Introduction

Acute stroke is a leading cause of morbidity and mortality, affecting ~ 800,000 people in the United States per year [1]. As the vast majority (80–90%) of strokes are ischemic as opposed to hemorrhagic, this review focuses primarily on emergent neuroimaging of acute ischemic stroke (AIS). The field of stroke imaging continues to rapidly evolve, with many recent clinical trials and increasing public awareness. This review article provides an overview of stroke neuroimaging in the emergent setting, with emphasis on recent updates in the stroke literature in the past 5 years.

Role of Imaging in Acute Stroke Management

The primary objective of acute stroke imaging is to identify patients most likely to benefit from rapid reperfusion of potentially salvageable brain tissue. Treatment generally consists of intravenous thrombolysis or intra-arterial therapy (IAT), including mechanical thrombectomy. There is strong evidence of favorable treatment responses in patients receiving intravenous thrombolysis within 4.5 h of symptom onset [2, 3]. The administration of intravenous tissue plasminogen activator (tPA) in moderately symptomatic patients increases positive patient outcome rates from ~ 50 to 75% [4]. IAT can be performed at even later intervals [5•], with recent trials showing beneficial effects of thrombectomy in appropriately selected patients up to 24 h after symptom onset [6••, 7••]. As the benefit of IAT decreases with time by approximately 6% per hour of reperfusion delay [8], prompt evaluation of treatment candidates remains critically important. The 2018 stroke guidelines jointly published by the American Heart Association (AHA) and the American Stroke Association (ASA) now recommend that median door-to-imaging times be at most 20 min among treatment candidates [9••].

As timely accurate interpretation of neuroimaging is crucial for clinical decision-making, it is important for radiologists to be familiar with imaging assessment of AIS. The emergent imaging evaluation of stroke has the following goals: (1) rule out intracranial hemorrhage, mass lesions, or other stroke mimics; (2) define the ischemic core and ischemic penumbra (“tissue-at-risk”); and (3) assess the vasculature. The role of imaging in evaluation of these objectives and their impact on AIS management are discussed below under the subheadings for each modality.

Computed Tomography

Computed tomography (CT) remains the standard of care for initial AIS imaging due to its widespread availability, short scan time, and ability to identify contraindications for thrombolytic therapy. CT can reliably detect intracranial hemorrhage or intracranial mass lesions, which would preclude thrombolytic therapy. In additional to hemorrhagic transformation of an ischemic stroke, acute intracranial hemorrhage may represent a hemorrhagic stroke secondary to various causes, such as hypertension, coagulopathy, vascular malformations, aneurysms, or mass lesions (Fig. 1).

Fig. 1
figure 1

Intracranial hemorrhage in a 32-year-old male presenting with unresponsiveness. a Axial head CT imaging demonstrates hyperattenuating subarachnoid hemorrhage in the sylvian fissures and basal cisterns (arrowheads) and a focal hyperdense clot near the anterior aspect of the suprasellar cistern (arrow). b SWI also depicts subarachnoid and intraventricular hemorrhage (arrowheads) as low signal. A focal clot near the anterior aspect of the suprasellar cistern (arrow) is also visible. c A 3D volume-rendered image from a CTA study shows an anterior communicating artery aneurysm (arrow) as the likely cause of hemorrhage

Full size image

CT can also identify early signs of ischemia, including parenchymal hypoattenuation, loss of gray–white differentiation, sulcal effacement, and other evidence of mass effect (Figs. 2a, 3a). Sulcal effacement with preservation of gray–white differentiation, i.e., isolated sulcal effacement, is occasionally seen in stroke patients with proximal vascular occlusion and robust collaterals [10]. If present, this finding may indicate preserved underlying parenchyma that could benefit from reperfusion [10]. CT findings of AIS typically become more conspicuous over time as magnitude of hypoattenuation increases (Fig. 2b, c). Growth of infarct volume is common in AIS, and infarct volume at 24 h or at 1 week predicts functional outcome [11].

