So, I'm going to start off on neonatal brain MRI cases. Just to quickly introduce you to my hospital, this is an X-ray from 1896 of a child getting an X-ray done in Toronto at my hospital, so we have a long tradition and a long history of radiology at SickKids in Toronto. My objectives today are very basic, to show you some MR imaging cases of neonatal encephalopathy. You remember that the most important cause of neonatal encephalopathy is hypoxic ischemic injury. So when you think of clinical neuroimaging in the neonate, amongst the techniques that are available, ultrasound is a great bedside tool, it can show hemorrhage and hydrocephalus very well. CT is used for more acute things like trauma or acute hemorrhage. And the standard that is used, which is most reliable for causation and prognosis is MR, and that's what we'll focus on more today. And this question is regarding the conceptual patterns of birth-related hypoxic ischemic injury. Which of the following is most closely associated with a sentinel event? By sentinel event, we mean something that has happened acutely and over a short duration to cause significant hypotension and decreased flow to the brain. Is it white matter predominant injury, is it total anoxic injury, is it basal ganglia predominant injury, or a large artery infarct? And the answer is basal ganglia predominant injury. Note that when you look off, when you think of conceptual patterns of hypoxic ischemic injury, there's often a spectrum, so look for the dominant pattern. And if you have basal ganglia predominant pattern, which often goes along with the perirolandic pattern, and usually the posterior putamen and ventral lateral thalami, then that indicates a sentinel event. And in the next few slides, I'll show you the different patterns and how to think about them. So this is our first case, an example of profound injury due to severe blood loss due to vasoprevia. Note the involvement of the basal ganglia in the perirolandic cortex on diffusion and T2 images, and also note the abnormal T1 signal over here. The occipital lobe involvement in this case was presumed to be associated severe hypoglycemia at the same time. So this is what it looks like. These are magnified T1 images, which shows you the increased signal in the ventral lateral thalami, the posterior putamen, and the increased T2 signal. Note that flare at this stage is often useless because there's bound water in unmanneted white matter, which does not get suppressed. So while diffusion restriction is obviously an important sign of acute ischemia, another important sign to look for is on T1 images, where you lose the posterior bright stripe of the posterior limb of the internal capsule. This is normal. And you start seeing more increased T1 signal in the posterior putamen and the ventral lateral thalami. And this, when you see it, is a very grave prognostic sign. The only caveat is diffusion imaging or diffusion restriction occurs within two to three to four days, whereas these T1 changes occur more than a week after the insult. So this is our second case. In this case, there was intermittent type of hyperperfusion insult. And you can see that the watershed type of injury or the white matter is more dominantly involved over here, with diffusion restriction and increased T2 signal. The same child a month later shows this multicystic encephalomalacia in the white matter. And a year later, you can see extensive and severe atrophy of the white matter. Note also how the thalami are small over here. And that is because the thalami contains a lot of white matter tracts connecting with the cortex. And once you have injury in the white matter, because of axonal and valerian degenerations, the thalami also lose volume. This is our third pattern of injury, the acute and total anoxic injury. And this is an injury which occurs both that can occur with acute profound or a prolonged intermittent pattern of ongoing ischemic injury. Both of which, if they last for a longer time, will recruit more areas of the brain. And then ultimately, you can have the entire cerebral hemispheres involved. The cerebellum is often spared. And coronal DWI images sometimes make it very easy to make this diagnosis. So this is a slide which shows you the pattern of ischemic injury in newborns. What I like to think about is the first three over here are the diffuse ischemic patterns. And these two, multifocal white matter injury and the arterial ischemic strokes, are the more focal ischemic patterns of injury. So if the basal ganglia is dominantly involved, then you think of acute profound. If it is watershed dominant, you think of an intermittent, slow, insidious onset which occurs over many hours, like hypercontractility of the uterus, or if there's ongoing hypotension which lasts for a long time. And then, of course, both of these, if they last very long, can go on to a total anoxic pattern of injury. Let me show you examples of multifocal white matter injury. These are seen as dots of increased T1 signal in the deep white matter. And they do not correlate with hemorrhage on gradient echo images, and they're often diffusion restricted in the acute phase. And in this single pathology report, which is available for us to see what is the cause of these, they are thought to be due to increased microglial activated accumulation in those areas of ischemia. And that's how this is explained. These occur often in preterms or even in term in funds with congenital cardiac anomalies. Let's come to neonatal arterial ischemic stroke, and Susan is going to speak a lot more about pediatric stroke. These often present with focal seizures within one, two, or three days after birth. These children are born normally and then have the problem. You can see that this large arterial infarct over here, there is acute diffusion restriction in the corticospinal tract, which is often a gray prognostic sign. And this acute restriction in the corticospinal tract is also often a characteristic finding of very early childhood stroke. An interesting thing that occurs is we often do not see the large clot in the involved vessel. We will see hyperperfusion and a larger caliber of that vessel as seen in these two cases. If you're lucky, or rather unlucky for the child, you may see a clot in the large vessel on SWI images like in this case. So here's our ARS question number two. When is the best time to image with MR to show true extent of diffusion restriction in acute birth-related hypoxic ischemic injury? Right after birth, 24 to 48 hours, 70 to 84 hours, or after 10 days? So the correct answer is at 72 to 84 hours. And the reason for that, there are two reasons for that. One is logistically, you cool for 72 hours, so that's good. The other important reason is diffusion restriction maximizes at that time. And I'll show you this case to point that why and how it sort of does that. So this was a newborn with severe asphyxia. We did an MR right on the first day of life because there was consideration for alternate care and