There you go. So as you can see, this is a very broad topic, right? While people have specific assignments, mine is, oh, just talk about the newborn and pediatric brain. So I decided to choose one topic, which is the most anxiety-inducing topic in pediatric neuroradiology, which is the MRI in neonates. And to just start and get situated, it's important to know that the patterns that we're going to see on the MRI depend on a bunch of factors, such as what is the state of brain development of the time of injury? Are you looking at a premature baby or a full-term newborn? The severity of hypotension and hypoxia. And usually, hypotension is the one who will define the dominant pattern of injury. Is this moderate or profound hypotension? Usually, brief periods of hypotension will not cause a brain injury detectable on MRI. What was the duration of the event if recorded? Was it a brief event or a prolonged event? And then, of course, the timing of MRI imaging acquisition. Since these patterns are dynamic, as I will show here, the same patient image at different days, the pattern will evolve over time. And that has to be understood so it's not confused with new episodes of injury. So what do I have to say about premature babies? I have to say that the incidence of difficult birth, just difficult delivery, is smaller, is lower since these babies are smaller. But also because of the immature or the pre-oligodendrocytes that are vulnerable to injury, they have extremely vulnerable white matter. So without clear, well-defined events, you can have a dominant pattern of white matter injury in premature babies. So for example, here, a classic picture on MRI and auto-ultrasound of a two-month-old born at 29 weeks, so very premature, and below 28 weeks, we'd call extreme premature. And you can see this pattern of cystic white matter injury. And also on the flare here, how the cysts in the corona radiata and central semiovalley, they all lose signal. This pattern, this honeycomb appearance or grade four white matter injury is now much less common since the intensive monitoring at neonatal intensive care units. Just keeping the babies under homeostasis, blood pressure, blood sugar, treating infection, and preventing inflammation seems to decrease the prevalence of severe neonatal white matter injury. This is what we see later in children that we call children with cerebral palsy, meaning a neurological deficit that is predominantly motor, that is non-progressive, and it is a result of injury during the prenatal or perinatal period. That's the definition of cerebral palsy. And notice that although the cortex is normal, I'm sorry, you can't see what I'm pointing at. Let's see. How about now? Yeah. Although the cortex is normal, there is decreased volume of the white matter, which causes undulated appearance of the ventricular walls. And a very important sign, notice how the posterior sylvin fissure in the cortex, it almost contacts the ependyma. This is an extreme case or a severe case. These children are usually quadriplegic. But in spastic diplegia, very mild cases, you're going to find the same findings. So always look at this region between the cortex and the atrium of the ventricle for loss of that white matter. OK. Now, the first thing to identify or to try to identify hypoxic systemic brain injury is to understand the normal neonatal pattern on imaging. And it's essentially the reversal of the adult pattern, which is essentially at one year of age on T1 and two years of age on T2-weighted imaging. So these two-year-olds looks like a normal adult brain in terms of the progression of myelination with bright white matter and dark white matter on T2. First, you have the T1 signal increase from myelin formation. And the T2 signal decrease is because as the myelin sheath tightens, you decrease the extracellular water. So the brain starts very watery and ends up myelinated. Notice the newborn has a reversal pattern where a term newborn we basically only see here at the level of the basal ganglia, the posterior limb of the internal capsule, as bright signal. That's important because when we lose that bright signal, it is a sign of poor prognosis. And you can see that too on the T2-weighted image, just a dash of myelination in the posterior limb, some hypo-intense signal in the ventral lateral thalamus, which is normal, and the homogeneous signal intensity of the putamen similar to the cortex. These are all the normal findings that we see. Now, if we think about the severity of hypotension in moderate or profound and the duration of the event, so severity and duration. Let's look at the patterns that we get. First, just a quiz question. This is not in your audience response system just to save time. But what is the most common location of brain involvement after a profound and brief hypoxic ischemic episode or something we call profound anoxia? These are placental abruption and kids that need to be resuscitated immediately, but are reanimated and resuscitated very quickly. They will have, usually, injury to these structures, the dorsal putamen, ventral lateral thalamus, and periurelandic cortex, which are the areas of higher energy consumption in the brain. So there is no chance for them to compensate their energy consumption. They are the first ones to suffer the hypoxic ischemic injury. If you have moderate and brief hypotension and hypoxia, you get watershed or border zone infarctions. So it's predominantly the hypotension and low blood flow that is causing that. I just asked you about the profound but brief. Notice that if it's profound and prolonged here, this is total brain injury or cerebral devastation. Then it's too profound, and it's also too prolonged, and the brain will not have any selective involvement. It will have global involvement. Now, what we call partial asphyxia, so not in the umbilical cord that is not too tight, or the umbilical cord around the neck, or shoulder dystocia, and there is moderate hypotension that is prolonged, then you can have diffused cortical injury. But in this situation, there is time to shift the energy consumption to the noble structures of the brain, the brachial ganglia and the periurelandic, meaning the motor and sensory cortex. The areas of high energy consumption are then spared, while the rest of the cortex is involved. So it's the reversal pattern from profound and brief hypoxia. So here's an example of watershed infarction, similar to what we see in patients with carotid stenosis, but in this case, bilateral. Notice on the diffusion-weighted images in ADC, ACA, MCA territories, there is some MCA, PCA, but there's also involvement of the optic radiations, splenium, and dorsal thalamus. This occipital lobe involvement here in occipital parietal is also from superimposed neonatal hypoglycemia. Another teaching point is that hypoglycemia involves these posterior regions of the brain in neonates, different than the pattern in older children and adults, where it involves the basal ganglia. Neonatal hypoglycemia, usually below 26 milligrams per deciliter, and may be superimposed on sepsis, on HI, or on neurometabolic disorders. So always look for that posterior pattern as well. Now here's your profound anoxia pattern, very selective involvement of the dorsal putamen on the ADC map of the ventral lateral thalamus, posterior limb of the internal capsule, and the periurelandic cortex. Why the periurelandic cortex? This is a very high, highly active area of the brain. Babies are not thinking abstract thoughts with their parietal lobes, or trying to make judgment decisions with their frontal lobes. They're just trying to feel and move around. So the sensory and motor cortices are the most active, and these are the ones that will suffer first. As you can see here on the diffusion, and then there are certain areas of the brainstem, like the colliculi, that may suffer as well because they are actively myelinating. What I showed you was in the first few days after the event, if I go back here on diffusion, notice day three after the event. Now looking at a different child, day eight after the event, the diffusion will have normalized, typically between day six and 12, more delayed on the children that have been cooled, and we're gonna talk about that. But eight days after the event, now notice heterogeneous signal intensity on T2, and very bright T1 signal intensity in the ventral lateral thalamus, globus pallidus, and putamen, particularly dorsally, while we not see that dash of myelination in the posterior limb of the internal capsule. So that is what pathologists would call status marmoratus, where the model appearance of the brainstem would show on a necropsy exam. This is profound, brief hypoxia, or an anoxic episode. Now what is the most important imaging sequence to evaluate in the first three or four days after hypoxic ischemic event? And you can see the options, and the correct answer would be DWI and ADC. We're first looking at this sequence before pseudonormalization, where the brain starts to look normal, and the value starts approximating normal again on the ADC. So let's look at a moderate hypotension that is prolonged. So this is the partial asphyxia pattern. Notice this is a five-month-old who suffered abusive head trauma. It also had undergone