Fig. 2
figure 2

A 64-year-old female with acutely altered mental status. a Early ischemic changes on the initial non-contrast head CT include subtle hypoattenuation in the left MCA territory (arrows), with indistinctness of the left insular cortex and basal ganglia. b Follow-up CT imaging 24 h after the first head CT shows greater conspicuity of the region of hypoattenuation (arrows), with increased mass effect. c 7 days after presentation, the final infarct volume is more clearly delineated (arrows), with even greater mass effect. d–f MRI acquired 2 days after symptom onset demonstrates diffusion restriction and elevated FLAIR signal within the left MCA territory infarct. g SWI shows small foci of low signal intensity consistent with magnetic susceptibility from petechial hemorrhage (arrows) in the left basal ganglia within the infarcted territory. h Parametric maps from a CTP at initial workup show matched areas of markedly diminished CBF, low CBV, and elevated MTT and Tmax, consisted with a large infarct core without a significant penumbra. i A maximum intensity projection (MIP) of the brain CTA obtained at initial workup shows an LVO of the left M1 segment (arrows). The anterior left temporal lobe shows diffusely decreased vascularity (arrowheads) without appreciable collateral filling of the distal left MCA

Full size image
Fig. 3
figure 3

A 67-year-old male presenting with right-sided weakness and aphasia. a Findings on head CT include subtle hypoattenuation with loss of gray–white matter differentiation involving the left insula and frontal operculum (arrows), consistent with acute ischemia. b Post-processed CT perfusion data utilizing RAPID software yields fully automated volume estimates of the infarct core (28 mL) and the ischemic penumbra (124 mL). The perfusion data meet DEFUSE 3 criteria for predicting a favorable response to endovascular intervention, including small infarct core size (< 70 mL), high ratio of ischemic tissue to infarct core volume (≥ 1.8), and penumbra volume ≥ 15 mL. c Digital subtraction angiography shows an occlusive thrombus in the proximal inferior division of the left MCA (arrow). d After mechanical thrombectomy was performed, DSA shows recanalized flow within this vessel (arrowhead). The patient subsequently showed a complete functional recovery

Full size image

A hyperdense artery sign, often of the proximal middle cerebral artery (MCA), may occasionally be seen in large infarcts and is relatively specific but insensitive for intravascular thrombus (Fig. 4a). Although this sign independently predicts poorer functional outcomes [12], its presence or absence does not affect the benefit of tPA or thrombectomy on clinical outcomes [12, 13]. In some studies, lower thrombus density on CT was associated with lower recanalization rates in AIS [14, 15], but one study found no effect of thrombus density on clinical outcomes [16]. The 2018 AHA/ASA stroke guidelines recommend that the hyperdense MCA sign not be used to withhold tPA in an otherwise eligible patient [9].

Fig. 4
figure 4

Acute stroke imaging workup in a 62-year-old female. a The initial non-contrast head CT shows a hyperdense left MCA (arrow). b, c A coronal MIP and a 3D volume-rendered image from a brain CTA show absent contrast opacification of the intracranial left ICA (arrows) and MCA (arrowheads). Neck CTA imaging (not shown) also revealed absent contrast filling within most of the cervical left ICA. d DWI obtained 14 h after the initial head CT confirms the presence of an acute infarct involving the left basal ganglia

Full size image

Although not sensitive for small or early ischemic changes, CT is nonetheless useful in assessing extent of ischemia, either qualitatively or using semi-quantitative methods such as the Alberta stroke program early CT score (ASPECTS), a 0–10 scale with higher scores denoting smaller infarcts [17]. Extent of ischemia on CT has prognostic implications. Early imaging signs on CT involving greater than one-third of the MCA territory have been associated with greater risk of hemorrhagic transformation and overall poorer outcomes [18,19,20].

Extent of abnormalities on baseline CT may also influence treatment eligibility, as larger baseline infarct volumes typically have poorer treatment responses, and many clinical trials showing benefit of IAT included only patients with small infarcts [21•, 22•, 23•, 24•]. However, the literature is still evolving, with recent data suggesting that strokes with moderate CT-visible infarcts (ASPECTS 5–7) may also benefit from IAT [25]. Although past guidelines had specified involvement of a large hypodensity involving over a third of the MCA territory as a contraindication to tPA, secondary analysis of IST-3 trial data found no significant interaction between baseline CT hypoattenuation and tPA administration on clinical outcomes [26]. The most recent AHA/ASA stroke guidelines state that extent and severity of hypoattenuation or early ischemic changes should not contraindicate tPA in otherwise eligible treatment candidates [9••].