they wanted to withdraw care. So you can see the diffusion restriction involving the basal ganglia and the dorsal brain stem. So if you do it too early, you might miss the true extent. If you do it too late, you might miss the true extent of diffusion restriction. In this case, parents were still not happy with whatever evidence was available and they wanted more proof to withdraw care. So on the second day, you can see that diffusion restriction has evolved. And on the fourth day, that is after 72 hours, you can see extensive involvement of the brain. You can see the cerebellum is spared. So this was like a total anoxic pattern and it showed the true extent of diffusion restriction. And unfortunately, this child did die and there's this pathology proof that those areas of diffusion restriction really led to necrosis in those areas of the brain, including the dorsal brain stem, as shown over here. So why do we time it at 72 hours in most cases? It's because of these pathways of cell death and the neonatal ischemic injury. The reason for that timing is because in most cases of HIE, there's recovery after an initial insult, but that recovery is transient. And then you have secondary ATP energy failure, which leads to delayed cell death and diffusion restriction. And this is when it usually peaks at 72 hours and beyond, and that is when you should time your injury. Of course, if you have severe acute hypoperfusive insult involving the whole brain, you'll see diffusion restriction right away, like you would see in an acute arterial infarct. The third mechanism, which has been studied a little bit, is delayed apoptosis, but there's no good real imaging marker for that. Maybe some volume loss in the long term, but we really don't correlate it very well with imaging. So this is just to show you also how diffusion and T1 changes evolve over time. At day 4, you'll see the diffusion changes. This is in the posterior putamen. And the same patient at day 10, you'll see that the T1 changes are much more dominant and are useful to predict the bad prognosis in this particular case. So at most institutions and at SickKids, we do an immediate MR only for prognosis or to withdraw care or if you're suspecting metabolic or other problems. In our usual cases of suspected HIE, we would do it after cooling, after 72 hours. And if there are atypical findings or there are questions, then we'll follow up that MR at 7 to 10 days. So let's look at this case number 4, a 4-day-old term baby born normally and then present in status epilepticus at day 1. And you can see the extensive involvement on the CT scan. The MR shows diffusion restriction extensively involving the cortex, the basal ganglia, and the thalami. And this was actually a case of an autosomal recessive rare inherited metabolic disorder, molybdenum cofactor deficiency with homozygous mutation and MOX2. This is to remind you that sometimes rarely catastrophic inherited metabolic white matter disease can mimic HIE. Acute sulfite oxidase deficiency also can cause that. In this particular case, the baby was born normally. There were no signs of ischemia at birth. There was one clue. The other one was the cerebellum was already showing some volume loss and atrophy, which suggests that something was going on even before birth. Here are two other cases of HIE mimics. The important thing in both of these is that the encephalopathy occurred after birth, after a few hours or even after some days. So this first one is the classic MSUD, which shows diffusion restriction and increased T2 signal with swelling in the posterior of the internal capsule, the dorsal brainstem, and the deep cerebellar tissue. This one is a case of non-ketotic hyperglycemia, which on imaging looks similar to MSUD, but has a very characteristic MR spectroscopy signature. So this is the case for non-ketotic hyperglycemia, where you can see that it looks very much like MSUD, except that the T2 swelling is a little less. But on MR spectroscopy, you see this abnormal glyceme peak, which lives very close to where myonosatol lives. And the way to differentiate both is that on the short T, you should be able to see myonosatol. Myonosatol has a very short echo time. So when you increase your echo time, it should not be seen very well. So if you see what looks like myonosatol at 3.4, 3.5 on the intermediate or the long T, then you know that is glyceme, and you make your diagnosis of non-ketotic hyperglycemia. Here's our case number five. This patient presented normally at birth, but then had seizures on the right side, and you can see evidence of polymicrogiaria. A follow-up scan at 10 months showed that there were areas of the brain which were enlarged. This was hemimegalencephaly also. And you can see that increased hypo-intense signal and increased T1 signal of the underlying white matter due to hypermyelination due to ongoing seizures. Interesting finding over here was look at the septum pellucidum, which is thicker due to an aberrant tract crossing over. Very important for neurosurgeons to know because they need to resect that if they do a functional hemispheric tomy or a corpus callosotomy. So here are other causes of neonatal encephalopathy. This is HSV-2 encephalitis with asymmetric diffusion restriction. This is par-echo encephalitis. Enteroviruses and par-echo viruses can have this similar typical pattern. This is a vein of gallen malformation. So there are many other causes that can cause neonatal encephalopathy, but the most common is hypoxic ischemic injury. My last two slides and next 20 seconds, I just want to talk about other things that you need to consider when you're imaging the neonatal brain. So history is very important, timing is very important, and then technique is very important. Use the correct coil. If you use a normal head coil, your signal-to-noise ratio will be very poor because that depends on coil radius. Use the correct technique. For example, on T1 and T2 images, you can use much higher TRs because the neonatal brain has a lot of water and very long relaxation times, so you can use much higher TRs to increase your SNR. And also, you can use high ETL, echotrain lens. It does not cause any blurring, and you can say one time and do it in the same time. And with that, I want to thank my colleagues and everybody else, particularly the children whom we deal with. They stand tall in spite of all the things that they undergo with their problems. Thank you very much. Hello, everyone. Today, I'm going to talk to you about how my team and I at CHOA approach cases of unwitnessed trauma in young children. The objectives for this talk are to review an appropriate imaging approach to unwitnessed pediatric trauma and to discuss the strategy I would use when tackling these cases at the workstation. Okay, we're going to jump right into the first case. So case one is a six-month-old female that reportedly rolled off of her father while they were both sleeping on the bed. This short-fall scenario is a very common one that we see. The