So in the next few minutes, I think Bryce would be a hard act to follow, and looks like this is a very difficult group of intensivists to stump, but I'll try to share my perspective with you on how brain ultrasound can help simplify diagnosis. Do we have a, how do I advance slides? Thank you. So this is me, obviously. A few disclosures. The only one relevant for this talk would be some cranial ultrasound applications. I do have a intramurally funded study supported by one of the ultrasound vendors as a loan for the devices, which I'll introduce in the course of this talk. So the big objective of this talk is to share my perspectives with you, how brain ultrasound in particular can change management in a few cases, and provide you a little extra information about the cerebral perfusion, or blood flow to the brain, which is mostly the target of most systemic cardiopulmonary support. The two things I'm going to focus on is cases that show you where traditional neuroimaging was inaccessible, not feasible, or there were other priorities with patients' systemic resuscitation that made it difficult for a patient to be prioritized for CT scan or MRI, mostly screening unstable patients for neuroimaging, and prioritizing ICP lowering therapies for patients going for a head CT. In no way am I going to even try and make a case that ultrasound will replace CT scan, but here is the patient that cannot be moved, and has other priorities like hemodynamic resuscitation that can be managed with an ultrasound while awaiting the CT. And the second series of cases I will try to give you a perspective on how actually looking at the cerebral perfusion using a non-invasive modality like transcranial ultrasound can help you figure out patients' pathophysiology, rather than arbitrarily calculating it using an ICP monitor and a systemic map. So how many of you at this point in time either use or do yourself transcranial ultrasound or Doppler's in the ICU? So just a minority. So for the others, your skull is obviously the biggest impediment in taking ultrasound images of the brain. In most of the patients, 80 to 90%, at least a small part of the brain, temporal bone on the side of your head has skull that is thin enough to allow insonation of the brain. We call this temporal windows to the brain. 10 to 20% of the patients may not have a temporal window, but a significant proportion of these may have a window on one side, not the other side. And that is the biggest challenge about widespread scalability of transcranial ultrasound. We typically use the same probe that you use for echocardiograms. It's a low frequency probe and we try to take an image at the level of the basal cistern where you're seeing the opposite skull. This is the midbrain, the top of the brainstem, and this is the sphenoid wing outlining the middle cranial fossa. The circulophilus PCA will be around here and I'll show you some pictures down the cases. So anywhere when you see pictures of transcranial ultrasound in the following cases, this would be anterior of the patient, this would be posterior of the patient, and this would be the ipsilateral temple that is being insonated, and this would be the opposite skull. So with that orientation, the one place where I'm personally investigating the use of ultrasound is to look for intracranial pathology that requires emergent management in austere environments and resource-limited environments. Obviously, the current data validation that I'm doing is in our own ICU to look for intracranial pathology and we have a topography article that has been sidelined. So you can see the pointer is pointing towards the patient's ICH. This is the patient's opposite skull. You can see the fox cerebri here. This hyperechoic signal is the ICH and you can see the midbrain here. Only supratentorial ICHs that are more than a centimeter large and hyperacute, so after three to five days, the signal of ICH becomes oblivious to the ultrasound scanning. So fresh, big bleeds can be scanned with an ultrasound. Again, in no way we are propagating that this can replace CT scan, but in a place where having such a diagnosis can help you screen a patient, this is being investigated as a modality. This is another patient where you can compare it to the CT scan. You have opposite skull, the fox cerebri, midbrain is somewhere here, and you can see on the transcranial preset and the abdominal preset what an ICH looks like. The challenges of current application of this modality to your ICU is limited by the fact that we don't have a preset of ultrasound that is optimized for B-mode imaging. The transcranial Doppler is optimized for Doppler, so the next best is abdominal, and as a part of my research project, I'm working with the ultrasound vendors to create that preset. This ongoing study is also looking for field applicability of cranial ultrasound for pre-hospital ICH diagnosis, and hopefully the results will be available soon for you guys to pre-use. The one place you can use it is patients that are too unstable for getting a head CT, or the head CT might be perilous in that particular circumstance. One example is a patient on advanced hemodynamic support like VA ECMO, HFO, or CRRT on multiple pressors, where sometimes we do make an effort to get a head CT, but this might be difficult both for the patient as well as the staff transporting the patient. So a common