Magnetic Resonance Imaging

Diffusion-weighted imaging (DWI) offers greater sensitivity in detecting early ischemic changes (Figs. 2d, e, 4d), with restricted diffusion often apparent within minutes of stroke onset. In a prospective comparison of MRI and CT, sensitivity of MRI for detection of AIS increases over time from 73% at < 3 h after symptom onset to 92% at > 12 h, whereas CT sensitivities are in the 12–20% range at these time intervals [27]. The absence of DWI abnormality, in addition to resolution of symptoms within 1 h, is needed for distinguishing transient ischemic attack (TIA) from strokes [28]. It is important to note that diffusion abnormality is initially absent in some patients who ultimately receive a diagnosis of AIS, particularly in early or very small strokes. The extent of restriction diffusion may be used to quantify irreversibly infarcted tissue and has been used to determine eligibility for endovascular treatment in recent clinical trials [6••, 7••].

In addition to DWI, other MRI sequences are also helpful in AIS. FLAIR shows signal hyperintensity (Fig. 2f) but typically hours after stroke onset and to a greater extent in patients with poor collaterals [29]. The accuracy of MR imaging for evaluating acute intracranial hemorrhage is likely similar to that of CT and may be higher for detecting chronic cerebral microhemorrhages [30], particularly when using sequences sensitive to paramagnetic susceptibility, such as gradient-recalled echo or susceptibility-weighted imaging (SWI) sequences (Fig. 2g). The presence of microbleeds may be a potential concern when considering thrombolytic therapy due to a higher suspected risk of developing symptomatic intracranial hemorrhage, as supported by recent meta-analyses [31, 32]. However, some authors have argued that the risks attributable to microbleeds are likely small relative to the benefit of thrombolytic therapy [33]. The 2018 AHA/ASA stroke guidelines recommend against routine use of MRI to exclude microbleeds prior to tPA [9••].

Although some authors have proposed routine use of MRI as the initial imaging modality during AIS assessment, expense and timely accessibility are factors limiting MRI use compared to CT in most settings. MR imaging offers greater sensitivity in detecting ischemic stroke within 12 h of onset, but the influence on clinical decision-making or outcomes in the acute setting compared to CT may not be significantly different. A recent systematic review and meta-analysis concluded that routine MR imaging in all patients with AIS is not cost-effective [34].

Perfusion Imaging

Determining the extent of brain that is viable but at risk can be assessed by imaging, including CT perfusion (CTP) and MR perfusion, sometimes in conjunction with clinical neurological assessment using the National Institutes of Health Stroke Scale (NIHSS). CT perfusion, in which multiple CT acquisitions are obtained sequentially during bolus administration of intravenous contrast, is frequently used in emergent stroke assessment. It can detect strokes with 95% specificity and 80% sensitivity in a recent meta-analysis [35]. Limited spatial resolution likely contributes to false negatives, which consists mostly of small lacunar infarcts. Given poor sensitivity of non-contrast CT in detecting posterior fossa strokes, CTP can improve diagnostic accuracy of acute ischemia in the posterior circulation [36].

A major strength of CTP is its ability to characterize the dynamic perfusion status of the brain. The pathophysiology of ischemic stroke is typically based on hypoperfusion of a vascular territory supplied by a stenosed or occluded artery. A markedly hypoperfused ischemic core representing irreversibly damaged tissue may be bordered by a less severely hypoperfused ischemic penumbra. The penumbra is maintained by collateral circulation and is potentially salvageable with timely reperfusion. There could also be an adjacent area not at risk of infarction described as benign oligemia.

Time–density curves from CTP acquisition allow calculation of hemodynamic parameters such as cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), time to peak (TTP), and time to maximum of the tissue residue function (Tmax) (Fig. 2h). The ischemic core may be defined by low CBV or reduction in CBF to below a relative or absolute threshold. Similarly, the penumbra can be estimated as the area of prolonged MTT, TTP, or Tmax that is not included in the infarct core, or it can be defined by CBF values within a defined range. While various postprocessing software packages are available to generate these parameters from CTP data, the RAPID tool (iSchemaView), which has been validated in recent stroke trials, is one example of software that provides fully automated identification of ischemic core and penumbra in a standardized and easily accessible manner. A representative output from the RAPID software, shown in Fig. 3b, illustrates RAPID’s ability to estimate volume of ischemic hypoperfusion based on Tmax > 6 s, estimate core infarct volume based on CBF < 30% of normal tissue, and calculate values for mismatch volume and mismatch ratio based on the difference and ratio of these values, respectively. Automated CTP-based mismatch selection using RAPID software is robust in clinical practice and is associated with faster treatment decisions compared to MRI [37].

