Acute Ischemic Stroke Therapy Overview

Noncontrast Head CT

The noncontrast head CT (NCCT) has been a staple of acute stroke imaging for many years. Rapid acquisition time, applicability, and availability to all patients, regardless of comorbidity or setting, and easy detection of intracranial hemorrhage (ICH) has made it a universally invaluable tool. The critical distinction between underlying hemorrhage and ischemia immediately impacts patient management while determining eligibility for acute thrombolytic or endovascular therapy (EVT). With more patients being sent for potential EVT from peripheral centers that lack access to vessel imaging, it is often the NCCT that serves as the only source of radiographic clues for eligibility.

Early Ischemic Changes

These are signs of brain edema that develops in some patients early after onset.¹² Hypoattenuation of the x-ray beam occurs as water content increases, which leads to the appearance of a hypodensity and a reduced contrast between adjacent the gray and white matter. The latter phenomenon of blurred distinction between structures, depending on location, leads to radiographic signs, such as the loss of the insular ribbon or obscuration of the lentiform nucleus. These hypodensities are predictive of final infarct volume and demonstrate irreversible tissue damage, despite successful reperfusion.¹³ Another form of early ischemic changes (EIC) is cortical swelling, which leads to the appearance of sulcal effacement or loss of the subarachnoid spaces between gyri and local mass effect. Cortical swelling, when in isolation, has been correlated with compensatory vasodilation and increased blood volume and has a propensity to be reversed with reperfusion, perhaps representative of penumbral tissue.¹⁴

Detection of EICs is difficult and requires training and experience for accurate detection. Adjustments to the CT window and level (eg, W30/L35, W30/L50, etc) can increase the contrast of tissues and may augment the detection of subtle EICs.¹⁵ In general, EICs have decreased sensitivity for detecting smaller infarcts, hyperacute infarcts, and infarcts of the brain stem and cerebellum where CT beam artifact is a common confounder. The utility of identifying and quantifying EICs has evolved since the early thrombolytic trials. The Alberta Stroke Program Early CT Score (ASPECTS) was created and validated to help systematically detect and quantify EICs^16,17 (Figure 2A). Starting with 10 prespecified areas of the middle cerebral artery (MCA) territory, a point is deducted for each area in which a hypodense or hypoattenuating EIC is present. A lower score indicates more extensive EICs and a larger territory of tissue likely destined for infarction. A posterior circulation ASPECTS was also developed (Figure 2B). The ASPECTS may, therefore, provide a readily available method for determining if the ischemic core is small or large on a noncontrast CT scan that is essentially performed on all AIS patients.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

A, Example of an ASPECTS of 3 out of 10, with one point deducted for early ischemic changes in each territory: M2, M3, I (insula), L (lentiform nucleus), IC (internal capsule), M4 and M5 territories with sparing of the M3 and C (caudate) and M1 (not shown). B, pc-ASPECTS score of 4 out of 10. The yellow numbers indicate point deduction for areas of EIC. ASPECTS indicates Alberta Stroke Program Early Computed Tomography Score; EIC, early ischemic changes; and pc-ASPECTS, Posterior Circulation Alberta Stroke Program Early Computed Tomography Score.

Use of the ASPECTS in Clinical Practice and Trials

The ASPECTS has been independently associated with long-term functional outcome. A score of ≤7 is associated with dependent functional status (modified Rankin Scale [mRS] score ≥3), as well as an increased risk for symptomatic intracerebral hemorrhage (sICH).¹⁸ Lower scores (0–5) are more useful for prognosis than the scores in the higher range.¹⁹ When applied to thrombolytic trial data, intravenous (IV) tissue-type plasminogen activator (tPA) did not have a treatment modification effect on outcome, despite the score.^17,20 This is reflected in recent American Heart Association/American Stroke Association (AHA/ASA) guidelines, which conclude that there is no evidence to withhold IV tPA from those with extensive mild-to-moderate EICs and determined that no adequate threshold has been established where EIC within 4.5 hours from symptom onset should preclude the use IV tPA.²¹ Those with a clear, extensive hypodensity of more than one third of the MCA (≈ASPECTS ≤5) is a class III/A recommendation against therapy.²²