child, in this case, was lethargic in the emergency room with a GCS of eight and warranted head imaging. At CHOA, we start with a CT head without contrast in this situation. But before we get into the findings of the first case, I want to quickly review a few key CT elements that we utilize to help us out. We do ask our friendly CT techs to send us three plane reformats. We do get 3D volume renderings of the skull, and when possible, we obtain dual energy scans for bone-subtracted images. Okay, back to case number one. I will acknowledge that the findings in this case are not difficult to see, but this will be a good case to highlight some important points about how we interpret cases of unwitnessed trauma. Now, there are many approaches to beginning your search in these cases. I personally like to start from the outside and work in, looking at soft tissues, then bones, then intracranial contents. And I've gone ahead and skipped the soft tissues to go straight to our 3D volume renderings, which do show comminuted bilateral parietal and occipital skull fractures extending to multiple cranial sutures, and at this point, we already suspect that the mechanism of fall reported here is inaccurate, since we would not expect this level of comminution and extent of fractures in a minor fall. Moving forward, axial CT shows a large right cerebral convexity subdural hematoma causing leftward midline shift. The subdural hematoma in this case is mixed density. Now, while mixed density subdural hematomas are more commonly seen in abuse than in accidental trauma, we need to avoid using the term acute on chronic subdural hematoma in this setting, simply because there are too many confounding factors which affect the density of subdurals on CT. For example, there's always a possibility that CSF could mix into the subdural space causing a mixed density, and this can occur from a single traumatic episode. Now, this can get into a very lengthy discussion, which we won't have time for in this talk, unfortunately. But, I have included several references here that all recommend to avoid aging subdurals on initial head imaging. My colleagues in IHO do not age subdural collections on initial CT or MRI, unless there are clearly septations or membranes within the collection which are felt to be chronic in nature. This image also shows loss of gray-white matter differentiation in the right cerebral hemisphere compatible with parenchymal injury, and when I'm reading this at the workstation, my mind jumps to the possibility of this being hypoxic ischemic injury to the brain, and hence raises a suspicion for abuse in this case. We've accomplished a lot of things with this CT. We've determined the need for emergent craniectomy to save this child's life, and we have shown that the injuries sustained are out of proportion to the reported mechanism, suggesting the high probability of this mechanism of injury being due to abusive head trauma. With the findings presented and given clinical history, this child received a Child Abuse Protection Team consultation for further investigation and management. This is a good time for me to review our MRI protocol for NAT evaluation. These scans are ordered typically after the Child Abuse Protection Team has seen the patient and deemed that NAT is a reasonable possibility given the clinical history, physical exam, and imaging findings. Of course, the patient must be stable enough for a 45-minute long scan. In that 45 minutes of scan time, we obtained diffusion-weighted imaging, T1, T2, flare, susceptibility-weighted imaging, and a non-contrast MRV of the head. The MRV of the head is done to rule out venous sinus thrombosis, which is occasionally claimed by defense attorneys as a cause for findings in abuse cases. We also perform a screening MRI of the spine, which consists of sagittal T2 fat-sided images of the entire spine, usually divided into two fields of view. The images are then finally checked by the covering pediatric neuroradiologist for quality prior to letting the patient leave the scanner. Ideally, the MRI would be performed at 48 to 72 hours after the traumatic injury to maximize visualization of the diffusion changes, as this helps us determine prognosis. However, this is often impossible due to the critical condition of many of these patients. Now we can get back to case one. And this patient did require emergent hemicraniectomy and subdural hematoma drainage, so they were not stable for MRI until day 14. It is crucial that we keep in mind the timing of the MRI when we interpret the findings because of how the findings can evolve over a relatively short period of time. The diffusion-weighted imaging in this case shows restricted diffusion involving portions of the cerebral cortex, right greater than left, and deep nuclei on the right. Despite there being these changes on DWI at this point in time, we should consider the possibility that much of the hypoxic ischemic injuries in this case have normalized or pseudonormalized, as would be expected after 14 days. It is often helpful to perform a follow-up MRI a few months later to evaluate for parenchymal brain injuries, which may have been missed due to suboptimal timing of the MRI. Axial flare in T2 images show cavitations within the bilateral frontal lobes, which are suspicious for cerebral lacerations. These injuries are fairly specific for abuse and certainly add to our suspicion for abuse as the mechanism in this case. Notice the asymmetric involvement of parenchymal injuries in this case, which is something that we commonly see in the setting of abuse for reasons which are not clearly understood. Susceptibility-weighted imaging shows bilateral retinal hemorrhages posteriorly. It is important to mention that our routine susceptibility-weighted imaging of the whole brain does lack sensitivity for retinal hemorrhages, and it's critical that these patients get an ophthalmologic exam as soon as possible. SWI and T2 images at the vertex show artifact due to surgical staples, which limit the evaluation for cortical vein injuries in this region. This is unfortunate due to the fact that cortical vein injuries in the setting of abuse often happen in the parasagittal regions along the sagittal sinus. In this case, no cortical vein injuries were identified. These are our non-contrasted time-of-flight MRV images of the head. Now some scanners do produce better images than others when it comes to these. It is important to remember that non-contrasted MRVs of the head are prone to artifact and can occasionally cause confusion as to whether there's a dural venous sinus thrombosis or not, and performing these with optimal technique is very helpful. In this particular case, you can see that there is diminished signal at the inferior portion of the straight sinus and in some portions of the superior sagittal sinus. When viewed on other images, such as the axial and coronal T2, the flow voids were felt to be preserved and this was thought to be within normal limits. Occasionally we