Ischemic core sizes on baseline perfusion imaging can predict final infarct sizes [38] and are a major determinant of clinical outcome at 3 months [39]. In general, patients with smaller ischemic core sizes and growth are associated with more favorable outcomes. Patients with very poor collaterals tend to exhibit rapid infarct growth, while those with more developed collaterals typically have smaller ischemic core sizes relative to the hypoperfusing region and are generally favorable candidates for reperfusion therapy.

An inherent disadvantage of routine CTP use is the high radiation exposure related to acquisitions at different spatial levels over time. Typically images are obtained at lower tube voltages (e.g., 80 kVp) than for head CT. Further reduction in tube voltage to 70 kVp and use of variable sampling times have recently been shown to lower radiation dose without significantly compromising imaging quality or diagnostic accuracy [40, 41]. Ionizing radiation exposure can be avoided by using MRI to assess perfusion status, using such techniques as dynamic susceptibility-contrast MRI (DSC-MRI) and interpreted in a manner analogous to CTP. In combination with DWI to define the infarct core, this allows calculation of a diffusion–perfusion mismatch as a parameter for penumbral estimation. Although the concepts underlying CT and MRI perfusion imaging are similar, some perfusion parameters derived from the two modalities may not be concordant, but good agreement between CT and MR perfusion can be achieved when using Tmax [42]. Perfusion status using either CT or MR allows better prediction of clinical outcomes than recanalization assessment [43, 44].

Imaging of the Vasculature

CT Angiography

CT angiography (CTA) is often included in initial stroke workup and can evaluate vessel patency, thrombus characteristics, collateral status, and other vessel abnormalities in the head and neck. CTA can usually be performed rapidly in the emergent setting, often immediately after the initial non-contrast head CT. The need for intravenous administration of iodinated contrast occasionally raises concern for subsequent renal issues, but screening serum creatinine is generally not needed prior to CTA in thrombectomy-eligible patients with suspected proximal occlusions without history of renal impairment [9••], based on growing literature supporting the safety of CTA in acute stroke workup with respect to renal function [45].

CTA is useful for emergent detection of large-vessel occlusion (LVO), a common finding in AIS, typically involving the internal carotid artery or proximal MCA (Figs. 2i, 4b, c), that may be amenable to emergent endovascular therapy. Detection of LVO on CTA is associated with poorer outcome in stroke patients receiving intravenous thrombolysis [46]. CTA can also evaluate thrombus size. In general, longer thrombi are associated with worse 90-day clinical outcomes compared to shorter thrombi but may have a greater relative benefit of thrombectomy over intravenous thrombolysis alone [47]. Although it has been proposed that thrombus length exceeding 8 mm may have lower recanalization rates with intravenous thrombolytics, some of these patients may still benefit from intravenous thrombolysis [47].

Classification methods exist for describing extent and location of thrombus on CTA, including the Boston Acute Stroke Imaging Scale (BASIS) [48] and the clot burden score (CBS) [49]. In BASIS, strokes are dichotomized as major or minor based on detection of proximal occlusions in major arteries or presence of a significant ischemic lesion on DWI or non-contrast CT. CBS assigns strokes to a scale from 0 to 10, and lower values, denoting greater clot burden, are associated with lower odds of reperfusion, larger final infarct sizes at follow-up, and worse functional outcome [50]. Modifications of these scoring systems have been described that better quantify the extent and location of thrombus and may have better prognostic accuracy [51].

CTA also permits evaluation of collateral flow, which contributes to maintaining viability of penumbral tissue and may therefore predict clinical outcomes. In general, the benefit of intra-arterial therapy is greatest in patients with good collateral supply on baseline CTA [52,53,54]. 4D-CTA or dynamic CTA, in which CTA images are reconstructed from whole-brain volumetric CT perfusion data, may also be used in AIS evaluation [55], allowing evaluation of both brain perfusion and vessel patency in a single CTP acquisition. Dynamic CTA may better predict clinical outcome by improved assessment of collateral flow compared to single-phase CTA, likely due to its superior temporal resolution [56]. Dynamic CTA may have other applications, such as evaluating extent and velocity of cortical venous filling downstream of an occluded artery to predict clinical outcome [57].

