- 773
Transcript
Bogdan: Hello, everyone, and welcome to our first Novalis Circle Case of the Month webinar. My name is Bogdan Valcu. I'm the director of Novalis Circle. I also serve as director of clinical affairs at Brainlab.
Today, we are introducing a new concept, reviewing a clinical case aimed at providing further in-depth knowledge into the expertise behind making clinical decisions for various radiosurgical treatments, review the clinical evidence for a treatment path that has been selected, and summarize relevant technological steps in generating an ideal treatment plan.
The Case of the Month webinars are a precursor to the Virtual Tumor Board Series that we'll introduce next year. And we hope to continue to provide you with valuable clinical expertise as you continue to expand your programs. For the first Case of the Month webinars, I have selected a few cases that I find interesting for the community at large. But please feel free to send me your suggestion for future events.
Today, I have the privilege to introduce you to our long-time partners at the University of California in Los Angeles. And the clinical team would be reviewing a spine metastasis patient. As best clinical decisions are derived from multidisciplinary interactions, we have today presentations from radiation oncology, neurosurgery, and medical physics.
Dr. Tania Kaprealian is a radiation oncologist and an associate clinical professor at UCLA. And she will begin with a case presentation. Dr. Kaprealian is the chief of the CNS Services and also serves as the medical director at the main campus in Westwood, LA.
Joining her is Dr. Luke Macyszyn from the department of neurosurgery. And he will be addressing the value of minimally-invasive surgical procedures. Dr. Macyszyn is an assistant professor of neurosurgery, radiation oncology and orthopedics, and he's also the associate program director of the Neurosurgery Training Program.
To complete the team and discuss treatment planning and IGRT considerations, we have Dr. Nzhde Agazaryan, who is a professor of radiation oncology at UCLA. Dr. Agazaryan is the chief of Medical Physics and the Dosimetry, and also serves as a quality and safety officer.
As true for the regular webinars, we continue to provide CE credits for the Case of the Month webinar as well. Should you need CAMPEP, MDCB or ASRT credits, upon successful completion of the live event, please email us at info@novaliscircle.org for further details.
And please register for our webinar during ASTRO's Online Industry Expert Theater on October 26. We have an exciting lineup of radiation oncology and medical physicist from Munich and Brussels presenting on their first treatments with ExacTrac Dynamic for both cranial and extracranial indications. And you'll have the opportunity to learn and ask questions on the new IGRT capabilities for the new system.
Lastly, please remember to use either Google Chrome or Safari. And should you have any internet challenges, simply refresh the page. Send us questions via the chat interface. We will answer all your questions upon completion of the three lectures.
Monitor the polling infrastructure as we may like to ask you some questions. And should you choose to follow us on social media, please utilize the #NovalisCircle.
And now, we'd like to turn it over to the UCLA team for their Case of the Month Review.
Dr. Kaprealian: Thank you, Bogdan. These are our disclosures. And today, we'll be discussing a complex metastatic spine case. We'll discuss the surgical considerations as well as the radiosurgical considerations, and then we'll review the treatment planning.
So, to start with the case, the patient is a 49-year-old woman with a history of metastatic left breast infiltrating ductal carcinoma. She presented with a progressive spinal T10 lesion with concern for cord compression.
Her symptoms consisted of worsening pain over the course of two months in the mid-low back at the T10 level that radiated to the mid-to-low abdomen. And it was very painful for her to walk. She denied any extremity pain or numbness. And she had no neural findings on exam and no pain on palpation of her spine.
An MRI was obtained and, as you can see here, there is a T10 lesion that involves the entire vertebral body and posterior elements. There's circumferential epidural tumor surrounding the thoracic cord and it causes at least moderate circumferential cord compression and contour deformity with cord edema.
So, on spinal assessment, she was determined to be a Bilsky Grade 3 with a SINS score of 9. And in the NOMS criteria, she had high-grade cord compression, she had a radiosensitive tumor. There was question of stability of the spine. And we knew that she would be able to tolerate a surgery.
I'll now hand it over to my colleague, Dr. Macyszyn.
Dr. Macyszyn: Hi, everyone. So, I guess, the first thing to discuss, which is kind of critical to all these cases, the decision making, a lot of it hinges on the epidural spinal cord compression, which is frequently called the Bilsky Grade from Dr. Bilsky that created this grading system, allowing us to more kind of quantitatively talk to each other to determine what the level of spinal cord compression is. This grading system goes all the way from 0 to a 3.
In a 0, basically, the tumor is confined strictly to the vertebral body and there is no epidural extension. In a Bilsky 1a, there is some epidural extension and there's tumor in the epidural space, but there's no deformation of the thecal sac and certainly not of the spinal cord. In a 1b, there is deformation of the thecal sac. In the 1c, there is a budding of the tumor or the epidural component of the tumor with the spinal cord. In a Grade 2, the disease kind of deforms the spinal cord and starts to push or displace the spinal cord, either to one side or the other or dorsally. And in the Grade 3, there's complete kind of obliteration of the CSF space with compression of the spinal cord.
Here on this slide, we have kind of clinical, you know, pictures, radiographic MRI showing all these grades all the way from a 0 to a 3. And you can see there's progressive involvement of the epidural space, the spinal canal, and progressive compression, deformation of the spinal cord.
Basically, we divide this up into low-grade epidural spinal cord compression and high-grade. And once again, this helps us with our decision making to determine if surgery is or isn't necessary and what kind of radiosurgery we would be able to do or not to do.
Basically, anything from a 0 to a 1b is considered low grade epidural spinal cord compression. And a grade from 1c to a 3 is considered high grade epidural spinal cord compression. Obviously, the most obvious ones are the grades 0, and the grades 2s and 3s. And probably, the one that's kind of the most debatable and kind of on the fence is the grade in 1C.
There's also the spinal instability scale, which is called the SINS classification systems, which is detailed here. It allows us to look at, evaluate the radiographic images of the patient as well as their symptoms. In this case, it's mostly just pain. And to determine how stable or unstable potentially a patient spine is due to the involvement of the spinal column with the tumor.
This goes over the location of the tumor, whether the patient has pain. If there's any bone lysis, alignment of the spine, vertebral collapse, and involvement of the spinal elements. Obviously, as you get more and more of these, you score more and more points, and making you more and more potentially unstable, therefore, requiring surgical stabilization either before and after radiosurgical treatment.
So, anything in the SINS scale from 0 to 6 is considered as stable spine, from 7 to 12 is the harder part. It's kind of indeterminant range. There might be instability or there might not be instability. And obviously, anything from 13 and higher usually is considered unstable.
This is a grading system, once again, that was designed so radiologists and physicians are like can communicate not only the radiographic cord compression regarding patients with metastatic spine disease, but also the potential stability of the spine. And we advocate a kind of a surgical consultation, obviously, for anybody who has greater than a SINS score of 7.
With regards to the NOMS criteria, this is kind of the overall picture and kind of paradigm that we use to treat patients. We've kind of get their neurological status, their oncological status. Once again, there's the mechanical stability of the spine, which we just talked about, with the SINS criteria. And obviously, systemically, how are they doing? Are they able to tolerate surgery? Are they not able to tolerate surgery? And I'll go over this in a little bit more detail in the following slides.
This is kind of the NOMS framework. And it's pretty easy and kind of straightforward. Obviously, first, I've kind of highlighted all the unstable, you know, mechanical pathologies that might present in the patients. Obviously, if a patient is considered unstable and they are able to tolerate surgery, we do surgery followed by radiation therapy. If they're unstable and radioresistant and they are not able to tolerate surgery, obviously, we will proceed with palliative radiation only. For patients who have a stable spine and has a low-grade epidural spinal cord compression, obviously, in that case, we would proceed with either classic radiotherapy treatment or stereotactic radiosurgery.
Finally, for the radiosensitive tumors that are stable, and even with a high-grade epidural spinal cord compression, so the things that come to mind is stuff like lymphoma and multiple myeloma. In those instances, we would select radiotherapy as the treatment without the necessity of spinal decompression, even though there's high-grade epidural spinal cord compression.