drainage of subdural collections. But what's striking here is, while the cortex is almost diffusely involved, the motor and sensory cortex are spared. So the partial asphyxia gave the brain time to readjust and shift the consumption of oxygen and glucose to the motor and sensory cortex, while the rest of the cortex has suffered. You see this pattern too in adults. So when you look at an ICU case of cardiac arrest, sometimes we see a similar pattern. And finally, profound and prolonged. These are usually not difficult to diagnose, and knowing the status of the child will make you never miss a case like that. This can only be a pseudonormal if a not-very-well-experienced reader is not paying attention or not looking at the clinical status of the child. This is the picture pre-brain death, where the T2 shows no cortex. This is the missing cortex sign, but completely diffused. Here's the diffusion, and here's the ADC. If you measure ADC values here, they will be very low, and we'll talk about that. This is a case on day six that was actually the preliminary read by a neuroradiology fellow called it normal. And this was the first HIE, or hypoxic ischemic encephalopathy case, read by this particular fellow, which is probably a very unpleasant experience to get this wrong, but it will probably be the only case that this fellow will ever get wrong, because once you start looking at the case, notice that the cortex looks brighter than the basal ganglia, but it's such a symmetric pattern. They're like, is this normal or abnormal? You can window the diffusion however you want, but once you start measuring a few ROIs, notice that the cortex is showing an average of 542. The normal, in a term neonate, in the basal ganglia in the cortex is about 1,000 to 1,100, okay? Anything below 800 or 700 is wildly abnormal. So I always measure the ADC values. The other thing to remember are the patterns of secondary cytotoxicity. Five days after the event, this child has now restricted diffusion in the corpus callosum, particularly in the splenium, and this is sometimes called a new infarction in the splenium or a second episode of hypoxic ischemic brain injury, but we're just looking at the evolution. First, the gray matter is affected on diffusion. After a few days, the white matter, and this is very common, irrespective of the etiology of the injury, even a large stroke or ischemic stroke, ischemic infarction in a neonate could cause this pattern of transsynaptic injury. We also call this acute wallerian degeneration, even though it's not a precise pathologic term, but this is actually glutamate release, secondary cytotoxicity, okay? Notice how on the ADC, you still see some restricted diffusion in the motor sensory cortex, in the basal ganglia, and also restricted diffusion in the splenium, which is the secondary finding. Again, day five after the event, markedly abnormal hypo-intense in the deep gray nuclei. Notice the markedly abnormal hypo-intense signal in the amygdala and hippocampus, and also on T1-weighted images, these are T2-weighted images in here, T1, super bright signal, very hyper-intense in the deep gray nuclei. Same child, six years after the event, so this was selective brain injury, so this child did not die, but instead got cavitary encephalomalacia in the deep gray nuclei, microcephaly, and tissue loss, encephalomalacia in the supratentorial brain while the brain stem and the cerebellum were particularly or relatively spare. They can be involved to a certain extent. These children will not have the spastic diplegia or the spastic cerebral palsy. If they have cerebral palsy, it will be the dystonic type due to the basal ganglia injury, okay? So just by the type of cerebral palsy that a child shows up, the neurologist or pediatrician can guess, or we can try to guess what was the imaging pattern. So here are some practical tips. So these are the big, small tricks to avoid big mistakes, or that was the name of the session, right? So know the age of the child, premature or not. Inquire about the known details of the event. You know, it's not cheating to know the history, and I don't like reading cases without knowing the detailed history. On DWI, a practical tip is most of the times, not always, but most of the times the cerebellum is normal, is spare. So window your DWI image on the cerebellum, then you go up to the supratentoral brain to avoid a superscan or pseudonormal scan where you can falsely adjust your window to not show the brightness. Measure the ADC in basal ganglia cortex and white matter, term baby, basal ganglia cortex about 1,000 to 1,100, white matter, anything between 1,500 to 1,800. And know also, and this is important for medical legal cases or for documentation, hypothermia, brain cooling delays the evolution of DWI changes. Usually most papers will mention six to eight days without cooling the brain, the interval to pseudonormalization of diffusion findings. Notice this interval is delayed with brain cooling. And when to suspect the neurometabolic disorders. In just a few minutes, why do we even talk about it? These are very rare disorders, but collectively they are not so rare. They mimic more common problems that occur in neonates, such as hypoxic ischemic encephalopathy and sepsis. Many of these are treatable just from dietary restriction or supplementation of amino acids. Most of these are amino acidopathies and organic acidurias. And rapid diagnosis may ensure proper management and may prevent irreversible brain injury. There is one classic one that we identify that usually has a clinical suspicion and I'll show it to you here. But this is the syndrome you have to remember. It's a fairly well-defined clinical syndrome. I do not like this term. This is an official term that you'll find in pediatric neurology books. The group of devastating neurometabolic disorders of the newborn. I don't know how a neurologist can come to a patient's family and say, I think this is the group of disorders that your child may have. But that's how you'll find these in the textbooks, okay? So unlike a baby with hypoxic ischemic brain injury that comes severely depressed and needs resuscitation, this is typically normal at birth. There's no association with premature, so a term neonates, that a few days later will refuse to feed and vomit. And usually gets a request for an ultrasound of the pylorus for hypertrophic pyloric stenosis. But the kids with hypertrophic pyloric stenosis are vomiting acid. They have metabolic alkalosis, while the kids with neurometabolic disorders have metabolic acidosis. And any neonate with metabolic acidosis should be suspected to have one of these neurometabolic disorders until proven otherwise. Then they will develop other organs failure, like acute liver failure, lethargy, coma, changes in tonus, maybe hyper or hypotonia, or alternating, and that may give a clue to the disorder. And then seizures, remember also the role of hypoglycemia that could be a contributing factor. This is a classic case where a radiologist can make the diagnosis even without much history. So it's a two-week-old hypotonia and seizures. You look at the brain on T2, and the brain just looks swollen. There are no cell site. The cortex appears intact, but all the white matter looks very swollen. Notice the internal capsule looks thickened and swollen as well, the cerebellar white matter. Then when you look at diffusion-weighted images, there is restricted diffusion in the motor tracts, see periurelandic and coming down, the internal capsule, in the thalamus, in the long tracts of the brainstem, and in the cerebellar white matter. So these are all the tracts that we saw that are myelinated or myelinating in a term newborn. So this is a myelination map. I didn't show you the ADC, but these are all restricted. So restricted diffusion in the myelinated white matter, while the rest, the remainder of the supratentor white matter, has edema. This is a classic finding for maple syrup urine disease. The name is because the cerumen and the urine smell like burned sugar or maple syrup. Actually, this was described by a resident in New England. That was Dr. Mankies when he was a resident at Boston Children's Hospital. I can't imagine a Latin American doctor like me diagnosing maple syrup urine disease. I just wouldn't know what it smells like. This is a leucine encephalopathy, so you have to restrict this amino acid and requires lifelong dietary restriction. If the diagnosis is made before the first crisis, this student can have a relatively better outcome. So in summary, for hypoxic ischemic brain injury, know the age of the child, the history of the event, use everything in our favor to get to the right diagnosis, and adjust the window for DWI for the cerebellum, then look at the supratentorial brain. Measure the ADC values. Know that there are 20% of cases that can be normal and the child may still have abnormal outcome. And for neurometabolic disorders, the history is also key. And the most important for us is to look for patterns that do not match any of these patterns I've shown you of hypoxic ischemia, okay? So I'd like to start by thanking the RSNA and the Refresher Course Committee for the invitation to speak today. Over the next 20 minutes or so, we're going to review a series of cases