surrogate we use for cerebral edema when we know there is a focal lesion is midline shift. You can use any of the midline structures, the fox cerebri, septum, pellucidum, third ventricle, depending on the pathology of the patient to assess midline shift. The correlations of ultrasound measured midline shift with the CT scan are pretty high, with hours of more than .85, .95, depending on the study you read. The caveat there is that whatever pathology you consider, so in this particular case, we're using the posterior part of fox to measure it. The technique needs to be consistent across serial measurements, and the patient's pathology needs to be relevant to the midline structure you picked up. So if it was a deep temporal structure, you should not be looking at the fox, because fall scene midline shift is not gonna happen in that patient until very terminal ICP increase happens. The second thing is the serial measurements help more than the absolute measurements themselves, because no matter how standardized your technique is, to get a perfect horizontal plane like a CT scan is almost impossible when you're scanning the ultrasound through a limited temporal window. So these are two different renditions of midline shift assessments, where serial assessments can help follow. We also have patients that may not have access to their pupillary anatomy because of orbital edema, orbital injury, or glabellar edema, post-craniotomy and trauma patients, and you all understand the importance of pupillary reactivity in cerebral edema assessment. So this is an ultrasound version of checking for papilledema. In the video where you can look at pupillary dilatation with both direct and consensual reflex with the light shown on the closed eye. You don't have to open the eye, and this will help you get pupillary reflex assessment when the patient's eye cannot be opened. For anybody who believes in fundoscopy, this is your ocular ultrasound showing you the ultrasound version of papilledema, where this is your optic nerve, optic nerve sheath, and the lamina cribrosa is bulging into the posterior chamber. The same significance as you would account for papilledema on fundoscopy would be the application of this. So this is where the anatomy of the brain can help you figure out patients' cerebral pathophysiology. The second thing I wanna focus on is the actual assessment of the patient's cerebral perfusion rather than measuring it. And in this particular case, we're going to be using Doppler. This is a color Doppler rendition of your basal cistern level of the skull. So this is opposite skull, this is your midbrain, and this is your circle of Willis with posterior cerebral artery around the midbrain, middle cerebral artery going towards the probe, and ACA going in the front. We're using pulse Doppler everywhere, and putting the pulse Doppler on the interesting, whichever segment of the circle of Willis we're interested in, and we're looking at the waveforms. There are five vital organs in the body that have low resistance circulation. So the classical cerebral circulation intracranially, well, this is one cardiac cycle, systole and diastole. You have robust flow during the diastole, which is a third to half of the peak systolic flow. That is very classical of renal, hepatic, coronary, and brain circulation. These patterns qualitatively change when the resistance of the vascular bed changes. So systemic circulation, mesenteric artery, peripheral arteries, femoral, brachial, they are high resistance circuits, and if you insulate these by pulse Doppler next time you cannulate somebody for an arterial line, their waveforms look more like this. There is decent systole, but the diastole is pretty minimal, sometimes zero, and this is a high resistance circulation. In cerebral pathophysiology, when the ICP increases, you basically create a peripheral high resistance vascular bed, so the waveforms change from low resistance to high resistance. And the grades of resistance can also be reflected both qualitatively and quantitatively in this waveforms. So this is a low resistance waveform, this is an extremely high resistance waveforms, and this is somewhere in the middle. People have validated these transitions in patients with progressive cerebral edema, progressing to cerebral circulatory arrest. So as your ICP increases, your normal low resistance waveforms with normal ICP increase to the point where the ICP is close to your diastolic pressure, so your diastole is getting obliterated. Eventually, your ICP exceeds your diastolic pressure, but not the systolic pressure, so any forward flow during the systole is being reversed back to the heart during the diastole. This oscillating pattern is also called no net forward flow, or oscillating flow to the brain, and typically would be non-sustainable for life unless it is reversed in the acute setting. This pattern ultimately will progress to only stump waveforms and no flow, or what we also known as cerebral circulatory arrest. So the progression of these patterns from normal to abnormal, and after institution of medical or surgical therapy, regression of these patterns can be a goal-directed marker for you to titrate your therapies. So say you have a patient that was found on the field, unknown downtime, cardiopulmonary arrest, ROS with two rounds of CPR, and sedated and paralyzed by the time the