For radioresistant tumors now, and this is kind of self-explanatory, that have high-grade spinal cord compression, we do something called spinal cord separation surgery first, followed by radiation treatments to once again enable, you know, higher doses of radiation to the tumor without negatively affecting the integrity of the spinal cord.
So, you know, there's a number of ways we can proceed here. Do we do surgery alone? Surgery followed by radiation therapy? Just considerations for the radiation therapy plan and as well as pain management and a patient who presents with once again metastatic spine disease, potential instability, and pain.
We discussed this patient in our radiosurgery conference and the recommendation was made to proceed with separation surgery due to the epidural spinal cord compression of this tumor, followed by adjuvant SRS. So, we performed basically a T9-T10 laminectomy, and resection of the epidural tumor.
So, why separation surgery? Separation to this is probably one of the, you know, the classic and most kind of talked about trials, certainly in spinal surgery, but definitely in our radiosurgery board as well. This is from an older trial performed in 2005 by Dr. Patchell. This was published in Lancet that basically spinal cord decompression, followed by radiosurgery, leads to a better outcome than either modality on its own.
And the caveat here, obviously, is that these patients, for the most part, we select the patients with one single, you know, epidural or one single lesion to the spine that had epidural compression. And, obviously, their ability to walk was assessed at the conclusion of this trial as various other neurological function.
And you can see here that these are the patients that were signed up. And patients that had surgery with radiation fared better than patients that just had radiation alone. So, in this picture, here you see once again a spinal column with a spinal tumor metastasis to one of the vertebrae and, in this case, leading to spinal cord compression.
In a separation surgery, in this case, demonstrating fusion as well. Pedicle screws are placed, a laminectomy is performed, and we performed a partial corpectomy, once again to circumferentially decompress the spinal cord, as indicated here, and remove this tumor away from the thecal sac to reestablish once again that margin of CSF around the spinal cord, and obviously enabling, you know radiosurgery to a much better extent following surgery.
Here, once again, a cage is placed and the patient's spine is stabilized. And in this case, once again we're just kind of showing the benefits of stereotactic radiosurgery treatment. This was published in the journal of neurosurgery, "Spine" as compared to regular or more conventional hypofractionated SRS. And you can see that patients with a single dose SRS have a lower cumulative incidence of recurrence at the site of their disease.
Dr. Kaprealian: As Dr. Macyszyn described, this patient underwent a separation surgery. And the reason we decided to go that route for her, she was symptomatic, she had cord edema with cord compression, and we wanted to give her a hypofractionated radiation course postoperatively. And as you can see here, she also did have a pretty aggressive tumor because even in the short time span between surgery and this MRI, she had an interval, we can see the interval decompression at the T10 level but she had increase in the ventral epidural tumor as extending superiorly from T10 and as well as extension to the bilateral T9-T10 neural foramina.
And in addition, she has the soft-tissue component. There's a large prevertebral tumor at T10, displacing the aorta, and persistent cord edema still between T9 and T11 levels. So, the patient underwent CT stimulation two and a half weeks later for planning for post-operative stereotactic radiosurgery.
What are some indications to treat spine mets with SRS? Radiosurgery has been shown to have long-term rates of tumor control, as high as 90%, both in the salvage setting, after receiving traditional external beam, and as primary management.
And what SRS can accomplish is pain control, improvement or preservation in neurologic function, as well as possibly, in situations with spinal cord compression, it can be of benefit. But the issue is what is the correct target for spinal SRS? And this is not only important for capturing the tumor itself but it also plays a role if you have to come back and re-irradiate. It's really important the choice of the initial target volume delineation.
And this is just a paper showing several studies of postoperative SRS and the local control rates. And you can see that it ranges from 80% to the mid-90%. So, it's really an excellent treatment modality postoperatively.
So, what are some of the issues with postoperative spinal SRS? There's the impact of surgical hardware on the CT and MRI as well as for treatment planning. And then, the target delineations are more complex, both not only the target itself, the tumor itself, but also the critical neural structures.
So, as most of you know, there are consensus guidelines that are developed for the intact spine and recently there are consensus guidelines that have been developed for spine SRS in the postoperative setting. And what this takes into consideration is the preoperative tumor plus adjacent anatomic compartments.
So 9 radiation oncologists and 1 neurosurgeon with spinal SRS expertise contoured 10 cases. And what they recommended is that the CTV should include the entire preoperative extent of bony and epidural disease, plus any immediate adjacent bony anatomic compartments that are at risk of microscopic disease extension. A donut-shaped CTV is used in cases of preoperative circumferential epidural extension, regardless of the amount of residual epidural extension that's left. And spinal instrumentation is consistently excluded from the CTV. So this is very important because this is a common question that comes up.
So, in addition to the written guidelines, they have image guidelines that you can use to help correlate based on where the tumor is located in your patients' vertebral body, and then use this to help you contour the vertebral body accordingly.
So, we planned our patient in Eclipse. However, we are now commissioning Elements Spine SRS. So we would like to review our preplanning steps using this Elements program.
So first, what you want to do is a cross-modal rigid image fusion. And as you can see, in this case, that there's a difference in the curvature between the MRI and the CT. And this is a challenge for all applications. And you really have to be careful when you co-register. So, what you would have to do in this specific situation is designate a region of interest manually and define that so that you can do the image fusion in that specific area. And it basically correlates to the bony anatomy in that tumor region only.
So, this results in great registration in that area, but the rest of the spine will be off. So, for patients with tumor involvement in multiple levels, you would have several regions of interest and multiple registrations.
So, to solve this problem, Brainlab provides a curvature correction elements. And so, for this patient, we use the CT as the reference because the treatment plan is generated on the CT. And we allow the distortion algorithm to modify the MRI to match the anatomical position of the CT. This process doesn't require a manual definition of a region of interest. But unlike what we've shown here and done here, it's best to use a full MRI scan rather than a partial MRI.
And so, then the elastic deformation is anatomically penalized to not allow the bone to be deformed. And as you can see, it manages to correct for the significant cord curvature, which has significant dosimetric implications on the spinal cord dose. And Dr. Agazaryan will go in this further in his portion of the talk.
Another very important time-saving step for the physicians is the normal anatomy contouring. So, this program uses the universal Atlas element and it contours the normal anatomy organs. And then, the physician just has to come and check and approve the contours. Of course, manual contours can be added, if appropriate, based on the spinal cord level. But a lot of the organs are included in this automatic segmentation.
And then, finally, what's really critical for the physicians is to define the target. And so, with the Brainlab Spine SmartBrush Elements, the physician will contour out the GTV. And then, the program has the capability of developing the CTV expansion automatically using the intact spine control guidelines. So, in the postoperative setting, you might have to do manual adjustments, but for the most part, it's following the consensus guidelines and they're pretty similar.
So even though Brainlab provides this tool for single level, it is possible to use this workflow for a multilevel disease, as our patient had with T9 and T10. So, first, you would need to define the GTV at your first level. And you can do this with quick contouring in the axial view and in one reconstruction, and then the software will generate the CTV. This segment interactive then allows you to either add or subtract components as necessary without having to manually contour any further.
And once this level is complete, you then have to define the initial CTV as an object type called undefined, and then you exit and reenter the SmartBrush Element. So then you would have to repeat this step for the second level, manually contouring the GTV, and then the program will develop the automatic CTV and then you would again mark it as undefined. And then, finally, the last step would be to basically merge or union those two contours and you have the Object Manipulator elements to help you perform that.
And then, the last step is the PTV definition. But before we get to that step in our patient, remember that this patient also had a soft-tissue component. And that's not part of the consensus guidelines because those guidelines focus on the bony anatomy. And so, but what you can do is you can use this program to manually incorporate that soft-tissue components into your CTV. And then, what we do is we do a one-millimeter PTV extension.
Here is the patient's plan in Eclipse, and this is the final plan that we generated. We gave 6 gray times 5 for a total of 30 gray. And here again, you can see the patient's plan at the multiple different levels.
So, the patient tolerated the treatment well. As I mentioned, she received 30 gray and 5 fractions. And over time, she had significant improvement in her pain and she was able to taper off her steroids and pain medications.
So, this is the two-month follow up thoracic spine MRI. She had interval decrease in size of the epidural mass at T10 as well as improvement in that paraspinal soft tissue mass that was from T9 to T11, and an improved cord edema at those levels as well, and then resolution of the cord compression.
