I'm going to spend about 10 minutes talking about screw fixation. And so why is there a potential for IRT to really enter into this traditionally surgical space? So surgeons have this established workflow. They know what they're doing. They know their equipment. But for us, we have this advanced imaging equipment that we can really leverage a lot of experience using coaxial techniques under CT and thoroscopy. And we can also add on different other minimally invasive technologies, like ablations or embolizations, and really exercise a lot of creativity. So for us, the keys to success are really meeting those surgical standards. We have to select the correct patient. We have to learn the equipment. We have to understand the principles. We have to set realistic clinical expectations, not just for the patient, but for the referring providers. And then we have to meet and really exceed those expectations because we're entering in a field that is surgically dominated. So we're already fighting an uphill battle. So we really have to do things that they cannot do, and they value our support. So some of the core principles that I've kind of boiled everything down to is knowing when to place a screw. And screws really counter torque or shear stress, tension stress, in a way that cement cannot. So we really have to understand that that's the first basic principle, when to apply a screw. So this is just a very basic example. If you imagine this patient standing up, the force is coming down their spine, hits their sacrum, and you have this type of shear force. And there's also some torque force as they walk. And so here, cement itself is not going to do the trick, and you really need a screw fixation. And the addition of cement is very helpful. Next basic principle is that screws are much better when you have a large ISIS defect. And this is a short presentation from Deschamps that really showed that greater than 3 centimeters, it's very helpful to have that additional support. And then another side comment I have is to consider that when you're spanning a great distance, you really need solid anchor points on either side. So if you imagine a bridge suspended over a crevice or a large defect in the land, you really need a solid foundation on either side. And if you don't have one, you have to create one. So this is an example where I had to create a solid anchor point in the anterior part of that iliac bone with the injection of a lot of cement. And then you start taking these into different permutations, and you'll see different IR physicians approach these differently. Some might put in one screw, some multiple screws. And you realize there is a lot of creativity here in how you approach it, but understanding these basic principles. Next one is screws provide stabilization when cement leakage is expected. So in this example, I really anticipated cement to leak quickly into the neuroforamina or that L5S1 disk space. And so I put the screws in first and then injected the cement. And sure enough, the cement leaked really quickly, but the screws held up very well. And this patient was able to go from being bed-bound to walking within a few days. Why is cement even necessary? Deschamps and a lot of the European colleagues have really shown a lot in the literature that cement, the addition of cement with screws is very helpful in this case. It's a basic example, one of my examples where I placed a screw, but unfortunately, after a few weeks of walking, the screw did migrate out. And so the addition of cement would greatly help in that location. We do have recent publications that show the combination of screw plus cement provides greater stiffness and provides dispersal of force across a greater area so that you really get a better outcome. And this is the graph in that article there, very helpful. You also have to approach by setting really good clinical expectations with the patients and with referring providers, and also just with your team, the radiologist team, when they're gonna evaluate this. I wouldn't expect this to decompress or reduce, rather, this fracture, but stabilizing these components would allow that fracture to hopefully heal better. We do have to select appropriate patients. They have to have a lot of mechanical pain, no resolution of pain with conservative measures, and we really do have to have a clinical imaging finding that correlates with that significant pain. And again, a lot of these things are basic principles that I went over earlier. And then this last part is, I think, very important as we get started in this field. There has to be no surgical alternative or reasonable surgical alternative, so that's something to really consider. And then these are the basic, realistic clinical expectations I have with patients, really how their pain's gonna resolve. I say approximately 50%. You are gonna be able to walk. You will be able to work with PT, OT the next day. There's no contraindication, really, with radiation therapy or any chemotherapy, so this works in seamlessly with what they already are doing for their oncological care. Getting access to the equipment is pretty easy. All the surgeons in your hospitals use what we use. And then you just have to get familiar with it. If you've done anything in the garage, it comes in very handy, and you'll pick it up very quickly. A few basic principles, again, going back to certain key papers that I've listed here. A lot of these, I'm not going into great detail. I'm posting them just so that you have them, and if you review these slides in the future, you can pick up these articles and go into it. But knowing how the pelvic biomechanics work really helps us select your screw corridors and how you stabilize these different