Issam A. Awad , MD, presents clinical and research perspectives on cerebral cavernous malformations.
[MUSIC PLAYING] ISSAM A. AWAD: So, what I thought I would do here is try to take what we have learned about the science of cavernous malformation and try to project how we use it in clinical practice every day. For the residents who are with me in clinic, they know exactly how we do that, and some of this will be familiar to you. But the idea here is, how do we take the science of formation, apply it as best we to all our patients' care, and the other side of the presentation is, how do we take the clinical information back to the science and try to develop the translational strategy through Novel imaging, biomarkers of disease as potentially treatment? So, it is a spectrum and a continuum. So the first concept-- and I think what was addressed already by Greg-- is the question of differential diagnosis and, although the MRI is very specific, sometimes the appearance of the CCM can be hidden by blood. When it's not obvious what you have, this turned out to be a CCM despite this massive hemorrhage. Of course, when you have multifocal pathology, in general, you can understand this as being a CCM, although we have to also realize that a hemorrhagic microangiopathies, multifocal melanoma metastases, et cetera, that can mimic that appearance. But with very careful imaging, combining contrast enhanced imaging and also sometimes just letting the blood clear and repeating the image, is really the way to go. The high field and susceptibility imaging have special roles, which we will discuss later and, for this entity, angiogram is rarely needed, even with this appearance, unless we are questioning an atypical appearance of an AVM with a large draining vein. So, the natural history of cavernous malformation has been a very puzzling story, and almost all the studies-- I will show you on the next slide, a meta analysis of all the studies-- have actually generally agreed with the population-based report from Scotland that was published in Lancet Neurology in 2012, that lesions that have not bled before, have a very small risk, under 1% per year, of bleeding the first time. Now, that can add up in a lifetime to something that's not insignificant. Could be 10%, 15%, et cetera. However, a lesion that has bled has a much more likely risk of re-bleeding, and this differential risk stays up to five years from the time of hemorrhage. So, it is very rational for us to tell the patient, if you had had the clinically relevant hemorrhage, you have a different type of disease that if you have a silent lesion. And if you have to live with a lesion that has not bled, we don't want this natural history to disable an individual and make them not want to do athletics, not be healthy, not take aspirin when they have heart disease, et cetera. So, basically there are very few things we restrict them from doing. I typically in my practice tell them to avoid extremes, like don't go bungee jumping the first time now that you know you have a cerebellar cavernoma. You know, just leave that. But otherwise live a healthy life. Now, the natural history is not always so benign, and the groups that tend to have shown higher rates of hemorrhages, as much as 25% to 30%, have mostly been lesions that had bled previously and, particularly if they are in deep brain locations. So, deep brain location and recurrent hemorrhage is bad for the natural history. There is a theme here, because you're going to hear that also in the natural history of AVM. When we talk about management options, we're talking about quality of life, and 0 is horrible, 1 is great, I'm living my life normal, and you have to try to assess, is the treatment going to give you double vision and make you unable to do the things you want to do, or is the first bleed going to give you double vision, for a pontine lesion. And you have to decide if the bleed is not going to happen for some time, then live your life with binocular vision for as long as you can, rather than having the operation early up front and cause a deficit. In other situations, the price of the treatment is very low. It's a low-hanging fruit, and therefore the decision is made to cure the patient, particularly if they have a solitary cavernoma, and then they don't have to live with the threat, no matter how small, of altering their life the later on. Now, what we do well, in general, and because this is now the clinic part, I will talk more about this. It's basically to try to remove the lesion while sparing brain function, and brain function includes cortical structures, subcortical pathways, arteries and veins. And in the veins category, we have the venous anomaly this is associated with the cavernoma, which you must preserve unless it is tiny, so if it's a very tiny one, effectively you're resecting the lesion and its draining vein, that's OK, but if the draining vein is draining a significant component of brain, then it should be left alone, or you will have that brain at risk. So, this is in general what it looks like to be removed. We can do it very well with intraoperative mapping of motor function. Sensory motor sulcus is very nicely mappable with evoked response at surgery. This patient could not have functional MRI, because her hand was paretic already, so you cannot have functional MRI if you hand doesn't move. You can do brushing of the hand and see where the sensory function is, but the motor will be lost but this is