Magnetic Resonance Angiography

Magnetic resonance angiography (MRA) is an alternative to CTA in assessing the intracranial and neck vasculature. Although useful for evaluating large-vessel stenosis or occlusion, its sensitivity and specificity are less than for CTA [58], and sensitivity of MRA decreases when evaluating smaller caliber intracranial vessels. Although time-of-flight MRA has the advantage of not requiring contrast media, it may occasionally overestimate vessel stenosis, particularly in regions of high and/or turbulent vascular flow. MRA also requires more time to acquire, which may increase likelihood of motion artifact.

Digital Subtraction Angiography

While digital subtraction angiography (DSA) had historically been considered the gold standard for evaluating cerebral vessels due to its superior spatial and temporal resolution, it is infeasible as a fast diagnostic tool as it is resource-intensive and associated with higher complication rates compared to non-invasive imaging. Furthermore, one study found helical CTA to be superior to DSA for assessing posterior circulation patency in the setting of severe stenosis [58]. DSA is useful when concomitant intra-arterial therapy is contemplated (Fig. 3c, d). As a dynamic study, DSA can also provide information on tissue viability. The capillary index score is a method to assess perfusion and identify non-viable brain tissue, based on the observation that brain areas that lack capillary blush on DSA before treatment correspond to non-viable tissue [59]. Successful revascularization is more likely to produce favorable functional outcomes when capillary blush is present [60].

Other Modalities and New Techniques

Relatively new or emerging techniques in stroke imaging include arterial spin labeling (ASL), blood-oxygen-level-dependent (BOLD) MRI, and machine learning. ASL allows assessment of CBF without need for gadolinium contrast and may be comparable to DSC-MRI perfusion in assessing reperfusion in AIS [61]. Combining ASL with SWI sequences may be useful in predicting hemorrhage or functional outcome in AIS patients [62]. A modification of ASL known as territorial arterial spin labeling, in which major individual arteries are labeled separately, can be used to evaluate collateral circulation and yield results comparable to DSA [63].

Perfusion status can also be assessed by evaluating changes in BOLD signal. Resting-state functional MRI temporal analysis using BOLD allows non-invasive evaluation of cerebral hemodynamics, including CBF, without requiring contrast administration. Studies have shown that BOLD delay can be used to detect hypoperfusion in acute stroke [64,65,66], subacute stroke [67], and chronic cerebrovascular disease [65, 68].

In recent years, artificial intelligence (AI) has received increasing attention in radiology, and applications of AI to stroke have been described in recent reviews [69, 70]. For instance, supervised machine learning is a form of AI that can be used to learn complex structures from a training data set and apply that learned knowledge to predict outcomes. Deep learning is a more recently developed and specialized form of machine learning that mimics the structure of the human brain using multiple layers of artificial neural networks (ANNs). One study demonstrated that an ANN may be effective in recognizing acute ischemia and accurately differentiating stroke from mimics [71]. The ability of AI techniques to detect patterns from large amounts of complex data may enable them to eventually play an important role in imaging evaluation of AIS patients. Machine learning and other AI techniques continue to rapidly evolve and become more accessible to clinical users.

Impact of Recent Stroke Intervention Trials on Diagnostic Neuroimaging

There are numerous ongoing and recently completed clinical trials in stroke intervention. This section briefly discusses a few recent randomized clinical trials evaluating endovascular intervention that have implications particularly relevant to diagnostic radiologists. The role of endovascular therapy in AIS has evolved rapidly over the past 5 years, with current evidence now supporting a benefit of IAT in selected patients with LVO. Previously, multiple clinical trials published in 2013 reported no benefit of IAT compared to standard care with intravenous tPA [72,73,74], possibly explained by lack of imaging selection criteria (e.g., CTA confirmation of LVO or perfusion evidence of salvageable tissue), low reperfusion rates, and use of older mechanical thrombectomy devices in contrast to newer stent retrievers that may offer greater recanalization speed and efficacy. The superiority of IAT compared to intravenous tPA alone in treatment of selected patients with AIS secondary to LVO has subsequently been proven in multiple randomized trials published in 2015 and 2016 that include Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands (MR CLEAN) [5•], Endovascular Treatment for Small Core and Anterior Circulation Proximal Occlusion with Emphasis on Minimizing CT to Recanalization Times (ESCAPE) [23•], Extending the Time for Thrombolysis in Emergency Neurological Deficits—Intra-Arterial (EXTEND-IA) [24•], Solitaire With the Intention for Thrombectomy as Primary Endovascular Treatment (SWIFT-PRIME) [22•], Randomized Trial of Revascularization with Solitaire FR Device Versus Best Medical Therapy in the Treatment of Acute Stroke Due to Anterior Circulation Large Vessel Occlusion Presenting Within 8 Hours of Symptom Onset (REVASCAT) [21•], and Trial and Cost Effectiveness Evaluation of Intra-Arterial Thrombectomy in Acute Ischemic Stroke (THRACE) [75•]. In all 6 trials, a large majority of patients in the endovascular treatment arms underwent thrombectomy utilizing retrievable stents within 6 h of symptom onset in addition to standard care or intravenous tPA. These endovascular arms demonstrated improved clinical outcomes compared to standard care or intravenous tPA alone, deemed to be predominantly due to reduction in final infarct volume [76]. The HERMES (Highly Effective Reperfusion evaluated in Multiple Endovascular Stroke Trials), a meta-analysis of pooled data obtained from 5 of these trials, established that the number needed to treat in order to achieve a one-point improvement in the modified Rankin Score, a 7-point scale for assessing post-stroke neurological disability, at 90 days was 2.6 [77].