fractures or weak points here. This is a second article from a different group in France. So they both understand the same basic principles, and I think those two articles in combination are really helpful. Moving on, then, to how we provide an added benefit to these patients with our knowledge. We really can leverage our advanced imaging technology. We have a combination hybrid CT fluoroscopy machines. We've got DynaCT. We've got image guidance. This is a great paper that went through some of those different things that we can look at and apply in this field. And then we also have to anticipate the complications. These are the more common complications that we might see listed here, and I've certainly had every single one of these happen to me. In general, they're not a big deal. You just have to anticipate them, look for them, and prepare for the potential and manage the complications that arise. This is a great paper that came out about a year ago that went into the complications from one of the French groups, and I think it's a great review. And lastly, just consider when you're approaching these patients that you really can provide additional treatments by combining your screw fixation with other interventional oncology techniques, whether it's a biopsy or a steroid injection or embolization or ablation. We all have these different examples, and really, what we're applying here, you go back to your basics for every single modality. What is the ablation doing in this situation? What will the embolization provide? Am I doing the embolization to decrease my bleeding risk for the screw placement, or does it have some other function? And you start getting these different permutations that get quite complex and quite fun. I think that really kind of shows us how this group that we have in MSK gets to be very creative, and that's why I really think our group is really neat, because we get to get creative and approach these complicated situations. And there's no right way or wrong way to do it, as long as you maintain your basic fundamental principles and communicate those to the patients and their referring providers. Thank you very much. Great to be involved in this, and this is a very innovative group, and one of the things that's coming forward is how we use different approaches in terms of both planning, guidance, and treatment. So I'll talk a little bit about mixed reality approaches for doing MSK-type treatments. So there's components around device tracking. There's questions around accuracy as you start to remove maybe the direct feedback in terms of what you're doing in terms of placement of devices. Question, is it worth the effort? There's a fair bit of setup work that's involved with some of this, so it takes time to become acquainted with some of these technologies. And I'll just show briefly where the world might be going in terms of education simulation as things get more complex. So a little bit of background first is where did we start with this sort of approach and where are we going? And really, there's a continuum between how you do these sorts of tracking approaches, and the differences is often just in how it's displayed. So Medtronic Stealth System is one that really revolutionized placement of devices in a way, particularly in the OR, so this, if you look at the system that's used, it's basically a circular fluoroscopy system, binocular cameras with indexed devices. And this allowed a fusion of planning with placement of devices, and it's used a lot in the OR. I think it's really important to recognize that there's optical, there's electromagnetic tracking. IMACTIS is one that we use in our practice. Very simple, and you can see a field generator with a sensor that's placed on devices, and here it's placed on a device, trocar device that we're going to be using. You can see, as you look, you can kind of move the needle around and change the angle of orientation that you might be doing for a procedure. Gives you a little bit more confidence in start point and how you're gonna place the device. And I think it adds accuracy, particularly at high-angled approaches. So here, you have tracking visualizations to a screen. If you think about where this is being used a little bit, and most of the innovation's actually happening in the OR. A couple systems I'll talk about. One is NOVARAD. They've done basically an approach where they develop a headset and do augmented reality for use of these systems. In particular, they were the first to be able to use pedicle screws in the OR with FDA approval, now more than four years ago. And here, they start to bring image data into the OR as their approach, and this is presented to the surgeon for what they're doing, or here, to do a cryoablation of the spine. Probably one of the more commonly used systems by AugMedix. They have a custom-built headset. It's a lot like loops that are worn, where, now I'll just explain how this works. Here, the systems are pretty similar, actually, to what Medtronic actually developed, where you have a track system in here. The device is anchored onto a spinous process in the spine so that you can stay oriented in terms of placement. And what the surgeon sees, or whoever might be using this, is two different projections on, basically, a transparent viewer that the proceduralist is wearing. And with this, the devices are also indexed, so it's very similar to the stealth system in terms of the approach. Now, it's just, where does the image information guidance get presented? Now, it's on a headset instead of a screen. These are things that we could actually bring into the IR space. You might argue about putting an anchoring device on a spinous process, but it's certainly possible. They've done