And then, again, at the five-month mark, she had no evidence of progressive bony metastatic disease, no compromise of the spinal canal, and an interval resolution of that soft-tissue mass. And now, I'd like to hand it over to my colleague, Dr. Agazaryan.
Dr. Agazaryan: Thank you for the opportunity to incorporate my comments for this Case of the Month. These are my disclosures. Dr. Kaprealian and Dr. Macyszyn already presented on the details of Eclipse treatment plan.
Here I highlight some of the TG101, those endpoints that will be used to compare with the Brainlab Spine Element later on. The cord was redrawn to end six millimeters superior and inferior of the overall target. The inferior part of the cord extends further down in this coronal view because of the inferior extension of the tumor seen on a 3D view here.
We have also created two other cord structures that correspond to each of the vertebral bodies, T9 and T10. The dosimetry of the cord counter corresponding to the overall target meets TG101 constraint.
This, for my display here, is done with scripting after the Eclipse plan is finalized. Similarly, T9 and T10 cord subvolumes, individually taken, meet the TG101 constraints. Clearly, the cord sections, separately evaluated, are more conservative approach for the 10% criteria. While experimenting with the Spine Elements, we experienced excellent graphical user interface and results.
I showed the deformable fusion for this particular patient with the deformation heatmap. Deformable registration uses anatomical mapping and tissue classification with deformable and nondeformable tissues. This is very similar to biomechanical registration. Obviously, this is a multilevel spine co-registration that can be done with MRI-CT or CT-CT while compensating for different spine curvatures.
The basic principle here is to automatically calculate individual rigid fusions for each of the vertebra. And from these results, interpolate a single 3D deformation field that simultaneously matches all vertebra in the fused images. Tissue model is also used in this process. It is essentially an iterative process between these two, where the integrity of the bony anatomy is maintained.
So, we just discussed the ability of doing a deformable fusion in Elements. We have transferred all the images to Elements and redrawn the spine contour on the T2 image in Elements. You can see that an appreciable difference is seen in the dosimetry of these two objects, one based on a rigid fusion and the other is based on deformable fusion. Deformed caused those maximum was greater in this case. We will come back to this discussion in the tissue tolerance section of the presentation.
Once the structures are redrawn, the planning process commences. The optimization and the calculation can be done both with Pencil Beam as well as with the Monte Carlo. I display the Monte Carlo toggle switch here.
This toggle serves two purposes. First, it does forward calculation with two algorithms. So, the distributions, the DVHs, 3D Display can be displayed either with Pencil Beam or Monte Carlo. The optimization itself does not change. Or, if you toggled during the optimization, then corresponding algorithm is used for optimization process.
The first optimization should always be Pencil Beam algorithm. Monte Carlo optimization is recommended for the last refining stages.
Here I show the results from Elements plans. The first one is the Pencil Beam optimization and a Pencil Beam calculation. The Pencil Beam plan achieved 99% target coverage. Then, Monte Carlo forward calculation helps to check the Pencil Beam calculation with Monte Carlo algorithm. Since we find, in this case, that the Pencil Beam underestimates the cord dose since we don't meet the criteria with Monte Carlo. Then, we proceeded with Monte Carlo optimization.
The Monte Carlo optimization itself meets the criteria with the coverage of 97%. Then we renormalize the coverage so that the 99% of the target is covered with a prescription and it still meets the TG101 criteria.
So, all in all, keeping the cord endpoints approximately the same, 99% target coverage is achieved with Elements plan. It is difficult to make general conclusions with selected cases. In fact, plan comparisons are often arguable, even with multiple cases, because of variations in planning techniques.
However, since we are discussing this specific case, it was decided that presenting plan comparison is a fair approach. Here's the Monte Carlo optimized, a Monte Carlo calculated treatment plan.
The planner gets real-time feedback from the software during the planning process versus running the script at the end of the planning process shown during the Eclipse calculation. So, here with the traffic light approach, you can incorporate the TG101 constraints and give a real-time feedback to the planner.
Here we show Eclipse plan versus Spine Elements Monte Carlo calculation. The Spyglass shows the Eclipse plan. The Element plan is more conformal and has a rapid fall off in this particular case towards the cord. More detailed studies with more patients need to be conducted for conclusive results.
I should mention that the Spine Elements conformity is achieved by PTV splitting and arc-duplication approach. The user can restrict the total number of arcs included in duplications. Our plan started with one arc, and then it was duplicated to three arcs. Spine metastases are located within vertebra, which are concave structures presenting challenges for VMAT optimization.
Splitting PTV allows for creation of sub PTVs with much less concave regions. As I show here in this example. Even by splitting into two parts, the concavity of the structures are minimized. This allows for much better conformity, the sparing of the cord.
In our specific example, PTV splitting and arc duplication resulted in three arcs covering different segments of the PTV. Brainlab Element Spine Radiosurgery Module also has a radiosurgical prescription mode, which does not restrict the tumor dose heterogeneity. Hence, it allows for greater gradient fall-off into the spinal cord.
So, in the normal mode, planning generates highest homogeneity while maintaining conformity indices and gradient indices. In a radiosurgical mode, it allows for hotspots. So, you're not constraining and you're not asking for homogeneous dose distributions. And that itself allows for better sparing of the cord. We have seen these phenomena in other sites, including pancreas and prostates.
Another aspect of cord sparing during the radiation therapy is motion management. Since the planned dose distribution is not exactly what the patients get considering patient motion.
Here we show patient motion data from cervical targets. As you can see, vertebral anatomy movement varies as much as three millimeters and it can occur in as little as five minutes of time.
Here we show the patient motion data from thoracic targets, more relevant in this case. Each color line represents a unique patient. Top row is the translational misalignments, and the bottom row is rotational.
On this slide, the ExacTrac data for this particular patient is shown for all five fractures. Although the clinical plan was created with Eclipse, ExacTrac at our institution is always utilized for spinal radiosurgical cases.
The first line of each of the fractions is the initial X-ray setup. This is post infrared positioning. Hence, this can be disregarded for actual patient motion purposes. During the fraction number five, however, we can see a large movement.
If you look into details, it happened prior to the second arc delivery. Of course, the impact of this movement is much smaller when IGRT intrafraction monitoring is used. And we deploy ExacTrac prior to each of the arcs.
If you misaligned the isocenter position by one millimeter, or by two millimeters, or three millimeters for this particular patient plan, the high-dose areas of the cord will be significantly hotter, as shown on the DVHs here. To investigate the dosimetric impact of positioning errors on target coverage, as well as cord sparing for a spinal radiosurgical patients, we have analyzed data from 62 spinal radiosurgery patients.
For each treatment plan, the direction of misalignment was selected to minimize the target in cord geometric separation. The movements were in that worst-case scenario direction. To start with, none of these patients exceeded 12-grade criteria because this is cases with single fraction. The same idea applies for five fraction cases.
With 1-millimeter motion, about 50% of the patients violated that criteria. With 2-millimeter motion, about 75% of the patients violated that criteria.
So, I would argue that it is not uncommon for the spinal cord to actually receive maximum doses above planning maximum doses. In fact, max those value used in planning tissue constraints is not actual tissue constraint at all. More on this to come.
Similar trends hold for rotational misalignments. Here, I'm showing you rotational misalignments. And again, the trend holds that even with 1 degree rotation, 40% of the patients don't meet the criteria. Same thing can be said for 10 gray, 10% criteria with translational layers, and in this case, with rotational errors.
This is why planning those constraints are generally lower than the tissue constraints or tissue tolerances. There is a lot of evidence in the literature that the spinal cord can tolerate more dose than the constraints we use for planning purposes. Our good colleague Paul Medin has demonstrated in a single-fraction treatment that no appreciable neurological deficit is seen below 17 gray in swine models, similar results are seen in mouse model experiments.
So, essentially, by using a lower dose constraint, we are utilizing safety margins for possible patient motion. The argument holds true for one fraction cases as well as for five fraction cases. One could possibly use higher prescription doses with a rigorous motion management plan.
Our current institutional protocol requires intrafraction motion management by stereoscopic imaging prior to each treatment field. This is in addition to CBCT, confirmation of the correct vertebral body. Patients are repositioned when one millimeter or one degree tolerance is exceeded.
