Okay, thank you very much. And then next we have Dr. Volker Kohnen to discuss supralateral MFB-DBS for depression. Well, good morning. Thank you for having me. Thanks for the invitation to the committee. I feel very privileged about that. I have the chance to talk about our results with supralateral mediophore hormonal stimulation in depression, something that is still in the open-label phase. I'd like to show my disclosures here briefly. And these are because typically at the end of the talks you typically don't have a lot of time to talk about who you're working with. So I just tried to show that at the beginning I'm working with Horst Obacht from neuroradiology, who's helping me a lot with the imaging. And then, of course, Tom Schlepfer, who's the psychiatrist who trusts me enough to give his patients into my hands to do deeper in stimulation, and who's probably one of the most important figures here in psychiatry leading the field here. Depression itself is a common mental disorder. We think that, or the WHO thinks that more than 320 million people are affected. It's a leading cause of disability work worldwide. More women are affected by depression than men. And they revised that a little bit. 800,000 patients per year commit suicide based on depression. So it's a gruesome disease. There are effective treatments for depression, drugs, psychotherapy, ECT, transcranial magnetic stimulation. But if you look at the results from the STAR-D trial from the early 2000s, you will notice that even if you only give one drug in the beginning, you only have a remission rate of up to 35%. So that's remission and not response. So it's a hard-to-treat disease, actually. And more than 20% of those patients are going to be treatment-resistant in the long run. Obviously, this is a world of depression. And the 5% incidence in a population is shown here with a lot of countries. And you cannot read that. So the Americans are down here. So somewhere between 4%, maybe the Germans are a little bit higher. And I've been thinking a lot about that, what that might come from. So maybe we're a little bit more stubborn. You know, the German machine, you know, like to work, not to, you know. So, but I couldn't make my mind up. But the interesting part, actually, is that I couldn't go back. But right on the left side where the depression is very high, that's the Dutch. And I do not, I do really not get why the Dutch are up in the high depression. But they are, obviously. When we're talking about deep brain stimulation or treatment for depression, we have to acknowledge that the psychiatrists themselves, they basically have different forms of depression. We're only talking about major depressive disorder, single or recurrent episodes. But if you look at the 322 million that are affected, that's that whole group, right? So I'm not saying that 20% of those 322 million are prone to be treated with DBS or are treatment resistant. But a subgroup of them, and the numbers are really difficult to take. Very nice paper from the London group that I show. And it's basically the best summing up. And it shows in how many regions, basically, DBS has been tried to work for depression. I'll just show them. That's the set cingulum of Hal Mayberg and Andres Lozano. Lozano, lateral hyphenula, our mediophobic mononucleus accumbens. And essentially, a lot of the things, and you see that from the crosses, how they come here, a lot of the targets that are used for depression are actually used for OCD. And that is very likewise the case with medication. Now, where are we? We have two stopped pivotal multicenter trials for DBS and depression. And all the whole field went into turmoil and, oh, God, how could that happen and all that? You have to acknowledge that when you look at drug studies in the phase 2, phase 3 region, more than 95% of the studies fail. So it's not like we have two trials and they failed. So it's just a very, very expensive therapy that is typically driven by researchers, not by the industry. Those two trials were early driven by the industry. And the RECLIM trial from Medtronic researching BCVS stimulation was stopped in 2010 because a futility analysis showed that this trial would not reach its primary goal. Same happened to the BROADEN study. That's CG25 or SCG25 stimulation. First ANS and shoot medical. Now that's Abbott. And that was stopped in 2012 for the very same reason. So you cannot say that therapy is not effective. But what you can say is that those trials didn't work. That's a slide that I have by courtesy of Tom Schlepfer. And it basically shows what happened in the RECLIM study. So those patients were implanted. And you see the Madras percent improvement here, 0 to 40%, which is already interesting to show it because the response rate is there. So that whole group would have been responsive if you would have shown that. So it's not even going down to the response rate. And what you see is basically that the actively stimulated group in green here does not significantly differ at week 16 to the control group. And if you then look at the long-term open label and everybody says, yeah, there's a trend to be efficacious in the long-term and all that, you again see that whole group gets better over the time. They are handled a lot because they're still seen very often by psychiatrists who deal with them. And that is very untypical for this population. This population is typically sitting at home. They're not seen by psychiatrists anymore because the psychiatrists tend to not see them and then forget about them. And then they're very similar to surgeons. Then that's a great success because they don't see patients that are treated in the refectory. And again here you see that they're not, even after 24 months, they're not reaching, not the whole group is reaching the response criteria. So this is a stimulation, or that's the paper from Lanza Psychiatry for Alzheimer, Andres Lozano and Helen Meiberg. 