patients who have small tumor equal or less than 3 centimeters and non-serviceable hearing were usually selected for SRS 1 fraction of 12 to 13 Gy. For hypofractionation, the author prescribes different fractionation schemes, 3 by 10, 5 by 4, 6 by 3, or 5 by 5.
In this study, hearing was categorized as serviceable, poor, or no hearing. Serviceable hearing was defined as the capacity to use the phone unaided. Regarding the dose received by the cochlea, a significant difference in patients with preserved hearing was seen with a minimum cochlear radiation dose less than 5 Gy. Anthony Tolisano reported that tumor margin dose less than 12 Gy and cochlear dose less than 4 Gy have also been demonstrated to maximize hearing preservation rate. However, no consensus has been reached regarding the dose below which radiation poses no adverse effect to the cochlea.
The author classified changes in tumor size after SRS into five categories: enlargement requiring surgical intervention, which means local failure, stable or reduced tumor, initial enlargement with secondary reduction, initial enlargement that remains stable, and enlargement of 1 to 3 millimeter on last MRI with uncertain progression, suggesting a new assessment in one to two years.
This study investigates the efficacy and toxicity of a de-escalation of the marginal prescribed dose from 16 to 11 Gy. Ninety-seven patients were consecutively treated for a vestibular schwannoma with LINAC-based SRS from November 1995 to April 2019, with an 8-year follow-up. The results of local control at 3, 5, and 10 years comparing the different doses, retrospectively, was not significant under the non-inferiority criteria. It may be said that the weakness of these results of the de-escalation is a small number of patients analyzed in different periods, but the strong point of the trial is the long follow-up.
Our experience. From 2010 to 2020, we treated 120 patients who had diagnosis of vestibular schwannoma using SRS techniques, but I am going to refer only to 18 patients who were planned with Cranial System brand.
Here, you can see that 61% of the 18 patients planned with Cranial Systems received 13 Gy in 1 fraction, 28% 6 by 3, and 11% 5 by 5. These figures show the PTV volume and tumor size in cc and millimeters according to fractionation. These figures show the mean dose received by the cochlea where patients were treated with one or three fractions. Here, you can see the maximum and mean dose received by the brainstem in 18 patients. The dose was prescribed to 95% of the PTV. In this figure, you can observe that the symptoms analyzed were hearing, tinnitus, dizziness, and facial paralysis. Dizziness was the only symptom that improved at one year forward. Changes in tumor size after SRS in 15 patients having MRI: Stability or reduction of the tumor, 73%, initial enlargement that remains stable, 20%, initial enlargement with secondary reduction, 7%.
We will present three patients. In our practice, the organs at risk considered for each fractionation scheme are the following: the cochlea for SRS 1 fraction, the cochlea and brainstem for 3 fractions, only the brainstem for 5 fractions.
The first patient was a 54-years-old man with right vestibular schwannoma. SRS 1 fraction. Symptoms began a few months prior to consultation in our facility. Right hearing was serviceable. Tinnitus and dizziness were present. MRI in September 2019 showed a tumor of 11 millimeters in the right IAC. SRS 13 Gy in 1 fraction was delivered in October 2019. After one year of follow-up, the patient has a normal activity, dizziness disappeared, and hearing and tinnitus remained stable. No toxicity in facial or trigeminal nerves was observed. MRI in August 2020, the image showed initial enlargement. Volumes: GTV, vestibular schwannoma, PTV, 1 millimeter, cochlea received less than 4 Gy. MRI, August 2020, changes in the image at 10 months. We can observe the initial enlargement of the tumor.
The second patient presented here is a 66-year-old man with a schwannoma in the right side. SRS 3 fractions. Symptoms began in 2014. Tinnitus, no hearing, mild dysfunction of the facial nerve, no dizziness. MRI showed an expansive lesion measuring 30 millimeters in the right IAC. SRS 6 by 3 fractions given from April 22 to 24, 2019. After one year of follow-up, he is working and leading a normal life. He remains deaf. Mild dysfunction of the facial nerve and tinnitus remain stable. There are no new symptoms. The PTV is in contact with the brainstem and near the cochlea. Change in the image at 8 months, MRI, December 2019, necrosis or cystic formation is a question.