So again, thank you for the opportunity to incorporate my comments for this Case of the Month.
[00:36:50]
[Silence]
[00:37:37]
Bogdan: I apologize, everyone, I was on mute. I was actually trying to start with a question. So, clearly, we selected a patient that received combined therapy for the spine disease. Are there any parameters for this patient that would have triggered a contraindication for a monotherapy, either surgery or radiosurgery?
Dr. Kaprealian: I can address that. So, even if we just did... We don't recommend just surgery alone. Even if a larger surgery was done, we would still follow it with postoperative radiation therapy. So, I wouldn't recommend monotherapy in the surgical setting. And if Dr. Macyszyn wants to chime in afterwards, we can see what his thoughts are.
With regards to just doing radiation alone, we certainly could have done that. We were just trying to relieve the tumor that was abutting the cord to kind of give a little bit of space so that we could do a hypofractionated course. But certainly, if we weren't meeting cord tolerances, and the patient couldn't tolerate a surgery, we could do a 10-fraction course, like 320 times 10. And that would be appropriate option as well.
Dr. Macyszyn: Yeah, I completely agree with Dr. Kaprealian. And I think there's an option to do just radiation here in the manner described. But surgery alone is, you know, rarely ever an option for metastatic disease, at least.
Bogdan: In terms of what kind of radiation plan can you generate for this kind of patients, given that this was a Bilsky 3, are you typically able to come up with a plan that meets your constraints for either three or four fraction SBRT or single fraction radiosurgery for this kind of cases?
Dr. Kaprealian: You can. You just have to be mindful of the dose to the cord. So, you might be under dosing that area. And the number one area where we currently see recurrences or progression is that epidural space. So, you just have to be mindful that you are getting in enough dose. If you're too cold in that area because of the cord tolerance, then you would want to do a fractionated course.
Bogdan: Okay. Nzhde, we have a few questions for you. And maybe we can start on the treatment planning side. So, what type of immobilization is used for this kind of patients? And then linked to this, what are the arc geometries or beam geometries that you typically use?
Dr. Agazaryan: Yeah, thank you for that question. So, the immobilization used for these patients are Medical Intelligence BodyFix, essentially a very minimal immobilization without the body wrap option of it. We started the program many years ago using the body wrap as well. But then we abandoned that as we started using intrafraction motion management with the ExacTrac for all patients. So, that's the mobilization that we use.
In terms of geometry, both plans presented here were all coplanar arcs, essentially full arcs. The Eclipse plan had two full arcs, and there you have to specify the arcs, obviously. But in Elements, we started with one arc and then the program itself split that into three arcs. They were all coplanar arcs in this case,
Bogdan: Does the immobilization device at all restricts arc geometry that you can select for these patients?
Dr. Agazaryan: The immobilization device itself does not, but just because we are treating the spine, there are couch kicks that restrict the gantry motions, yes, but not the immobilization device itself.
Bogdan: Okay, let's talk a little bit about margins and then IGRT. So, we had the majority of the people responding, they assign a one-millimeter CTV to PTV extension for post-op spine cases. Obviously, I think, in your practice, that is also one millimeter for this patient. In your experience, would you or have you assigned larger margins in the past? And how is this linked then to how you check for intrafraction motion?
Dr. Agazaryan: I'll give a try and maybe then Dr. Kaprealian can pitch in. So, yes, correct. In this case, we have used one millimeter margin extension on the CTV to create the PTV. And correspondingly, when we employ ExacTrac intrafraction motion management, the tolerances for repositioning of the patient, when images are required, are one millimeter and one degree for each of the arcs.
Dr. Kaprealian: Yeah, I mean, I agree because we have the capability of doing image guided radiotherapy. We try to minimize the margin that we need to use. And one millimeter is typically appropriate for our patients. I think, very rarely, we might use two millimeter, if we're worried about the patient's setup or ability to, you know, lie still or for various reasons, if there's any issues. But for the most part, we primarily use one-millimeter margins.
Bogdan: And Nzhde, perhaps you can provide a clarification, we have a question in terms of how this one millimeter is being derived, whether it's imaging or IGRT data. Being that you do utilize the ExacTrac to check every arc for intrafraction motion, what is the residual extent of deviations that you see for these patients or that you allow clinically?
Dr. Agazaryan: So, when you employ one millimeter and one degree, when we look at the residual extent in motion, typically, those are below one millimeter. But there are scenarios that shown in this piece actually where you would exceed that. Where do you come up with a one-millimeter margin?
First, like Dr. Kaprealian mentioned, because of the IGRT [inaudible 00:44:42], but also there has been publications in the literature that we have analyzed covert of patients and their movements. And there has been a suggestion of using one millimeter to improve the margin based on that. So, I wouldn't say that a very strong [inaudible 00:45:02] for using that, but there is a pretty good sort of evidence out there to follow that guideline.
Bogdan: Okay, thank you for that. You have another question in terms of how does the nonuniformity in the dose distribution help with a better spinal cord dose there?
Dr. Agazaryan: Yes. So, this is a very interesting phenomena, if you to think about this. We, of course, learned about this when planning other cases for research purposes. So, it turns out that most planning systems are designed to deliver uniform distribution at the target and delivering uniform distribution itself [inaudible 00:45:51] constraint that the system needs to meet.
In other words, when you're asking the system to create a uniform distribution, you have to compromise somewhere else and that compromise could be over the cord. But when you allow for, you know, [inaudible 00:46:09] distribution, especially you are taking away that constraint. And the optimization has more freedom to create faster, full or higher gradients towards the cord. It's a well-known phenomenon. We have seen it in prostate. We have seen it in pancreas. And I think it's an excellent implementation by [inaudible 00:46:31].
Bogdan: Here's an interesting question as well regarding the output dosimetry, and one that we actually investigated over the years with you as well. What is the preferred technique for this kind of patient static? Well, IMRT Beams or, I would say, arc modulated? I guess modulated arcs will be the question
Dr. Agazaryan: Modulated arcs are typically, prefer to put here, because of many reasons. But one of them is dosimetry is better. But also, the delivery is faster. In terms of the output, we have validated the output or the results of the software in many different ways. And there's a pretty good agreement. I'm not sure if I did answer the question, Bogdan, but is that where you were going with it?
Bogdan: I think it's very good. How long was the actual treatment for this patient?
Dr. Agazaryan: I haven't looked up the number. But if I have to guess, it's probably starting from the first part all the way until the end, including the intrafraction motion, probably somewhere close to 10 minutes, if not less.
Bogdan: Dr. Macyszyn, I have a question for you. What percentage of your spine patients receive separation surgery? And what are some of the benefits of separation surgery over a complete or gross total resection for this kind of patients?
Dr. Macyszyn: Yes, I would say probably the minority of patients receive separation surgery or cases like this, where we have kind of complete circumferential kind of tumor and compression of the spinal cord is where separation is helpful, basically, to reestablish once again that CSF flow between the tumor, two to three millimeters of CSF in the spinal cord.
I would say, above and beyond that, especially, we typically try to get more than just a couple millimeters resecting the rest of the, let's say, vertebra or any kind of paraspinal tumor tissue. I feel there's a little benefit of that, especially in the context of metastatic disease, because that is much better and more efficaciously treated using post-operative SRS.
Bogdan: Thank you. And last question, and maybe Dr. Kaprealian, you can answer this as well. But what are the benefits of utilizing both cone beam CT and ExacTrac for these patients? And how do you evaluate the daily IGRT for the final setups?
Dr. Kaprealian: So what I'm looking at what the cone beam CT is the localization to the correct vertebral body. And that's particularly challenging in the thoracic spine, if you just use ExacTrac because there aren't any other bony landmarks to visualize. And so, that initial cone beam is done to localize to the correct vertebral body. And then, the ExacTrac imaging is done, just as Dr. Agazaryan described. And I don't know, Dr. Agazaryan, do you have any further comments on that?
Dr. Agazaryan: No, that was an excellent answer. No.
Bogdan: Great, well, thank you all for your presentations. And thank you all for joining the webinar and we'll see you on the next one. Thank you and bye bye.