with a theme of disorders of CSF circulation, and your job is going to be able to identify the abnormality and then answer the provided question. So get your phones out now. We will be using the audience response system. So here's case number one. So for each of these cases, you're going to get a series of images. I want you to take a look at the images, make a diagnosis or identify the finding, and then we're going to answer the question. So here you've got three axial images from an MRI of the brain, and the first question here is, what is the most likely underlying ideology for this finding? And so here are your choices. Is the most likely underlying ideology subarachnoid hemorrhage, B, congenital aqueductal stenosis, C, idiopathic intracranial hypertension, or D, normal pressure hydrocephalus? So go ahead and take out your phones. We're going to take a look here. And go ahead and enter your answers. We'll give everybody some time to do that. Another five seconds. Okay, we've got a bunch of answers here. Let's go ahead and take a look. So most people thought it was B, congenital aqueductal stenosis, but we had a couple of choices for normal pressure hydrocephalus as well. Well, what's the right answer? The right answer here is indeed B, congenital aqueductal stenosis. So good job to the group here. And so what are the findings on this first case? Well, the first thing that we're seeing is that there are enlarged lateral ventricles on that image all the way to the right. The other thing that we're seeing, of course, is that we've got an enlarged third ventricle here on that middle image. But the real key here is to identify that you have a normally sized fourth ventricle on the image all the way to the left. So that suggests that there's some sort of obstruction within the ventricular system that's causing upstream dilation of the ventricular system. And so what we're looking at on this case, so this is that case. This is a patient with this patient's sagittal steady state free procession image. And if we kind of zoom in into the region of the cerebral aqueduct of sylvius, you can see this little web here, which is causing the congenital aqueductal stenosis. And so what we've got here, this is demonstrating an example of obstructive hydrocephalus, which is occurring because you've got most of your choroid plexus being created within the lateral ventricles, extending through the foramen and roe into the third ventricles, and then being blocked here at the cerebral aqueduct of sylvius. So you have to understand the general course of CSF. And from here, of course, it's going to go into the fourth ventricle and then out foramen and luschka and magende. Now whenever you have a stenosis at the level of the cerebral aqueduct of sylvius, you do have to entertain other possible causes of that. And one other cause is a tectal plate glioma, as we're seeing here in this other patient. So you have to consider that and look for that as well. So we'll take an opportunity here to talk about hydrocephalus in general. And there's two broad categories, right? We've got our communicating hydrocephalus and our obstructive hydrocephalus. So in this case, what we saw was obstructive hydrocephalus, and that was due to obstruction of the ventricular system and led to the dilation of the ventricles proximal to the point of the obstruction. So causes of that, of obstructive hydrocephalus, include our congenital aqueductal stenosis, but also neoplasm or things like hemorrhage leading to a clot sometimes within one of the outlet foramina. On the other type here, our communicating hydrocephalus, we've got a really different pathophysiologic mechanism. Here we've got failure of CSF resorption at the level of the arachnoid granulations. And this leads to a diffuse dilation of the ventricular system. So you might as well think of this as too much CSF. And so the causes of this include subarachnoid hemorrhage, meningitis, particularly pyogenic meningitis, and then leptomeningeal carcinomatosis. So note that all of these causes involve things like cellular or protonaceous material within the CSF, which disrupts the arachnoid granulations. OK, so great job on case one, the room one on the first case. So let's take a look and see how you guys do on case two here. So this is a patient that you're reading in the reading room. And the indication for this cervical spina mar is neck pain. And so my question for you here is, in which quartile of the image is the causative pathology? Is it A, B, C, or D? So go ahead and take out your phones, and let's go to the votes. All right, we're going to give you about seven more seconds. So go ahead and enter in your votes, make a decision. All right, that's about a critical mass there, so let's see. What did everybody say? Most people said C. And then there was a pretty good split with the rest, but about 21 of you got D. So let's take a look. Let's look at C. And so we said neck pain. Well, the answer here is actually D, OK? And so I understand why most of you said C, because there's this disc here. But actually, the pathology is being pointed to here by this yellow arrow. And what we're seeing is a collection in the ventral epidural space of the spinal canal. And this black line that's being demarcated here by the yellow arrow is actually the dura. So this is a ventral CSF leak, OK? And in this patient, I did not give you this, but this is their accompanying thoracic spine. So you can see that leak to even better advantage here. So here's the black line that's the dura and our ventral epidural CSF leak in a patient with neck pain. It is really common for patients with spontaneous intracranial hypotension to end up with neck pain or pain between the shoulder blades. So this is what this looked like on an axial image. And you see the even better advantage here, this kind of thin black line, which is the dura and the ventral epidural CSF leak. And so that's what the diagnosis was. It's SIH or spontaneous intracranial hypotension caused in this case by an osteophyte spur. So what happens with this subtype of a spinal CSF leak is you've got an osseous spur from a disc bulge that penetrates the dura and results in a tear of the ventral dura and then spilling of CSF into that ventral epidural space. So here's an example from Duke. And this is an intraoperative photo here. So this person had a ventral CSF leak. And so this was the dural hole here because of the osteophyte spur. This is actually the spur right there. And this is an intradural approach. So our surgeon is going through the dura dorsally, displacing the cord, and taking a look at that hole in the dura ventrally. And so that was caused by this large osteophyte spur that's shown right here. We've got a pretty neat video here. If it works, this is a lateral digital subtraction myelography actually with the patient prone position. And as we put contrast into the normal subarachnoid space, boom, you're going to see it split out right through that hole here, confirming the site of the CSF leak. And that was at the exact location of this disc osteophyte spur. So what does this look like? So usually we go hunting for the CSF leaks with myelograms. And at Duke, we do our myelograms with the patient on the CT table. So we do a lumbar puncture followed by a myelogram and a quick rapid scan to see if we can capture the CSF leak. And so this is what we end up seeing. So in this case, we've scanned the patient. And here's normal contrast on our postmyelogram CT within the subarachnoid space. Here's spinal cord. What's not normal is this collection of fluid in the ventral epidural space. Now note that this fluid, over time, or rather as we increase, extend rather cranially, we see that the density within that ventral fluid collection increases, right? And so the density is really isodense to the normal CSF-containing contrast up here in the cervical spine, but it's relatively hypodense down lower. And so that's called the contrast gradient sign. It can be really useful because it suggests that the hole in the dura is up higher, somewhere here in the cervical thoracic junction, and can help us pinpoint where the potential source of the leak is. So if we took this middle image and then we went back to the image of the MR, you can see there's a really similar appearance here, right? So I want you to remember this MR appearance because even those of you who are not spinal eventualists and are going to be treating SIA patients, you may have the opportunity to be a hero in the reading room when you read that next cervical spine MRI for neck pain and you end up identifying a CSF leak. I've seen a bunch of these just reading normal cases, and you can really help a patient out because people are not thinking of this diagnosis. So this is important for everybody to remember to tell your friend Sam later. And there's a couple of different types of spinal CSF leaks that we need to know about. And one is a nerve root sleeve diverticulum here. There are these areas of dura that get denuded. The arachnoid pooch is through at the axilla or armpit of the nerve root sleeve, and you end up with a CSF leak. That's a nerve root sleeve diverticula. Type 2 is an osteophyte spur here, so that's the one that we saw already. And type 3 is probably the most interesting