do offer CT venogram or MR venogram with contrast to further evaluate for troubleshooting purposes. Sagittal T2 screening MRI of the spine shows posterior atlantoaxial and upper cervical intraspinous ligamentous injuries. These injuries would be much more difficult to see or impossible to see without fat saturation. Anterior ligamentous injuries happen to be much more common in the setting of abuse. However, in more severe cases of spinal trauma, you can also see anterior ligamentous injuries and possibly even spinal cord injury. To summarize case one, we had a six-month-old who reportedly fell two to three feet from a bed onto a hardwood floor. The clinical presentation and constellation of CT and MRI neuroimaging findings are not compatible with the reported mechanism of injury and are highly suspicious for abusive head trauma. Okay, let's jump to case number two. A five-week-old presents to the ED with a reported apneic episode at home and presents with hypothermia and hypotension on exam. There was no reported trauma. Axial CT shows right parietal scalp swelling. The 3D volume rendered image of the skull shows a right parietal skull fracture extending to the sagittal suture. The cranial sutures also appear to be splayed open, a finding that can be seen in the setting of increased intracranial pressure. Again, in this case, we are already seeing evidence of traumatic injury despite no reported trauma, which is immediately suspicious and raises red flags for the possibility of abuse. This then helps us direct our search pattern to look for other potential sequelae of abuse. Coronal reformatted bone-subtracted image shows subdural hemorrhage at the left cerebellar tentatorium. Axial CT images show scattered loss of gray-white matter differentiation throughout the bilateral cerebral hemispheres. These findings have significantly increased our suspicion for abuse, and we notify the ED physician, triggering a CAP team consult. The MRI gets done six days later after the patient is stabilized. At day six, we see extensive diffusion restriction throughout the bilateral right greater than left cerebral hemispheres involving white matter and cortex. The pattern of diffusion restriction is compatible with diffuse profound hypoxic ischemic brain injury. At this point, without even seeing other images, we know that the prognosis is poor. Axial flare in T2, so a typical appearance of cerebral laceration in the right temporal lobe with a small volume of layering hemorrhage. Susceptibility-weighted imaging near the vertex shows bilateral parasagittal cortical vein irregularities compatible with bridging vein injuries. Interestingly, our susceptibility-weighted imaging did not clearly show retinal hemorrhages in this case. However, retinal hemorrhages were confirmed by ophthalmology on their subsequent exam. So this highlights an example of when our susceptibility-weighted imaging is limited in the evaluation of orbital injuries. Spanning spine MRI shows T2 hyper-intense signal at multiple cervical interspinous spaces compatible with ligamentous injuries. There can actually be a normal T2 hyper-intense signal from the venous plexus anteriorly within this space, but this signal abnormality was felt to be more than what we would typically see from the venous plexus. Lumbar spine imaging shows subdural hemorrhage layering in the lower thecal sac, which may be tracking from above or which may be due to primary spinal insult. To quickly summarize case two, we have seen CT and MRI findings that are compatible with traumatic brain and spine injury, especially in the absence of reported trauma. At this point, the CAP team determined that given the patient's reported history, physical exam findings, and constellation of head, spine, and body imaging findings, non-accidental trauma was the most likely etiology. So we've now reviewed two cases of abusive head trauma and shown how our imaging technique allows us to clearly see the findings we need to see to assist our colleagues in optimizing patient care. It's important that we as radiologists are armed with the proper knowledge to be able to suspect abuse based on a mismatch between the imaging findings and reported history and to know what additional findings we need to look out for that support the conclusion of abuse. I would like to end my talk by adding some context to all of this and to show the gravity of abuse in the U.S. So traumatic brain injury is the leading cause of death and lifelong disability in children affecting more than half a million children per year in the U.S. alone. Now the majority of these are due to falls or accidental trauma. However, abuse is the leading cause of traumatic brain injury related death in children less than four years old. Overall, abusive head trauma has a very high mortality rate, estimated at 15 to 25 percent. And 53 percent, over half of fatal traumatic brain injury in children under the age of two are due to abuse. And this seems to peak in percentage at the ages of one to two months. With that, I conclude. Thank you for your time. So I'm going to be speaking on the imaging approach to the child with stroke-like symptoms. So I'm going to start off with the common scenario, child arrives to the ED with a new focal neurologic deficit. The differential for this presentation is going to be a long one. In it, it includes stroke mimics and also the diagnosis of possible acute ischemic stroke. So our imaging has to be focused to the diagnosis of potential acute ischemic stroke. So you're on call and you're probably going to get a call from your ED doctor who's just going to order a STAT imaging study. Not sure which one, but you're going to get that call. So when you're there as a neuroradiologist, what are you going to do? Are you just going to do whatever the ED physician orders? Are you going to do head CT? Are you going to do it with or without contrast? Are you going to do an MRI? Lots of things to think about. The first thing you should do is ask for a neurology consult because your neurologist is going to help you weed out the cases that truly need that emergent imaging for the stroke protocol to be activated. They can confirm focal neurologic deficit is present and that emergent neuroimaging is indicated. If we're going to proceed with our emergent imaging, I'm sure all of you at your institutions have something similar. This is what we have at Lurie Children's. It's our stroke protocol and it just delineates who's gonna get called, how that alert's gonna go to neurosurgery, neurology, how that patient's gonna move from the ER through medical imaging. MRI is the preferred modality in pediatric stroke and this is different from adults where the first modality is going to be CT. And there's several reasons for that. First, diffusion imaging, we all know, is just exquisitely sensitive to ischemia. There's no radiation involved, which is a great thing when we're imaging children. And also keeping in mind that non-enhanced CT, if you're just gonna go with that, that