90 patients implanted. And actually the most important thing here is this graph that you see here. The stimulated patients start with the mean score, the Madras mean score up here, go down there, and that's the control group. And actually after the sham condition, from here, the whole group is stimulated. And you don't see anything reacting. There's no basically little ditch in that curve. So it appears to me that nothing is really happening. And again, and I like to show that, that's the response criteria. So actually that is not an effective treatment. In the long run, you then see that they're somewhere going into the response over time. But if you look at the recent publications, even the open publications in the Atlantic recent paper, you can see that Helen Meiberg now is saying that the response criterion for DBS should be 40 percent because the patients feel that they are doing so much better, which is, I think, a difficult thing. The interpretation was the study confirmed safety and feasibility of SCG DBS. And it showed significant statistic antidepressant efficacy in a six-month double, failed to show this significant antidepressant effect. And further studies are needed to investigate factors such as clinical features or electrode placement that might improve efficacy. And then in the longer discussion, it says it is notable that the sample on average had more chronic depression than in previous studies. Okay. And then targeting use in diffusion temperature imaging could significantly improve the efficacy. There might be some truth to that. And I'm the butcher who is selling the meat, so probably that is something that I would say as well. So what about mediaforum bundle? We have done a little study simulating what the effects of the stimulation would be if we would basically overlay that with mediaforum bundle imaging technology and simulated electric fields. And from that study, basically, this picture comes. And we know that we have a 60 percent in open-label trial. That's the sub-causal cingulate here. We have VCVS in open-label trial, 53 percent in that first paper with 15 patients. Nucleus accumbens, 13 patients, 50 percent response. And then we are here with an even smaller number in a pilot trial just looking at is there a signal and can you do it in seven patients, 85 percent response, which essentially means six out of seven responded. One patient did not respond at all. The reason why we are looking at the mediaforum bundle is because we think that there is an overlap between the mediaforum bundle that you can basically see here in a heat map. And if you stimulate with a medium current of 4.5 volt, so that's a voltage of 4.5 volt at an amplitude of 1,000 or 1 kilo ohm, you would basically get electric fields that would cover in alec of the interior limb of the internal capsule, VCVS, nucleus accumbens. And in SCG, we think, we at that moment in time thought that the results would interpret that towards that the mediaforum bundle would reach that. We do not think that anymore. And here's a map of where the mediaforum bundle can be found just lateral of the ventral tegmental area, then going in the frontal vein. And I'd like to point out that we probably have to reshape our thinking about what the reward system in the human really looks like. It's not only hypothalamic. It's not only dopaminergic driven and all that. But obviously, while the neocortex has developed evolutionary, other expressions of reward system and motivated learning basically have been, have found their expression in the frontal lobe. And we're going to see a little bit of that in the next results. So that's rodent studies. That's a very nice paper from Russo and Nestle in Nature Reviews Neuroscience in 2013. And they were discussing basically what the role of the reward system in affective disorders were. So obviously, a lot of scientists, especially basic scientists, come there. From, when you look at the ventral tegmental area, so just media of the mediaforum bundle, there's a heavy glutamatergic accelerating drive to dopaminergic neurons down there. And a little bit applies that also to the nucleus accumbens. And there's a direct connection between the VTA, the nucleus accumbens, and direct connection of the VTA to the mediaprofrontal cortex, which is then later interpreted as orbitofrontal cortex. However, if you look at DTI study, which we heavily do and look at it, you have to acknowledge that there's not only a connection between the ventral tegmental area and the orbitomedia prefrontal cortex, but also to the dorsolateral prefrontal cortex. A lot of the regions with reward-related learning are basically associated to the mediaforum bundle. And thanks to Ludwig and his group, we now know that basically the ventral tegmental area is really posterior to the hypothalamus. And he has done some nice work on posterior hypothalamus, or I must say VTA stimulation to treat cluster headaches and other headache forms. That's a study that we recently published, 55 normal controls with a multi-shell imaging protocol that we've used. We seeded in lateral to the lateral and in the ventral tegmental area and looked what the mediaforum bundle would look like and where most of those fibers