patient comes to you in the ICU, and while you're figuring out a patient's pathophysiology, the CT scan looks full, but a relatively young patient, you perform the transcranial dopplers, and their waveforms look like this. So this is middle cerebral artery velocities, and you have forward flow during systole and backwards flow during diastole. This by itself tells you that the patient's intracranial pressure is significantly higher than the patient's diastolic pressure. Now what you're gonna do next, obviously all of you guys know, compared to a patient who has normal waveforms, that can change your management. So this is a patient that you need to prioritize for ICP lowering therapies if they have underlying salvageable injury, or it will affect your prognostication in this particular patient. We use this corollary on the opposite side. So for VA ECMO patients, their waveforms are non-pulsatile, and in this particular case, if you look at the mean flow velocities, which for MCA should be about 40 to 80 centimeters per second, although the flow is non-pulsatile, on both sides, you see pretty decent flow. So you can safely say that at least this patient is perfusing, both MCAs normally. And a significant decrease in these flow on serial assessments can be a surrogate of progressive cerebral edema. This is another example where looking at cerebral perfusion rather than just measuring it or assuming it can be beneficial. This is a patient that was undergoing an elective T-bar therapy for ascending aortic aneurysm repair. Patient had an intra-op rupture and intra-op cardiac arrest. As you can see, the patient is on significant amount of hemodynamic support. Although the patient's maps have been restored to 60s or 70s, which in our case would be the target that we are looking for essentially to maintain cerebral perfusion, this patient is also on intra-aortic balloon pump counterpulsation. So you see two wave cardiac pulsations here. One is the patient's native cardiac pulsation and the second is the balloon pump. But either of these pulsations have oscillating flow. So this hemodynamic support, as advanced it is, is not maintaining the patient's perfusion to where it will sustain. Obviously, in a couple of hours, when we were able to do a head CT, the patient had massive global cerebral edema and eventually progressed to death. So these waveforms could really give you an index whether your systemic resuscitation is meeting its target or not. Obviously, it needs to be validated on an evidence-based level. This would be an example where the TCDs came to rescue. This is a four-year-old kid that had H-influenza pneumonia and was resuscitated on VV ECMO, suffered a small intracranial bleed as a result of the heparin anticoagulation, so had to be taken off anticoagulation. The bleed itself was not deemed to be life-threatening at that point, was maintained in osmotic therapy. Intracranial pressure monitor was placed to continue patient's neuromonitoring, but because of the heparin anticoagulation not being able to continue, the reason being the ICH, he had to be transitioned to high-frequency oscillator. So this patient is on high-frequency oscillator, and the oscillations are preventing the EVD waveform from being transduced enough to get a good ICP. So that created a significant concern for a paralyzed, sedated patient, and the team asked us to give a non-invasive surrogate for patient's ICP to continue guiding osmotic therapy, and we were able to get very good waveforms despite the high-frequency oscillator in all circulation, and use that as a non-invasive surrogate to continue escalating osmotic therapy. Once the patient was able to be weaned off HFO to a regular mode, the CT scan showed stable ICH, and this could have been a trigger for an earlier imaging or even escalating therapy. So the other way I've been using ultrasound, which I'm trying to validate on a prospective scale is does actually looking at the CPP change the way you manage a patient rather than calculating it? So I'll remind you the Mundro-Kelly hypothesis which all of us are familiar with. So we have three components in the brain. You have the brain component itself, the structural component, the CSF component, the fluid component, and the blood component. All three of these segments can contribute to intracranial pressure. Traditionally, when we talk about cerebral edema, we're talking about expansion of this brain component, either vasogenic or cytotoxic, global or focal, depending on the patient's pathology, and the assumption we make is that this global increase in cerebral edema is compromising the patient's cerebral perfusion, and we try to have ICP-lowering therapies, medical or surgical, to restore the cerebral perfusion to rest of the healthy brain. In this particular phenotype, the transcranial Doppler's is the one that I described to you where patterns have been described, that when cerebral edema compromises the patient's cerebral perfusion, you will see these perfusion-limiting waveforms where the ICP is affecting the CPP, and you have no net forward flow or the oscillating flow. Now, I will remind you that the other component, the blood component, can also contribute to intracranial pressure increase, and I'll share some cases with you. So in this particular case, if you were to actually perform the transcranial Doppler's, whether invasive