you will miss a large percentage of acute stroke. So all of us should have some sort of a rapid brain MRI protocol. This is what we use at Lurie. And we always start off with our diffusion imaging as a first sequence and then follow that with susceptibility-weighted imaging because that's gonna be helpful for us to differentiate stroke mimickers. After that, we can, obviously we're gonna do T2-weighted imaging fast. This does not need to be super high resolution. We can go faster if we have thicker slices. And then we've incorporated ASL into our standard rapid imaging protocol. And so we'll do that as well. It's not a beautiful image, as you can see, but it can be really helpful, give us good information. And then if we're suspecting that we're actually dealing with a stroke, we wanna finish that off with a time-of-flight MRA. And we wanna keep our imaging down to less than 10 minutes because if it is an acute stroke, we want to move quickly on that. So I'll start with the first case study. A child arrives to the ED with a new focal neurologic deficit. We're gonna start off with our diffusion imaging. And in this case, we can see that we have no evidence of diffusion restriction so we're not dealing with an acute stroke here. The second sequence is gonna be our SWI. We can see that it's asymmetric. The veins are more prominent on one side of that hemisphere. If you're doing ASL, it's really nice because it can confirm that asymmetry. And we can see here that the ASL is asymmetric as well. Now it can be hard, which side is the abnormal? You know, it can be challenging to figure out. That's where you're gonna wanna talk to your referring, to your ordering provider and find out which side the neurologic deficit is on. So we're dealing with right-sided weakness. We're gonna be looking to the left cerebral hemisphere. And that means we can see that we have this increased conspicuity of cortical veins. That's due to the increased deoxyhemoglobin that's present causing increased susceptibility due to the increased O2 extraction from that hypoperfused hemisphere that we can see on the ASL. And then you can not only exclude the diagnosis of acute stroke, but you can make the diagnosis of hemiplegic migraine, in fact, in the aura phase where we have hypoperfusion. So you can give all that information to your ordering provider. What about this case here? This is a different case. Starting off with diffusion imaging, it's negative. But look at the SWI, asymmetric. And the ASL is asymmetric, but which side is abnormal? So we're gonna wanna talk to our ordering provider. This child has right-sided weakness. So we're gonna look to the left cerebral hemisphere. We can see decreased conspicuity of cortical veins on that side and increased perfusion on the ASL. So we're dealing with something different here. Again, hemiplegic migraine typically in the headache phase is when we're gonna get that hyperperfusion phenomenon. So we can give all this information. And hemiplegic migraine is the most common stroke mimic in children. It can account for up to a third, 28% of children coming to the ED with stroke-like symptoms. It has a typical onset in childhood. It may or may not present with a headache. Typically, the child will come with hemiparesis or a hemifacial weakness, numbness. The episode is typically brief and resolves completely spontaneously. But we can still, on imaging, see perfusion abnormalities, and we can detect those up to 24 hours post-symptom onset. Okay, so next case. Again, same history. Child arrives at the ED with a new focal neurologic deficit. You're gonna start with your diffusion imaging. In here, you can clearly see we have an abnormality involving that left cerebral hemisphere. There's diffusion restriction, but it's not really what we expect to see in acute arterial ischemic stroke. It's involving that entire hemisphere. We can see that it's really involving that white matter. This is something that we see with seizures due to the cytokine injury that we can see with seizure activity. It's not conforming to a specific arterial territory. We're gonna follow up with our SWI, and we can see decreased conspicuity of those cortical veins in the left cerebral hemisphere. And on our perfusion imaging, which is right there, we can see the asymmetric increased perfusion, and that explains why we're not seeing that cortical vein prominence in that left cerebral hemisphere. We're gonna ask for some history. We're gonna see what side is the neurologic deficit on, and then always in these kids, we're gonna ask if there was a history of a seizure, a preceding seizure, or does the child have a history of seizures. Sometimes seizures will show up with diffusion restriction, as in this case, sometimes they don't. So we can see we have non-territorial diffusion restriction and asymmetry of the SWI in our perfusion imaging. Right-sided weakness. So this is a seizure in the periictal and ictal phase. So it really sort of follows what we expect to see on scintigraphy. So during a seizure in the ictal phase, we're gonna have hyperperfusion, and postictally, we're gonna have hypoperfusion, and this is a seizure. So seizure can also be a stroke mimic in children. It can present as a symptom of stroke in up to 25% of children. So strokes in and of themselves can present as a seizure. Seizures can also be non-convulsive in children. So just because there's not clinical signs of a seizure doesn't mean that there's not EEG evidence of seizure activity. And then Todd paralysis refers to a postictal, temporary hemiplegia secondary to cerebral perfusion abnormality. So let's go to the third case. Again, child derives the ED with a new focal neurologic deficit. In this case, we can clearly see there's abnormal diffusion restriction within the basal ganglia on the left, and we can immediately make the diagnosis of an acute ischemic stroke. But are we going to stop there? We shouldn't. We should go to SWI, because it can give us so much more information. As Manu showed us, it can actually show us where a thrombus lies, but can also show us what area of the brain is hypoperfused. It can show us the ischemic penumbra without us even having to give gadolinium. So you can see the prominence of these cortical veins in that MCA territory, and we have this mismatch. Well, here's the ASL, and we can see the decreased perfusion corresponding to the area where those veins are prominent. And what SWI and ASL is showing us here is brain at risk. So the DWI-SWI mismatch is sort of a really good, convenient way of seeing the brain at risk, and it can guide us in further management. Then we're going to want to do our T2-weighted imaging, in this case, the flare. And flare is going to be really helpful in combination with diffusion imaging to give us a timeframe of when that ischemic insult occurred because it can be, sometimes we don't know exactly when it occurred. So if we have a positive diffusion scan and a negative flare, then we know that we're in the timeframe of