her. [VIDEO PLAYBACK] -Hi everyone. I'm Sarah P. I'm doing OT. We're four and half weeks out from surgery, and I wanted to send you a little video of me moving my hand. This is a scanned resection. So, I have full movement, just some tingling in my pinky and ring finger. So, here I am picking up buttons. I can touch each fingertip with my thumb. Full movement. So, I'm getting stronger and stronger, and it's because of you. Thank you. Have a great day. [END PLAYBACK] ISSAM A. AWAD: So, this was her plegic hand and so be nihilistic if you're operating in a hand motor area and you have a plegic hand. It will recover. This is another case that was a bit more challenging in mapping, because the lesion was subcortical, it had already grown, had posed significant growth, so we determined that for this young woman we would resect it, because of the demonstrated growth and it had the venous anomaly with it, and it was nestled behind the sensory fibers, but we did the mapping of the functional MRI, and we projected the fibers that are coming from the motor cortex and from the primary sensory cortex, and these are the associated visual pathways. So, we determined, based on that, a corridor for approach that would actually allow us to stay away from these important structures, and the famous anomaly was right here in front of the lesion, so we obviously left it alone as we teased the lesion away from it. Epilepsy, when we are dealing with epilepsy and cavernoma, the first question we ask, is it a solitary lesion or multiple lesions. Solitary lesions, the answer is usually very simple. It's the lesion causing the seizures until proven otherwise. m we'll ask for proof that it's not the lesion. Otherwise, it's usually the lesion, resecting the lesion can be very helpful in helping control the seizures. About 30% of cavernoma epilepsy becomes intractable, so not only will the patient now be having occasional seizures-- cannot drive, cannot fly a plane-- many professions have been already stopped because they've had a seizure, so the idea of removing the lesion and curing them, allowing them to get off medications, the better results are with shorter duration of seizure. So, don't let the patient have seizures for 10 years and then expect them to get off Keppra and dilantin. If you want them off medication, you have to do the operation earlier. If you want control of recalcitrant epilepsy, that's another level of discussion. Multiple lesions are very challenging, and I was cautious about doing epilepsy surgery with multiple cavernomas, because you can have situations where you resect the part of the brain, only to have a cavernoma bleed or develop in the opposite part of the brain, and you have Kluver-Bucy all over again. Lesionectomy. This is in the primary speech area. This is an engineer who runs, actually, a big company and had cholesterol seizures, and this in a primary speech area, and he very much wanted to try to get seizures stopped and to be off medications. We mapped speech and speech was removed from this area by at least a gyrus. The connective fibers between the broca and Wernicke were also deep to the lesion, so we felt we could resect this lesion safely. And the challenge to this lesion was actually the vein of labbe and a big angular artery that was running right in it. But with microsurgery techniques, this can be done. And this case illustrates all the different structures that we have to to preserve. And he has been seizure free and, actually, one year later we took him off anticonvulsants. Cases that are intractable for many years warrant a more aggressive resection. This is a patient who had to to three seizures a week for the last 10 years. We mapped that he was having frontal lobe epilepsy and very dysmorphic gyrai, so in this case we resected all the spiking tissue in this relatively less eloquent part of the brain in an effort to really-- and we map to be sure there are no residual spikes, in an effort to give him maximal epilepsy surgery that will benefit. He has been seizure free, except when he tried to get off medication, so this guy we were able to convert into a control, but we will never be able to get him off medications, and remember how long he was with the seizure. The brain stem cavernoma is very expensive real estate. Everything runs there, so the natural risk is worse, and the treatment risk is worse. Why is the natural that this worse? Because even a tiny bleed in the brain stem will be symptomatic bleed. But a tiny bleed in the frontal pole may be a migraine headache, and the patient would not symptomatic. Therefore, we have a lot higher rate of bleeding cavernomas in the brain stem. It may not be through bleeding, it may be symptomatic bleeding, because of the eloquent location. But, here again, we have to think of the function, much more crowded structures, the veins-- if the venous anomaly looks like this monster, you're not going to do a supracerebellar approach to this region, you're going to try to go from the side in a subtemporal approach, like we ended up doing in this case. Obviously, you don't want to wait. This patient was flown to us from New Orleans after he became respirator dependent and quadriparetic. This is too late. He had had symptoms for the prior year and a half. So, this is, I think, the right timing for this lesion. This gentleman had diplopia on two occasions from which he recovered, then he had diplopia that was permanent. With a third bleed, we decided to go in because