I'm Gary Steinberg from Stanford University, and I'll be discussing microsurgical management of brainstem, thalamic, and basal ganglia cavernous malformations. This was originally scheduled as a 2020 AANS annual scientific meeting breakfast seminar, but because of the COVID pandemic, we've rescheduled this as a distance learning e-module. I'll be joined by Ed Smith and Mike Lawton on this seminar to discuss management controversies for cavernous malformations. When cavernous malformations are located in these deep structures, they pose unique, challenging problems because of the high-priced real estate, including the cranial nerve nuclei, thalamic nuclei, cranial nerves, and the long tracts which course through these very small structures. The natural history of these deep cavernous malformations is worse than for supertentorial cavernous malformations. You can see the bleed rate, and particularly the re-bleed rate annually is much higher. In our series, we found a re-bleed rate of 31.5% annually. We tried treating these patients with stereotactic radiosurgery in the 1980s and early 1990s, but with poor results. Here, you can see an example in the upper left of a patient with a thalamic cavernous malformation who developed significant radiation necrosis after treatment, and these patients continue to have multiple bleeds after radiosurgery. I've resected 13 patients, one to nine years following radiosurgery, and pathologically, we found that none of these cavernous malformations were completely thrombosed, and they all had radiation necrosis. This is different from what we find with radiosurgery for high-flow AVMs. Several series, including our series, have shown that the symptomatic re-bleed rates are still quite high after radiosurgery, and the complication rate also is approaching 9 to 26% for symptomatic complications with a significant risk of permanent deficits and mortality. This was the first patient that I operated on in the brainstem back in 1990. She was a 14-year-old girl from Los Angeles who had bled four times, had quadriparousis, multiple cranial neuropathies, and with some trepidation. I explained to the parents the risks involved in the surgery, and I explained to the doctor that I had managed this by separating the cerebellar tonsils. The brainstem was extremely thinned, made a small incision here. You can see the hematoma, worked through the hematoma cavity, and resected the lower portion of the malformation. She had some improvement immediately in her quadriparousis and cranial neuropathies, and then two months later I came back through a transcalosal approach and resected the midbrain component, and although she wasn't normal, she did make a significant recovery and never bled again. This gave me some confidence in operating in these structures. We now have 314 patients where I've resected 323 cavernous malformations in the brainstem thalamic region or basal ganglia. You can see the distribution in terms of gender. Patients range in age from three months to 73 years old, and 65% of the patients suffered multiple hemorrhages before operating. We published our results in the first 176 patients in 2013. The vast majority of patients presented with multiple signs and symptoms, including cranial nerve deficits, motor deficits, sensory, and cerebellar signs. You can see that preoperatively 38% had a MRS of greater than 2. The location was three quarters in the brainstem and half of all the malformations were in the pons, but every area was represented. I prefer to operate between four and eight weeks since the last hemorrhage since that gives some time for the clot to liquefy and I can work through the hematoma cavity, but if one waits more than two months, then the brainstem or thalamus can reconstitute and it's sometimes more difficult to find a safe corridor. We use electrophysiological monitoring, SEPs, motor revoked potentials, brainstem auditory revoked potentials, and cranial nerve mapping for nuclei. Of course, navigation is critical and I still employ mild hypothermia. Various skull base approaches are critical in order to access these difficult lesions in a safe way. In the medulla, the majority were approached through a midline suboccipital approach with 20% through a far lateral approach. Here's an example of a 36-year-old woman from Colorado who suffered four clinical hemorrhages, had severe brainstem dysfunction, multiple cranial neuropathy. She's aspirated several times, was wheelchair bound, and when she came out, we did a prophylactic tracheostomy before resecting the malformation, was able to achieve an excellent resection, did not set her back in terms of motor sensory function, but interestingly, while she was breathing fine during the day when she was awake, at night when she fell asleep, she did not breathe and had to be kept on the bed for seven weeks, and then she finally recovered and we were able to take her off the ventilator and ultimately get her tracheostomy out. The failure to breathe due to injury in the medulla region is known as Andi's Curse. However, patients do ultimately recover. In the pons, the most common approach was the midline suboccipital, but all these other approaches were also quite valuable in certain instances. Here's a six-month-old, one of our younger patients, who presented with a right hemiparesis and left six nerve palsy from this cavernous malformation, a bleed. Approach this from the left side, far lateral suboccipital approach. The feet are down here. The head is up here. Here's the vertebral artery, the brainstem. And you can see, mobilizing the verbal artery and retracting it immediately, stimulating the side of the brain stem medulla to ensure there are no critical pathways, making a small incision. Right here is the clot cavity. Remove the clot cavity and then take out the malformation. And I'm always amazed at the way postoperatively the brain stem can reconstitute itself compared with the preoperative example. He was one of the minority of patients who had immediate improvement in his symptoms, his hemiparesis fosophila, though you can still see the six nerve palsy on the left side. This patient was a learning experience for me. This was a 40-year-old male with