Data from the PROACT-II trial (Prolyse in Acute Cerebral Thromboembolism) suggested less benefit for patients with a lower ASPECTS despite recanalization. The MR-CLEAN trial (Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands) did not exclude patients based on EICs but did perform a prespecified subgroup analysis on ASPECTS ranges of 0 to 4, 5 to 7, and 8 to 10. The 8 to 10 subgroup had a primary outcome (shift of mRS) odds ratio of 1.61 (1.11–2.34) and was the only group to maintain statistical significance.²³ The ESCAPE trial (Endovascular Treatment for Small Core and Proximal Occlusion Ischemic Stroke) and the SWIFT PRIME trial (Solitaire With the Intention for Thrombectomy as Primary Endovascular Treatment for Acute Ischemic Stroke) excluded patients with ASPECTS ≤5 and the REVASCAT (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), excluded patients with ASPECTS ≤6. The ESCAPE, SWIFT-PRIME, and REVASCAT trials all showed better outcomes patients with ASPECTS ≥8.²⁴ A meta-analysis of trial data showed that the subgroup of dichotomized ASPECTS of ≤7 versus >7 had no interaction for improved functional outcome.²⁵ An ASPECTS cutoff of ≥6 is a class IA AHA/ASA recommendation for selecting endovascular candidates. Those with scores <6 are relegated to a class IIb/B-R recommendation because a majority of these patients were excluded from these trials.²⁴ The posterior circulation ASPECTS has been validated and can be useful for predicting futile recanalization in those with acute basilar artery occlusion. Those with posterior circulation ASPECTS of <8 have been shown to independently predict poor functional outcomes (mRS score ≥4), despite complete recanalization.²⁶

Hyperdense Vessel Signs

The NCCT can provide clues to the presence of a large artery occlusion (LAO) when a hyperdense vessel sign is present. These are segments of blood vessels that have high signal attenuation and that represent intraluminal thrombus. The sensitivities and specificities for predicting confirmed vessel occlusion on angiography vary by location, often with high specificity but poor to moderate sensitivities (Table 1).^27–34

Intracranial LAO

When compared with the gold standard of digital subtraction cerebral angiography, CTA has a sensitivity and specificity of nearly 98% to 100% for detecting intracranial LAO.^35,36 CTA is also sufficient for detecting acute intracranial nonocclusive thrombi, a rare though potential predictor of subsequent clinical decline.³⁷ The clot-burden score has been developed to quantify extent of intracranial thrombosis visualized on CTA. It is a 10 point score that is calculated by subtracting points for the presence of occlusive or nonocclusive clot in certain locations: 2 points are deducted for the presence within supraclinoid ICA, proximal MCA trunk/MCA–first segment (M1) or distal M1; 1 point deducted for presence within infraclinoid ICA, anterior cerebral artery or MCA–second segment division (Figure 3).³⁸ A clot burden score of 0 indicates a complete multisegmental vessel occlusion, with lower scores validated to independently predict lower recanalization rates with IV tPA, larger infarcts, poor outcome, and risk of hemorrhagic transformation.^38–40 Based on the data acquired from recent endovascular trials, the AHA/ASA has granted a grade I/A recommendation to endovascular candidates only with an occlusion located in the M1 or terminal ICA. The remaining large-vessel arterial locations are designated class IIb/C.²⁴

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Example of a left carotid occlusion involving the infra- and supraclinoid carotid artery (left). The Clot Burden Score is 0 out of 10. Infraclinoid ICA =1 (purple), supraclinoid ICA =2 (red), proximal M1 =2 (light green), distal M1 =2 (dark green), A1=1 (yellow), and M2=1+1 (light blue). ICA indicates internal carotid artery.