he was confident now he will not recover again. The key to this approach is to go where the lesion presents to the surface, which is right here, which means you have to split the cerebellum across the vermis, and the lesion is now at the surface of the brain stem and eminently resectable. When the lesion comes to the surface on the side of the brain stem, we like this approach a lot more, because the back of the brain stem is more eloquent. The front of the brain stem is more eloquent. The best corridor to approach the brain stem is from the side. So, you have here the fasciculus gracilis and cuneatus in the back of the brain stem. It would be horrible if we opened the medulla right there, because this patient would lose proprioceptive sense. But, going from the side at the level of the olive, we were able to reach in and resect the lesion, and this young woman has done phenomenally well. CCMs associated with VMs. So, we leave alone if the predominant lesion is a horrible venous anomaly, like this. But, if you have a dominant lesion that is symptomatic that is a cavernoma, we would make an approach, either from the side or medial to this vein, and get this cavernoma out while preserving the venous anomaly, because if you take the venous anomaly with this, you have to be prepared for a cerebellar infarction. So, it is very important to strategize where the vein is when treating these lesions. So, this is one where the patient presented with dystonia and hand paresis because of this growing lesion. She has a big -- you see the vein right here. This is a very eloquent location with the lenticulostriate arteries, the vein, and obviously going through the Sylvian fissure to get this out. We did not do this lightly. We did this because the woman was literally losing hand function and dystonic postures that were making her useless in social situations and as a worker. So, eventually, we did this fairly serious operation that involves a trans-Sylvian resection. This is the intraoperative fluoroscopic angiogram showing the preservation of the vein and, of course, the middle cerebellar arteries. And this lady has done incredibly well. This is at her two-year follow up. [VIDEO PLAYBACK] -Set your arms in front of you, both of them. Close your eyes. Touch your nose with your left index finger. [END PLAYBACK] ISSAM A. AWAD: It is perfectly normal function, not even a hint of a tremor, and we were really very, very pleased with this outcome. Spinal cavernomas are not different from the brain stem, except it's even more crowded and more dangerous, and I think what we say here, is that just because it looks ugly on imaging, if the patient is functional, you want to be very careful with these, not to operate on highly functional patient with very, very serious lesion, because some patients will be made worse that otherwise may have a number of years good function and may never have a bad event. This particular lesion here, this patient had had two prior small bleeds and now presented with a that he was starting to use a walker. When we resected this lesion, unfortunately he woke up paraparetic, completely paraplegic, unable to move. The only thing he had was toe sensation. And I said, oh my god, what did we do, you know. At least he was using the walker before. But a little sensation, you can build on. This is at six weeks in rehab, so it's an incomplete injury to the cord, which is really what's great. So, you know, even if the cord was stunned, this was an incomplete injury, and you could see here the dura in front of the cord. I thought he would never come back from this, but he did very well. Pregnancy is a controversial topic. Epidemiologically, there is no detectable increased risk during pregnancy, but every specialist in this disease sees patients who bleed during pregnancy, and it's a complicated thing. And if you look only at CCM bleeds, many of them are in pregnant women, more than you would expect in the general population. We think that there is a very small added risk of misbehavior of a lesion during pregnancy than in any other year of the woman's life, but that in the big epidemiologic picture of pregnancies, it's lost-- that little increase is lost. Nevertheless, we let these patients carry their pregnancies to term. My practice is to image them without gadolinium, with MRI, in the second trimester, and if the lesion has not changed allow them vaginal delivery. If the lesion has changed or become symptomatic, then obviously go to cesarean delivery at term and then treating the symptomatic lesion. Radiosurgery is controversial but occasionally inevitable in these lesions. This particular woman with a hamartoma-- not hamartoma-- with a cavernoma of the hypothalamus, in the same place where you have hamartomas, kept bleeding every six months to eight months. She would have multiple hemorrhages and finally I could not make her, in any way, nondisabled if I resected this. At the very least, she would become obese and a hypothalamic cripple if I removed this tumor. This beautiful young woman was just losing function gradually but I didn't want to harm her more, so we ended up doing radiation at very low-dose emitry, and she stopped bleeding. She had one bleed at six months after and then stopped bleeding. It's been now about six years. Very happy, living in Montana. The natural history shows that if you select patients that were bleeding a lot before, you do radiosurgery, in the followup period, eventually these lesions