HomeStrokeVol. 53, No. 4Should Primary Stroke Centers Perform Advanced Imaging? Free AccessArticle CommentaryPDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessArticle CommentaryPDF/EPUBShould Primary Stroke Centers Perform Advanced Imaging? Michael D. Hill, Steven Warach and Sara K. Rostanski Michael D. HillMichael D. Hill Correspondence to: Michael D. Hill, MD, Department Clinical Neuroscience, Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Foothills Medical Centre, HBA 2939, Health Science Centre, 3330 Hospital Dr NW, Calgary, AB T2N4N1, Canada. Email E-mail Address: [email protected] https://orcid.org/0000-0002-6269-1543 Departments of Clinical Neurosciences, Community Health Sciences, Medicine, and Radiology, Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Alberta, Canada (M.D.H.). Search for more papers by this author , Steven WarachSteven Warach https://orcid.org/0000-0003-3125-7979 Department of Neurology, Dell Medical School, University of Texas at Austin (S.W.). Ascension Healthcare‚ St. Louis‚ MO (S.W.). Search for more papers by this author and Sara K. RostanskiSara K. Rostanski https://orcid.org/0000-0002-1162-317X Department of Neurology, NYU Grossman School of Medicine, New York, NY (S.K.R.). Bellevue Hospital‚ Manhattan‚ NY (S.K.R.). Search for more papers by this author Originally published1 Mar 2022https://doi.org/10.1161/STROKEAHA.121.033528Stroke. 2022;53:1423–1430Other version(s) of this articleYou are viewing the most recent version of this article. Previous versions: March 1, 2022: Ahead of Print Should the primary stroke center (PSC) do advanced imaging? There is a remarkable lack of standardization and agreement globally on how exactly suspected acute stroke patients should be imaged. Furthermore, it is completely unclear if this is a problem to be solved or irrelevant to improving outcomes. Rather than stake out polar opposite positions and argue for and against this question, we address the nuanced considerations of clinical practice, where the capabilities and available expertise of PSCs are broad and the optimal choice of imaging depends as much on institutional and clinical circumstances as opposed to one-size-fits-all algorithms. Although determination of thrombolysis and thrombectomy eligibility and treatment is arguably the most important decision in the acute ischemic stroke assessment—and we will spend the majority of consideration with this requirement—most patients with stroke are not candidates for these interventions,1 and most of the hospital-based stroke care is not related to reperfusion therapies. Therefore, the potential added value of advanced imaging must be considered in the context of the management of all stroke patients arriving at the PSC.For patients that present as potential thrombolytic or endovascular thrombectomy (EVT) candidates, time is a determining variable. Both physiologically and clinically, there is clear evidence that the speed of reperfusion after ischemic stroke onset modifies the treatment effect.2,3 Thus, the choice of imaging paradigm is also dictated by speed of performance which most often means that a standard prespecified stroke imaging protocol is used. The stroke chain of survival4,5 has a complex set of links and depends on highly coordinated care in a team setting by highly trained personnel. In all areas of medicine, teams perform well on average if they perform repeatedly; the volume of cases is associated with better team performance.6 While medical treatment with thrombolytic agents can be more decentralized, EVT is necessarily available at specialized hospitals only where technology is available, teams of highly qualified personnel are present, and adequate patient volumes can be maintained. In systems of stroke care, PSC can administer intravenous thrombolysis but typically cannot provide EVT; a thrombectomy-capable center can provide EVT but perhaps only under time frames or other constraints; comprehensive stroke care centers (CSCs) can provide both. Transfer of patients from primary to CSC hospitals is necessary for EVT. In cases where the transfer involves great distances or expenses, confirming EVT eligibility with advanced imaging before transfer decisions may be desirable.There are 2 treatments for acute ischemic stroke both of which minimally require simple imaging on the basis of current guidelines7–10 in the early time window. Intravenous thrombolysis within 4.5 hours of symptom onset is minimally dependent upon a noncontrast computed tomography (CT) or simple magnetic resonance imaging (MRI) only.11 Endovascular treatment in the first 6 hours is additionally dependent upon a single-phase CT angiography (sCTA) or magnetic resonance angiography (MRA) or angiogram to define a target vessel occlusion.12–17 Even sCTA is not necessarily simple for all PSCs. Although there is a general agreement on the minimal imaging requirement, there is a considerable range of practice and advocacy for