The last patient is 84 years old man with hearing loss in the right ear, which began 10 years before. In June 2019, he underwent surgical partial resection, and in August 2019, a residual tumor was observed. Symptoms at consultation in our facility were right facial nerve dysfunction due to surgery. MRI, May 2019, showed an extensive injury with cystic areas in cerebellopontine angle involving the 7th cranial nerve with bone erosion of the petrous bone and mastoid. The tumor displacing the brainstem and cerebellum measured 46 millimeters. The patient was treated with SRS 5 by 5 delivered from December 5 to 11, 2019. After one year of follow-up, the patient has normal activity, he is working, no changes in right facial nerve dysfunction, no hearing, and no new symptoms. This is the large cystic tumor displacing the brainstem and invading the bones. MRI, March 2020, we can observe changes in the image at 4 months.
In this publication, the enlargement of a preexisting cyst was observed in three patients.
In conclusion, the criteria for choosing SRS in 1, 3, or 5 fractions were tumor size and proximity to critical organs. In spite of the fact that this is a small Cranial series with a very short follow-up, we can report that, in all these patients, dizziness was the only symptom that improved. Tumor volume remained stable or reduced in 73% of the patients. Initial enlargement was observed in 20%. Enlargement with secondary reduction in 7%. Thank you for your attention.
Bogdan: Thank you for that great review, Dra. Zunino, and we'll switch now to Dr. Venencia for the follow-up talk.
Dr. Venencia: Thank you very much for inviting me to give this talk. I do not have any conflict of interest, and this would be my agenda. I'm going to go through the Elements Cranial SRS workflow, and I'm going to show some physicist aspects related to vestibular schwannoma radiosurgery, some practical consideration using Elements Cranial SRS. I'm going to show patient-specific result using different tools and some physicist aspects of the case presented in Dra. Zunino's talk.
First of all, I would like to mention that the radiosurgery program in our institution started in 2010. We use Primus LINAC, with micro-MLC, ExacTrac, and iPlan TPS. From 2012 to 2017, we installed 4 Novalis systems, 3 Tx and 1 TrueBeam STx, all LINAC beam-matched and all of them with ExacTrac capability. We move forward to BrainLab frameless system and image guidance for all our radiosurgery treatments. And from 2016, we are a Novalis-certified institution.
Okay. Well, the radiosurgery of vestibular schwannoma with MLC has the advantage of reducing the treatment time due to the use of a single isocenter. A dynamic conformal arc is one of the most used treatment technique with MLC because it is fast and with a very simple verification patient-specific grade. By using dynamic conformal arc, the MLC will play an important role, especially for a small target, as you can see in this comparison between 2.5, 5-millimeter, and 10-millimeter leaf width. The reduction of the MLC leaf width will improve quality in the system. And then you are able to see here a comparison in the DVHs.
For treatment planning of vestibular schwannoma, the cochlea and the brainstem are the main organ of risk, and in those treatments where the target is in touch with the organ at risk, inverse planning with modulated things will help to reduce the dose event and it also could improve the conformity and gradient images.
The Elements Cranial SRS is a workflow to manage all the treatment of several cranial diseases using a VMAT treatment technique approach. And the software is more than a planning tool since it has all essential steps in the process, which include images visualization, image fusion with cranial distortion correction, automatic organ at risk delineation, a SmartBrush drawing tool, object manipulation, and of course, dose planning.
The planning varies on clinical protocol and beam template. And for vestibular schwannoma radiosurgery, we have treatment protocol that you can see here, depending on the number of fractions. For them, we have to define the dose prescription. Generally, we use the same dose for the target and the boost. And also, you have to define the dose volume constraint for your organ at risk.
During the initial setup, the software will take into account in the objective function the superposition between the PTV and organ at risk and the radiological depth of the target. This will allow to select the most suitable arc setup dependent on the target localization. Then the couch angle and the gantry START and STOP angle will be modified using a 4Pi optimization objective function.
Here, we have an example. As you can see, the modulation helps to reduce the dose, in this case, to the left cochlea. Once we have the plan, the Monte Carlo volts calculation can be used, and if it is necessary, the optimization could be run again to account dose differences coming from the dose calculation algorithm.