over 2,000 commercial cases now, and it continues to grow in terms of utilization. Question, does it work? So I'll just talk about a couple different studies that have been done here. Osteoporotic fractures with the infertile cleft was studied in here. This is a group out of China, and this shows you how complicated it can be, and where we really need to go, and which is segmentation's part of it. So you can see that the spine is segmented in the different levels of different colors. Here they're actually doing this with a VR headset. Raises the, maybe the risks profile a little bit, as you're not looking directly at the patient or at any other information. They published how well does it work relative to just doing these with fluoroscopy, and I'll just point out a couple. One, there must have been a learning curve, because they were able to get their approach to be faster than fluoroscopy, and obviously if you use a non-radiation-based system, the fluoroscopy time should be lower. But if you actually look down the line, this system was perceived as better for all aspects of this procedure. So maybe, certainly people that are really committed to this technology might have been users of this. So that's one approach, which is AR. We looked at kind of we're in a loop-type system. VR, presentation to your headset. Another one, and this is now a different imaging system, where now you've got sensors attached to a fluoro unit, and this is one that was, it's called Clarified by Philips. The team in Strasburg looked at this, and here you can see you basically put sensors down on the skin, our tracking systems, map the surface of the body to the imaging. When this is done, you have basically a multi-view approach to placement of devices so that you can track and place systems, and here you can see segmentations part of this again. Maybe a little idealized, and then you can see the orientation for placement of a trocar or other device into the spine, as shown here. This might have been a little bit more realistic in terms of how well you could do these procedures. Took less time to do a fluoroscopy than it did to use the augmented reality approach. So you can see a factor of two. With this, though, you were able to reduce the amount of radiation exposure for patients, potentially proceduralists over time, and with proficiency, presumably, the speed would go up. So this is another report appeared in the surgical literature where they use the same system. What's interesting is you consider accuracy. It's much more complex, obviously, in the upper T-spine, cervical spine, but gets easier as you go down into the lumbar spine. The red line is the size of the pedicle, so the pedicles get bigger as you go lower, obviously, and when there was basically accurate placement of the device screws was the dark blue or purple line, and anything that deviated from the pedicle is shown in light blue or green. So pretty reasonable accuracy overall in the placement of pedicle screws. So we looked at loops, have VR, you have basically projections onto screens. Another one that was very interesting came out early, which was basically to use a holographic type image. Now use a system that pushes the image down onto a screen laying over the top of the patient, this holoportal system. This screen is what the proceduralist is looking at. They have a whole segmentation suite for doing this work, so you can identify where you're going, and then you can translate what you see on the screen to what you're seeing in terms of imaging with that projection. So this is the projector, basically, over the patient. There was a lot of interest. The initial stock offering came out at $378 when they launched an IPO. I looked just last week, and it's down to $2. So they're continuing to focus their efforts, still in business, but it's super challenging to get into this space. Adoption's critical. So they basically have to live off of selling these systems. They thought, though, these are clearly people that were committed to this, and this is what they published. How hard is it to do? Did they feel like it's accurate? They thought it was, they would use it frequently. It wasn't too complex. Didn't need too much support. They actually liked the interface and felt reasonably confident. Didn't feel like they had to train too much, and it wasn't too difficult to control. So everything basically suggested that they liked this. And here you can see what the screen projection looked like at the time of procedures for doing lumbar-based procedures. One of the problems, background on the problem, is kind of hard to do this in a reasonable way, is you have to either bring imaging in and deal with imaging as we see it, or do a segmentation work. There's a lot of work that's going on in AR and AI to develop basically segmented imaging so it's easier to do these procedures. And a lot of this is actually coming out of the RADONC literature. A lot of teams are building models so they can do auto-segmentation and labeling, and that gets ported into these systems in order to be able to more reasonably use these different approaches for procedures. You also have to be able to, if you're doing ablation, you have to be able to predict outcomes. And we've all known that this has been a difficult path for companies to survive. We actually published that you can actually basically simulate what an ablation's gonna look like. Here's a cryoablation of a rib. Visible line, circled or segmented in red. You can actually predict it with models now. Here you can see a zero degree C line really outlines it well, and the difference was pretty minimal. So you can actually start to blend guidance, placement, and prediction with some of the tools that are