the added value of more advanced imaging within these early time windows. Because the speed of treatment is so critical to outcome what imaging is necessary for best practice decision-making at the time of patient presentation, at both the PSC and the CSC?Types of Stroke CentersWhat defines a PSC? Definitions vary internationally and in some jurisdictions classification as a PSC is defined by the requirements of certifying bodies. In the United States, the 4 main hospital accreditation organizations (Joint Commission, DNV-GL, Healthcare Facilities Accreditation Program, and Center for Improvement in Healthcare Quality) offer PSC certification, as do some state-level Departments of Health. The hallmark of a PSC is the ability to provide thrombolysis 24/7 and to provide specialized care to uncomplicated stroke patients. While 24/7 availability of rapid diagnostic brain imaging (CT or MRI) is a requirement for all PSCs, certifying bodies in the United States vary in requiring simple vessel imaging (CTA and MRA). This is true despite guideline recommendations that all PSCs should be able to rapidly perform and interpret CTA or MRA 24/7.18 Broadly, PSCs should be able to promptly and reliably identify patients eligible for intravenous thrombolysis and for EVT in all time windows. They should be able to treat patients with intravenous thrombolysis and transfer those that cannot be treated on site, including those patients eligible for EVT and those that require a higher level of critical care or surgical intervention. To fulfill this role, at a minimum, they must perform vessel imaging with CTA or MRA. The question of the utility of perfusion imaging depends on the specific PSC capabilities and local stroke systems of care factors.Types of ImagingHistorically, as acute stroke multimodal imaging with magnetic resonance (MR) emerged in the 1990s, the ability to understand tissue physiology acutely with diffusion-perfusion mismatch imaging dominated thinking in acute stroke care. However, this focus on MRI led the field away from speed of treatment as the sole factor in optimal acute stroke care. Additionally, because of its greater resource investment, multimodal MRI for stroke was largely reserved for CSCs. With the development of CT perfusion (CTP) imaging and automated image interpretation, physiological imaging is more rapidly and widely available and is now a consideration for PSCs.Several imaging sequence types include (1) noncontrast brain CT only; (2) sCTA; (3) multiphase/triple-phase CTA; (4) CTP; (5) simple MRI with diffusion, fluid-attenuated inversion recovery (FLAIR) and a T2* sequence for blood and intraluminal thrombus detection; (6) MR with FLAIR-diffusion mismatch; (7) MR with diffusion-perfusion mismatch using dynamic susceptibility contrast imaging; and (8) MR with MR angiography. Commonly used CT stroke or MR stroke protocols combine parenchymal, vascular, and perfusion imaging with postprocessed perfusion maps all with image acquisition and processing times <15 minutes. In experienced centers, routine use of advanced stroke imaging protocols, including perfusion imaging maximizes average speed. Infrequent use at smaller or less experienced PSCs is associated with excessive delays. There is a tradeoff between time taken to gather more imaging information and process those images for interpretation and the speed of transfer and treatment. Greater imaging information increases the specificity of diagnosis and the theoretical accuracy of therapeutic decision-making at the risk of reduced treatment efficacy due to treatment delays.Technical Issues in ImagingImage acquisition technique and parameters matter and can be optimized for both CT and MR. In all modalities, movement artifact confounds acquisition and image processing with longer acquisition times being more susceptible to movement. Noncontrast CT has high sensitivity and specificity for intracranial hemorrhage in all intracranial compartments but poorer sensitivity for an inclusive diagnosis of ischemia. Interrater reliability for Alberta Stroke Program Early CT Score (ASPECTS) is modest, but better among experts and with automated image processing.19 sCTA has high sensitivity and specificity for large proximal vessel occlusion, but lesser reliability for more distal occlusions, something which is ameliorated with multiphase or triple-phase CTA.20,21 Pial collateral filling is ideally assessed on multiphase or triple-phase CTA but can also be assessed on source images from CTP. CTP provides tissue level blood flow imaging with greater resolution representing not only the pial collaterals but also capillary and venule level filling. Importantly both CTA and CTP image blood flow and should be interpreted together with the noncontrast CT scan which images the parenchyma.MR diffusion-weighted imaging (DWI) has very high sensitivity for acute ischemia.22 DWI hyperintensity and apparent diffusion coefficient hypointensity are observable