Here, we have another example of the right target, and here, we have the conformity index and gradient index using 6X and HDMLC. If we go with the plan with dynamic conformal arc and compare them, we will see that the Elements plan improved the sparing of organ at risk and also the quality index. Now, if we go with the same plan but using a 5-millimeter leaf width MLC, we will see a detriment of the quality indices.
We started using Elements Cranial SRS in our institution on September 2018, and until today, more than 150 plans had been done. Vestibular schwannoma corresponds to 33% of the treatment site. And here, I am showing some statistical data corresponding to the conformity indices. This is the obvious value. And the following graph corresponds to the gradient indices value of all plans.
In our institution, for vestibular schwannoma radiosurgery, we use CT slice thickness of 0.6 millimeter and MRI images, T1 with contrast and T2. We fuse the CT and MRI using cranial distortion correction. The organs at risk are defined by anatomical mapping. The GTV is delineated on MRI images. And we use a PTV of 1-millimeter margin from the GTV. We use 6X proton beam, depending on the machine with type of 6X proton beam. And our desired prescription isodose is 95% and 99% for the PTV and GTV, and the tolerated isodose is 90% and 95%, respectively. We choose to have an SRS prescription with a homogeneous dose distribution. And the most important organ at risk generally is the cochlea for one fraction and the brainstem for three and five fractions. During the arc setup, we consider the eyes, the cochlea, and the full optic system. And we use a dose grid resolution of 1 millimeter.
During the arc setup, the accuracy threshold for X-ray verification is 0.5 millimeters and 0.5 degrees. We force verification after couch rotation. And you can see, in this video, we do the first setup using the lasers. Then we place the infrared head frame. And then the technologist use the hand frontal to get the initial patient position. After that, we do an X-ray acquisition, and we fuse the images, excluding anatomical regions with possible movements. While the fusion is verified by the radiation oncologist, the final shifts are calculated and transferred to the machine. And then we apply safety couch movement. You can see here, the movements are done by the machine, and while that gets ready, we do an X-ray verification, and we can start the treatment for that couch angle. And we'll repeat this process for each couch angle.
VMAT plan, the disadvantage often is a patient-specific QA. Due to the size of the target, we need a high-resolution dose distribution verification method. Patient-specific is a tool included into the Elements workflow. And we got to use several methods, and this is our tolerance criteria.
As you know, GAFchromic film is a method with a very high resolution. We use always film dosimetry for small size targets, typically in one fraction. We use axial or coronal film configuration, using the original treatment plan couch angle.
We use also Delta 4 phantom for medium and large size tumors, as long as enough number of detectors are included in the high-dose region. All arcs using this device should have the couch angles set to zero. And we also export the structures to the phantom's analysis software so we can compare measured and calculated DVHs.
We do also an independent 3D dose calculation on Eclipse, with AAA dose calculation algorithm, and both 3D dose distributions are exported and compared by Gamma analysis on the PTW VeriSoft software.
Portal dose image prediction calculated using Eclipse could be used as a patient-specific QA. Here, we have the portal dosimetry results on three Novalis TX with beam matching. And here, you have the passing rate, which are very similar. This confirms the reproducibility of the plan on different LINACs with same dosimetric characteristics.
As part of the Elements workflow, we can export the plan to an independent MU calculation software. In our case, we have RadCalc, and we use a tolerance level of 5% and an action level of 10%.
Okay. Now, this is the first case presented by Dra. Zunino corresponded to a right vestibular schwannoma in one fraction. Here, we used a 6X FFF beam, with a dose rate of 1,400 MU per minute. In this case, we used four arcs, and in this patient, the PTV is in touch with the right cochlea. So the right cochlea is the most important organ at risk and is a very small organ. Here, we have another view of the plan with the arc trajectories, and in this plan, as you can see here, only five leaves are involved in the beam modulation. So it's very important, the left width.
Here, we have the DRR setting and view on ExacTrac with target projected on it. Here, we define the mandible region as an area to take care of the fusion. And this is the final shift for treatment. The total treatment time for this patient was less than 20 minutes. We used film dosimetry for QA of this patient. Here is the result of the comparison between calculated and measured dose distribution for axial plane using the RIT software. And here, result, the same comparison with the film in the coronal configuration. Of course, both of them within tolerance. Of course, the use of Delta4 Phantom is not recommended for small target due to the insufficient number of detector included in the target region.