Dr. Kaprealian: Thank you. Bye-bye.
Dr. Macyszyn: Thanks, everyone.
Today, we are introducing a new concept, reviewing a clinical case aimed at providing further in-depth knowledge into the expertise behind making clinical decisions for various radiosurgical treatments, review the clinical evidence for a treatment path that has been selected, and summarize relevant technological steps in generating an ideal treatment plan.
The Case of the Month webinars are a precursor to the Virtual Tumor Board Series that we'll introduce next year. And we hope to continue to provide you with valuable clinical expertise as you continue to expand your programs. For the first Case of the Month webinars, I have selected a few cases that I find interesting for the community at large. But please feel free to send me your suggestion for future events.
Today, I have the privilege to introduce you to our long-time partners at the University of California in Los Angeles. And the clinical team would be reviewing a spine metastasis patient. As best clinical decisions are derived from multidisciplinary interactions, we have today presentations from radiation oncology, neurosurgery, and medical physics.
Dr. Tania Kaprealian is a radiation oncologist and an associate clinical professor at UCLA. And she will begin with a case presentation. Dr. Kaprealian is the chief of the CNS Services and also serves as the medical director at the main campus in Westwood, LA.
Joining her is Dr. Luke Macyszyn from the department of neurosurgery. And he will be addressing the value of minimally-invasive surgical procedures. Dr. Macyszyn is an assistant professor of neurosurgery, radiation oncology and orthopedics, and he's also the associate program director of the Neurosurgery Training Program.
To complete the team and discuss treatment planning and IGRT considerations, we have Dr. Nzhde Agazaryan, who is a professor of radiation oncology at UCLA. Dr. Agazaryan is the chief of Medical Physics and the Dosimetry, and also serves as a quality and safety officer.
As true for the regular webinars, we continue to provide CE credits for the Case of the Month webinar as well. Should you need CAMPEP, MDCB or ASRT credits, upon successful completion of the live event, please email us at info@novaliscircle.org for further details.
And please register for our webinar during ASTRO's Online Industry Expert Theater on October 26. We have an exciting lineup of radiation oncology and medical physicist from Munich and Brussels presenting on their first treatments with ExacTrac Dynamic for both cranial and extracranial indications. And you'll have the opportunity to learn and ask questions on the new IGRT capabilities for the new system.
Lastly, please remember to use either Google Chrome or Safari. And should you have any internet challenges, simply refresh the page. Send us questions via the chat interface. We will answer all your questions upon completion of the three lectures.
Monitor the polling infrastructure as we may like to ask you some questions. And should you choose to follow us on social media, please utilize the #NovalisCircle.
And now, we'd like to turn it over to the UCLA team for their Case of the Month Review.
Dr. Kaprealian: Thank you, Bogdan. These are our disclosures. And today, we'll be discussing a complex metastatic spine case. We'll discuss the surgical considerations as well as the radiosurgical considerations, and then we'll review the treatment planning.
So, to start with the case, the patient is a 49-year-old woman with a history of metastatic left breast infiltrating ductal carcinoma. She presented with a progressive spinal T10 lesion with concern for cord compression.
Her symptoms consisted of worsening pain over the course of two months in the mid-low back at the T10 level that radiated to the mid-to-low abdomen. And it was very painful for her to walk. She denied any extremity pain or numbness. And she had no neural findings on exam and no pain on palpation of her spine.
An MRI was obtained and, as you can see here, there is a T10 lesion that involves the entire vertebral body and posterior elements. There's circumferential epidural tumor surrounding the thoracic cord and it causes at least moderate circumferential cord compression and contour deformity with cord edema.
So, on spinal assessment, she was determined to be a Bilsky Grade 3 with a SINS score of 9. And in the NOMS criteria, she had high-grade cord compression, she had a radiosensitive tumor. There was question of stability of the spine. And we knew that she would be able to tolerate a surgery.
I'll now hand it over to my colleague, Dr. Macyszyn.
Dr. Macyszyn: Hi, everyone. So, I guess, the first thing to discuss, which is kind of critical to all these cases, the decision making, a lot of it hinges on the epidural spinal cord compression, which is frequently called the Bilsky Grade from Dr. Bilsky that created this grading system, allowing us to more kind of quantitatively talk to each other to determine what the level of spinal cord compression is. This grading system goes all the way from 0 to a 3.
In a 0, basically, the tumor is confined strictly to the vertebral body and there is no epidural extension. In a Bilsky 1a, there is some epidural extension and there's tumor in the epidural space, but there's no deformation of the thecal sac and certainly not of the spinal cord. In a 1b, there is deformation of the thecal sac. In the 1c, there is a budding of the tumor or the epidural component of the tumor with the spinal cord. In a Grade 2, the disease kind of deforms the spinal cord and starts to push or displace the spinal cord, either to one side or the other or dorsally. And in the Grade 3, there's complete kind of obliteration of the CSF space with compression of the spinal cord.
Here on this slide, we have kind of clinical, you know, pictures, radiographic MRI showing all these grades all the way from a 0 to a 3. And you can see there's progressive involvement of the epidural space, the spinal canal, and progressive compression, deformation of the spinal cord.
Basically, we divide this up into low-grade epidural spinal cord compression and high-grade. And once again, this helps us with our decision making to determine if surgery is or isn't necessary and what kind of radiosurgery we would be able to do or not to do.
Basically, anything from a 0 to a 1b is considered low grade epidural spinal cord compression. And a grade from 1c to a 3 is considered high grade epidural spinal cord compression. Obviously, the most obvious ones are the grades 0, and the grades 2s and 3s. And probably, the one that's kind of the most debatable and kind of on the fence is the grade in 1C.
There's also the spinal instability scale, which is called the SINS classification systems, which is detailed here. It allows us to look at, evaluate the radiographic images of the patient as well as their symptoms. In this case, it's mostly just pain. And to determine how stable or unstable potentially a patient spine is due to the involvement of the spinal column with the tumor.
This goes over the location of the tumor, whether the patient has pain. If there's any bone lysis, alignment of the spine, vertebral collapse, and involvement of the spinal elements. Obviously, as you get more and more of these, you score more and more points, and making you more and more potentially unstable, therefore, requiring surgical stabilization either before and after radiosurgical treatment.
So, anything in the SINS scale from 0 to 6 is considered as stable spine, from 7 to 12 is the harder part. It's kind of indeterminant range. There might be instability or there might not be instability. And obviously, anything from 13 and higher usually is considered unstable.
This is a grading system, once again, that was designed so radiologists and physicians are like can communicate not only the radiographic cord compression regarding patients with metastatic spine disease, but also the potential stability of the spine. And we advocate a kind of a surgical consultation, obviously, for anybody who has greater than a SINS score of 7.
With regards to the NOMS criteria, this is kind of the overall picture and kind of paradigm that we use to treat patients. We've kind of get their neurological status, their oncological status. Once again, there's the mechanical stability of the spine, which we just talked about, with the SINS criteria. And obviously, systemically, how are they doing? Are they able to tolerate surgery? Are they not able to tolerate surgery? And I'll go over this in a little bit more detail in the following slides.
This is kind of the NOMS framework. And it's pretty easy and kind of straightforward. Obviously, first, I've kind of highlighted all the unstable, you know, mechanical pathologies that might present in the patients. Obviously, if a patient is considered unstable and they are able to tolerate surgery, we do surgery followed by radiation therapy. If they're unstable and radioresistant and they are not able to tolerate surgery, obviously, we will proceed with palliative radiation only. For patients who have a stable spine and has a low-grade epidural spinal cord compression, obviously, in that case, we would proceed with either classic radiotherapy treatment or stereotactic radiosurgery.
Finally, for the radiosensitive tumors that are stable, and even with a high-grade epidural spinal cord compression, so the things that come to mind is stuff like lymphoma and multiple myeloma. In those instances, we would select radiotherapy as the treatment without the necessity of spinal decompression, even though there's high-grade epidural spinal cord compression.
For radioresistant tumors now, and this is kind of self-explanatory, that have high-grade spinal cord compression, we do something called spinal cord separation surgery first, followed by radiation treatments to once again enable, you know, higher doses of radiation to the tumor without negatively affecting the integrity of the spinal cord.