about 30 minutes to four and a half hours if the diffusion is positive and the flare is positive, then we know that we're in about a six to 12 hour timeframe. And if our flare is showing edema, as in this case here, then we're at about 24 to 48 hours. Now with the new DAWN and Diffuse Trials in the adults, the treatment window for pediatric stroke is really extended so we can treat stroke in children even up to 24 hours, which is very different from before. I think we had like a six hour window so a lot of these kids were excluded from potential therapy. And then we know that we're dealing with a stroke, we're gonna wanna finish our evaluation with an MRA and we can clearly see here that we have that abrupt termination of the middle cerebral artery. So case four, child arrives with the ED, again, focal neurologic deficit, but it's after hours and MRI is not available. So what are you gonna do? Are you gonna call your MRI tech in or are you gonna do a CT CTA? We're dealing with a child, pediatric imaging, everybody's concerned about radiation, but that's not the time for us to be concerned about radiation. If there's a stroke, we wanna be able to diagnose that and get potential management and treatment in place as soon as possible. So here's that child, we're gonna start with a non-enhanced CT. It's not really very revealing, but we always wanna do CTA in these cases. We don't just wanna stop at a non-enhanced CT. And the CTA showing us that abrupt cutoff of the middle cerebral artery. And then if we go back and we look at the CT again, you can see that very subtle hypoattenuation in the basal ganglia that you might miss on the first pass. So this is an acute ischemic stroke diagnosed with CT and CTA you can call your neurointerventionalist. Sometimes they'll require that we do a perfusion study. A lot of times they don't and they'll just intervene without having to do that. So this is the pre-thrombectomy angio with the stent retriever in place. Clot was removed. Good perfusion was reestablished. This is four hours from the time of event. So this is great. And a follow-up CT a month later really shows us that their brain was nicely preserved. So our imaging in acute stroke is gonna go down two pathways, an MRI pathway. If MRI is available and the patient can have an MRI, that's the way that you wanna go. You do a rapid brain MRI, start with your diffusion, and do an MRI if you're dealing with a stroke. If MRI is not available or the patient cannot have an MRI, maybe they have a ventricular assist device, you're gonna wanna go CT and CTA. So pediatric acute ischemic stroke, its incidence is about one to two per 100,000. If we add in hemorrhagic stroke in children, that incidence really approaches the incidence of pediatric brain tumors. So it's not something that is that uncommon. It's one of the top 10 causes of death in children, and it has a long list of etiologies, which is what makes it difficult for us to diagnose. But the three big things to consider in the setting of pediatric stroke are focal arteriopathy of childhood, cardiac disease, and sickle cell disease. 60% of the children will have permanent disability, and this is why we want to make that diagnosis of stroke as rapidly as we can to offer them brain-saving treatment. The recurrence rate can be quite high, 12% in the first year and as I said, treatment options are now available for children even if the symptoms have been present for more than 24 hours. So some of the take-home points is if stroke is in the differential and you're in the window of treatment, time is gonna be of the essence. Get a neurology consult, have a rapid brain MRI sequence available. Always start with diffusion imaging followed by susceptibility-weighted imaging, and go to CT and CTA if prompt MRI is not possible. Just know that the majority of the cases that you're gonna be doing are gonna be stroke mimics, but we need to image to the diagnosis of acute ischemic stroke. So I do have a question here. So which sequence is the most helpful in making the diagnosis of hemiplegic migraine? Is it diffusion-weighted imaging, susceptibility-weighted imaging, gradient echo imaging, or MRA? The correct answer is susceptibility-weighted imaging. Thank you so much for your attention. All right, good morning, thank you. So I'll be talking about challenging cases in the pediatric orbit today. So I'll be focusing on two cases today, one involving an orbital mass and another with optic nerve swelling. We'll go through the differential diagnoses for these entities in children and we'll walk through the imaging workup. Our first case was a six-year-old boy who presented with right eye swelling that was noted by his family for the past week. The initial clinical suspicion was for some sort of vascular lesion. He arrived at our institution with this outside CT that showed this well-circumscribed mass-like lesion in the right superior orbit that was causing proptosis and hypoglobus. And this actually came to us billed as a hemangioma. Before we go any further with this case, I want to briefly review the differential diagnosis for orbital masses in children. This list is specifically extraocular, non-osseous orbital masses. So it includes true vascular malformations like lymphatic or venous malformations. Hemangiomas, although they're vascular in nature, are technically benign vascular neoplasms. Other neoplasms include rhabdomyosarcoma, teratoma, peripheral nerve tumors like neurofibromas and schwannomas, and tumors of the optic nerve and nerve sheath, including gliomas and meningiomas. Other lesions that can appear mass-like in the pediatric orbit include hematomas, abscesses, congenital inclusion cysts, and encephaloceles. So as neuroradiologists, of course, the next step we would suggest is an MRI to better characterize this mass. One question we're commonly asked is whether the exam needs to be a brain in orbits or if just orbits is sufficient. So our usual recommendation, especially for the initial MRI, is to do both brain in orbits to better evaluate for intracranial extension of the orbital process and also to look for associated intracranial lesions or malformations that might help guide the differential diagnosis. This is especially relevant if the child's sedated since we don't want to have to re-sedate because of an incomplete workup. But in select cases where we're confident that the abnormality is isolated to the orbit or if the intracranial findings have been previously characterized, scanning just the orbits may be appropriate. So this is our standard protocol at our institution for an MRI brain in orbits with and without contrast. It's a fairly comprehensive protocol that lets us evaluate for the vast majority of orbital pathologies that we commonly encounter. It includes most of the standard imaging of the brain, diffusion, T1, T2, flare, susceptibility-weighted images. In the orbits, a coronal fluid-sensitive fat-saturated sequence, like a STIR, is usually helpful. On our Philips scanners, we run coronal