or non-invasive, cerebral blood flow monitor, you will see normal or even hyperemic waveforms, which are diastolic component, would be more than half of the systolic component. So let me illustrate this difference in the phenotypes with a few examples. A 19-year-old kid with urea cycle disorder, non-compliant with his therapy, comes with hyperemia in 400s, with seizure-like movements. We do an EEG, start dialysis, nitrogen scavengers, the typical hyperemic treatment. A patient's imaging initially, this is a 19-year-old kid, has a pretty, what we call is full brain, and in an older brain, you could call it cerebral edema, but in this patient, obviously, it's difficult to ascertain, and the MRI shows pretty significant cytotoxic edema caused by hyperemia, and hyperemic encephalopathy by itself is not irreversible. So in this particular case, because the patient has no exam, GCS is three, and he's having active symptoms, and we are very concerned about cerebral edema related to the hyperemia, he gets an intracranial pressure monitor, and in this particular case, you can see the patient's ICPs have multiple crises through his hospital stay, but he's auto-regulated, so every time the ICP goes up, his blood pressure goes up. This goes on for about two weeks, to the point that patient's having very recurrent, multiple ICP crises through the day, up to 50 to 60 of ICPs, and we are unable to get any follow-up imaging. We are on maximal osmotic therapy, deep sedation, therapeutic coma for ICP management, but the ICP crisis continue. So we are extremely concerned that by this time, the patient's probably had an irreversible brain damage, but because on lying flat, his ICPs go to 90, and family still wants full care, we don't have a follow-up imaging. When I come on this patient's service, two weeks after this refractory ICP, I'm also convinced by the sign-out provided by my colleague that the patient's probably brain dead, we just need to prove it. So he goes into this refractory 90 ICP crisis, nothing, I have literally nothing more to work, there's no bucket-handled craniotomy that's gonna happen in this patient, but the other thing I notice is that patient is on two micrograms per minute of norepinephrine to maintain a CPP of 70, but the blood pressure fluctuations in response to ICP have stopped over the last three to five days. So the patient has stopped auto-regulating. I perform these patient's bedside TCDs with an intent that I'm going to see oscillating patterns, which I sometimes use to try to convince the family that this is a done case, but hold behold, I see absolutely normal waveforms. So this is a patient with an active ICP crisis of 90 to 100, which you can see on the flow sheet before, and the normal ICPs. So this changes kind of your, that this is not a structural toxicity being caused by the cerebral edema component. It is possible that the CPP of 70 that we are trying to maintain with LevoFET in a patient with a dis-auto-regulated brain is the one that's driving the ICP crisis, and this patient's 90 of ICPs is a manifestation of reperfusion or hyperemia, not a truly perfusion-limiting cerebral edema which we traditionally are used to thinking. So in this particular case, assuming that this is reperfusion or hyperemia driving patient's ICP, we actually back off patient's osmotic therapy and sedation, and seven days later, we extubate the patient successfully, and he still had cognitive deficits. By no way he was normal, but GCS of 15 PEG tube, and he has very prominent positive functional involvement with the rest of his family, and is still surviving and followed by our hyperamidemia clinic. This is not just something that I'm the only crazy one. Other people have noticed these phenomena, mostly with regard to seizures. So in TBI population, if you have managed global TBI, we often have this patient that has global TBI, no structural lesion, and the nurse will basically say the patient's ICP is high, but by the time I get mannitol, ICP becomes normal, and these pulsatile increases in ICP keep on happening, which resolve by themselves without the patient getting any osmotic therapy. Columbia Presbyterian has published this case where they did multimodality monitoring with invasive electrodes, and every time the patient is having an ICP crisis, they noticed that the patient's blood flow was high, and the patient was actually having seizures. So having subclinical seizures, nonclinical seizures, was driving the increased cerebral blood flow, and that cerebral blood flow increase in a non-autoregulated brain was driving these ICP crisis. And I've personally, we don't do invasive multimodality monitoring, but we personally now use transcranial Doppler to look at the CPP during these ICP crisis and try to distinguish between a perfusion-limiting ICP crisis that needs escalation of osmotic therapy or structural escalation versus a perfusion-driven ICP crisis, which actually needs you to go back on the CPP targets, and that is something that has been recognized and elucidated by the Brain Trauma Foundation guidelines, that in the autoregulated brain, it is okay to target a CPP of 70 to 90, but a non-autoregulated brain, it's actually CPPs of 50 to 70 that might be better, and higher CPPs may be abnormal, and this could be the reason why. The next case I would present is using these waveforms as a target for