This is the comparison between the Elements 3D dose calculation and the Eclipse 3D dose calculation. And here, we have the comparison using a Gamma analysis, which shows an agreement of more than 98%. Here, we have the portal dosimetry result or the arcs are really intolerant. And it is important to emphasize that the configuration of the portal dose prediction algorithm on Eclipse should be done for small field sizes. And here, we have the result for the RadCalc software.
This is the treatment plan of a medium-sized right target in three fractions. Here, we use a 6X SRS beam with a dose rate of 1,000 MU per minute, and in this patient, the PTV is in touch with the brainstem generating a visible deformation. So the modulation is very important for this patient. Here is another view of the plan with the arc trajectory, and you can see here, in this 3D view, that in one of the arcs, the gantry start positions are different than the others due to the overlapping between the PTV and the cochlea, and also, for this arc and for the eyes. Here, we have the ExacTrac with the target projection and the final shift for treatment. And the total treatment time was less than 15 minutes for this patient.
We used film dosimetry for patient-specific QA, and here, the result of the axial [inaudible 00:31:36] elevation within the tolerance limit. It is important to use a threshold. We set it to 20% to focus the analysis in the high-dose region. But this third zone, it should be high enough to account the dose to the target but also to the organ at risk, in this case, the brainstem, that should be in this bridge. Keeping the film into axial setup with the treatment couch angle if a similar dose distribution that one would give to the patient. For medium-sized targets, the Delta4 Phantom can be used. Here, we can see that several detectors are in the target region. This device has diode detector every 5 millimeters in the central region and in two orthogonal planes. And the software calculates several 3D dose distribution based on dose measurement. Here, we have a Gamma comparison and the calculated versus measured DVHs. This is the comparison between the Elements and the Eclipse 3D dose distribution, and here, we have the Gamma analysis, which shows an agreement of more than 99%. Here is the portal dosimetry results or arcs within tolerance. And the RadCalc result.
This is the last case. It's a treatment plan of a large-sized right vestibular schwannoma in five fractions. We use 6X beam with a dose rate of 500 MU per minute. The most important organ at risk is the brainstem, with a high sparing and weighting to that organ at risk. And in this plan, we obtained 24.9, which is greater than the optimal dose constraint of 23, that's why this number is in red, but it's less than the mandatory constraint of 31. So in Elements, if you use this button, this configuration, you are allowed to modify the percentage of dose coverage or dose-volume [inaudible 00:34:06]. Generally, we want that to have 95% of the PTV with the prescribed dose, but this value could be modified following the radiation oncologist's requirement for the specific patient. Here is another view of the plan with the three arcs. For five-fraction treatment, we try to reduce the number of arcs, and as you can see, one of the arcs show a different gantry start position due to the overlapping between the PTV and the right arc. Here is the ExacTrac view. We defined the teeth and the mandible region as an area to take down all the fusion. And in this patient, the total treatment time was about 10 minutes or less than 15 minutes, including all couch positions.
Here is the film dosimetry result within the tolerance limit. This is the result using the Delta4 Phantom. As you can see, several detectors are in the target region. This is the Gamma comparison and this is...here, you can see the calculated and measured DVHs. This is the comparison between Elements and Eclipse on the result. We have a Gamma analysis which shows an agreement of more than 99%. Finally, this is the portal dosimetry result, which all arcs with passing rate higher than 99%.
So in conclusion, the use of Elements Cranial SRS for vestibular schwannoma allow us to follow an smooth workflow which include all necessary steps. The plan quality for one, three, or five fraction SRS meets the expected results, and the planners, generally, do not need to adjust any planning parameters during the optimization, and it is very important because the generated plan, less user dependent. The quality results are within tolerance level for radiosurgery treatment, and the QA results are reproducible. Small target requires high-resolution verification method. And patient-specific QA using independent 3D dose calculation is a very fast method and can replace in some case mentioned. Thank you very much.
Bogdan: Thank you so much for your talk, Daniel. And let's now go to a live question and answer session.