Well, thank you all for staying and thanks to our moderators for this lovely session. It's been great. So I will try to go through these slides quick so cocktail hour can be upon all of us. So many of you have seen various series from various institutions that look at subdural electrode outcomes and SEG outcomes and increasingly there have been several groups that have published their SEG outcomes. I was at the cusp of this transition pretty early on and I thought as now years have gone by it would be a good time to look back and really see how these two populations have changed what we do. Nothing relevant to this particular presentation. So of course this has been our standard in many institutions is still the standard of placement of electrodes and we all have the CT scan in some patient or the other that shows a lot of shift, a lot of problems with electrode placement as well. And this is not something that's unique to any one of us. This is a very nice analysis, meta-analysis of epilepsy complication rates that compiled 61 papers with 5,600 patients. And to spare you all the effort of reading it, if you just pay attention to this one table that I generated from their data, you can see also in years past the complication rates for definitive epilepsy surgery were not insignificant. But even today in the current era, which is from 1996 onwards, where most of these risks have dropped substantially, the risk of electrode placement is in published series, which therefore means large volume, reputable centers who are proud of the work they've done enough to publish it. There's approximately an 8% complication rate of what is essentially a diagnostic procedure. So John Taylorak, of course, many of us know him from various things that he did, but important to remember that he really started this whole thing. The psychiatrist originally who decided that psychosurgery was far too important to leave to neurosurgeons and became a neurosurgeon himself, developed then this technique of three-dimensional space and in this three-dimensional space worked to implant electrodes. I bring all this up because this traditional way of implanting electrodes, SEG, guided by fluoroscopy, guided by arteriography, is highly lateral to medial based, is very orthogonal placements to the midline, and has a tradition that has limited in some ways its broader applicability as well as the amount of time that it takes to accomplish this operation. And I think many of these individuals played a big role, I think Hans and Patrick certainly at different times at the Cleveland Clinic, and then Jorge and John, they were the two neurosurgeons I think in North America who started doing SEG before I did. And so I'd certainly, you know, look to both of them for their contributions in this. And what they did was they used imaging and they got rid of arteriography. In addition to that, I think we've all made some innovations. I started using the 0.8 millimeter electrodes instead of the 1.2 millimeter electrodes. I think that's now mainstream. And of course this azimuth based placement, idea being that you no longer have to be constrained by orthogonally based placements and the electrodes can go in any trajectory necessary. This also opens us up to the idea of 3D grids with John propagated. And then things that many of us have since done, staged implants and also combination of SEG with grids. There's various ways of doing SEG. I'm going to talk mostly about the ROSA. This is of course not unique to our center. This is just a graph to represent the amount of volume of change of publications relating to SEG, especially in the past three years. So our own experience now, the only other large series of similar size that was historical was the MNI series. Cleveland Clinic has a very similar result as well. And 160 patients with 2150 electrodes, we have the same results. Giorgio was in the room. He has a similar series as well. I certainly can't bring up everyone's data here. But the point is that we can reach that holy grail with SEG of zero complications, I think. And that should be the goal for what is, as I said before, a diagnostic procedure. So how precise is it and how precise does it need to be? So these are the X, Y, and Z. Sorry, that should be a Z there. Errors, the entry point error, the site deviation, the target point error for a cohort of the first 86 patients we measured this in. And if you now look at angle trajectories, another thing that was not done in the past, when you really change how steep the angle is, angle of incidence to the skull, including angles that are greater than 45 degrees, very oblique trajectories, the error rates are not enormous. They certainly go up, but they're not enormous. And it's still acceptable if that is the goal that you need to meet to target the areas that you want to study. So this is the comparison that we made. This is 260 patients, 139 of these were subdural cases, and 121 were SEG cases. Three of these were patients that had more than one implant for GRIDs, and five were SEG patients that had more than one implant. Now, as you can see, like in many centers, some patients might get SEG, then switch to GRID, some might get GRIDs, then switch to SEG, and this happens. And there's a certain amount of noise in this, if you will. And so then we finally use a, let me come back to this. We use this method of allocation of each type of patient to see which group they should actually be studied in, what was the last procedure that led to the definitive operation that was done on them. Let me go back here, let's see, yeah, there we go. And so, as you can see, this was our chart for subdural electrode implantation. We were on this nice upward trajectory until SEG came along and disrupted it, and now this is the SEG curve. 