within minutes of critical ischemia, although are less obvious within the first 3 hours if the lesion is small, located in the brain stem, or if the acquisition technique does not have sufficient signal to noise, which may give a false-negative interpretation. FLAIR MRI is useful for detecting early ischemic changes and hyperintensities starting to become evident beyond 4.5 hours, which has made it useful for determining which patients with stroke-on-awakening may benefit from intravenous thrombolysis (DWI-FLAIR mismatch).23 FLAIR may also identify arterial occlusions, dissections, and slow flow as hyperintensities in the artery and can be correlated with T2* or susceptibility-weighted images as a signal loss within and beyond the artery lumen at the site of occlusion.24 Acute, subacute, and chronic hemorrhage, as well as cerebral microbleeds are diagnosed on T2*-weighted MRI and seen with greatest sensitivity on susceptibility-weighted image and at higher field strength.25 Contrast-enhanced MRA of the head and neck in a single sequence is an excellent screen for large vessel occlusion (LVO). The main disadvantage of MRI is contraindications in ≈15% of patients and the need to complete safety screening before imaging. Both CT and MR images can be postprocessed to produce qualitative perfusion maps from which tissue status (penumbra and ischemic core) can be inferred.The Role of Imaging in Stroke Diagnosis and ManagementStroke imaging is required to make the correct diagnosis, including diagnosis of stroke type and as a direct corollary from correct diagnostics, to select subgroups of patients for specific treatments. We use progressively sophisticated imaging techniques to achieve these aims.Broadly, brain and neurovascular imaging are applicable to all stroke types providing value for both acute treatment decisions and for the investigation and secondary prevention approach. This is true for transient ischemic attack and minor stroke, stroke that is not eligible for thrombolysis or EVT, and for hemorrhagic stroke.Because statistically, stroke is the most common cause of a sudden onset neurological deficit, stroke is suspected primarily for all such clinical presentations. Basic parenchymal brain imaging (noncontrast CT, MR [DWI, FLAIR, gradient recall echo/susceptibility-weighted image sequences]) helps to exclude less common alternate causative structural lesions (eg, tumor or subdural hemorrhage). Clinically severe stroke will be more likely to be associated with imaging signs confirming ischemic stroke (hypodensity on CT, low apparent diffusion coefficient on DWI). Minor stroke or transient ischemic attack may have an associated normal CT and, therefore, CT in this setting has a very low negative predictive value. MR with DWI has much greater sensitivity and consequently a very high negative predictive value.22,25,26Neurovascular imaging with sCTA or time-of-flight MRA confirms or excludes a proximal arterial occlusion. Sensitivity falls more distally in the circulation. Neurovascular imaging with triple or multiphase CTA, increases sensitivity for both distal arterial occlusions and venous occlusions.20,21 A proven arterial occlusion relevant to the symptoms helps confirm ischemia as the diagnosis and identifies a target for treatment.Tissue level perfusion imaging can further refine the subgroup of patients likely to show a large treatment effect with reperfusion therapies by identifying predicted tissue at risk. Less commonly, perfusion imaging may be useful for inpatient management, such as assessing the severity of cerebral vasospasm or informing blood pressure management in patients with evidence of persistent ischemic penumbra. Perfusion imaging can also help identify a high-risk group of patients with transient neurological symptoms, thus potentially averting neurological deterioration with expedited workup and monitoring in this group.27,28Treatments for Acute Ischemic Stroke and Imaging and the Appropriateness of TransferIn real life, the ideal stroke workflow is designed to maximize speed and the clinical assessment occurs in parallel with acute imaging, rather than in series. Using the principle of obtaining the minimum necessary information for decision-making, the clinical assessment can be deterministic for decision-making, although our experience is that this is not usually true until it is combined with the imaging assessment. Simple imaging with CT or MRI is enough for intravenous thrombolysis decision-making within 4.5 hours of last known well, and additional simple neurovascular imaging with sCTA or MRA is enough to decide on moving to endovascular thrombolysis in many patients. Advanced imaging may refine the triage and treatment decision by providing physiological imaging and is recommended in current guidelines for consideration of EVT for patients >6 hours from last known well.7–10 Acute MRI is recommended in guidelines for consideration of intravenous thrombolysis