migraine headaches. And you can see he bled from a pontine malformation, did not come to a peel or a pendible surface. And so I did not operate. He had only migraines. However, two years later, he suffered two clinical hemorrhages and now had deficits. And this was the first patient that I operated on where the malformation did not present to a surface, either a peel surface or a pendible surface. And there are safe corridors to enter the brain stem if the malformation doesn't come to a surface. I came in from the right side under the temporal lobe. Navigation is, of course, critical here, and this was the approach that I took. You can see we are now under the right temporal lobe, which is retracted here. This is the tentorium. And this is anterior. This is posterior. Here's the fourth cranial nerve over here. This is the fifth cranial nerve. And working through normal brain stem but non-eloquent area, there's the malformation. And now we've taken out the malformation through a small opening, a few millimeters. And during the case, toward the middle of the resection, my colleague monitoring the motor evoke potentials told me that we were losing the somatosensory evoke and the motor evoke potentials. You can see here a decrease in the left leg SEP. The left arm SEP was lost but came back. However, the motor evoke potentials were totally lost from the left side. And how do you think the patient woke up? He woke up completely paralyzed on the left side, which was discouraging. However, within about 10 hours, he was moving extremely well on the left side. And you can see one week later, all he had was a mild left ankle weakness. Within four months, he was back to completely normal and shooting his single digit handicap golf score again. So what we've learned is that you can set patients back. In fact, you usually do set them back immediately, but that they will recover over three to six months and patients and families have to understand. This patient bled from a cavernous malformation in the pons symptomatically that did not come to a peel or a pendamal surface. I waited and three years later, she had a second bleed. Now it comes closer to the surface, but since she had minimal symptoms, I waited until a third bleed four years later. And at this point, I thought it should be resected. This approach was a trans-petrous approach. here drilling with a diamond a bit the pedris bone now we're very anterior in fact there's the fifth nerve and we're working anterior to the fifth nerve and taking out the malformation in the anterior pons. She did quite well went back to teaching. I wanted to make the point that cavernous malformations can often be associated with DVAs or venous angiomas and how important it is to preserve the DVA which don't bleed when resecting the cavernous malformation and that's what we did in this case you can see in the upper right. This particular patient had multiple bleeds including facial numbness and a six nerve parasis. He did not have a seventh nerve palsy. You can see the malformation it looks like it comes right to the appendable surface of the fourth ventricle here and in fact it did you can see here's the malformation however every time I stimulated over the malformation were presented to the appendable surface the seventh nerve fibers were very active and I decided not to reset this for fear of giving him a permanent seventh nerve palsy and we're following him he has not re-bled so far. Here's a 30 year old young patient asymptomatic this would be straightforward to reset however she's asymptomatic and I don't recommend operating on asymptomatic patients. In the midbrain there are a number of surgical approaches which could be used one of the most common that I prefer is a supracerabellar infratentorial. I use a far lateral supracerabellar infratentorial approach now and then there are other approaches anterior or posterior transcolosal as well as occasionally using occipital transtentorial. This patient bled from a midbrain cavernous malformation and this was another learning experience for me you can see on the lower right image the t2 image the malformation appears to present to the cerebral aqueduct surface however when I operated on the patient you can see in the upper right no evidence of malformation on the floor of the fourth ventricle or cerebral aqueduct and so I did not reset the malformation what I learned is that you have to look at a t1 image and you can see on the left that in retrospect there is brain tissue between the cerebral aqueduct and malformation you cannot utilize the t2 or GRE images as a true representation of where the malformation is because the hemosiderin blooms malformation so on her we waited and she had a fourth hemorrhage 19 months later and now you can see on the t1 it does present to the surface and this was a straightforward approach super cerebellar infantorial here's a nice example of a dorsal lateral midbrain cavernous malformation that bled symptomatically on the right side here you can see the fourth nerve just below the inferior colliculus here's the malformation these are branches of the superior cerebellar artery and you can see the malformation resected using a far lateral super cerebellar approach for these anterior midbrain lesions there are several approaches these are challenging and for this woman who had bled twice had a left hemiparesis and numbness I decided to come in ipsilesionally on the right side around the peduncle here and here's the exposure from the right side the structures from left to right are the right optic nerve internal carotid artery basilar artery apex and then the right third nerve here I'm dissecting between the basilar artery apex and the third nerve dissecting the third nerve from the uncus very important to avoid stretching the perforating arteries coming off the basilar tip here's the malformation resecting it gently and there's the final result and you can see that immediately postoperatively I did give her a new right third nerve palsy but this resolved after four months and her other preoperative symptoms we've resolved quickly as well this is a similar anterior located midbrain cavernous malformation but deeper than the prior one and