one. It's a CSF to venous fistula. So here at the level of the nerve root sleeve, you get an aberrant pathologic connection between the normal CSF-containing space around the nerve root and an adjacent paraspinal vein. This leads to unregulated egress of CSF right back into the circulatory system, and so you end up with low CSF volume and spontaneous intracranial hypotension. What's important to note is none of these are iatrogenic. That's the spontaneous component of SIH. Case number three. You guys are doing really well so far. Here we've got two images. One's a CT, one's an MR. The question here is what's the most likely diagnosis? I'm going to give you guys a second to take a look at that, and then we're going to put up the question choices. Is it A, infection? B, spontaneous intracranial hypotension? He wouldn't do that twice in one session, right? C, obstructive hydrocephalus? Or D, benign enlargement of the subarachnoid spaces? Okay, let's go to the voting. It's like we've got our answers coming a little slower this time than before. You guys are doing great. OK. Looks like we've reached a critical mass there, so let's see what everybody said. So here, it looks like the majority of the responses was A, infection. Slightly eking out D, benign enlargement of subarachnoid spaces, but we had a pretty good spread here. Let's go ahead and take a look and see what the answer is. Well, it was A, so the room wins. Good job, guys. So what are we looking at here? This is a case of ventriculitis, and the first thing that we needed to recognize is that on the left, this is actually a post-contrast CT, right? So we've got contrast within our torcula, but also in the internal cerebral veins, so we know that here, we've got a post-contrast image. And then what we're noting, and this is admittedly a subtle finding helped out by the adjacent diffusion image, right, is that we've got this little focal area of enhancement within the dependent aspect of the left occipital horn. And we see that again on the diffusion-weighted imaging. So we've got this area of enhancement within the left occipital horn that is corresponding to an area of restricted diffusion. So that's concerning for ventriculitis and infection, right? And so this patient, I didn't show you this because this would make it a little bit easier, but this health patient also had pretty florid meningitis, so here you're seeing enhancement of the leptomeninges, and this was a case of meningitis that led to ventriculitis. So ventriculitis typically is caused by a pyogenic source, and the causes or the sources of the pyogenic infection in the ventricles can be either due to iatrogenic causes like neurosurgery or ventriculostomy placement, but can also be secondary to contiguous spread from adjacent brain abscesses or hematogenous spread to the choroid or sub-epidema. Most commonly, though, it's secondary to a complication of meningitis. Ventriculitis, in general, has a very high mortality rate, so early diagnosis is critical. So what are some of the findings that we see? Well, on non-contrasted CT, it is somewhat nonspecific, but if you see layering, hyper-dense debris that is irregular within the ventricular system, you would want to raise that as a possible diagnosis, and then the other thing that we see is periventricular hypodensity or edema, that's a reactive edema. After you give contrast, a telltale sign is appendable enhancement along the ventricular margin as shown here on this CT, and then if you get an MRI, restricted diffusion within that layering debris is highly concerning in one of the more specific findings, which is suggestive. And then we also, again, see this appendable enhancement shown to better advantage here in our post-contrast T1-weighted image. FLAIR, again, you can see layering debris, but also that periventricular edema as we saw on the CT. Okay, so here's case number four. So we've got two MR images, axial images, and we want to know which of the following is the most likely cause of the pathologic finding. Your first step, of course, is to figure out what the pathologic finding is. Okay, is it A, high-flow CSF leak, B, acute intraparenchymal hemorrhage, C, leptomeningeal carcinomatosis, or D, idiopathic intracranial hypertension? So go ahead and get out your phones. Let's throw in our answers here. Seven more seconds, three, two, one, okay, let's see what everybody said. So pretty even spread here. Looks like our winning answer was acute intraparenchymal hemorrhage, followed closely by leptomeningeal carcinomatosis. Okay, so let's see. What's the answer? It's high flow CSF leak. And so the minority of you that answered that, that's really impressive. And what's going on in this case? Well, the first thing is we have to recognize what type of image this is. And these are axial T2 weighted images, right? So CSF is bright in this case. And then what we need to note is that there is low T2 signal or T2 hypo intensity within the cerebellar folia and along the brain stem. And so what is that? That's deposition of hemocytorin, right? So that's hemocytorosis. So now that we know there's hemocytorosis, we have to then know the following. And I'm showing case two again. Why would I be doing that? Well, on this case, we saw that ventral CSF leak. But what I didn't show you is up higher in this patient, we've got hemocytorin deposited within the posterior fossa. And that's because there is an association between this finding of hemocytorosis within the posterior fossa and a high flow CSF leak. In fact, this one series from our colleague, Wouter Shevink, demonstrated that about 3% of patients with cerebellar hemocytorosis or 3% of patients with SIH had cerebellar hemocytorosis. So if you see this, you should raise the question of the diagnosis of SIH. All of the patients who had this finding had that subtype of SIH that I showed you earlier where they had that large ventral CSF leak. So the thought is that there's repeated bleeding that leads to chronic hemorrhage, depositing in that location, and hemocytorin deposition. And yeah, so that was about 19% of those patients. In the interest of time, let's move forward here. Here's case number five. So here, I want to know what's the correct diagnosis. And here are our choices. Is it A, giant arachnoid granulations, B, SIH, I seem to like that diagnosis, C, dural venous sinus thrombosis, or D, idiopathic intracranial hypertension? All right. So let's enter our votes. Another five seconds here. Okay. So let's see what everybody said. So here, there's an overwhelming favorite in D, idiopathic intracranial hypertension by far. So let's see. Is that the right answer? Yes. All right. Everybody got it. So it's D, idiopathic intracranial hypertension. And so what are we seeing? Well, this is a time of flight MRA, right, or MRV, rather. And what we're noting on our MRV is that there are areas of loss of flow-related enhancement or signal at the junction of the transverse sinus and sigmoid sinus, that narrowing right there. And it's bilateral. And that's actually quite sensitive and specific for idiopathic intracranial hypertension, formerly known as pseudotumor cerebri or high CSF pressure. And we're seeing that kind of on this 3D rotating MIP here as well. So here's a paper by Farb et al. demonstrating that sensitivity and specificity for IAH with this finding is about 93%. And there is some debate about whether the cause of this finding is intrinsic, meaning the sinuses are stenosed intrinsically, or extrinsic due to that higher CSF pressure. But these two papers right here do show that there is some component of reversibility. If you take fluid off, you can make this sign go away. So it is clearly a dynamic process. And one reason why this might be a dynamic process is something called the Monroe-Kelley hypothesis. And so this states that the intracranial space is fixed, and the internal contents remain constant in constant total volume. So what that means is if you've got a calvarium or a skull, it's going to stay one size, right? And you've got three main components in there, brain, CSF, and venous blood. Now the brain really isn't going to change shape very much. It's the other two compartments that are going to be more pliable. So in the scenario where you've got CSF hypervolemia or too much CSF, everything shifts in this direction, right? And your venous compartment decreases. The opposite is true with intracranial hypotension, which is why we get pituitary hyperemia, venous distension, and enhancement of the dura. All right, final case here. Case number six. So for this one, I'm going to ask you a straightforward question. Which is the Chiari 1 malformation, A or B? That's too easy. Or is it both or neither? All right, let's go and vote. So which is the Chiari 1? Is it case A, case B, both of them, C, or neither of them? I'll give you 10 seconds here. Okay, let's see what everybody thought. Okay, so the majority of people by far thought it was case A. And that's correct. Good job, guys. That was our best one of the day. And I hope we got this right for the right reason. So here's Chiari 1 malformation, and here is spontaneous intracranial hypotension. I want to point out a few differences here. One of the challenges in differentiating these two is that there is an