emerging. So it can be accurate. So finally, one is doing procedures. The other, hopefully at some point we'll actually be able to plan procedures in a meaningful way, rather than just have a ton of experience to be able to do these safely. One report by a team that looked at how well would it work to start planning these tools, and this is basically a VR-based approach where you can use different tools. Some of the gaming industry is gonna move this forward faster than we do. And here you can see that they're trying to simulate a procedure for how you might do this. I wish it was that easy to place the probes. But this is kind of going in this direction over time. So some of these tools would be really great for training, for example, at least initially. So in conclusion, a lot of new tools are coming forward. A lot of it's coming out of the surgical environment for device placement. AI is really part of the story, so you can do this in a rapid way and get segmentation as part of the solution. I think we'll see ablation planning start to emerge a little bit more as it becomes possible to do these things. And then as a result, I think we'll have new tools for interventional oncology, particularly for hemoscale ablation. Thank you. And during the next 10 minutes, I will try to discuss how robotics can help us. I think when you start with a robot, the question is what you want to do with your robot or what you expect from your robot. And basically, you're gonna think you're gonna do my job and you're gonna drive the car. Or you're gonna call mom when I'm driving the car, which is another option. So he'll do something else when I'm doing something different. Or he can do what I don't like to do, which is always another option as well. And in our minds, we have the idea, and patients, when you discuss with patients, they have the idea that he's gonna do better just than physician. And maybe in the future, but we're not there, he's gonna interact with us and he can take his own decision. I think we are far from this part now. So if we go back to our bone intervention and needle guidance, I think that what we like with our robot is to go out of the actual plants and to go in plants that are difficult to reach for us, difficult tracks. We want him to be accurate, very low part of impact, short learning curves, and no radiation or decreased radiation, which are options. So, I mean, I'm talking about robotics, but in fact, we are not really doing robotics. We are doing cobotics or cobots because our robots are not working by their own. They are extension of our arms, we operate the robots, but they are not doing the jobs as a Roomba is doing at home. And I think this is an important discussion. And most of our device are kind of tele-operated device, remote control device, or whatever you want. And this is a nice example from the industry. This is called a robot, but it's cobotic. It's just helping human to do something it cannot do by its own. Of course, this one is very well known. And if you ask patients, I get surgery by the robots. They never get surgery by the robots. They get surgery by the surgeon using the robots, which is a totally different story. And this is basically more or less what we are doing. We're using tele-operated device to help us to stick needles. And we are not at the point where we have really clever device that's gonna do the job for us. Going back to my needle guidance topic in MSK, there have been a lot of effort during many years to use robot to stick needles in different part of the organs. Today, I will discuss mostly the robot I'm using, which is basically an optical driven robots. And basically, this camera is looking to the arm of the robots and to the patient reference and the skin of the patient. It's trying to work with those landmarks. And we are using computer on the screen to define what are the tract and what we want to do. I will take the example of pelvic osteosynthesis because we know that it's kind of complex stories. There is oblique axis in every direction. It's sometimes hard to place your needles. And I will run you through this case of this 50-year-old man that's got a metastatic ARCC with a large meds to the right iliacs. It's another meds over the acetabulum going down to the posterior part of the pelvic. And this is a case we'll go through with the robot. So trying to summarize, this is a patient we are dealing with. And the plan we want to do in this patient is to insert two screw on the right iliac. It's kind of this way. And we want to do what we call the T. Basically, it's one screw above the acetabulum and one screw for the ischialic access. And if you think about it, those three screw will replace the patient being supine at the table. Then you have to rotate the patient to access the ischialic parts to insert the screw. So how does it works? You get images. From those images, you're gonna do what we call the planning phase. So basically, you decide what is your target. You decide what is your entry point. Then you scroll down. You remember I want to place two different screws. So I'm going a little bit down. I will decide again what is my target or what I want the tip of my screw to be placed. I will landmark it. And I will decide about the second tract. And you can, of course, always adjust. You can always move the tracts. And in this case, of course, and you notice that I've been to be careful not to slip on the side of the bone because I'm not perpendicular to the cortical bone. So remember, we have a yellow tract. We have a purple tract. You can check those tracts in every direction. So you basically are going around and you can check in the axial, sagittal, coronal plane, whatever, what is your tract. And of course, you