for patients >4.5 hours of last known well.7 Beyond the guidelines, randomized trials assessing both thrombolysis or EVT treatments (WAKE-UP [A Randomized Trial of MRI-Guided Intravenous Thrombolysis in Stroke Patients With Unknown Time of Symptom Onset],29 EXTEND-tPA [Extending the Time for Thrombolysis in Emergency Neurological Deficits],30 DEFUSE-3 [Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3],31 and DAWN [DWI or CTP Assessment With Clinical Mismatch in the Triage of Wake-Up and Late Presenting Strokes Undergoing Neurointervention With Trevo]32) in later time windows, including a high proportion of patients classified as late-window because they had unwitnessed stroke onset, have all used more specific imaging protocols to select patients. Imaging protocols have been both CT-based (non-contrast CT + CTP primarily but also non-contrast CT + multiphase or triple-phase CTA) or MR-based (FLAIR-DWI mismatch).We should judge transfer appropriateness at the decision point, which is at the sending hospital, rather than in retrospect after time has elapsed at arrival or after evaluation at the receiving hospital. The decision to transfer a patient for EVT is naturally based upon the idea that if successful reperfusion is achieved, the patient will benefit. If the duration of transfer (the door-in-door-out plus the transfer time) is long, ischemia may progress to irreversible infarction making reperfusion ineffective or possibly harmful. Alternately, successful reperfusion with intravenous thrombolysis is asymptotically (to an imperfect ceiling of recanalization rates <100%)33,34 time-dependent resulting in removal of the target occlusion for EVT or an uncommon complication such as intracranial hemorrhage may occur all making the patient no longer a candidate for EVT at arrival at the CSC.Current State of the Art: Pros and Cons of Imaging ChoicesThe principal advantage of simple imaging with CT and sCTA (or multiphase or triple-phase CTA) is speed. The principal disadvantage is reduced specificity of diagnosis. The converse is true for multimodal imaging which has longer imaging acquisition and processing time but greater specificity. We illustrate the pros and cons of each modality (Table) with clinical examples.Table. Pros and Cons of Acute Stroke Imaging ApproachesFactorSimple: parenchymal+vessel imaging (CT and CTA or MR and MRA)Advanced: perfusion imaging (CT perfusion or MR perfusion)ImpactTime to acquire imagesLonger time for MR vs CTLonger time for MR vs CTNear neutral with CT modalities as automated processing leads to smaller time differentialsTime for automated image processingNo processing requiredAutomated perfusion map construction only 2–3 minNear neutral as time differentials become smallerPatient prep time for imagingLonger time for MR vs CTLonger time for MR vs CTFavor CT imaging but neutral between CT imaging optionsAccuracy for identifying thrombectomy eligibility (sensitivity and specificity)Greater sensitivityLower sensitivityVaries by health system needs.Lower specificityGreater specificityImaging artifacts (eg, motion)Reduced susceptibility to motion artifactGreater susceptibility to motion artifactFavors parenchymal+vessel imagingExpertise for interpretationExpertise more widely distributedExpertise less widely distributed but increasing due to automated image processingFavors parenchymal+vessel but nearing neutralSafety (radiation dose, contrast dose)Lower radiation and radiocontrast media exposureHigher radiation and radiocontrast media exposureFavors MRI but exposure is a small risk compared with the impact of strokeVolume effects on speed and image processingHigher volume of cases associated with greater speed of imagingHigher volume of cases associated with greater speed of imagingNeutralCT indicates computed tomography; CTA‚ computed tomography angiography; MR‚ magnetic resonance; MRA, magnetic resonance angiography; and MRI, magnetic resonance imaging.Clarifying clinical examples as follows:1. A 55-year-old woman develops global aphasia and severe right hemiparesis 90 minutes earlier and presents to a CSC. MR stroke protocol finds normal DWI, MRA, and MR perfusion. Diagnosis is concluded to be not stroke. She does not receive thrombolysis. Over the course of the next hour her deficits resolve, and she develops a headache like her usual migraine. Here, the greater specificity of diagnosis using multimodal MR leads to the correct clinical plus imaging diagnosis of a complex migraine and the correct treatment is provided.Consider the counterfactual where a noncontrast CT and sCTA is completed only at a PSC, the patient is treated with intravenous thrombolysis and transported to the CSC for possible EVT, despite no LVO being observed on the sCTA. At the CSC, 60 minutes later her deficits are resolving, and she has normal repeat multimodal CT imaging and no evidence of ischemia on a follow-up MR 24-hour later. The final diagnosis is a stroke mimic, illustrating