actually go. And this is basically where the mediaforum bundle goes in the frontal cortex. And if you look at that number, it does not really tell you. It looks like it's going everywhere. But if you then look at some parcellations, there's a superior frontal parcellation, the rostral middle frontal parcellation, which is this blue guy here, and you see the lateral orbital frontal parcellation, which is basically that red guy down here. Then you can see that basically 97% of the fibers from the VTI actually ending there. So there's a clear role for the dorsolateral prefrontal cortex, not so much for the media prefrontal cortex, and certainly the established role for the orbital frontal cortex. And if you look at this heat map here, you can see how heavily Brotman's area 10, a little bit 9 and 8 and 11 are addressed by the fibers that are coming from the ventral tegmental area. A brief update on what we do during surgery. We try to place our electrode essentially at the same spot where we do the target seeding and the seeding regions for the DTI. And we try to intercalate our electrodes in front of the red nucleus, just media to STN and SNR, just lateral to the ventral tegmental area. And we do electrophysiology intraoperatively. I'm not going to talk in too detail about that, but that is basically excluding electrophysiology. We want it to show on the lateral trajectory, we want it to show STN and SNR, and being in the region where cellularly there's the least activity, and that's where we think then the media formal bundle is. We do intraoperative stimulation. We always see, if the patient is not on beta blockers, we always see an increase of heart rate by 8 to 10 beats, which is there for 2 to 3 minutes, then goes down. Even the blood pressure can go up a little bit, but then comes down again. We like to see an acute effect, which you most of the time see. At the bottom of the implantation, at the deepest point of the implantation, we can actually see, we always see that there's an oculomotor activation. The lateral pigmented nucleus, which is a bipartite part of the ventral tegmental area, that's where the oculomotor nucleus basically exits the brainstem, and that's why you have that effect, and we like to see that at 1.5 milliamp. And if you then do a little bit of calculations on atlases, a typical implantation should look like that, and then probably that's going to work. If you look at where the effective contacts are actually located, you see there's a lot of variability. I put that on a Schaltenbrand plate just to show if you normalize the ACPC coordinates, you see how much variability is that. Look, that's around 3 or 4 here, and that goes up to a laterality of 12, right? And that's the lowest part, that's the targeting points, the black dots that we put in here. That's from the original publication. If you then look again from the other way around, if you look at the MRIs, and then try to find out, so where is those electrodes sitting? We arbitrarily basically designed this therapeutic triangle on three levels, and then the yellow ones are the responders and the gray ones are the non-responders. So you somehow want to be in that triangle, but we do not know where in the triangle we need to be, and we think the reason for that is, or you cannot predict that only with the DTI information. So very likely, the DTI information drags you to the correct point in that triangle. What we think we're doing is, so there's glutamatergic nerves coming, projections coming down in Arnold's bundle to the ventral tegmental area, the medial forward bundle projects up there into the nucleus accumbens, and if you stimulate that, you drive orthodromically and anthodromically the activity in the ventral tegmental area, nucleus accumbens, and the dorsolateral prefrontal cortex, and the orbitofrontal cortex. We do not know exactly how that is working, but that's our prediction. So that's the first publication that we had. That's from 2013. We see how those basically, within weeks of stimulation, they jump up. In the first week, they basically jump up with their Madras and Hamilton scores. This guy is not responding at all, and she stayed a non-responder. This guy is responding, then going down, and then reaches the response criterion again, and you can see here how much the single patient contributes to the mean, which is plotted here. We have had an independent replication from the Houston group around alphanoid, and they essentially show the same thing. They have the same target region, and they see how that jumps up, and bear in mind, at that moment in time, we were not clear about the placebo effect, which leads us to the next study that we basically have done. The placebo effect and the stun effect, or the microlesion effect, plays a certain role in here, like in every good target, obviously. So that's already included here because they start to stimulate basically right from the start here. That's the long-term results of the first cohort that we implanted, and it shows that basically the therapy, if it's working for patients, it keeps on working, but of course those patients are handled again on a monthly basis. And I wanted to show that we have significant problems with the study designs in those situations. And very, very importantly, we need a phase after the implantation. We need a phase of stimulation parameter optimization. We need a stabilized response. And then probably a discontinuation of stimulation with rescued criteria. And the problem is that we do not know how long this is going to take. So that is really going to