can do the same for the other tract, which is a purple tract, to be sure that you are going through the bones where you want to be. And I would say now you have done your tract, you have done your planification, and you can move to the next step as far as can move to the next slide. Maybe, yeah. So we are in the room. We are using the robots. I've done my planning. I'm gonna send it to the robots and trying to understand my videos, the arm of the robot will move, but I place what I will call a virtual needle or an optical needle on the robots. So the green line you will see on the screen is where is it going inside of the space when it's moving. So now we are moving the robots. So I'm sending the robots to the yellow trajectories, the first one we designed. And you can see how the robot is bringing the needle on track to the yellow trajectory. At this stage, I'm supposed to insert the first needle or pine, but I'm sending the robot to the purple trajectory for demonstration or we can move the robot from one place to another. So now we're gonna treat the patient. So next step is sending the robots. So I place a pine on the robots and you can see there is a small mark that is a depth mark where you have to stop at the skin to be sure you are in the correct location. In fact, this is not a robot, it's designed for bones, so we don't have the real tools in real life. In the liver, you stop automatically and you don't have to make those kind of, I would say, fine tunings. So we are now sending the robots to the location. So it's moving, if you remember, to the yellow tract. The next step is, of course, to insert needle, pine, whatever you want to insert. Remember, we have those small marks to stop at the skin to be at the correct depth inside of the patient's, robot is move away, and then you are ready to move again the second part, and you are signing to the purple tract. Once again, reaching the tract, I've done a small cut on the skin to insert the pine, and you are hammering the pine. To be honest with you, most of the time in the bone, we don't go from the entry to the target. We do a CT in between to be sure. When we are working in the abdomen for liver, you don't have this luxury because neither are moving. In bones, you are very stable, so you can make control or call them control in between. We keep moving. We are now going to the other side. So remember this screw we want to place above the acetabulum, and the robot is going in the correct direction. You know the story. We are going there. We are hammering, always the same story, and placing the pine. So basically, we are not far from what's happened in real time. So you can see how fast I can place three needles in a single patient without fluoroscopy, which is an interesting point. The job has been done. We have placed the pine. We have placed the needle. We have placed the cement. We have inserted the screw. We are done, and we are turning the patient in the prone positions. And then it comes to the ischialic access, which the guys doing those stuff know it's a complex one because it's very difficult to go in the bull's-eye view because it's a posterior access, and most of the time, you cannot access bull's-eye view. So before the robot, there were kind of all radiologists sitting on my right side, which were doing those kind of techniques. So basically, he's doing fluoroscopy, placing the pine, drawing a line on the skin of the patient, and rotating the tubes, aligning the pine, and drawing another line on the side of the patient, and basically, he's doing sextant. So basically, the image is from the past century, and we are working like in the past century. He's a young guy, but he's working like in the past century. I mean, the new kid instance is here, and I'm using GPS and robots. And this is how it works with the robots. In fact, he didn't know about those slides. And this is what it goes today. So basically, the patient is in a prone position. I want to go in the ischialic access. I'm sending the robot to the ischialic access, so the robot is going inside of the direction. Of course, there is limitation. Of course, the robot cannot go everywhere, anywhere, anytime, and you have to learn how to use it. But basically, you see that we are now in line with the ischialic access, and you know the story. You take the hammer, you push on the pine, and you're in the correct locations. I have to disclose, and I have to be honest, this is the second attempt. We failed the first attempt, and I will show it to you now, because it's not an everyday, everything is nice in life, but it's hard, but there is, let's say, drawback. And one of the drawbacks was this one, that we slipped on the bone, and we were saved. This is why I think it's very important to have the manual feeling of what you are doing. And if you discuss with robotic companies, some are telling you, I want the needle to be inserted by the robot, but I think you need feedback. Or I want the needle to be inserted by the man, and we have a kind of feedback. And this is what's happened at the second attempt. This is the final job which has been done, and this is how we get this ischialic access with the robot guidance. And this is a kind of comparison of what we plan and what we do, so we are very close to what we plan by doing this fixation. How accurate it is, I don't have the answer in the bone, to be honest, because I don't have enough experience. What we learned when we do animal studies in the liver, we were within four millimeters from the target. And very interestingly, it was true for novice, because one guy had never stick any needle in any animals or human before. He was a veterinarian, and get the same result as we do after 15 years of practice. This