the tradeoff among specificity of diagnosis and type/speed of imaging workup. Although‚ no harm is ultimately done to her health, she has endured the risk of an unnecessary treatment and transport involving many more personnel at greater system cost.2. A 68-year-old man developed hemispatial neglect and left hemiplegia and presents to a rural PSC 4:10 hours after last known well. The hospital performs advanced multimodal CT imaging with perfusion with automated image processing to build perfusion maps. Imaging shows an ASPECTS score of 6, confirms a proximal M1-middle cerebral artery (MCA) occlusion and only a modest degree of predicted salvageable tissue. At the end of imaging, he is not thrombolysed because he is beyond a 4.5 hours threshold and he is not accepted for EVT at the comprehensive center because there is minimal predicted salvageable tissue based upon CTP imaging. The patient’s family later declines hemicraniectomy and the patient dies.Consider 2 possible counterfactuals. The patient undergoes rapid simple imaging identifying an ASPECTS score of 6, a M1-MCA occlusion. He is rapidly treated with intravenous thrombolysis and transferred to the CSC for EVT. He has a persistent M1 occlusion at arrival, undergoes successful EVT and makes a modest recovery with modified Rankin Scale (mRS) score 3 at 90 days. Alternately, he has a persistent M1 occlusion at arrival, his ASPECTS score has worsened to 3 over 1 hour in transfer due to progression of infarction that was predicted by the CTP study, and despite intervention he has the same outcome with a malignant MCA infarction and dies.Both counterfactuals are possible today because of the imperfect predictive value of perfusion maps. From the population perspective, simple imaging is preferred because more patients with stroke in the population will get treatment and improve. However, this will occur at the expense of the health system and CSC because they will treat more patients who will have poor outcomes. From the CSC perspective, advanced imaging may be preferred because only patients with a higher probability of good outcome are selected for transfer; however as we can see in this case this could exclude this patient from life and morbidity saving treatment.3. A 75-year-old woman presents with aphasia and right hemiplegia to a PSC 9 hours after last known well. She undergoes advanced imaging with multimodal CT with CTP imaging. Her ASPECTS score is 8, a left M1-MCA occlusion is identified, and she has evidence of a large region of predicted salvageable brain based upon CTP maps. In teleconsultation with the stroke neurologist at the comprehensive center, she is not thrombolysed, is transferred to the CSC where she proceeds directly to the endovascular suite, undergoes successful EVT for a residual M2-MCA occlusion and makes an excellent recovery (mRS score 1 at 90 days).The counterfactual is illustrated with simple imaging identifying the same ASPECTS score of 8, the M1-MCA occlusion but no perfusion imaging is obtained. The patient is not thrombolysed. She is transferred to the CSC where she undergoes repeat imaging before successful EVT for a persistent M1 occlusion. Reperfusion is achieved 30 minutes later compared with the above and she makes a reasonable recovery (mRS score 2 at 90 days) but remains with moderate aphasia. The case illustrates the opportunity to use perfusion imaging at the PSC combined with a telestroke expert consultation to provide rapid, guideline-based treatment to a wide swath of patients.4. A 79-year-old man with SARS-CoV-2 infection was brought from home to the hospital 1 hour after a family member observed confusion, that is, his speech was nonsensical, and increased breathing difficulty. He was intubated in the field by paramedics because of hypoxia, after being given a neuromuscular paralytic. At the hospital, a CT stroke protocol was performed. It showed an ASPECTS score of 10, no visible occlusion on CTA but a region of focal ischemia with a predicted salvageable pattern in the territory of a left M3-MCA branch. He was treated with an intravenous thrombolytic 1 hour after arrival. MRI stroke protocol 24 hours after treatment, found a small left temporoparietal cortical infarct with normal perfusion. He recovered from his SARS-CoV-2 infection and at day 90, his Wernicke aphasia had virtually resolved but for an occasional paraphasic error (mRS score 1), and he returned to his work as a college English professor.The counterfactual is a noncontrast head CT on arrival based on the family member’s report of confusion (perhaps with a sCTA to rule out a possible LVO). The results are read as normal. Because of the neuromuscular paralytic, a clinical assessment could not be performed in a timely manner and a clinical diagnosis of probable stroke could not be made, and the confusion was attributed to hypoxemia. He does not receive the thrombolytic. He recovers from his SARS-CoV-2 infection but is left with