escalation of therapy. So this is a patient that I learned this phenomena retrospectively. This was a non-dominant intracranial hemorrhage in a GCS of 15. It was an IC aneurysm that ruptured in the parent chyma. There was no subarachnoid hemorrhage. We knew the patient was a candidate for minimally invasive surgical evacuation, but she was GCS of 15 and doing well, so we were waiting for her either to decompensate to take her for surgery or hope that she doesn't decompensate. During the time, the day she was decompensating, she's maximized in osmotic therapy. Because of GCS of 15, she did not receive an intracranial pressure monitor. As she's decompensating, she's supposed to go to the OR around 4 or 5 p.m. as the next case, knowing that we don't want to do this in the middle of the night when she herniates. I'm doing her transcranial dopplers, and in the morning, she had these non-compliant waveforms, but very normal cerebral blood flow. And about an hour before she was supposed to go to the OR, she's becoming more and more lethargic. The GCS is 12 now. It's taking more stimulus to give more exam. So we are certain that we made a good decision of deciding to taking her to the OR, but her waveforms changed to very resistive waveforms. So the diastole is pretty low now, and then systole has become very sharp. And when she comes back after that OR, after evacuation, her waveforms are all positive, but they're still pretty resistive. So we maintain osmotic therapy. She had her flap restored back, and we used this as a target further on to continue osmotic therapy until her waveforms came back to normal, and that's when we de-escalated osmotic therapy. People have started questioning this phenomena in post-arrest patients. And if I can propose, it's a crazy idea, but what if there are phenotypes of brain injury in these patients? What if there's one phenotype that had good CPR, good ROSC, and the patient did not have any cerebral injury? No matter what you do, till you make sure that you're not harming the patient, they're gonna do well. Then there is maybe another phenotype where the cerebral perfusion has been compromised so badly that it doesn't matter what you do, these are the patients that will progress to cerebral edema in the next few hours, and nothing that we're gonna do is change the outcome. And the third phenotype where maybe there is indeterminate cerebral injury, and maybe these are the patients that we can intervene and do something. So there are two separate studies that have shown. One study has shown that very resistive waveforms on the transcranial dopplers, patients are going to have unfavorable outcome no matter what you do. These are the patients that have cerebral edema enough that the perfusion is limited, and they will progress. Then this other subset of patients that have hyperemic waveforms, these are the ones that will develop delayed cerebral edema three to five days down the line, and these are the possible patients for targeted therapy. What if we phenotype these patients based on these CPP parameters, and then targeted them to prevention of fever versus mild hypothermia versus deep hypothermia? Again, a crazy idea needs to be validated in a larger setting. So I hope that I gave you a little bit perspective that actually looking at cerebral perfusion using a modality like a non-invasive ultrasound can help you screen for the presence of malignant cerebral edema, which sometimes can be very helpful in resource utilization in our intensive care settings. And these are the patients that otherwise would not be able to get a traditional imaging. Obviously, austere environments and resource-limited environments is a place where this can be helpful. And I hope some of these cases gave you some food for thought that complementing neuroimaging to actually look at cerebral perfusion rather than just calculating it in suspected cerebral edema cases helps you figure out the phenotype of what's driving the patient's intracranial pressure, and hence target the right therapy for that patient, which could be very different than osmotic therapy and surgical therapy, as I showed you in the hyperaminemic patient. And my hope is that if enough number of people investigate and develop competencies in this, we could start investigating the possible role of looking at CPP using ultrasound as a goal-directed therapy for cerebral edema. These are the protocols that I use myself in point-of-care assessment of a systemically ill, critically ill patient that I'm happy to share if you guys send me a message. And for more resources, the SCCM ultrasound course now has neuroultrasound as a part of the advanced course, and my hope is that critically ill patients do get a good assessment of the cerebral perfusion, whether you believe in invasive, but if you believe in non-invasive therapy, you consider adding neuroultrasound as a part of your toolbox for hemodynamic resuscitation. And here are a few references that you can refer to where you're very proud to have neuroultrasound be a part of basic ultrasound skills recommended for general and neurointensive care population to manage the systemically critically ill population. I'll be happy to take any questions at the end of the session. Thank you so much.

brain ultrasound

diagnosis process

neuroimaging

hemodynamic resuscitation

cerebral perfusion

transcranial dopplers