occipital headache with both of them, and there can be cerebellar tonsillar ectopia, inferior cerebellar tonsillar ectopia with both of them. So that's important to remember. But so you can see the low tonsils here. But one thing that we're seeing is that there's a difference in the pathophysiology. With a Chiari 1 malformation, there's a small posterior fossa driving forces this way. With a CSF leak, you've got a downward displacement of the brain here. And that downward displacement is key. So if we look here in the midbrain, there is an upsloping third ventricle in the Chiari 1 malformation and a downsloping third ventricle floor in the SIH case, and so that is really key. Here we have a flattening of the pons with SIH. And here, again, we have our upsloping in Chiari 1, downsloping in SIH. And so this is our mammillary body right here in our mammillary body. And we see that there's reduction in the mammalopontine interval as well. So that's another sign right here. So the number cut off here is about 5.5. Lower than 5.5 is abnormal and suggests SIH. If you misdiagnose this and call it a Chiari 1, what happens to these patients? It's a little craniectomy, right? That does not cure the spinal CSF leak. I've seen this about 50 of my own patients at Duke, so don't make that mistake. So a couple take-home points. We reviewed some commonly missed cases of CSF-related pathology. Remember narrow dural venous sinuses bilaterally. Think of IAH. And remember the imaging findings of SIH, hemocyterosis, that extradural spine collection. And please, don't mistake SIH for Chiari 1. Thanks. Okay. Thanks for having me here. It's great to see everybody here tonight. So I'm going to switch things up a little bit and take you on a case-based tour of pearls and pitfalls in the land of acute hemorrhage. So when we're talking about acute intracranial hemorrhage, it's very important to pay attention to the clinical presentation and history, including the patient's age, the location of hemorrhage, and have a working differential in the back of your mind. So the primary causes, the common causes of parenchymal hemorrhages include hypertension and cerebral amyloid angiopathy. And secondary causes are usually secondary to some lesion, like a tumor, vascular malformation, et cetera. The workup of acute intracranial hemorrhage generally begins with a non-contrast CT, because it's a really fast and good screening tool, and then often will go to a CTA. At our institution, our CTA is composed of a non-contrast CT, arterial phase images, as well as delayed phase images. And the importance of the delayed phase images I'll talk about in future slides. And in lieu of, or in addition to the CTA, and often you'll see it's in addition to it, we get an MRI. And that's mainly to look for an underlying lesion as the cause for hemorrhage. And depending on what you see on these cross-sectional images, you may decide to get a conventional digital subtraction endogram to improve your diagnostic capabilities, or to treat any vascular abnormalities that you see on cross-sectional imaging. And just remember, because this has been a savior, repeat imaging can be your friends, because lesions or findings can declare themselves on future imaging. So let's get on to the cases. Case number one, we see, I'm going to withhold the history here, but we see a series of non-contrast head CTs showing multifocal hemorrhages. You see it in the midbrain, the deep gray matter, the left frontal subcortical white matter. Of course, we get an MRI. GRE shows multifocal GRE spots. And just like you should know the differential for acute parenchymal hemorrhage, you really need to know the differential for multiple GRE spots. And there are plenty, and they're listed here. But every time you see a multiple GRE spot case, you're not going to list all of these diagnoses because it's not helpful. So what do you do? Look at the clinical presentation and history, the patient age, and the distribution of the hemorrhage. For example, in cerebral amyloid angiopathy, it tends to occur in older patients, and the microhemorrhages are peripherally distributed. In contrast to hypertension, where the microhemorrhages involve the central part of the brain as well as the deep gray matter, with diffuse axonal injury, you're looking for a history of trauma. And often, these patients have low GCS scores. For melanoma or other hemorrhagic metastases, you're looking for multiple enhancing lesions throughout the brain. For fat emboli, again, you're looking for a history of trauma, long bone fracture, or if the patient has sickle cell disease. Vasculitis, it's usually an inflammatory picture. You have elevated CRP, ESR, as well as multifocal beating on vascular imaging. With radiation, of course, there's a history of radiation. And with a hypercoagulable state, there's some sort of history that will lead you to it, like cancer, thrombocytopenia, DIC, and the labs will point you in that direction. So if I was to go back to this case, you don't even have to use your brain power. But if I gave you the history that this patient presented after a fall from a horse with a low GCS score, you're going to slam dunk on the diagnosis of diffuse axonal injury. Which brings us to question number one. What percentage of DAI lesions are hemorrhagic? Less than 10, 10 to 30, 30 to 50, 50 to 70, or 70 to 90? We're going to get going here. Maybe I'll wait for a few more seconds. Pretty good response. OK, so let's see what everybody chose. Let's see. How do I? Oh, OK. And then you press link again. Oh, this on this side. Oh, perfect. OK, great. It's actually even spread, but a lot of people chose 10 to 30. So let's see what the answer. Sorry, I'm just figuring out how to use this on the job training here. So the answer is 10 to 30. So good job. So you can see the minority of lesions are hemorrhagic, where the sensitivity of DWI in picking up these lesions is a lot higher. So based on the location of the DAI lesions, the volume of lesions, and the number of lesions, that is associated with clinical disability and worsening prognosis at the time of discharge. And we can grade the DAI lesions based on where it's located with grade three, the higher grade being in the brainstem. So next case, we have a 21-year-old male with acute lymphocytic leukemia presenting with confusion. We see a left parenchymal hematoma. Of course, we get a CTA. What do we see? We see a defect in the left sigmoid sinus. And coronal post-contrast images show extensive involvement of the transverse sinuses. And of course, we get an MRI, but MRI is not helpful, because we just confirmed the hematoma and clot in the dural venous sinuses. And this is a classic imaging appearance of hemorrhagic venous infarct. 70-year-old male, left facial droop and slurred speech. We see a hematoma in the right frontal parenchyma. MRI confirms the hematoma. Time of flight angiography, we don't see any obvious arterial venous malformation or ruptured aneurysm. Then we get pre- and post-contrast T1-weighted images. And we see a tubular T1 hyperintense structure along the lateral aspect of this hematoma. After we give contrast, it does not fill with contrast compared to the contralateral side, where you see a normal cortical vein filled with contrast. And this is the cord sign that is suggestive of hemorrhagic venous infarct. And it essentially represents a cortical venous thrombosis. And actually, if I was to show you all these images that I previously showed you, the cord sign was present. So whenever you have a peripherally located lobar hematoma, look for the cord sign. I'm not going to get so much into this diagram, but it's a great diagram that shows the normal venous drainage territories of the brain. So you can go to that website to find that. This is a 48-year-old male with right weakness and speech deficit. We see some hemorrhage in the left lentiform and caudate nuclei. CTA, pristine, normal vessels. MRI confirms hemorrhage in these locations. So what does this do to? Is this hypertensive hemorrhage? Well, we need to look at the patient's jacket, look at the history. One day ago, we see that this patient suffered a left MCA territory infarct. The seclusion of the MCA, the patient actually went to the cath lab, got recanalized. And this is hemorrhagic transformation of infarct. This is a companion case where we see a lesion with T1 hyper intense rim in the left lentiform nucleus. On the sagittal T1 weighted image, you can see that involves the caudate, involves a little bit more superiorly to involve the caudate. ADC, minimal restricted diffusion, no deoxyhemoglobin on GRE, no obvious hyperperfusion, a little bit of minimal enhancement. So what is this? Is this hypertensive hemorrhage? Well, if you notice that it involves both the caudate and the lentiform, it has this comma sign that's suggestive of a subacute infarct. So this is a subacute infarct. And that comma sign can really lead you to making that diagnosis. But of course, all that glitters is not gold, and radiology is not 100%. So this is a 69-year-old female with diabetes and problems walking. We see very similar appearance of hyperdensity in left basal ganglia confirmed on T1 weighted imaging, minimal changes between pre and post contrast imaging. So in a patient with diabetes, problems