bleed less. This is another one where recurrent symptoms-- this patient had the carpal tunnel and tarsal tunnel because of recurrent numbness of her hand and leg. Well, she didn't have no tunnel. She had this lesion in her internal capsule. But I knew I would make her worse if I resected it. If every bleed was making her numb for two months, what is it going to be if I went in there? So we ended up radiating it, and she actually has done also very well. So very selectively, in a very busy cavernous malformation practice, I may radiate one every one or two years, so it's a very rare choice, but it's one that we keep in our armamentarium. So the balance of clinical decisions, factors favoring surgery are obvious. Solitary lesions, symptomatic, and factors against surgery, higher threshold. Obviously the role of surgery for any lesion in this particular patient is almost futile or here. Genetic substrate has been reviewed, so I will not go over that again, but I want to also confirm that our own data and this was reviewed very thoroughly by our group here, including Dr. Rubeiz from neurology, who helped us with this. In fact, we confirmed that the CCM3s like Dr. [INAUDIBLE] had reported a bleed at a much younger age than the other familiar genotypes. We have detected a hint of this as far back as 2005, but now we have very good data that show the bleed risk, the recurrent bleed risk, the number of new lesions per year in this genotype, which are much higher than they are in the CCM1 and CCM2 genotypes. We still don't know what makes a particular lesion proliferate or bleed. Even the worst CCM genotype that are families, or patients within families, that a very benign. So there are factors that protect you against this behavior, and there are factors that causes misbehavior, whether they are genetic or epigenetic, we do not know, and after the first phase, first 15 years, of looking at the root cause of CCM, the next 15 years will be looking at modifiers, I think, because we really need to understand modifiers. Patients want to know the modifiers. They want personalized medicine. Of course, hemorrhage is present in every lesion. So we talk about symptomatic hemorrhage, one that causes new neurologic symptoms and imaging of acute blood, and that's what we count. The lesion itself is a permeability disease. It's not just the permeability disease because the tight junctions are weak, but also in the mouse the tight junctions are defective, at the molecular level they are defective, and you can make them defective by putting siRNA on endothelial cells in the DISH. So this is a permeability disease par excellence. It is also inflammation disease, and whether the inflammation is related to permeability or not, is a big debate and whether inflammation is one of those factors that modulate disease severity is another factor. Our group [INAUDIBLE] in our lab has pursued this story for the last decade. This is his most recent publication showing elegantly the pathogenetic complex in the CCM lesion with complement activation and oligochronality of heavy chain and light chain that allow them to synthesize an artificial antibody that now will allow us to explore things like vaccine or what could brought the inflammation in this lesion and whether it will modify the lesion. You already that the loss of CCM gene, homozygous loss and embryonic lethal. That means if you have heterozygous disease I always joke about this-- do not marry your cousin. It's not good. But the heterozygous disease really depends on the [INAUDIBLE] hypothesis, where the second copy is likely lost. I say likely, and we are the group who described this hypothesis, because I'm not sure that every CCM is due to that, but in every type of CCM, we have been able to show biallelic, including most recently in sporadic lesions, which had never been described before. So we know it happens, if you have one copy in the germ line, the second identical copy is lost in the lesion. We also know it is only in the endothelial cells of the lesion when you microdissect the different cells, and we also know that it can be biallelic in the somatic lesions. So if we construct a different model that is, in our opinion, more like the human disease, when you do a heterozygous loss, the mouse brain is too small and the life of the mouse is too short for them to have stochastic events like the human and form CCMs. Except in CCM3 model, these heterozygous CCM1 and CCM2 do not develops lesions. CCM3 is aggressive enough, they don't need any help. They get the lesions even in the heterozygous state. But CCM1 or CCM2, we have to encourage somatic mutations by one way or another. We can radiate the animal, we can give knock p53 or we can knock another very cool gene that is called Msh, which is a point mutation repair gene. And if you knock that gene, you're basically increasing somatic mutations, and you can now create lesions is the heterozygous normal-looking animal and it's a very complimentary model to the one that Elizabeth shared with you, because one is a hyperaggressive phenotype, the other one is almost too slow of a phenotype, and some therapies will need to be tested in one or the other. This is a mouse in a coil for MRI, and we can also detect the lesion and then treat it and see what happens to it later. The final thing, you heard about the story of ROCK which, because of work of several groups, knocking down any one of the three CCM genes, results in hyperpermeability of the