so I felt a contralateral approach would be better for the angle of exposure I was concerned because she had a left third nerve palsy and I was coming in from the right side I didn't want to give her bilateral third nerve palsies but the exposure worked well and here you can see the exposure of the left-sided midbrain lesion from the contralateral right side using an orbital zygomatic approach again working around the basilar artery apex and resecting the lesion fortunately she did not suffer a new right-sided third nerve palsy here's the postoperative imaging with a complete resection the patient did quite well this 31 year old software engineer presented a challenging problem as well he was cognitively intact but had bilateral third nerve palsies and gait ataxia and how do you think this should be approached very difficult question of how to access it but it comes closest to the surface of the third ventricle and so here you can see some of the fiber tracks more anteriorly motor sensory and so for this one I decided to use a transcalosal approach through the foramen of Monroe into the third ventricle and that nicely it came right to the surface was able to resect that and five months later he already had improvement is bilateral third nerve Paris's still had a left ptosis was walking normally though working as a programmer and 14 months later his left eye began to open for approaching cavernous malformations in the thalamus the anterior and posterior transcalosal approaches are very useful however some of these other approaches transylvian trans cortical either frontal or temporal or super cerebellar infertentorial can also be used here's an example of a man who bled into his cavernous malformation in the left thalamus pulmonary region and extending to the midbrain I decided to approach this supertentorial para falcine adjacent to the parietal occipital lobe and it's important here to work around the deep venous system often sectioning of the inferior surface of the Falks and the edge of the tentorium can help and work under the splenium of the corpus callosum and here's the intraoperative view this is on the left side with the feet towards the top of the photograph and the vertex of the head at the bottom this is left parietal occipital parasagittal approach Here, we're working around the deep veins. There's the surface of the pulvonar making a small incision and then malformation is exposed and working in the plane between the malformation and the hemocedrin stained thalamus resecting it and I've never seen a deficit related to operating in the pulvonar. This young medical student represented another difficult problem. She had several clinical hemorrhages, gait ataxia, memory loss and intermittent drowsiness as a result. She recovered from her last bleed except for mild memory loss. She sought consultation throughout the country, decided to come to Stanford and I approached this interhemispheric transcolostal through the foramen of Monroe where it presented to the third ventricular surface and here you can see the surgery resected it and it went quite well actually. Here's the postoperative result and we didn't set her back but she still had memory problems to the extent that she was never able to finish medical school. For the basal ganglia, several approaches can be used including transsylvian and transcortical approaches or interhemispheric. This patient bled from a left-sided basal ganglia cavernous malformation and you can see it's important to map out language function. We did this preoperatively first with functional MR scan and you can see the language function on the operculum surrounding the malformation and the insula. I like to operate on these patients awake so we can stimulate at surgery the insula to decide how we're going to approach the basal ganglia vascular malformation and avoid critical areas. Approach in this patient, this is Transylvian working around the middle cerebral artery vessels stimulating the insula non-eloquent area making a small incision through the insula Here's the clot which was drained then working through the clot and resecting the malformation and There's the final result Through a small opening How about this one, what does it say? How about this one, what does it say? It's a It's a fungus So we obviously would not have operated through that eloquent area where we were stimulating The Omniguide CO2 laser is something that I've been using over the last decade I find this extremely useful in terms of resecting these malformations, and I'll show you how we use this it has a tip diameter of 0.55 millimeters And I've used this CO2 laser in 227 cases mostly cavernous malformations in deep eloquent areas Here's a 42-year-old female with familial cavernous malformations and you can see the malformation that bled in the left lateral pons anteriorly Using this laser, we can see the malformation that bled temporal approach. So we're retracting the temporal lobe now. There's the tentorium. Navigation is critical. Here's the fourth nerve. Cutting the tentorium. Here's the fourth nerve. Here's actually the fifth nerve. So we're working anterior to the fifth nerve. And using this laser as a probe, it seals the vessels as well and cuts the smaller vessels. And you can get this into very small, tiny areas. Another example, this patient who had radiosurgery for an acoustic and then bled from a new cavernous malformation likely related to the radiosurgery. You can see it in the dorsolateral midbrain. So came through a far lateral infratentorial supracerrabellar approach and make a small incision with the laser. Again, the laser is nice because it doesn't interfere with the electrophysiological monitoring parameters like the bipolar. It's a much smaller profile than the bipolar forceps. And you can resect these difficult malformation areas. This example shows you how I was able to approach his pontal medullary cavernous malformation. It comes closest to the surface on the right side anteriorly and laterally. And here we are working a right suboccipital just far lateral on the right side here's the brainstem mapping out the brainstem for critical pathways and you can see in this tiny space how this laser fiber-optic