T2M Dixons. We also run post-contrast T1 fat-saturated images. I personally like a thin section, heavily T2-weighted sequence, like a T2 Drive, Kiss, or Fiesta, depending on your vendor. I like these for looking at the optic nerve, optic disc, other cranial nerves. Sometimes they also come in handy for subtle retinal hemorrhages or infiltrative processes. I'll say, though, that there's a little bit of difference of opinion in our group on the T2 Drive, primarily because it's a fairly motion-sensitive sequence. In the setting of a suspected vascular malformation, we often add dynamic post-contrast, time-resolved MR angiography, which can help characterize the phase of enhancement of the lesion and can also evaluate for arterial venous shunting. This is just a normal example. On the sagittal images, we're seeing a little bit of physiologic enhancement of the lacrimal glands and nasal mucosa in the venous phase. So before we go back to our initial case, I'll show some examples of entities in the differential diagnosis. This was a two-month-old girl who presented with proptosis of the left globe. You can see this very T2 hyper-intense, avidly and uniformly enhancing mass, insinuating around the left optic nerve and extraocular muscles, but with well-circumscribed margins. On the dynamic imaging, you can see early arterial phase enhancement and subsequent partial washout. These findings are typical of an infantile hemangioma. I always find the terminology here a little confusing. I remember being taught that orbital hemangiomas in kids are capillary hemangiomas, and those in adults are usually cavernous hemangiomas. And while the term capillary hemangioma is still often used in children, most of these are technically infantile hemangiomas. And keep in mind, there's an association with phase syndrome, so be on the lookout for associated findings like posterior fossa malformations and arterial abnormalities. This is a companion case that in many ways looks very similar to the last case, but note the dark signal on the T2-weighted images and the less avid degree of enhancement. This was an optic nerve sheath meningioma in a child with NF2. Here's an example of a lymphatic malformation in a toddler with left eyelid swelling. In this case, showing intralesional hemorrhage, and this is a common reason for lymphatic malformations to present acutely. Note the multi-cystic appearance with only faint enhancement of the septae and cyst walls, but no internal enhancement. This was a newborn imaged on the first day of life for left proptosis and concern for orbital mass. There's a large cystic component within this mass, but note the solid enhancing components. This turned out to be a teratoma. Fat-saturated imaging to look for intralesional fat, T2 star-weighted imaging to look for calcifications, and vascular imaging can all be helpful to characterize these masses. This next case was a 19-year-old male with sickle cell disease who came in with left eye proptosis and diplopia after an episode of severe orbital pain. An MRI was ordered for further evaluation, which showed this large subperiosteal collection in the supralateral orbit. Note the intrinsic T1 shortening that identifies this as a hematoma. Orbital findings like these in sickle cell disease are uncommon, but they do occur. The subperiosteal hematomas are a result of vaso-occlusion and orbital wall infarction. Epidural hematomas can actually also rarely occur by the same mechanism, and similarly, lacrimal gland swelling can occur in sickle cell, likely also from vaso-occlusion. There are also many other findings in the head and neck associated with sickle cell disease, whether from occlusion, anemia, or infection, that you can see here. This next patient is a child who suffered an orbital roof fracture and a motor vehicle collision. You can see that on the CT on the lower right. He came back in 10 months later with hypoglobus on the right, and this follow-up MRI at that time showed a growing fracture with a new encephalocele that was causing the eye finding. Now, with all that in mind, we'll come back to our original case. This is the outside CT, just to refresh your memory. And this is the MRI that was done two days later. You can see that this mass is well circumscribed, has areas of restricted diffusion, and low T2 signal that correspond to increased cellularity, and it enhances avidly, but less so than the hemangioma we just saw. These are all imaging features that fit best with rhabdomyosarcoma in a patient of this age. So, rhabdos are the most common extraocular orbital malignancy in childhood. They can affect patients of any age group, but they have a peak age presentation of six to eight years old. They're essentially always unilateral and usually extracronal, but they can extend into the intracronal space. The embryonal subtype is by far the most common type encountered in the orbit, and it seems to have a predilection for the superior orbit. They tend to be well circumscribed with intermediate to low T2 signal and restricted diffusion on MRI. All right, let's switch gears now. This is our next case. This is a 13-year-old boy who presented with blurred vision. It was acute in onset and painful. So, optic nerve swelling was noted on fundoscopic exam, and the primary clinical concern at that time was for optic neuritis. So, the imaging study that we suggested was an MRI of the brain in orbits with and without contrast to look for optic nerve single-laterality and enhancement, and also to look for intracranial lesions that could help guide the differential diagnosis. So, I'll start with these post-contrast coronal T1-weighted fat-saturated images from his MRI. These show that the optic nerves are enlarged and enhancing. Note the bilateral long-segment involvement, including involvement of the posterior optic nerves that, in this case, extended all the way back to optic chiasm. So, before we go further in this case, I want to go through the differential diagnosis of optic nerve swelling in a child and go over some of the associated terminology. So, I like to think of optic nerve swelling as fitting into one of three categories. You can have nerve edema without increased intracranial pressure, papilledema, which specifically refers to disc swelling that's caused by increased intracranial pressure, and pseudopeplidema, which refers to elevation of the optic disc without edema. So, let's start with papilledema. Often, you'll see increased CSF around the optic nerves with extension of the nerve sheaths and flattening of the posterior sclera. And this can occur in a variety of settings, including obstructive hydrocephalus, for example, from a posterior fossa mass, other space-occupying lesions or hemorrhage, like a subdural hematoma or cerebral edema. You can have a cranial vault that's too small to accommodate the growing brain, as you can see in craniosynostosis. You can see overproduction of CSF, like you get with a choroid plexus tumor, or impaired resorption of CSF, as you might see