is our experience in September. Let's say today we have done like 15 or 20 tests with bones, some pros. Any plan, robot don't care, but the actual plans, out of plan, bull's eye view or whatever. You can compare what you plan and what you did, which is very interesting, but you can do that with CBCT as well. Very easy for multiple needle, you see I can send the robot to three different location within a few minutes. Short learning curves, really this is a short learning curves. Low operator impact, really this is. And I think there is no irradiation or very little irradiation, which is probably good news for us. There is plenty of counts. Setup time, it takes time, it takes your team to be experienced with the robots, because you are working with techs and nurses, and it takes time to get used to it. There is a footprint in the room. Pushing from skin to target is kind of stress for us, because we are used to check our needle along the path when you insert the needle. You need some fine tunings. The device I show is not tailored for bone. It's not working perfectly with bone, so we have to work on that. You have to be very careful when you rip on the cortical, because the robot does not know it, and if you keep pushing, you go anywhere, but you don't go to the place, and there is a cost. So trying to summarize, I don't have the answer, what I expect from the robots. It's L for complex locations. I think it can make a complex procedure easy for younger physicians, for more physicians, and I think this is a very important point. And the next question is, does my robot get me? Well, parts of it probably get part of me, but it's not because of design. It's because probably you're going to save radiation, and I might be faster in some cases. So thank you for your attention. I'm Guillaume Corr from Salzburg, and we will see today how to increase safety during ablation in the spine. During treatment of vertebral metastasis, the spinal cord and nerve roots are at risk of iatrogenic injury, and both structures can be directly injured during the intervention, and the spinal cord might also be damaged by ischemia. We have several tools to prevent these events to happen. We have, first, active protective measures, such as hydro- or carbodissection. These techniques allow to push back the structure from the ablation area, and it can also modify locally the temperature. We can also record the temperature using a thermocouple. This might be useful, but it will only monitor the temperature at one point, and it does not assess the integrity of the nerve. And the simplest way to assess the nerve function is to use neurostimulation. To use neurostimulation, we need to insert an electrode next to the nerve, upstream of the ablation zone, and an electric current is applied to induce depolarization. The evaluation of the nerve function is done only by visualization of the muscle contraction. This technique has the advantage of being easy to perform without a long learning period. On the other hand, it only allows to evaluate spinal roots and the motor pathway. The use of evoked potentials allow to explore the whole motor and sensitive pathways, as well as cranial nerve, if it's necessary. However, it takes a long time to set up, and there is a learning phase to be able to do this. So, how does it work? The theory is for somatosensory evoked potential. First of all, an electrode placed next to the peripheric nerve will create a depolarization that will migrate through the spinal cord to the parietal cortex using the posterior funiculus, and the depolarization will be recorded using scalp electrodes. For the motor evoked potential, the scalp electrode will create a depolarization of the motor cortex, which will go down along the spinal cord in the anterior and in the lateral funiculi, then to the peripheral nerve and induce muscle contraction, which will be recorded by muscle electrodes. So we need to insert electrodes, and usually we place electrodes in the tibialis anterior and in the abductor hallucis muscle for the lower limb. And for the upper limb, we put electrodes in the abductor pollicis brevis. For nerve electrodes, we usually place the electrode next to the tibial posterior nerve and also next to the median nerve for the upper limb. Usually for scalp selector, we use a template that allows us to position this electrode like this, and this electrode I position just in front of the primary motor area and in front of the primary sensoratory cortex. The placement of the electrode is done during the general anesthesia, but it takes some time at the beginning to be complete. Once the electrodes are in place, the recording can begin, and it will allow the production of a submatosensitive and of a motor curve. A baseline is recorded at the beginning of the procedure, and then a real-time recording is made during the ablation and is superimposed on the baseline curve. It is not necessary to analyze the shape of the curve, but the modification of the real-time curve compared to the baseline. Two parameters should be monitored, latency and amplitude. The latency should not increase more than 10%, and the amplitude should not decrease more than 50%. For the motor curve, we only need to focus on the amplitude, and it does not have to decrease more than 80%. We will now see examples that use evoked potential. First, this case of this man with a single L2 vertebral metastasis of kidney cancer. After placing cryoprobes and performing a hydro-dissection, evoked potential was used during the ablation. And as you can see, there was no change in amplitude and in latency during the ablation. This allows us to perform a safe and complete ablation of the metastasis. Here is a second example of a man with a sacral