permanent language deficits that disable him from continuing to work as a college professor. This case illustrates the decisive value of perfusion imaging for thrombolytic decision-making when a clinical diagnosis of stroke is uncertain.Health System Structure, Mobile Stroke Units, and Population DistributionChoices of imaging depend upon health system structure. With a focus on population health, high sensitivity for LVO is desirable so that no patient is denied an opportunity for highly effective therapy. The drawback is that the system and its agents at the CSC have to tolerate a higher proportion of acute stroke activations that turn out not to be good candidates for treatment. When the focus is a single CSC and its catchment through a defined referral network, best use of resources may favor higher specificity through perfusion imaging at the PSC, thereby decreasing the number of transfers to the CSC. This tradeoff needs to be determined based on the specific characteristics of the given stroke system of care.35 The use of perfusion imaging at a PSC can enable more specific triage of patients meeting the criteria for endovascular therapy, and thus warranting transfer to the CSC. This information at the PSC level may be especially desirable from a hospital system perspective where unnecessary transfers prove both resource and time consuming. However, this may not be best on an individual level. A recent single-center study on direct-to-angiography observed that reliance on perfusion imaging lead to over selection of patients for EVT and that perhaps, we should be considering the opposite approach, with limited imaging at the PSC and transfer based on clinical trigger.36 This requires further study but the mere question demonstrates the clear tradeoff of imaging approaches.The population approach to specifying which set of patients should undergo a stroke imaging workup is evolving. Any strategy to increase the number of CTAs being performed on patients presenting with acute stroke symptoms must be balanced against the added burden of performance and interpretation of these diagnostic neuroimaging studies on a PSC level. The National Institutes of Health Stroke Scale with a threshold of ≥6 has been suggested as perhaps the best predictor of an LVO, however, even at this level, LVOs will be missed and false positives will be high (87% sensitivity and 52% specificity),37 leading some to advocate a CTA for-all approach.38 At the current time, the specific imaging strategy should be guided by PSC level factors and resource availability, but this remains an area in need of further study.Future: Can We Have Our Cake and Eat It Too?The increasing availability and standardization of treatment for acute ischemic stroke means that it may be possible to conduct a proper randomized assessment of imaging strategies. Diagnostic imaging trials with clinical outcomes are immensely challenging by design because treatment lies sequentially between the intervention (diagnostic imaging) and the outcome (mRS score at 90 days), as a known potential confounder. As treatment becomes standardized, this methodological concern fades.It remains an hypothesis that advanced imaging can determine best who to treat and transfer on a population basis.36 Today, judging outcomes based upon imaging choices are not observable for individual patients. We can observe from the examples that differences in outcomes may only be modest. Reports of cohort studies comparing modalities,39 and even reports from randomized trials of treatment, are and will be biased by selection.Alternately, progression of technology may overtake this proposed need for a randomized comparison. Advances in image acquisition and processing may increase specificity and sensitivity and speed of both approaches, essentially resulting in a convergence of imaging techniques. The potential rising implementation of CT-equipped mobile stroke ambulances can change the imaging discussion.40–43 The mobile stroke unit essentially becomes an extension of the CSC and patients requiring EVT are brought directly to the neuro-angiosuite, obviating any repeated multimodal imaging and maximizing speed of treatment. Finally, advances in flat-panel CT allows for the potential for a direct-to-neuro-angiosuite approach which has been and could be implemented more widely, if the costs of imaging patients with stroke presentations not due to LVO can be justified.44Using criteria from positive treatment trials, advanced imaging is not recommended for early time windows but is recommended for later time windows.29,31,32,45,46 For now, we favor re-emphasizing 2 ideas: (1) simple brain and neurovascular imaging (non-contrast CT and CTA, or MR with MRA) is the minimum standard that has broad utility for all stroke types and; (2) reperfusion treatment for ac
Search from 250 M+ research papersSearch from 250 M+ research papers
Open Access
Just Published
Apr 29, 2024
Atya Baidaa A Ibrahim 