walking, movement disorder, you're going to be making a diagnosis of hyperglycemic injury. So the clinical presentation is really helpful. Moving on, 51-year-old male with new left visual changes. We see a right thalamic deep temporal lobe hematoma. On andrographic image, we see a spot of contrast. And on the delayed images, this is why we get delayed images, it blooms. So this is pretty wide. Everyone must know this. This is known as the CTA spot sign. Of course, we get DSA and MRI, which show no abnormality. But this brings us to question number two. The presence of a CTA spot sign predicts hematoma expansion, infarct growth, shorter hospital stay, future independent living, or aneurysm formation. OK, so we're working on this. Okay, so wait for a few more responses. Okay, so everyone chose hematoma expansion for the most part. So the correct answer is hematoma expansion. So the CTA spot sign is a marker of active contrast extravasation, and it's associated with larger hemorrhage, poor prognosis, high rates of early clinical deterioration and mortality. And if we were to look at this patient, you can see one day later, the hematoma expanded in two separate patients with CTA spot signs. Again, you see that the hematoma is bigger on follow-up imaging. Companion case, 84-year-old male presenting after a fall down the stairs. We see an acute right-sided subdural hematoma and CTA spot sign. So what do you think this hematoma will do on follow-up imaging? Well, on serial follow-up, the hematoma was stable. Why is that? It's because the underlying pathophysiology of subdural hematomas is often related to rupture or tearing of bridging veins and not arterial extravasation. So you can get the CTA spot sign in different acute settings, and depending on the underlying mechanism, there's a different clinical prognosis. 82-year-old female found down. We see a right parenchymal hematoma in the temporal lobe. Of course, we get a CTA. The vessels look pristine. Next we get an MRI, which confirms the hematoma, but we also see a hematoma in the right occipital lobe here. Then on post-GAD and perfusion imaging, we don't really see an underlying lesion. So no underlying lesion, and this is an 82-year-old. So what is going to be your diagnosis as a cause of the hemorrhage? Before I give that away, this is another patient with the same diagnosis for more classical imaging findings, where there are multiple microhemorrhages in a peripheral distribution. Also there's leptomeningeal hemocytosis in different stages, and that is more classic for cerebral amyloid angiopathy. And the other case was path-proven CAA as well, which brings us to question 3. What is the most important risk factor for development of CAA? Hypertension, increasing age, diabetes, smoking, or male gender? Okay, let's see. So great, so everybody said increasing age. So the correct answer is increasing age. Wow, you guys are really smart, or I just need to make my questions harder. But so this diagram we got from Stroke, it's a really good diagram, not going to really get into it in the interest of time, but there are several variants of CAA. The CAA we're talking about didn't have any associated inflammation, but there are inflammatory variants of CAA. Here's a companion case, 64-year-old male with left hemiparesis and seizures. We see an MRI, lots of edema in the right temporal lobe, as well as the bilateral parietal lobes, left greater than right. On GRE, there are cortical hemocytin staining, as well as peripheral microhemorrhages. Post-scatalinium, we see enhancement in the left parietal lobe. Again, that's a little bit nonspecific. One month later, we can see evolving injury in the right temporal lobe, but it's getting worse in the left parietal lobe with more microhemorrhages. In this patient, when you have peripheral microhemorrhages that's associated with edema, this should clue you into the diagnosis of inflammatory cerebral amyloid angiopathy. The patient was placed on steroid therapy, and six months later, did okay. 64-year-old male with left hemiparesis and seizures, we have a posterior fossa hemorrhage. You can go down your differential of so many different things causing this hemorrhage, but if you look through the patient jacket, we notice that the patient just had lumbar spine surgery. So based on the previous talk, we would suggest that this is remote cerebellar hemorrhage. And that brings us to the next question, which is, what is the underlying pathomechanism for remote cerebellar hemorrhage after spinal surgery? I'm just going to wait for a few more responses here. All right. So most people chose venous bleeding. And the correct answer is venous bleeding, presumably due to rapid CSF loss during surgery resulting in cerebellar sagging and rupture and tearing of the bridging veins. Now, the most important thing to know is this is benign and self-limited. And it can occur. It's rare, but it can occur after lumbar spine and supratentorial surgery. So let's move on to subarachnoid hemorrhage really quickly. 37-year-old female with headache and confusion. This is obvious subarachnoid hemorrhage. Get a CTA. Very classic finding of ruptured aneurysm. Why am I showing this? Because I just want to show the utility of dual energy CT, especially with these iodine overlay maps. We can differentiate between iodinated contrasts and hemorrhage. So the iodinated contrast is in red. And you can see that the aneurysm is red, whereas subarachnoid hemorrhage is not. And more to follow on that. 64-year-old male with left weakness and difficulty speaking. We see a right MCA territory infarct. We see a clot and occlusion of the right MCA. There's tissue at risk. We bring the patient to the cath lab. We confirm occlusion. There's successful mechanical thrombectomy and recanalization. And then we get an MRI one day later. The infarct size is about stable. There's a little bit of hemorrhagic transformation of the right basal ganglia. But what do we notice in the subarachnoid space? Is this subarachnoid hemorrhage? Well, unfortunately, this was called subarachnoid hemorrhage. But we got a follow-up head CT the same day. And curiously, there's no hemorrhage in the subarachnoid spaces. And on the follow-up MRI five days later, again, we don't see any hemorrhage. So this is what is known as hyper intense acute reperfusion marker, otherwise known as HARM. And our last question of the day, what is responsible for harm seen on MRI? Infectious meningitis, chemical meningitis, acute subarachnoid hemorrhage, blood-brain barrier disruption, or artifact? OK, so let's see what everybody said. And everybody said blood-brain barrier disruption. OK, so the answer is blood-brain barrier disruption. Exactly. So you can also detect this phenomenon with CT. For example, this is a non-contrast CT. We see an occlusion of the right MCA, delayed Tmax in the right MCA territory, suggesting tissue at risk. This patient is brought to the cath lab. There's an occlusion of the MCA. There's successful recanalization. And we see hyperdensity in the sole side. Again, to an untrained eye, you would say that has potential subarachnoid hemorrhage. But it could just be blood-brain barrier disruption, which in this case, it was. So this is a really exciting companion case. 83-year-old male with chronic renal failure and meningitis. So this patient was admitted and had altered mental status and had this first head CT here, which we didn't see anything acutely. But then six hours later, we get this head CT. And everyone freaks out, because this looks like diffuse subarachnoid hemorrhage. But it's curious, because it'll only happen six hours later. The patient ruptures in six hours. Anyways, we look through the patient jacket and notice that this patient had contrast-enhanced CT, giral infectious source. But we were so freaked out by the CT, we went ahead to work it up, got a CTA. It was negative. The MRI was essentially negative for any acute processes or hemorrhage. The patient had a ventricular catheter replaced. And we sampled the CSF. There was only low RBCs in the CSF. So at this point, we conjectured that there was blood-brain barrier disruption, because as you remembered, the patient has a history of meningitis and chronic renal failure, which can exacerbate blood-brain barrier disruption. And the patient got dialyzed. And three days later, all that contrast was removed. OK, we're winding down here. 77-year-old female with dysarthria and right-sided weakness. NIH stroke scale of 18. Essentially normal CT. We see an M2 occlusion, tissue at risk. Bring the patient to the cath lab. Got successful recanalization, and we see this picture. Interventricular and subarachnoid hyperdensity. Is this true subarachnoid hemorrhage? Well, in contrast to the last case, this patient clinically deteriorated really rapidly, whereas the last patient wasn't so clinically sick. And this was true subarachnoid hemorrhage. So can we differentiate hemorrhage from iodinated staining? Well, this is where the utility of dual energy comes into play. For example, a 71-year-old female, left hemiparesis, had a right MCA territory infarct, no hemorrhage, ICA occlusion, brought to the cath lab, recanalized, had the CT. And with dual energy CT, you can see that the subarachnoid hyperdensity corresponds