2017 is incomplete data, so it's, the growth rate is accelerating there as well. So with that assignation, what do we know? Well, here is firstly the amount of time that it takes in the OR. So these are the SEG patients. As you can see that the red curve is lower than the blue curve, which is the GRID patients. As that trends down, there is a learning curve, maybe the first 30 or 40 patients, and then it declines quite significantly to the point where the curves are almost parallel to each other, but they're separated by about a two-hour difference. I'm going to allude to this later in oral morphine equivalence in terms of the pain medication requirements that these patients need. So here it is, here is the comparison. So this is the implant OR time that is significantly different. This is wheels-in to wheels-out time, so obviously that includes registration. In my case with the ROSA, it includes placing skull fiducials, a trip out of the OR to the CT scanner, because we get a CTA with every case, and then back to the OR. So this is about five hours with some noise, and it's about seven hours for GRID placement. The actual implant time is much, much shorter, so it's about two hours for SCEG, and give or take about five hours for GRIDs. Of course, with SCEG, the electrodes are sometimes more closely spaced. We can also implant more electrodes more efficiently, and so the number of electrodes implanted actually goes up. And I do feel that this has something to do with the outcomes that we will talk about in a bit. Nineteen patients in the GRIDs population out of 139 got transfusions. I think that's pretty standard across many institutions. This is transfusions not just from the GRID operation, but transfusions during that hospital stay, because many of them would then come back a week later for their craniotomy, and now they've had two craniotomies or two cranial procedures within a seven- to ten-day interval, and they've lost enough blood to receive platelets and or packed cells or both. And the number of days of monitoring is the same. And of course, it's important to point out also that the age at surgery, the number of medications failed, the age at onset of the epilepsy, the duration of the epilepsy, the gender distinctions are all the same. So it's a very similar cohort in that sense. So here's the next point of comparison, the morphine consumption. So this is a huge difference in the amount, in the milligrams of morphine consumed by these patients, which is an indicator of how much pain and discomfort they suffer from one operation or the other. And then all of us are very fascinated with this operation, but I still had nine complications in this GRID cohort. Most of those were minor. Five of them were removal of hematomas, where we continued the monitoring. Two were infections that were superficial, needed antibiotics for a period of time. And one was an intracerebral hematoma that had an effect on cognitive function. This is the next obvious difference between these groups, that the percentage of patients who then undergo a definitive resection and or ablation is about 90% in the GRIDs population and about 73% in the SCEG population. And keep that in mind. We'll come back to that issue later. But what was done eventually was not that different. The number of people who had either temporal or neocortical resections were the same. The numbers that did not undergo intervention, of course, are slightly different, but that was not significant. So this is then sort of the flowchart of the outcomes, and I'm going to actually just jump to the next slide because that will illustrate this better. So here are the outcomes for the patients who had GRIDs. Class 1 or Class 2 at six months is 65% for the GRIDs and about 84% for the SCEG. And then at 12 months, it's about 72% for SCEG and 54% for GRIDs. I think these numbers for GRIDs are pretty standard relative to most other series at one year where it's about a 50% seizure-free rate, and the numbers for SCEG are significantly better. Now you could argue not everybody that you evaluated with SCEG had an operation. What if you do sort of an intent-to-treat analysis where you take all patients who had the procedure done regardless of whether or not they had the definitive operation? Well, in that case, there is no difference in outcome at all. So if you take all comers, you have no worse outcome with SCEG, and if you take the patients you actually operate on, you do better. And we'll come back to that issue in a minute. Also if you plot this now as Kaplan-Meier survivor curves, you can see the difference between the two. So this is the summary at one year, 71% versus 54% with SCEG. And I would still say before I move on to my sort of conclusions of this, is that there are still indications for GRIDs. You know, while these are many nails in the coffins of GRID implantations, this is still not the end of subdural electrode placement, and certainly we do still implant subdural GRIDs in patients who need language mapping, who cannot be done awake, or who have very posterior neocortical temporal onsets. But in our practice, most of those patients would have had an SCEG evaluation first to make sure that it's not an anteromedial temporal lobe epilepsy, ovary frontal epilepsy, or some other deep-seated epilepsy that we haven't thought about, like insular remedial parietal. So to summarize, and to also address this question of the selection of patients, I would submit that the patients who come to us are selected for needing an intracranial evaluation. It is our job to decide which of those needs an operation, and indeed the procedure that does better by not subjecting those who would likely not benefit from surgery to an unnecessary resection is a good thing, and this is indeed what the test should do. And if you think of this, again, as a diagnostic test, and we have a test that actually yields the results that would lead to a better outcome more effectively. So to summarize, SCEG methods are better tolerated, it's certainly less resource-intensive, we're also looking at a direct cost issue with these patients, it's more accurate in placement of the electrodes, it's more predictable where the electrodes will be, it's less morbid, less committed to an intervention, and then the outcomes associated with SCEG are at least similar and superior in those who actually undergo a resection or an ablation. So I think we can say today that if there is any equipoise in whether GRIDS or SCEG could be used in a patient, then we must use SCEG rather than use GRIDS first. Thank you so much. Thank you. You were talking about these electrodes that come in more diagonal with more than 45 degrees and they have more error. Are they associated with more complications by poking things we don't want to poke? So we've looked at, as I said earlier, we've had luckily, and I'm sure one day this will change, we've had zero hemorrhages and zero infections so far. And I think a lot of this has to do with tolerance that you allow around each trajectory. So I assume that my electrode is going to be off by two millimeters in some direction and so I leave a four millimeter zone of tolerance around each electrode. And so I think that's why it doesn't matter so much. Well, thank you so much. Thank you.

stereoelectroencephalography

SCEG

subdural grid

electrode placement

epilepsy surgery