So, you know, there's a number of ways we can proceed here. Do we do surgery alone? Surgery followed by radiation therapy? Just considerations for the radiation therapy plan and as well as pain management and a patient who presents with once again metastatic spine disease, potential instability, and pain.
We discussed this patient in our radiosurgery conference and the recommendation was made to proceed with separation surgery due to the epidural spinal cord compression of this tumor, followed by adjuvant SRS. So, we performed basically a T9-T10 laminectomy, and resection of the epidural tumor.
So, why separation surgery? Separation to this is probably one of the, you know, the classic and most kind of talked about trials, certainly in spinal surgery, but definitely in our radiosurgery board as well. This is from an older trial performed in 2005 by Dr. Patchell. This was published in Lancet that basically spinal cord decompression, followed by radiosurgery, leads to a better outcome than either modality on its own.
And the caveat here, obviously, is that these patients, for the most part, we select the patients with one single, you know, epidural or one single lesion to the spine that had epidural compression. And, obviously, their ability to walk was assessed at the conclusion of this trial as various other neurological function.
And you can see here that these are the patients that were signed up. And patients that had surgery with radiation fared better than patients that just had radiation alone. So, in this picture, here you see once again a spinal column with a spinal tumor metastasis to one of the vertebrae and, in this case, leading to spinal cord compression.
In a separation surgery, in this case, demonstrating fusion as well. Pedicle screws are placed, a laminectomy is performed, and we performed a partial corpectomy, once again to circumferentially decompress the spinal cord, as indicated here, and remove this tumor away from the thecal sac to reestablish once again that margin of CSF around the spinal cord, and obviously enabling, you know radiosurgery to a much better extent following surgery.
Here, once again, a cage is placed and the patient's spine is stabilized. And in this case, once again we're just kind of showing the benefits of stereotactic radiosurgery treatment. This was published in the journal of neurosurgery, "Spine" as compared to regular or more conventional hypofractionated SRS. And you can see that patients with a single dose SRS have a lower cumulative incidence of recurrence at the site of their disease.
Dr. Kaprealian: As Dr. Macyszyn described, this patient underwent a separation surgery. And the reason we decided to go that route for her, she was symptomatic, she had cord edema with cord compression, and we wanted to give her a hypofractionated radiation course postoperatively. And as you can see here, she also did have a pretty aggressive tumor because even in the short time span between surgery and this MRI, she had an interval, we can see the interval decompression at the T10 level but she had increase in the ventral epidural tumor as extending superiorly from T10 and as well as extension to the bilateral T9-T10 neural foramina.
And in addition, she has the soft-tissue component. There's a large prevertebral tumor at T10, displacing the aorta, and persistent cord edema still between T9 and T11 levels. So, the patient underwent CT stimulation two and a half weeks later for planning for post-operative stereotactic radiosurgery.
What are some indications to treat spine mets with SRS? Radiosurgery has been shown to have long-term rates of tumor control, as high as 90%, both in the salvage setting, after receiving traditional external beam, and as primary management.
And what SRS can accomplish is pain control, improvement or preservation in neurologic function, as well as possibly, in situations with spinal cord compression, it can be of benefit. But the issue is what is the correct target for spinal SRS? And this is not only important for capturing the tumor itself but it also plays a role if you have to come back and re-irradiate. It's really important the choice of the initial target volume delineation.
And this is just a paper showing several studies of postoperative SRS and the local control rates. And you can see that it ranges from 80% to the mid-90%. So, it's really an excellent treatment modality postoperatively.
So, what are some of the issues with postoperative spinal SRS? There's the impact of surgical hardware on the CT and MRI as well as for treatment planning. And then, the target delineations are more complex, both not only the target itself, the tumor itself, but also the critical neural structures.
So, as most of you know, there are consensus guidelines that are developed for the intact spine and recently there are consensus guidelines that have been developed for spine SRS in the postoperative setting. And what this takes into consideration is the preoperative tumor plus adjacent anatomic compartments.
So 9 radiation oncologists and 1 neurosurgeon with spinal SRS expertise contoured 10 cases. And what they recommended is that the CTV should include the entire preoperative extent of bony and epidural disease, plus any immediate adjacent bony anatomic compartments that are at risk of microscopic disease extension. A donut-shaped CTV is used in cases of preoperative circumferential epidural extension, regardless of the amount of residual epidural extension that's left. And spinal instrumentation is consistently excluded from the CTV. So this is very important because this is a common question that comes up.
So, in addition to the written guidelines, they have image guidelines that you can use to help correlate based on where the tumor is located in your patients' vertebral body, and then use this to help you contour the vertebral body accordingly.
So, we planned our patient in Eclipse. However, we are now commissioning Elements Spine SRS. So we would like to review our preplanning steps using this Elements program.
So first, what you want to do is a cross-modal rigid image fusion. And as you can see, in this case, that there's a difference in the curvature between the MRI and the CT. And this is a challenge for all applications. And you really have to be careful when you co-register. So, what you would have to do in this specific situation is designate a region of interest manually and define that so that you can do the image fusion in that specific area. And it basically correlates to the bony anatomy in that tumor region only.
So, this results in great registration in that area, but the rest of the spine will be off. So, for patients with tumor involvement in multiple levels, you would have several regions of interest and multiple registrations.
So, to solve this problem, Brainlab provides a curvature correction elements. And so, for this patient, we use the CT as the reference because the treatment plan is generated on the CT. And we allow the distortion algorithm to modify the MRI to match the anatomical position of the CT. This process doesn't require a manual definition of a region of interest. But unlike what we've shown here and done here, it's best to use a full MRI scan rather than a partial MRI.
And so, then the elastic deformation is anatomically penalized to not allow the bone to be deformed. And as you can see, it manages to correct for the significant cord curvature, which has significant dosimetric implications on the spinal cord dose. And Dr. Agazaryan will go in this further in his portion of the talk.
Another very important time-saving step for the physicians is the normal anatomy contouring. So, this program uses the universal Atlas element and it contours the normal anatomy organs. And then, the physician just has to come and check and approve the contours. Of course, manual contours can be added, if appropriate, based on the spinal cord level. But a lot of the organs are included in this automatic segmentation.
And then, finally, what's really critical for the physicians is to define the target. And so, with the Brainlab Spine SmartBrush Elements, the physician will contour out the GTV. And then, the program has the capability of developing the CTV expansion automatically using the intact spine control guidelines. So, in the postoperative setting, you might have to do manual adjustments, but for the most part, it's following the consensus guidelines and they're pretty similar.
So even though Brainlab provides this tool for single level, it is possible to use this workflow for a multilevel disease, as our patient had with T9 and T10. So, first, you would need to define the GTV at your first level. And you can do this with quick contouring in the axial view and in one reconstruction, and then the software will generate the CTV. This segment interactive then allows you to either add or subtract components as necessary without having to manually contour any further.
And once this level is complete, you then have to define the initial CTV as an object type called undefined, and then you exit and reenter the SmartBrush Element. So then you would have to repeat this step for the second level, manually contouring the GTV, and then the program will develop the automatic CTV and then you would again mark it as undefined. And then, finally, the last step would be to basically merge or union those two contours and you have the Object Manipulator elements to help you perform that.
And then, the last step is the PTV definition. But before we get to that step in our patient, remember that this patient also had a soft-tissue component. And that's not part of the consensus guidelines because those guidelines focus on the bony anatomy. And so, but what you can do is you can use this program to manually incorporate that soft-tissue components into your CTV. And then, what we do is we do a one-millimeter PTV extension.
Here is the patient's plan in Eclipse, and this is the final plan that we generated. We gave 6 gray times 5 for a total of 30 gray. And here again, you can see the patient's plan at the multiple different levels.
So, the patient tolerated the treatment well. As I mentioned, she received 30 gray and 5 fractions. And over time, she had significant improvement in her pain and she was able to taper off her steroids and pain medications.
So, this is the two-month follow up thoracic spine MRI. She had interval decrease in size of the epidural mass at T10 as well as improvement in that paraspinal soft tissue mass that was from T9 to T11, and an improved cord edema at those levels as well, and then resolution of the cord compression.