lead into adaptive study designs, which is very likely the reason why those two other trials have failed for depression research. We did the 4C2 trial, which we call an explorative gateway trial, with an eight-weeks placebo phase, and that's the plot of it. We assessed 300 patients for eligibility. We excluded 284, included with 16. We did IPG implantations and also SLMF BDBS. And then basically for eight weeks, one group was left off and the other group was basically directly stimulated. And this is basically paradigmatic about a lesioning or a placebo effect, because you see how the complete group, so the stimulated group and the sham group, directly reacts and then declines. And then the stimulation group recovers a little bit. But if you look then here, that's the response criterion again with the 50%. So the stimulation group at eight weeks barely reaches the response criterion, but it does, and the other does not. But that's not statistically significant. So the problem is we really think, and if you look at those patients, that's the lesioning effect. So something is going on there. They are psychomotorically different than they used to be before surgery, and that's a huge discussion with the reviewers about the differences between placebo and lesion effect. If you then look in the open phase, so then they're not blinded anymore, everybody is stimulated, you actually see again the same effect until 52 weeks, and that's the response criterion. So obviously that again, if you're in the open label phase, that works again. But why did we do that? So that was a trial that we did to better appreciate the placebo response, to better appreciate what is actually, what's the magnitude of the effect, and we totally overestimated the magnitude of the effect because of our first seven patients who basically were so dramatic. And now we're actually doing a third trial, 4C3, which is again going to be oligocentric, so it's just two centers at the moment. Maybe we'll recruit two more centers into that to get an idea. I'd like to say that we had some adverse events. We had one intracranial bleeding. That's actually the non-responder from 4C1 who had an intracranial bleeding, never responded to that. She recovered, but never responded to stimulation with a suicide attempt, which you have to reckon on if you have patients with suicidal ideation. And this down here, which is named as an adverse event, blurred vision, double vision, strebism, you want to see that. That's idiosyncratic basically for the target. You want to be close to oculomotor. If you're close to oculomotor and you stimulate down there, you're close to VTA, and then you very likely have a good result. And as a comment to, because I forgot to say that, to placebo response and lesion effect is those who have a good lesion effect or placebo response, they will in the long run react better than the others. So that's very, very similar to what we see in movement disorders. I'd like to show this as the last slide. What we're working with now at the moment is because the DTI technology is something that can be very arbitrarily used. You can carve the fibers as long as you want, and they will show up as you would like them. And this is basically a machine learning algorithm that has learned on a manually labeled MFB and the other structures like fornix around it in three anatomical controls and then did an fully automated tracking in one patient that you see here. And we hope that this is going to lead us into a future of better targeting for fiber tracks, which is at least in that part of the procedure is more objective. I'd like to thank you for the attention, and if we have time, I'm open for questions. Thank you. Well, thanks for a wonderful talk, and it's good to see, of course, the sustained interest in DBS for depression. Do you feel that somehow medial forebrain stimulation is just qualitatively different than other DBS targets that have been tried, either clinically or somehow neurophysiologically? Yes, I think, because how rapid those patients respond, basically, we can reflect on the interoperative acute effect on how is that going to work later, a little bit taking the individual patient's situation into response. The point that I guess I'd like to make is that we, as neurosurgeons, we are in control of that target, because if I basically show each and everybody of you who has used microelectrodes, who has used test simulation and all that, at the end of the procedure, you will know that you're in the right position because you have biomarkers surrounding you, you have an autonomous effect, you have an acute effect, you have micro-recording sound, you have the oculomotor effect. So, at the end of your operation, you know that you did your job right, right? And we can define that. We can define that by imaging where you want to be. We can define by the interoperative effect, by the post-operative effect. That's different than the other targets. The other targets are electrophysiologically on the OR table. They are silent, so you cannot really understand what you're actually doing. And this brings depression surgery, if that really holds true multicentrically, that might bring DBS surgery for depression into the realm of how movement disorders can be treated. So, that's, I think, is the biggest difference. And what, so if you were to posit the primary mechanism, do you think it's somehow enhanced dopaminergic transmission, something like that, if you had to put your finger on it? I don't know what. I think it's enhancing the glutamatergic transmission. I think the