metastasis of thyroid cancer. We can see that the metastasis is really close to the S1 nerve root. To treat this lesion, we inserted cryoprobes inside the lesion and a thermometer next to the S1 root between the tumor and the root. During the ablation, the temperature against the nerve dropped below 10 degrees Celsius, but the ablation was stopped only when evoked potential began to change. And as you can see, here is a nerve and here is the ablation zone. It was possible to perform a thermoablation as close to the nerve as possible while preserving its safety. And this, in order to increase the efficiency of the thermoablation. The use of evoked potential has some limitations. A pre-existing neuropathy, or if there is a tumor invasion of the nerve, we won't be able to produce curves in order to use this technique. Another limitation comes from the clavarium ablation itself, and the inertia of the ice after the gas has been stopped is a problem. And we need to keep this in mind when we decide when to stop the ablation with the information we have from the evoked potential. These techniques are complementary, but will not replace the other protective measure. Another problem comes from the vascularization of the spinal cord. Arterial vascularization of the spinal cord comes from medullary artery, and the flow usually comes from a segmental artery between T8 and L1 on the left side. But if we do a cryoablation next to this artery, this may lead to vasospasm and to a risk of spinal cord ischemia. Here is an example of a man with a paravertebral mesothelioma that is on the right side between T6 and T8. Before performing the thermoablation, an arteriography was performed to search the anterior radiculomedullary artery. In its absence, allows us to perform a secure treatment by cryotherapy. And the last example of a man with right paravertebral metastasis between T1 and T3, pre-op angiocity revealed the presence of radiculomedullary artery at this level. And due to its presence, we did not perform the ablation due to the risk of ischemia. In conclusion, to increase safety during thermal ablation, it's possible to use evoke potential for lesions that are localized in the posterior part of the vertebra next to the spinal canal, but always in association with thermocouple and hydro-dissection. And in case of anterior or paravertebral metastasis, arteriography should be performed to search for an anterior radiculomedullary artery to prevent spinal cord ischemia by vasospasm. Thank you for your attention. This talk is about electrochemotherapy of metastatic epidural extension. As you know, for vertebral metastasis without epidural involvement, there are several options that we can have to destroy the tumor. We can offer thermal ablation. We can perform stereotactic radiotherapy or surgical vertebrectomy. There are less options for vertebral metastasis with limited epidural extension, meaning vertebral grade 1 on epidural spinal cord compression scale or grade 2. Thermal ablation is usually contraindicated because it's too risky for the spine and for the cord. And SBRT is probably the only option to have a complete destruction of this metastasis, but it is also quite risky for the spine. So usually the risk of local failure is relatively high in the epidural space. And finally, for vertebral metastasis with an extended epidural involvement, meaning grade 3, grade 4, or grade 5, there are no curative options, and we can only offer conventional external beam radiotherapy in association with surgical decompression. The bad news is that after this treatment, the progression is frequent and represents a challenge because re-irradiation is usually the only second-line treatment available with even poorer outcome. So what can we do, like in this case, when we have a recurrence after radiotherapy and surgery? This patient had a melanoma, a metastasis in the central canal, that had been treated by stereotactic radiotherapy six months before. He recurred in January, and then he progressed. So we performed a surgical resection of this metastasis in June. Unfortunately, one month after the surgery, he had a very huge and rapid recurrence for which we don't have any options. So what can we do except looking at the paraplegia? In this patient, now we offer electrochemotherapy. Electrochemotherapy is a treatment that refers to a combination of two treatments, reversible electroporation and a systemic administration of chemotherapy to enhance the systemic effect locally of the bleomysin. Electroporation is a phenomenon that occurs when cells are exposed to high-intensity electric pulses. And depending on the amplitude and the number of the pulses applied, electroporation can be reversible with membrane permeability recovery, or irreversible with loss of cell homeostasis and cell death. And the electric field threshold is over 400 volts per centimeter for reversible electroporation, and over 800 volts per centimeter for higher E. So what about electroporation close to the spinal cord? We have this numerical feasibility study that has been published and demonstrates that we can cover the tumor almost entirely, even if there is some epiduritis. And the author says that hypothetical damage to the spinal cord exists, especially if the electric field is above the assumed irreversible threshold that is over 800 volts per centimeter. We know also that there can be some damage after electroporation on the nerve, but it seems that the endothelium architecture is preserved, and there is probably a potential for regeneration. So we go for this patient for electrochemotherapy by inserting probes from each part of the central canal and delivering the pulses and the chemotherapy to have a high concentration of