And then, again, at the five-month mark, she had no evidence of progressive bony metastatic disease, no compromise of the spinal canal, and an interval resolution of that soft-tissue mass. And now, I'd like to hand it over to my colleague, Dr. Agazaryan.
Dr. Agazaryan: Thank you for the opportunity to incorporate my comments for this Case of the Month. These are my disclosures. Dr. Kaprealian and Dr. Macyszyn already presented on the details of Eclipse treatment plan.
Here I highlight some of the TG101, those endpoints that will be used to compare with the Brainlab Spine Element later on. The cord was redrawn to end six millimeters superior and inferior of the overall target. The inferior part of the cord extends further down in this coronal view because of the inferior extension of the tumor seen on a 3D view here.
We have also created two other cord structures that correspond to each of the vertebral bodies, T9 and T10. The dosimetry of the cord counter corresponding to the overall target meets TG101 constraint.
This, for my display here, is done with scripting after the Eclipse plan is finalized. Similarly, T9 and T10 cord subvolumes, individually taken, meet the TG101 constraints. Clearly, the cord sections, separately evaluated, are more conservative approach for the 10% criteria. While experimenting with the Spine Elements, we experienced excellent graphical user interface and results.
I showed the deformable fusion for this particular patient with the deformation heatmap. Deformable registration uses anatomical mapping and tissue classification with deformable and nondeformable tissues. This is very similar to biomechanical registration. Obviously, this is a multilevel spine co-registration that can be done with MRI-CT or CT-CT while compensating for different spine curvatures.
The basic principle here is to automatically calculate individual rigid fusions for each of the vertebra. And from these results, interpolate a single 3D deformation field that simultaneously matches all vertebra in the fused images. Tissue model is also used in this process. It is essentially an iterative process between these two, where the integrity of the bony anatomy is maintained.
So, we just discussed the ability of doing a deformable fusion in Elements. We have transferred all the images to Elements and redrawn the spine contour on the T2 image in Elements. You can see that an appreciable difference is seen in the dosimetry of these two objects, one based on a rigid fusion and the other is based on deformable fusion. Deformed caused those maximum was greater in this case. We will come back to this discussion in the tissue tolerance section of the presentation.
Once the structures are redrawn, the planning process commences. The optimization and the calculation can be done both with Pencil Beam as well as with the Monte Carlo. I display the Monte Carlo toggle switch here.
This toggle serves two purposes. First, it does forward calculation with two algorithms. So, the distributions, the DVHs, 3D Display can be displayed either with Pencil Beam or Monte Carlo. The optimization itself does not change. Or, if you toggled during the optimization, then corresponding algorithm is used for optimization process.
The first optimization should always be Pencil Beam algorithm. Monte Carlo optimization is recommended for the last refining stages.
Here I show the results from Elements plans. The first one is the Pencil Beam optimization and a Pencil Beam calculation. The Pencil Beam plan achieved 99% target coverage. Then, Monte Carlo forward calculation helps to check the Pencil Beam calculation with Monte Carlo algorithm. Since we find, in this case, that the Pencil Beam underestimates the cord dose since we don't meet the criteria with Monte Carlo. Then, we proceeded with Monte Carlo optimization.
The Monte Carlo optimization itself meets the criteria with the coverage of 97%. Then we renormalize the coverage so that the 99% of the target is covered with a prescription and it still meets the TG101 criteria.
So, all in all, keeping the cord endpoints approximately the same, 99% target coverage is achieved with Elements plan. It is difficult to make general conclusions with selected cases. In fact, plan comparisons are often arguable, even with multiple cases, because of variations in planning techniques.
However, since we are discussing this specific case, it was decided that presenting plan comparison is a fair approach. Here's the Monte Carlo optimized, a Monte Carlo calculated treatment plan.
The planner gets real-time feedback from the software during the planning process versus running the script at the end of the planning process shown during the Eclipse calculation. So, here with the traffic light approach, you can incorporate the TG101 constraints and give a real-time feedback to the planner.
Here we show Eclipse plan versus Spine Elements Monte Carlo calculation. The Spyglass shows the Eclipse plan. The Element plan is more conformal and has a rapid fall off in this particular case towards the cord. More detailed studies with more patients need to be conducted for conclusive results.
I should mention that the Spine Elements conformity is achieved by PTV splitting and arc-duplication approach. The user can restrict the total number of arcs included in duplications. Our plan started with one arc, and then it was duplicated to three arcs. Spine metastases are located within vertebra, which are concave structures presenting challenges for VMAT optimization.
Splitting PTV allows for creation of sub PTVs with much less concave regions. As I show here in this example. Even by splitting into two parts, the concavity of the structures are minimized. This allows for much better conformity, the sparing of the cord.
In our specific example, PTV splitting and arc duplication resulted in three arcs covering different segments of the PTV. Brainlab Element Spine Radiosurgery Module also has a radiosurgical prescription mode, which does not restrict the tumor dose heterogeneity. Hence, it allows for greater gradient fall-off into the spinal cord.
So, in the normal mode, planning generates highest homogeneity while maintaining conformity indices and gradient indices. In a radiosurgical mode, it allows for hotspots. So, you're not constraining and you're not asking for homogeneous dose distributions. And that itself allows for better sparing of the cord. We have seen these phenomena in other sites, including pancreas and prostates.
Another aspect of cord sparing during the radiation therapy is motion management. Since the planned dose distribution is not exactly what the patients get considering patient motion.
Here we show patient motion data from cervical targets. As you can see, vertebral anatomy movement varies as much as three millimeters and it can occur in as little as five minutes of time.
Here we show the patient motion data from thoracic targets, more relevant in this case. Each color line represents a unique patient. Top row is the translational misalignments, and the bottom row is rotational.
On this slide, the ExacTrac data for this particular patient is shown for all five fractures. Although the clinical plan was created with Eclipse, ExacTrac at our institution is always utilized for spinal radiosurgical cases.
The first line of each of the fractions is the initial X-ray setup. This is post infrared positioning. Hence, this can be disregarded for actual patient motion purposes. During the fraction number five, however, we can see a large movement.
If you look into details, it happened prior to the second arc delivery. Of course, the impact of this movement is much smaller when IGRT intrafraction monitoring is used. And we deploy ExacTrac prior to each of the arcs.
If you misaligned the isocenter position by one millimeter, or by two millimeters, or three millimeters for this particular patient plan, the high-dose areas of the cord will be significantly hotter, as shown on the DVHs here. To investigate the dosimetric impact of positioning errors on target coverage, as well as cord sparing for a spinal radiosurgical patients, we have analyzed data from 62 spinal radiosurgery patients.
For each treatment plan, the direction of misalignment was selected to minimize the target in cord geometric separation. The movements were in that worst-case scenario direction. To start with, none of these patients exceeded 12-grade criteria because this is cases with single fraction. The same idea applies for five fraction cases.
With 1-millimeter motion, about 50% of the patients violated that criteria. With 2-millimeter motion, about 75% of the patients violated that criteria.
So, I would argue that it is not uncommon for the spinal cord to actually receive maximum doses above planning maximum doses. In fact, max those value used in planning tissue constraints is not actual tissue constraint at all. More on this to come.
Similar trends hold for rotational misalignments. Here, I'm showing you rotational misalignments. And again, the trend holds that even with 1 degree rotation, 40% of the patients don't meet the criteria. Same thing can be said for 10 gray, 10% criteria with translational layers, and in this case, with rotational errors.
This is why planning those constraints are generally lower than the tissue constraints or tissue tolerances. There is a lot of evidence in the literature that the spinal cord can tolerate more dose than the constraints we use for planning purposes. Our good colleague Paul Medin has demonstrated in a single-fraction treatment that no appreciable neurological deficit is seen below 17 gray in swine models, similar results are seen in mouse model experiments.
So, essentially, by using a lower dose constraint, we are utilizing safety margins for possible patient motion. The argument holds true for one fraction cases as well as for five fraction cases. One could possibly use higher prescription doses with a rigorous motion management plan.
Our current institutional protocol requires intrafraction motion management by stereoscopic imaging prior to each treatment field. This is in addition to CBCT, confirmation of the correct vertebral body. Patients are repositioned when one millimeter or one degree tolerance is exceeded.