biggest effect is really the anti-drumming activation of Botman's area 10, 9, 8, which basically do not nicely communicate with the ventral tegmental area. The problem, very likely, is that we have degenerative processes in the ventral tegmental area where some of the dopaminergic neurons, and that's the tonic ones, the tonic firing ones, are basically dying off. And what we probably do with our gamma band stimulation, which 130 hertz basically is, we basically transfer some of the phasic ones into tonic ones, basically, but that have that nice effect. And that would very much explain some of the effects, which we also see actually in OCD, very, very rapid effects that you can also see on the OR table. So, but we're doing PET studies on that now to get an idea. I don't know. It's just speculation. So, and then it's the last question. So, if that's the case, then what would explain the potential lesional effect that you see with sham stimulation? The, I think the problem is that, so the idea is, and that's highly speculative, as in OCD, you have a very high activity of VTA neurons, of dopaminergic VTA neurons, that makes you susceptible to stress. And that leads to anxiety, at least in the rodent model, if you have a high activity of VTA neurons, that leads to anxiety. Some of those patients have a lot of anxiety and OCD components on top of it. If you put in those electrodes, you silence those highly phasic neurons for the moment, and it might be that this is actually leading to that lesion effect. So, I could, and that's all speculation, because we have no, but that's the route that we're actually following. Hi, Volker. Thanks for an excellent talk. I mean, this is a huge clinical problem that we need to try and address if we can. I have a question, a comment, if I might. I noted that in the 4C2 trial, you reviewed 300 patients to select, what was it, 16? So, you've clearly got some very stringent exclusion, inclusion criteria, and if we've learned anything from movement disorders, we know that it's not just the surgery, it's mostly the patient selection. Can I ask, what were the, why have you whittled out most of those patients? Was there a primary reason, or was it a mixed bag? I tried to reflect a little bit on that, on the 322 million on the planet that suffer from, a lot of them have personality disorders. So, there's a lot of personality disorders in there that leads you to a wrong interpretation of how your life should be, that secondarily leads to depression. But if you have personality disorders, and then we occasionally have, because you cannot detect them, because you haven't seen the patient's pre-morbid personality, if you have a narcissistic character in there, and most of the male patients have a combination with a narcissistic personality disorder, you will later not be successful, and that's why we exclude so many, typically, personality. And the other one is a comment, where, you know, when you see the 4C2 data, clearly what we need to do is leave patients off for longer, to see how long that stun effect is likely to work. And it's really a comment on the whole field, where in the field of Parkinson's disease, we'd operate on thousands of patients before we started doing randomized controlled trials. Yet now, if you apply for funding, it has to be a randomized controlled trial, otherwise you won't get the money. And I think this is misguided, you know, until we have a better feel for our therapies in an open-label fashion. We will continue to construct badly designed studies, not because we want to conduct badly designed studies, but because we don't understand the therapy well enough, and I think that has to be exploratory. So I don't know what the solution is, but I think it's a comment to the field, that I think we need a little bit more leeway to look at therapies in an open-label fashion before we start designing randomized controlled studies that are bound to fail, not necessarily because the therapy is bad, but because we don't understand the therapy properly. It's exactly like that, Ludwig, and really the problem with the publication of 4C-2 at this moment is that the reviewer says, hey, why is that not an RCT? And we are like, because we were not there yet. This is an explorative trial that, you know, helps us to learn about the therapy before we do it, before we go into a multi-centric, you know, with all that single center, you know, bias that you put in there, and we hope that it gets published based on our argumentation now. Thank you for a very nice talk. As you know, I have been in the area of your target lipid, and I have one question to you, and that is, do you think that the middle forebrain bundle is equally well-suited for patients with a more melancholic component as for patients that have a more anxiety-driven depression? Based on the data that I have, I cannot answer that. My feeling is that the melancholic ones will be better off in the long run. It is a little bit like beating the dead horse. If you are down already into anxiety and all that, that depression is very far down and probably previously has been more of the melancholic type, and it probably is a disease process, and you cannot reactivate that. So, but that is only, I do not have the data. I am looking on 27 patients that we have treated with that therapy, so there is no reason to speculate on that, but my feeling says it is like you say. Well, a good guess might be valuable. Thanks. Yeah.

MFB-DBS

depression treatment

treatment-resistant depression

DBS targets in depression

study design challenges

patient selection