to hemorrhage, whereas the basal ganglia corresponds to iodinated contrast. But if you don't have dual energy CT, just get a follow-up MRI or CT, and you can see that the hemorrhage, again, localizes to the subarachnoid space. So in the interest of time, I'm going to skip over these cases. But in summary, the acute intracranial hemorrhage workup really starts with a non-contrast head CT. Sometimes you go to a CTA or an MRI to look for a spot sign or vascular abnormality or an underlying lesion, respectively. And depending on what you see, you can go to DSA for improved diagnostic capabilities or to treat vascular abnormalities. I've shown you a bunch of stroke cases. And remember, repeat imaging can be your friend. Thank you. OK, I'm happy to speak in the session and tell you about some do not miss lesions of the spine. We're going to start with a quote from a great American philosopher, former President George W. Bush, who said, fool me once, shame on. Shame on you. Fool me. You can't get fooled again. So hopefully, once you guys see these cases, you won't be fooled again. So what type of errors do radiologists make? It's an interesting paper from Kim and Mansfield in 2013, where they look at errors that were picked up in retrospect. The most common error by far is what they call under-reading, just the finding is simply missed. That's about 42% of cases. Second most common error is satisfaction of search. So one finding is made, but a second, perhaps more important finding is not identified. And third most common is so-called faulty reasoning. A finding is made, but the correct significance isn't applied to it. Maybe an incorrect differential diagnosis is given. When you look at lawsuits, by far the most common reason that radiologists are sued is so-called failure to diagnose for a variety of reasons. But that's much more common than, for example, procedural complication or failure to communicate some of these other issues. But when you look at anatomic subtypes, spine imaging is actually the third most common for a claim against a radiologist, so relevant to everyone in this room. OK, I just want to start with a case here. I have a question to follow. This patient presented not too long ago with back pain, and so they ordered a lumbar spine MRI. And I'll give you a moment just to look at this. So we've got a sagittal T1, sagittal T2. And we're going to go to the question here. So based on those images, what is the most likely diagnosis? Is it a disc extrusion? Is there an epidural abscess in the image? Is it a retroperitoneal mass? Is there subarachnoid hemorrhage or evidence of a vertebral fracture? Give you guys a little bit of time to weigh in. Okay, so let's look at the answer here. So a split in terms of the responses. Let me show you some other images from this case. So here's the axial from that same patient, and you can see there's something here in the retroperitoneum that's replacing the right psoas muscle. But if you go and look at the localizer view, it's really not subtle. There's a very large abdominal retroperitoneal mass, encasing the aorta, lifting up off the spine, secondarily invading the spine. And that's what's responsible for this patient's back pain. This was biopsied, and it came back as lymphoma. So my first tip for you is scrutinize your localizers. Look at your scout images. You're responsible for everything on the film. And at the different spinal stations, there's a lot of anatomy that you can catch and see. Here are just three cases I had recently come to the reading room. This first patient here had a cervical spine MRI, and we noticed all the cervical lymph adenopathy. That was biopsied, and she had undiagnosed lymphoma. The next patient in the middle presented with neck pain, had a cervical MRI. The canal and neuroframing were open, but had this abnormal right vertebral artery flow void, diagnosed with a vertebral artery dissection that was responsible for her neck pain. And on this axial lumbar image, the spinal canal is essentially normal. But if you look outside of the spine, you can see there's bilateral aortic and iliac dissections. All right, next case here. So we have a sagittal STIR and an axial T2 through the conus. You notice there's some high signal in the conus. And on the axial T2, it really conforms to the central gray matter. You have that butterfly appearance. So hopefully, everyone's thinking about a diagnosis here. But once I show you the diffusion, that should help you. So here's an axial diffusion, axial ADC. It's bright in diffusion, dark in ADC. Different patient. Up in the cervical spine, we got this vague intramedullary T2 Edmonton signal. On the axial image, it is central gray. It's conforming nicely, kind of that butterfly appearance. Here's the diffusion image. So these are both examples of spinal cord infarct. And my second tip for the lecture is, use diffusion imaging in the spine. We usually get a sagittal diffusion. It's very helpful for a cord infarct, if you're trying to diagnose epidural abscess, or discitis, for those of you that are familiar with the claw sign. I think it's a useful modality, particularly for inpatients. Spinal cord infarct usually presents with acute on-site weakness, loss of sensation. You get this snake eyes or owl eye pattern on T2 from the central gray involvement. And it's positive on your diffusion scans. Here's kind of a companion or corollary case is the pediatric patient has increased T2 signal in the cervical cord. On the axial T2, it's also that central gray pattern. But the diffusion on this patient is negative. And they had acute flaccid paralysis. They were positive for Enterovirus D68. So the diffusion can help distinguish those two diagnoses. OK, next case here, this is a young woman who presented with an orthostatic headache. So we had an excellent lecture on this topic before. We see the cervical cord looks normal. The thoracic cord is also normal. But when you look at the brain, the tonsils are low. So that's one thing we're going to note. But then you have a lot of secondary findings. Pituitary gland is very big. You have this down slope in the third ventricle. The mammillary bodies are almost touching the pons, the pre-ponting spaces of the face. So there's evidence of sagging. So this patient has intracranial hypotension. She saw a few different doctors. She's in her 20s. She had a blood patch. She had a myelogram that was negative. Eventually she got a follow-up imaging study about 18 months later. And what's happened? Her brain looks almost identical, but she's developed a very large cervical and thoracic syrinx. She saw enough neurosurgeons until someone agreed to decompress her posterior fossa, which as you heard earlier, doesn't help with this diagnosis. So now she's had a suboccipital craniectomy, a C1 laminectomy, but she still has a very large syrinx. CT myelogram was negative. She saw us eventually. We did an MR myelogram. So this is an axial T1 image. It's a T1 fat set, a little bit of intrathecal gadolinium. And that really lights up the spinal canal. What's interesting about this image in particular, though, is that you have contrast in the renal collection system and in the ureters. We've known for a very long time, since the days of nuclear cisternography, et cetera, that early excretion of intrathecally administered contrast is a sign of CSF leak. And when we go up a little bit farther on this same study, again, axial T1 fat set, you have the subarachnoid gadolinium. We have this irregular left-sided perineurosis, you know, pacifying left paraspinous vein draining into the asthma system. So we diagnosed this patient with a CSF venous fistula. They went to the OR to have this level clipped. And here's them just localizing the level in the OR. And I want to show your post-op imaging because it's pretty dramatic. So here's that presentation scan. Eventually, she developed a syrinx. And here she is just a few weeks post-op. So normal configuration as opposed to your FASA. The tonsils came up. And you can see that the syrinx is almost completely resolved. So intracranial hypotension, you know, the tip here is not all low tonsils are Chiari. It's something you will see not uncommonly as a radiologist. And it's often misdiagnosed both clinically and on imaging. These patients may develop a syrinx, even a really bad syrinx like this case, that still does not make it a Chiari malformation. The brain imaging mnemonic is SEAPS. It's all based on this Monro-Kelley hypothesis. The brain's trying to fill in the spaces. So you get some dural collections, macumin and gel enhancement, big pituitary, sagging. And findings are very similar, dilated epidural veins, dural enhancement. One finding that I want to highlight here is this so-called C1C2 false localizing sign because I've seen some of my colleagues get tripped up on this. Basically what it is is on a STIR sequence, you can see this high signal dorsal to C1C2 extending to the paraspina soft tissues. The idea here is that these patients, because they have orthostatic headache, they like lying flat. They like lying on their back. And the CSF leak comes up and pools in the dorsal epidural space. The ligamentum flavum ends at C2, so then you get the CSF pooling out at C1C2. So this is a marker of spinal fluid leak, but