cable is used and now I'm removing the final portion of malformation through a very tiny opening. We've been able to use the laser through extremely small peel surface exposures. Here's a patient who was neurologically normal and had bled twice from this malformation which should not present to a peel or appendable surface entered through the side corridor mapping out any critical structures and you can see was able with the laser to resect this through a one millimeter opening. This patient taught me a new approach this was someone who bled from a midbrain hypothalamic cavernous malformation you can see it's located very anteriorly and so I decided to approach this contralaterally here you can see the mammillary bodies this through a pterional translaminar terminalis approach was the first time I operated through the laminar terminalis to get to the midbrain and hypothalamus here you can see the left optic nerve here's the carotid artery third nerve And here's the A1 anterior cerebral artery. So now opening the lamina terminalis. There's the third ventricle. Navigation is critical to avoid injuring important structures. And now the laser is used to make a small incision and working between the malformation and the hemosiderin-stained hypothalamus and midbrain. Take it out. Don't think that this could have been resected using the bipolar because the space was so confined and that worked nicely. We published a study comparing the laser resection to bipolar resection and you can see with the bipolar there's significant thermal injury, disruption of the elastic lamina and soft tissues of the malformation. And you can see here minimal thermal injury with the laser used at low wattage. The advantages, as I mentioned, are that it's atraumatic, much better visualization. Use it as a dissecting tool as well as a cutting tool. Minimal interference with neuromonitoring. The disadvantages, it does not seal the larger vessels and you need a bipolar for that and you can't operate. So if you look at the results, you can see that immediate postoperative results in these patients do not look that great. There's a 30% chance of worsening. However, if you look at the long-term results at six months, almost 90% of the patients are either improved or unchanged and only 11% of the patients are worsened and some of these are worsened due to problems not related to the cavitation. So the new deficits, as I mentioned, are common in about a third of the patients. And here you can see the kind of deficits we see. I do not like to operate in the midline as opposed to the spinal cord where we make a incision right in the midline dorsally. In the brainstem, that can cause a bilateral intranuclear ophthalmoplegia. So always try to lateralize your incisions, openings through the floor of the fourth ventricle. We had four immediate deaths. Some of these were from patients who were in poor shape and one was a unrecognized heart disease in patients. I wanted to mention this phenomenon, hypertrophic olivary degeneration. And here's an example of that. You can see preoperatively a patient who bled into the pons. It usually occurs in the pons. Postoperatively resected the lesion. Here's the medulla. Before surgery, immediately postoperatively looks fine. And then in a delayed fashion, you can develop this degeneration has to do with this triangle of guillain and molarré. It can occur even without surgery just from a bleed. We had it occur in nine patients, 4% at six months postoperative, usually in the pons, once in the midbrain. And it can cause palatal myoclonus nystagmus or osillopsia. Two patients that suffered tremor, hemibolismus, and four of the nine were worse neurologically. Long-term outcomes, as I mentioned, at six months or more are encouraging. 82% of the patients had good outcomes by MRS score. Only 18% had poor outcomes by this criteria. And if patients come in in good shape prior to surgery, only 2% had poor outcomes, whereas if they are in poor shape, initially 42% had poor long-term outcomes. On multivariate analysis of predictors of good outcome, we found, not surprisingly, age less than or equal to 40 years predicted a better outcome for the brainstem, cavernous malformations. Preoperative MRS also was a high predictor of good outcome. Patients in better shape did better in long-term outcome for all cavernous malformations. And as I mentioned before, surgery within eight weeks is a predictor of better outcome. This girl, who was four years old and bled twice, came out from Texas. She has a large thalamic vascular malformation. I resected this completely and she was a good outcome showing immediate improvement in her hemiparesis. The post-operative MR scan show complete resection of the malformation. However, she came back two years later with a new bleed at the same site and analysis of our results shows that 2% of patients had a re-bleed after complete resection was confirmed both visually at surgery and on post-operative MR scans. So I now follow all my patients with annual MR scans. In conclusion, these deep calve mouths have a high bleed and re-bleed rate. A third are in poor shape neurologically. The surgical approach uses many skull base approaches according to where the malformation comes closest to the peel or appendable surface. A third of the patients were worse off immediately after surgery, but long-term morbidity mortality is quite acceptable at about 11%. 98% of patients who were in good neurological status before surgery had good post-operative outcomes. However, interestingly, many patients in poor condition preoperatively also had good outcomes. Good preoperative status, early surgery, and young age were predictions of good outcome. And I believe that in experienced hands, deep cavernous malformations can be resected with acceptable morbidity and mortality. I wanted to thank our entire cerebrovascular team and Stroke Center for the success of our program over the last 30 years. I hope all of you will stay healthy and safe during this unprecedented COVID crisis. Thanks very much for your attention.

microsurgical management

brainstem

thalamic

basal ganglia

cavernous malformations

surgical resection

electrophysiological monitoring