bleomycin inside the tumor cells. And look at the results. This is one month after the electrochemotherapy of this tumor that is a fast-growing tumor compressing the cord. And one month after, we have a complete response with a grade 0 metastasis and a huge improvement of the symptom. Here is another case, a 39-year-old male suffering from a metastatic paraganglioma. He had this pathological fracture that compressed the spinal cord. Unfortunately, he has already been treated with radiotherapy and also with cement. So there is no other option except this electrochemotherapy that we perform in September 2021. And look at the results. At one month, we have a partial response. We have still this area of enhancement in the central canal. But the symptoms were really improved. And this is three months after the procedure. We have here a complete response of the metastasis without any viable tumor visible. The patient has no systemic chemotherapy during the whole time. And now we are one year after the treatment, and he's still in complete response 12 months after the local treatment of this huge epiduritis. So it's not only about treatment after radiotherapy failure. It is a kidney metastasis that we have treated with radiofrequency ablation and cementoplasty to avoid a local progression. Unfortunately, as you know, we have some limits close to the central canal. And here is a small recurrence close to the spinal cord at one month. This small recurrence progressed rapidly. And nine months after our terminal ablation, we have this spinal cord compression. And here is the MRI. So we perform electrochemotherapy for this metastasis using six probes. And we insert the probes from each part of the central canal and the spinal cord. Here is the procedure. It is performed under general anesthesia. We insert the probes under CT guidance. And then we deliver the pulses, which are 600 volts per centimeter for pulse of electrode located from each part of the spinal cord, and 1,000 volt for electrode located in the same side of the spinal cord. It's quite a fast procedure. It takes, let's say, one hour to do that. And look at the result. We are here CT two weeks after with a complete devascularization of this kidney metastasis. And here is the MRI at one month. We have a complete destruction of this metastasis that compressed the spinal cord. And of course, a huge improvement of the symptoms. We can perform electrochemotherapy at two or several level in the same procedure. Here is one spinal cord compression for paraganglioma that compress the cord at T8 and T9. So we perform electrochemotherapy in a single session at these two levels. Look at the result at one month. We have a partial tumor response. There is still some contrast uptake at T8, but also at T9. But the spinal is decompressed. So we planned a second procedure, but the patient said that is fine. The symptoms are really improved. So we get for two more months and follow the patient. And look at the MRI at three months. The tumor keep going and shrink during the following months, probably because of the chemotherapy inside the tumor cells that result in cell death. We have here almost a complete treatment at three months at T8 and also at T9. So for now, we have performed the treatment with electrochemotherapy in 40 consecutive patients over the last two and a half years. The metastases were located or compressed the spinal cord between C5 and L1. Most patients had compression at one or two levels, and most of them have an extended epidural involvement, meaning at least three, four, or five grade. We had complication. Most of them are acute pain during the following days, about 25% of the patient. And some have prolonged radicular hypoesthesia, about 10%, probably secondary to acute electrochemotherapy induced inflammation close to the nerves root, like in this case of metastasis from lung cancer located at T2 levels. We have also three severe complication where a patient become paraplegic, so 7.5% of our population. Here is one where we have performed probably a too powerful electroporation with several probes and too many impact close to the central canal. So this is a configuration that we don't use anymore because it's too close. This electrode is too close to the spinal cord. And what happened is two weeks after the procedure, the patient become paraplegic. It takes two weeks, probably because the time of the valerian degeneration. And when we performed the MRI, we realized that the tumor is in partial response. What we have here, hypersignal in T2 at the spinal cord, probably related to the toxicity of our treatment, very similar to what we have with the myelopathy after radiotherapy. About the follow-up, now we have this 40 patient, and at one month, we have a huge decrease of the pain from 6.1 before the procedure to 1.9 at one month. And about the response at MRI, at one month, we have 46% of complete response, meaning grade zero, no epidurate at MRI, 31% of partial response, meaning at least a decrease of one point on the epidural scale, and 23% of a stable disease. So to conclude, definitely electroporation is a new option for metastasis with epidural extension or for metastasis that compress the nerves. Our preliminary results are very promising, at least for short-term local tumor control and long-term follow-up is needed. We have a few but severe complication in three patients, 7.5% of paraplegia in this very advanced patient, and better understanding of the electric field across the spinal cord is mandatory, and we are now working on it. Thank you for your time.

interventional radiology

robotics

spine procedures

screw fixation

minimally invasive

advanced imaging

biomechanics

oncology interventions

mixed reality