it is generally not the site of CSF leak. And you don't want to go blood patching at C1C2. I've actually seen that quite a few times. CSF venous fistula we heard about a little bit earlier. This is something that's increasingly described and diagnosed. Here's just a recent patient I had. I did a decubitus myelogram to try and have the fistula pop out. I put some arrows on it here. Some people prefer digital subtraction myelography. We've had a lot of luck with MR myelography. This is a different patient I had. Again, a little bit of intrathecal gadolinium, and you have this paraspina cyst, so pacifying paraspina vein draining to the asthma system. So you can see that quite well. If you can identify the fistula, then a surgical clip may lead to a cure. Generally, these are not responsive to blood patching. So here's a companion corollary case. This patient does not have sagging. Their tonsils are normal in position. They have an incidental pituitary, raspy cleft here. But on the axial susceptibility-weighted imaging, there's a lot of siderosis in the posterior fossa. All the cerebellar folia are full of hemocedrin. Patient had a myelogram, big CSF leak here. You can see the contrast getting diluted down here distally. It's a little tough to see on the sagittal, but there's a little dural reflection here. And on the axial, you can see a little bit easier, there's this ventral epidural collection. It's a high-flow leak. We couldn't figure out the site, so they were sent for a subtraction myelogram, digital subtraction myelography, and we thought it was here at the cervical thoracic junction. They eventually went to the OR, and they found this big ventral dural runt that was repaired. So the tip here is that CSF leaks are associated with siderosis in a minority of patients. There is a very strong association with ventral dural defects. It's probably secondary to friable vessels at the site of CSF leak that just ooze hemorrhage over time. The ultimate theory is that their brain sag causes some rupture of superior cerebellar veins, but there's not as much evidence to support that. So if you have siderosis that's unexplained, think about this diagnosis, particularly if you have an epidural collection in the spine. Okay, next case here. So subtle finding here on the sagittal T2 is a little bit of a contra abnormality of the upper thoracic cord. You see it better in the fiesta. The cord is kind of kinked forward up here in the upper thoracic spine, and really nice axial image. So we have a ventral epidural collection here. This is normal subarachnoid space. You can see the dural nicely, and you see the cord just herniating right through that defect there. So we diagnosed this patient with a cord herniation. They went to the OR, and indeed they found this large ventral dural defect, and here's my neurosurgical colleague repairing that. So the tip here is look for cord signal, but also look at the morphology of the cord. This is diagnosed more common in women. There's a good prognosis with surgery, and think about these echobalanced sequences. These usually occur in the upper thoracic spine. So companion case here, high signal at the cervical thoracic junction, and there's this kind of odd contour of the upper thoracic cord here. It kind of looks like a surgical scalpel, so this has been called the scalpel sign. We suggested this patient had a dorsal thoracic arachnoid web. They went to the operating room, and here's what they found. So this is an intraoperative video. This is spinal cord, and you can see this horizontal band of tissue dorsal to the cord just pulsating back and forth. This is the intraoperative ultrasound showing the same thing. So this was ligated, and I just want to show you the post-op imaging because it's pretty dramatic. Here's that pre-op, and this is them just a few days post-op. So the cord contour is normal, and that edema has resolved. So dorsal thoracic arachnoid web, think about this scalpel sign. There's a nice paper from AGNR in 2013 that describes it. It has something to do with incomplete or disrupted arachnoid formation, but it's got a good prognosis with surgery, so if you make this diagnosis, it can really help out your patients. Another companion case for this diagnosis, so the cord comes up here in sagittal T2, and then there's this abrupt kink, and it almost looks like there's something behind the cord here. We sent this patient for a CT myelogram. Here's that subarachnoid contrast, spinal cord coming up, and there's this ovoid structure here, dorsal to the cord, kind of pushing off to the right. So it looks like an axial cord, and then there's this structure here. So we diagnosed this patient with an intradural arachnoid cyst. Usually these are located in the thoracic spine. They can communicate freely or can be sort of isolated, like the case I just showed you. Myelography can help differentiate those. These become symptomatic once they grow enough to compress either the cord or the nerve roots. The reason they grow is unclear. Maybe it's a ball valve, something to do with the asthmatic gradient, or secretions from the cyst itself. Okay, next case we're going to look at here. So we have cervical and thoracic spine. Cervical cord looks pretty good. When we come down to the thoracic spine, you can see there's this abnormal T2 hyper-intense intramedullary signal. And hopefully you also notice that there are these T2 hypo-intense abnormal flow voids surrounding the cord. You can see them well here. So we're suspecting a dural AV fistula. I referred this patient for a spinal MR angiogram, and here's what we found. This is highlighting the right T10 radiculo-meningeal artery. It's coming in, and these are subsequent coronal cuts. You can see it feeding this fistula as a contrast to this MRA. Here's some 3D reformation. So we told our angiographers, we think it's right T10. Take a look. Indeed, they injected the right T10 level, and you can see the fistula filling, the dilated veins. The dampness is at a different level. Here's just a zoomed-up view showing the fistula's connection. And if you can get a super-selective catheter out there and glue across the AV shunt, you can give these patients a cure just with angiography. So spinal dural AV fistula. This is a type 1 dural AV fistula. It's the most common. There are other types. Much more common in men, usually present in 50s or 60s, usually lower thoracic cord. The venous hypertension leads to chronic hypoxia of the cord and myelopathy. The general workup for us is MRI to diagnose, MRA to guide, and DSA to confirm it and hopefully treat it. If you have edema without flow voids, so kind of like this case here, you may have a slow flow shunt. It doesn't exclude the diagnosis. And if you have flow voids without edema, maybe it's early in the disease process. So something just to keep in mind. Okay. We couldn't have a do-not-miss lecture without talking about trauma, so the last case I want to show you here is just a lateral and open mouth odontoid radiographs. I don't read a lot of plain film to me. This is a little bit subtle. Maybe it really jumps out at people in the room. But the C1 vertebral body is too wide. It's falling off of the C2 vertebral body here. So I think it's a little bit of subtle diagnosis. An axial CT, it's very obvious. You have a C1 burst fracture, Jefferson fracture of C1. No question about the diagnosis. So the reason I show this is my final tip here. Get cross-sectional imaging for cervical trauma. You can clear the cervical spine reliably on clinical grounds. There's a Canadian C-spine rule. There's something else called the Nexus. It doesn't matter which one you use. They're both very good at clearing a cervical spine. However, if you are going to image the patient, you probably don't want to get radiographs. This one study showed radiography for cervical spine trauma had a 43% sensitivity. We obviously don't want to accept that for a trauma patient. Whereas the sensitivity for CT is closer to 99%. This other study looked at the risk-benefit. So they said, okay, CT costs more. There's a theoretical radiation risk. Is this worthwhile in low- and high-risk trauma patients? And basically, the answer was yes. With pediatric patients, it's a little more complicated because this entity of sclera, spinal cord injury without radiographic abnormality. But I think the take-home point here is if you have a child with blunt trauma, even if their neurological exam is normal when they present to the ER, if they had any transient neurologic symptoms, they need an MRI because you can detect some abnormalities. And up to a quarter of those can have delayed symptoms, including paralysis. So the take-home points here, look outside of the spine. Use advanced imaging like diffusion. Not all low tonsils are archiari. Think about intracranial hypotension. And SIH can present with superficial sclerosis. Look for displacement of the spinal cord. Think about a dural AV fistula if you have flow voids and or chordedema. And cervical trauma requires cross-sectional imaging. Thanks very much.

neuroradiology

pediatric neuroradiology

neonatal brain

MRI imaging

brain injury

hypoxic ischemic injuries

white matter injuries

cerebral palsy

encephalopathies

diffusion-weighted imaging

metabolic disorders