Collateral Assessment

Collateral blood flow via leptomeningeal arteries and other vessels during an acute stroke is an important factor in the rate of ischemic tissue evolution and transition from penumbral tissue into irreversible infarct.⁴¹ It is highly individualized and is dependent on several factors, including variants of vascular anatomy, clot location, and clot burden. Several scoring methods were used to quantify the presence of collaterals and attempt to validate their ability to predict outcomes. Currently, standardization is lacking regarding the best method. The regional leptomeningeal collateral score assesses the presence and quality of collateral vessels in the 6 MCA cortical ASPECTS locations (M1–M6), parasagittal anterior cerebral artery territory, and the basal ganglia compared with the contralateral side with a maximum score of 20.⁴² Another example, the collateral score, grades collaterals as either absent (=0), present in <50% of the downstream territory (=1), >50% (=2), or in 100% of the territory (=3). Scores of 0 to 1 are considered poor. Both the regional leptomeningeal collateral and collateral score have been shown to independently predict poor outcome (mRS score >2), infarct volume, and propensity for hemorrhagic transformation in the setting of an acute occlusion.^39,42

The aforementioned evaluations of collateral status use a single conventional CTA maximal intensity projection image (single phase). Multiphase CTA is a technique that can be used to increase temporal resolution of collateral blood flow and more accurately depict the process when compared with the opposite hemisphere. This technique uses the initial CTA acquisition followed in rapid succession with 2 additional acquisitions in the midvenous and late venous phases, for a total of 3 maximal intensity projections.⁴³ Scoring collateral status takes into account asymmetry of individual phases and washout of contrast, conveying a particular advantage to single-phase assessments. Multiphase CTA collateral assessment, in addition to the ASPECTS, was used to select patients in the recently successful ESCAPE trial and remains the only collateral score validated in a randomized controlled trial.⁴⁴ Newer CT scanners capable of whole-brain perfusion allow for time-resolved (ie, 4D) dynamic CTA that allows for a temporally fused maximal intensity projection.⁴⁵

CTA Source Images ASPECTS

CT angiogram source images can be rewindowed (≈W75/L40) and used for an ASPECTS-like calculation. It has been shown that CT angiogram source images ASPECTS is superior to the NCCT ASPECTS for both the anterior and posterior circulation, can predict final infarct volumes and functional outcomes, and closely represents core lesions on diffusion-weighted imaging (DWI) on magnetic resonance imaging.^46–49

Extracranial CTA

The ability to detect an acute symptomatic total occlusion versus 99% stenosis (pseudo-occlusion) of the extracranial carotid artery may make a substantial difference in immediate management. The ability of CTA to detect this subtle difference has good accuracy for detecting a string sign when compared with the standard of digital subtraction cerebral angiography.⁵⁰ The ability to distinguish dissection rather than traditional atherosclerotic occlusion may also alter acute management and long-term prognosis. The absence of calcification and the appearance of the flame sign may give clues as to the presence of a dissection.

DWI and Apparent Diffusion Coefficient

DWI allows for rapid and highly sensitive identification of the acute ischemic tissue with a sensitivity of 77% to 97% and specificity of 95% to 100%.^51,52 Sensitivity is lowest for lesions that are hyperacute, small, and located in the brain stem. DWI hyperintensity represents restricted diffusion of water molecules, which are felt to occur as the cytotoxic edema of ischemic tissue traps water intracellularly while increasing the tortuosity and reducing volume of the extracellular space limiting the free motion of water. The apparent diffusion coefficient maps are derived from DWI to give a quantitative assessment of diffusion restriction (ie, increased restriction decreases intensity). Correlating apparent diffusion coefficient hypointensity with DWI hyperintensity increases the specificity for acute ischemia by eliminating artifact created by the T2 component of the original diffusion image (ie, T2 shine through). Apparent diffusion coefficient hypointensity does return to normal between 4 and 7 days (ie, pseudonormalization) and may eventually become hyperintense in chronic lesions. DWI is typically considered the gold standard for identifying infarct core irrespective of reperfusion. There may, however, be some area of reversibility of the DWI lesion with early reperfusion, but not much was seen in both animal data and human case series.⁵³ Apparent diffusion coefficient threshold values of ≤620×10^–6 mm²/s have recently been identified as a reasonable (sensitivity 69%/specificity 78%) marker for irreversible infarction.⁵⁴