So again, thank you for the opportunity to incorporate my comments for this Case of the Month.
[00:36:50]
[Silence]
[00:37:37]
Bogdan: I apologize, everyone, I was on mute. I was actually trying to start with a question. So, clearly, we selected a patient that received combined therapy for the spine disease. Are there any parameters for this patient that would have triggered a contraindication for a monotherapy, either surgery or radiosurgery?
Dr. Kaprealian: I can address that. So, even if we just did... We don't recommend just surgery alone. Even if a larger surgery was done, we would still follow it with postoperative radiation therapy. So, I wouldn't recommend monotherapy in the surgical setting. And if Dr. Macyszyn wants to chime in afterwards, we can see what his thoughts are.
With regards to just doing radiation alone, we certainly could have done that. We were just trying to relieve the tumor that was abutting the cord to kind of give a little bit of space so that we could do a hypofractionated course. But certainly, if we weren't meeting cord tolerances, and the patient couldn't tolerate a surgery, we could do a 10-fraction course, like 320 times 10. And that would be appropriate option as well.
Dr. Macyszyn: Yeah, I completely agree with Dr. Kaprealian. And I think there's an option to do just radiation here in the manner described. But surgery alone is, you know, rarely ever an option for metastatic disease, at least.
Bogdan: In terms of what kind of radiation plan can you generate for this kind of patients, given that this was a Bilsky 3, are you typically able to come up with a plan that meets your constraints for either three or four fraction SBRT or single fraction radiosurgery for this kind of cases?
Dr. Kaprealian: You can. You just have to be mindful of the dose to the cord. So, you might be under dosing that area. And the number one area where we currently see recurrences or progression is that epidural space. So, you just have to be mindful that you are getting in enough dose. If you're too cold in that area because of the cord tolerance, then you would want to do a fractionated course.
Bogdan: Okay. Nzhde, we have a few questions for you. And maybe we can start on the treatment planning side. So, what type of immobilization is used for this kind of patients? And then linked to this, what are the arc geometries or beam geometries that you typically use?
Dr. Agazaryan: Yeah, thank you for that question. So, the immobilization used for these patients are Medical Intelligence BodyFix, essentially a very minimal immobilization without the body wrap option of it. We started the program many years ago using the body wrap as well. But then we abandoned that as we started using intrafraction motion management with the ExacTrac for all patients. So, that's the mobilization that we use.
In terms of geometry, both plans presented here were all coplanar arcs, essentially full arcs. The Eclipse plan had two full arcs, and there you have to specify the arcs, obviously. But in Elements, we started with one arc and then the program itself split that into three arcs. They were all coplanar arcs in this case,
Bogdan: Does the immobilization device at all restricts arc geometry that you can select for these patients?
Dr. Agazaryan: The immobilization device itself does not, but just because we are treating the spine, there are couch kicks that restrict the gantry motions, yes, but not the immobilization device itself.
Bogdan: Okay, let's talk a little bit about margins and then IGRT. So, we had the majority of the people responding, they assign a one-millimeter CTV to PTV extension for post-op spine cases. Obviously, I think, in your practice, that is also one millimeter for this patient. In your experience, would you or have you assigned larger margins in the past? And how is this linked then to how you check for intrafraction motion?
Dr. Agazaryan: I'll give a try and maybe then Dr. Kaprealian can pitch in. So, yes, correct. In this case, we have used one millimeter margin extension on the CTV to create the PTV. And correspondingly, when we employ ExacTrac intrafraction motion management, the tolerances for repositioning of the patient, when images are required, are one millimeter and one degree for each of the arcs.
Dr. Kaprealian: Yeah, I mean, I agree because we have the capability of doing image guided radiotherapy. We try to minimize the margin that we need to use. And one millimeter is typically appropriate for our patients. I think, very rarely, we might use two millimeter, if we're worried about the patient's setup or ability to, you know, lie still or for various reasons, if there's any issues. But for the most part, we primarily use one-millimeter margins.
Bogdan: And Nzhde, perhaps you can provide a clarification, we have a question in terms of how this one millimeter is being derived, whether it's imaging or IGRT data. Being that you do utilize the ExacTrac to check every arc for intrafraction motion, what is the residual extent of deviations that you see for these patients or that you allow clinically?
Dr. Agazaryan: So, when you employ one millimeter and one degree, when we look at the residual extent in motion, typically, those are below one millimeter. But there are scenarios that shown in this piece actually where you would exceed that. Where do you come up with a one-millimeter margin?
First, like Dr. Kaprealian mentioned, because of the IGRT [inaudible 00:44:42], but also there has been publications in the literature that we have analyzed covert of patients and their movements. And there has been a suggestion of using one millimeter to improve the margin based on that. So, I wouldn't say that a very strong [inaudible 00:45:02] for using that, but there is a pretty good sort of evidence out there to follow that guideline.
Bogdan: Okay, thank you for that. You have another question in terms of how does the nonuniformity in the dose distribution help with a better spinal cord dose there?
Dr. Agazaryan: Yes. So, this is a very interesting phenomena, if you to think about this. We, of course, learned about this when planning other cases for research purposes. So, it turns out that most planning systems are designed to deliver uniform distribution at the target and delivering uniform distribution itself [inaudible 00:45:51] constraint that the system needs to meet.
In other words, when you're asking the system to create a uniform distribution, you have to compromise somewhere else and that compromise could be over the cord. But when you allow for, you know, [inaudible 00:46:09] distribution, especially you are taking away that constraint. And the optimization has more freedom to create faster, full or higher gradients towards the cord. It's a well-known phenomenon. We have seen it in prostate. We have seen it in pancreas. And I think it's an excellent implementation by [inaudible 00:46:31].
Bogdan: Here's an interesting question as well regarding the output dosimetry, and one that we actually investigated over the years with you as well. What is the preferred technique for this kind of patient static? Well, IMRT Beams or, I would say, arc modulated? I guess modulated arcs will be the question
Dr. Agazaryan: Modulated arcs are typically, prefer to put here, because of many reasons. But one of them is dosimetry is better. But also, the delivery is faster. In terms of the output, we have validated the output or the results of the software in many different ways. And there's a pretty good agreement. I'm not sure if I did answer the question, Bogdan, but is that where you were going with it?
Bogdan: I think it's very good. How long was the actual treatment for this patient?
Dr. Agazaryan: I haven't looked up the number. But if I have to guess, it's probably starting from the first part all the way until the end, including the intrafraction motion, probably somewhere close to 10 minutes, if not less.
Bogdan: Dr. Macyszyn, I have a question for you. What percentage of your spine patients receive separation surgery? And what are some of the benefits of separation surgery over a complete or gross total resection for this kind of patients?
Dr. Macyszyn: Yes, I would say probably the minority of patients receive separation surgery or cases like this, where we have kind of complete circumferential kind of tumor and compression of the spinal cord is where separation is helpful, basically, to reestablish once again that CSF flow between the tumor, two to three millimeters of CSF in the spinal cord.
I would say, above and beyond that, especially, we typically try to get more than just a couple millimeters resecting the rest of the, let's say, vertebra or any kind of paraspinal tumor tissue. I feel there's a little benefit of that, especially in the context of metastatic disease, because that is much better and more efficaciously treated using post-operative SRS.
Bogdan: Thank you. And last question, and maybe Dr. Kaprealian, you can answer this as well. But what are the benefits of utilizing both cone beam CT and ExacTrac for these patients? And how do you evaluate the daily IGRT for the final setups?
Dr. Kaprealian: So what I'm looking at what the cone beam CT is the localization to the correct vertebral body. And that's particularly challenging in the thoracic spine, if you just use ExacTrac because there aren't any other bony landmarks to visualize. And so, that initial cone beam is done to localize to the correct vertebral body. And then, the ExacTrac imaging is done, just as Dr. Agazaryan described. And I don't know, Dr. Agazaryan, do you have any further comments on that?
Dr. Agazaryan: No, that was an excellent answer. No.
Bogdan: Great, well, thank you all for your presentations. And thank you all for joining the webinar and we'll see you on the next one. Thank you and bye bye.
Dr. Kaprealian: Thank you. Bye-bye.
Dr. Macyszyn: Thanks, everyone.