with meningitis. And of course, there's also idiopathic intracranial hypertension. I have a couple examples of papilledema on MRI. This is an eight-year-old girl who came in with headaches and abnormal gait. And you can see a tectal glioma that was obstructing the cerebral aqueduct. She had marked enlargement of the lateral and third ventricles, and distention of the optic nerve sheaths with elevation of the optic discs. And in this 10-year-old boy with a one-month history of headache and vomiting, there's a medulloblastoma filling the fourth ventricle, resulting in obstructive hydrocephalus and the associated findings of papilledema. Pseudopapilledema, again, refers to abnormal elevation of the optic disc without edema. A common example is optic disc drusen, which are deposits of proteinaceous material in the optic disc. There are also some physiologic variants of the optic disc that can cause it to look prominent on imaging. A small percentage of people have just tilted discs. You can have atypical myelination of nerve fibers that extend through the optic disc and into the surrounding retina, or a small disc that results in crowding of the nerve fibers and consequently an elevated appearance of the disc. There are also a variety of systemic and syndromic conditions that are associated with this appearance. This is an example of calcified drusen, although we should note that drusen are much less likely to be calcified in kids compared to adults. These are fundus photos and fundus autofluorescence from a patient with drusen. The classic ophthalmology descriptor here is a lumpy, bumpy appearance of the optic disc. And clinically, this diagnosis is pretty straightforward in that drusen are superficial and calcified, but when they're deeper or non-calcified, drusen can mimic nerve swelling. And this is just a case of drusen on MRI. You can see slight elevation of the optic disc, but there's no posterior scleral flattening or nerve sheath extension like you often see with papilledema. These are fundus photos from a patient with myelinated retinal nerve fiber layer, worse on the left than the right. Normally, myelin extends along the optic nerve but stops at the optic disc. In some patients, though, myelin extends through the disc and into the adjacent retina, which can occasionally mimic nerve edema. And on MRI, this can sometimes show up as just a prominent or elevated optic disc. This finding can be associated with nearsightedness, can be unilateral or bilateral, and rarely is associated with syndromes like NF1 and trisomy 21. This patient had NF1. That leaves us now with true optic nerve edema that's not caused by increased intracranial pressure. And these are the entities where I think the radiologist can be most helpful. These include the various causes of optic neuritis, optic nerve masses, usually glioma, as well as other insults, including ischemia, hypertensive emergency, and various toxic and metabolic conditions. These are fundus photos from a five-year-old who described a few days of sparkles in her right eye. She also had decreased vision and proptosis on the right, and there was marked nerve head edema. Her MRI showed a large expansile enhancing mass of the right optic nerve consistent with glioma. So optic pathway gliomas can occur sporadically or in the setting of NF1. This patient did not have any clinical or imaging features of NF1. When gliomas do occur in NF1, they tend to be more indolent and progress more slowly. Now I'll focus on optic neuritis. Clinically, this presents as acute painful vision loss. Patients will often complain of blurred or foggy vision and may have decreased color vision, particularly for the color red. Listed here are three of the major players. Both NMO and anti-MOG disease are commonly bilateral and have long segment lesions, but NMO more often involves the posterior nerves and the chiasm, whereas anti-MOG is typically anterior with less chiasmatic involvement. MS, on the other hand, tends to be unilateral and short segment with sparing of the optic chiasm. There's certainly a lot more we could talk about with each of these and quite a bit of overlap in terms of imaging findings, but I think this provides a good framework for thinking about these disorders. So let's return to our original case, this teenager with blurred vision. We have imaging findings that fit with the clinical suspicion of optic neuritis. It's bilateral, long segment, and with involvement of the posterior nerves and although not shown here, the chiasm. So thinking back to that last slide, these features would fit best with an NMO spectrum disorder. Intracranially, there were also T2 hyper intense lesions in the hypothalamus and thalami in a distribution that's typical of NMO. He also had spine imaging that showed this cord lesion spanning more than three vertebral segments. So this was our suggested diagnosis. He ultimately tested positive for anti-aquaporin-4 antibodies. I'll finish with a few take-home points. Remember that not all prominent optic discs represent papilledema. There's broad differential diagnosis for both orbital masses and optic nerve swelling and the clinical history and exam findings are critical in helping distinguish the different entities. And finally, intracranial findings are often helpful when sorting through these differentials. So I have a couple audience response questions here. The first one is characteristic imaging findings of rhabdomyosarcoma include marked T2 hyper intensity and infiltrated appearance, cystic and solid components, areas of restricted diffusion, and extensive calcifications. So the correct answer was C, areas of restricted diffusion. Because of the high cellularity of these tumors, they're small round blue cell tumors. They tend to have low T2 signal. They are also relatively well circumscribed, so A would fit better with something like lymphatic malformation. When they get very large, you can see areas of necrosis, but they're typically solid tumors in the pediatric orbit and they're not typically calcified. Our next question, optic pathway lesions and neuromyelitis optica spectrum disorders are most often unilateral short segment and involve the anterior optic nerves, bilateral long segment and involve the anterior optic nerves, unilateral long segment and involve the posterior optic nerves and chiasm, or bilateral long segment and involve the posterior optic nerves and chiasm. So the correct answer was D, bilateral long segment involved the posterior optic nerves and chiasm. A is more typical of multiple sclerosis. B is what you see with anti-MOG disease. C can occur in NMO or anti-MOG, but it's not the typical or classic findings for either, so correct answer is D. So thank you again, thank you for your attention and thanks for letting me speak to you today.

pediatric neuroradiology

neonatal brain MRI

hypoxic-ischemic injury

non-accidental trauma

optic nerve disorders

pediatric trauma

diffusion-weighted imaging

retinal hemorrhages

optic neuritis

rhabdomyosarcoma