vascular bed, and this is accompanied by ROCK activation, rho kinase activation But what is important, is that if you inhibit rho kinase activation, you restore the normal phenotype of the endothelial cell, even in the loss of CCM protein. So we call that phenotype rescue. This is a cool therapeutic approach, because basically you can't change your genes but you can change what they do. If that approach works, it will not only be applicable to ROCK, but it may be applicable to many of the normal signals that are being detected. A combination strategy might be very useful But here the heterozygous capillaries have ROCK activity and the lesion itself has very high ROCK activity. And when we give ROCK inhibitor for the mouse, the equivalent of the first 15 years of life, which is I guess four months in the mousese' life, we in fact form-- the mice form less lesions and the lesions do not have inflammation and do not have ROCK activity. Iron leak in the lesions also is shut down, which is a very interesting thing, because this iron leak is the cause of many neurologic manifestations, like seizures, irritative phenomena, and there's also a marker of repetitive bleeds. So what this leads me to the idea of the last two, three slides, is about how do we take this information back to the humans? So counting lesions, we cannot slice the human brain and count it the way Cecilia does in the lab, like we do with the mice, but with susceptibility weighted MRI, we have a very accurate lesion count, and if you do the same type of MRI in the patient at multiple points in time, you can assess the number of new lesions per year, but the cannot have the MRI at Jack in the Box down the street. It will not be comparable. So you have to really have very strict protocol of how you do the MRI in order to take it to this level. On the other hand, counting T2* and T2 and T1 lesions can be done on any MRI, so eventually that can be generalized in the community. The lesions increase with age. We demonstrated this on retrospective analysis with Pierre [INAUDIBLE] and now we're looking at it prospectively, both in Montpelier and here in Chicago. So permeability is interesting, because you heard that it restores the phenotype in many ways, even the endothelial mesenchymal treatment restores permeability, ROCK inhibition restores permeability. So can we really measure permeability in man, and the answer is yes, and we are doing it. We have done it on nearly almost a hundred patients now, and what we see is that the familial cases, exactly as predicted in the mice, have higher permeability than the sporadic cases in normal brain. Nothing that lesions. A normal brain. The familial cases of higher permeability. Some of them do not, and some of them do. And, in fact, the ones with higher permeability have more lesions and the CCM3 has higher permeability, and patients who happen to be taking statin have lower permeability, which is a ROCK inhibitor, in a way. So the idea here that permeability might be a measure of disease background and disease activity, and the guide for permeability therapy is back to the human cases that we use. The other very exciting thing, and the last thing I want to tell you about is what Dr. Christoforidis told you is the ability to take this blood blotch that looks like black ink and tell you how much ink there is in it. So it's a quantitation of the amount of iron, and how good it is that? It is pretty darn good. If you did it in a test tube with any form of iron, you can do very well. And when I resect the lesion in the OR and we take it mass spectroscopy, the QSM predicts exactly the amount of iron that's in the lesion, so this is a very valid measure. You can see it in the animals. These are the mice. The iron is right here, the QSM is right here. So when you see changes in QSM, you have changed the iron. So if we're going to use a therapy like Fasudil or statin or the EGF modifiers that alter permeability, iron may be what we use as a biomarker of treatment and telling the patient your lesion has been shut down, has less iron, The more leaky lesions by permeability have more iron in them, which is a proof of concept that both techniques, looking independently at the problem, are measuring the real biology. So why do we spend our lives studying CCM? So for many of us, this will be the end of our mission. If we can solve CCM, we will be happy. But for medicine at large, I think it's a bigger issue. Many of the tools that we are applying to study CCM started in cancer and in vascular biology and we adapted them to stroke and CCM. Conversely, many of the discoveries from Dr. [INAUDIBLE] labs and others will now find applications in tumor genesis, in other areas. And there is a class of neurovascular diseases when there is iron in the brain, and those have remained dark boxes. And if we know how to measure iron, how to study it and modify permeability and brain iron, we can impact Parkinson's, we can impact the aging, and we can impact hemorrhagic strokes. So this is kind of the story of CCM from a clinician's perspective, and the next thing really, I cannot say that this can be done by any one individual alone. This is really a team sport. We are very lucky here in our group to have such a phenomenal group of both principal supporters and just people who make it easier for us to do the work. My colleague, what who me to spend the day in the lab is helping this research in a way as well. So thank you everybody for doing that.