The absolute proportions of patients achieving good outcomes after IAT vary among these trials, likely attributable to differences in patient selection and imaging inclusion and exclusion criteria. However, a common feature of these trials is the requirement of imaging confirmation of LVO, and therefore emergent imaging within 6 h of stroke onset in potentially eligible IAT candidates should generally include CTA or other vascular imaging. Although several of these trials included perfusion imaging or other assessment of the ischemic penumbra or collateral status, the MR CLEAN and THRACE trials demonstrated treatment benefit using only CTA and non-contrast head CT as imaging selection criteria. The AHA/ASA stroke guidelines recommend against additional imaging studies such as CTP for potential thrombectomy candidates presenting within 6 h of stroke onset so as to not exclude patients who could potentially benefit [9••].

Two recent trials published in early 2018 demonstrated efficacy of IAT beyond 6 h of symptom onset. The Clinical Mismatch in the Triage of Wake Up and Late Presenting Strokes Undergoing Neurointervention With Trevo (DAWN) trial reported better clinical outcomes in patients treated with thrombectomy 6–24 h from the last known well time, with 49% of patients achieving functional independence at 90 days in the thrombectomy group compared to 13% with standard medical therapy [6••]. The Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution 3 (DEFUSE 3) trial reported significantly improved 90-day functional outcomes, decreased 90-day mortality, and higher post-treatment reperfusion and recanalization rates in patients undergoing endovascular therapy 6–16 h after the time last known to be well [7••]. Both these trials had inclusion criteria that selected patients likely to have salvageable tissue and/or favorable collateral status. Both trials required imaging confirmation of LVO and used automated postprocessing software (RAPID) to identify key parameters on CTP or MRI diffusion/perfusion imaging. In the DAWN trial, enrolled subjects were required to have a discrepancy between clinical deficit (measured by NIHSS) and infarct volume on baseline CTP or MRI imaging [6••]. The DEFUSE 3 trial required an infarct volume on baseline imaging of less than 70 mL and ischemic penumbra sizes exceeding defined thresholds [7••]. The implication for diagnostic imagers is that MRI, CTP, or other methods for core and penumbral assessment will be increasingly used for identification of thrombectomy-eligible stroke patients presenting more than 6 h after symptom onset. Moreover, given evidence of a thrombectomy benefit at up to 24 h after symptom onset, emergent neuroimaging will likely be needed in a larger number of stroke patients, and therefore, radiology practices may need to plan accordingly so as to accommodate potential increases in volume of emergent imaging requests for acute stroke evaluation.

Conclusions

Acute ischemic stroke is a common and often treatable condition for which neuroimaging plays an increasingly vital role in determining appropriate treatment. The main objectives of imaging are to exclude intracranial hemorrhage, define extent of infarct, distinguish between ischemic core and penumbra, evaluate the vasculature, and exclude mass lesions or other stroke mimics. Unenhanced CT remains the modality of choice for initial brain imaging, particularly for rapidly excluding contraindications for intravenous thrombolysis, but MRI and CTP offer additional complementary information that could affect emergent stroke management and predict prognosis. Imaging of the vasculature, typically with CTA, allows detection of proximal vascular occlusions and is increasingly important in light of recent clinical trials showing benefit of endovascular treatment at up to 24 h after symptom onset. While stroke imaging guidelines have evolved over the past 5 years, reflecting the impact of multiple high-quality studies in the recent stroke literature, neuroimaging continues to have a crucial role in the initial workup of acute stroke.