Fluid-Attenuated Inversion Recovery

The use of fluid-attenuated inversion recovery/DWI mismatch is currently being studied in randomized controlled trials to identify early (ie, <6 hour) lesions in patients with unknown time of onset. Fluid-attenuated inversion recovery hyperintensity occurs as vasogenic edema develops and is not widely present <6 hours from stroke onset. Patients with mismatch who would normally be excluded from therapy based on unknown onset time may potentially benefit. Vascular hyperintensities on fluid-attenuated inversion recovery can also be indicative of slow-flow areas within an ischemic lesion likely representative of collateral blood flow.⁵⁵ Given its particularly high sensitivity for blood products and gadolinium in the cerebral spinal fluid space, fluid-attenuated inversion recovery can often detect disruption of the blood–brain barrier and serve a marker of increased risk for hemorrhagic conversion, especially in the setting of reperfusion. This finding of peri-infarct postgadolinium cerebral spinal fluid enhancement has been termed hyperintense acute reperfusion marker.⁵⁶

T2* or Gradient Echo

Gradient echo may identify the presence of cerebral microbleeds. A recent meta-analysis found that the presence of ≥10 cerebral microbleeds presents a significant increased risk of sICH with IV tPA. This knowledge may provide more informed risk versus benefit analysis in certain patients who are candidates for thrombolytics.⁵⁷ Prior to this meta-analysis, the last statement from the AHA/ASA did not consider microbleeds as a contraindication to IV tPA (Class IIa/B recommendation).²¹

T1-Weighted Images and Fat Suppression

T1-weighted images–fat suppression imaging is often ordered if arterial dissection is considered the stroke etiology, either by history or by other angiographic imaging that suggests this process. T1-weighted images axial fat suppression images allow for amplified distinction between surrounding soft tissue of the neck and intramural hematoma, specifically hemoglobin that has transitioned to methemoglobin in the subacute phase.^58,59 The confirmation of a dissection alters long-term management and prognosis but may alter acute management decisions depending on the circumstances.

MR Perfusion Imaging

Two different methods exist for perfusion-weighted imaging. The first is dynamic susceptibility contrast imaging, which uses a bolus of gadolinium contrast and uses many of the same principles as CTP. The thresholds and definitions are similar to those of CTP (Table 3). The second method is arterial spin labeling, which uses no contrast, is less available, and less well studied but can provide quantitative CBF data.

Intravenous Thrombolysis

The first reports of IVT in ischemic stroke emerged in 1958,⁶⁷ but stroke remained a condition without a proven treatment until 1995 when the first positive randomized trial was performed by the National Institute of Neurological Disorders and Stroke with recombinant tPA.⁶⁸ The National Institute of Neurological Disorders and Stroke Group sponsored the first set of phase II pilot studies,^69,70 which delineated the final tPA dose (0.9 mg/kg, with 10% administered as a 1-minute bolus and the remaining 90% as an infusion over 60 minutes) that could be safely administered in the treatment of AIS, leading to the pivotal phase 3 National Institute of Neurological Disorders and Stroke trial.⁶⁸ This 2-part study randomized a total of 624 AIS patients presenting within 3 hours from symptom onset to receive either tPA or placebo in a double-blinded fashion. Although the first part failed to show any significant difference between groups in terms of clinical outcomes at 24 hours, part 2 revealed at least 30% greater likelihood to have minimal or no disability at 3 months in those patients who received tPA versus those who did not (global odds ratio 1.7, 95% confidence interval 1.2–2.6; P=0.008), despite an increased incidence of sICH in the treatment group (6.4% versus 0.6%; P<0.001). No mortality benefit was noted. These results led to the US Food and Drug Administration (FDA) approval of tPA in 1996 and remains one of the most important breakthroughs in the management of AIS.

Additional trials were performed subsequently that focused on pushing the limits of the IV tPA time window. ATLANTIS trial (Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke) Part A and B,⁷¹ EPHITET (Echoplanar Imaging Thrombolysis Evaluation Trial),⁷² and ECASS trials (European Cooperative Acute Stroke Studies) 1⁷³ and 2⁷⁴ were not successful. A subsequent meta-analysis published in 2004 confirmed the strong association between rapid treatment and favorable outcomes and suggested a potential benefit of tPA infusion ≤270 minutes after stroke onset.⁷⁵ In 2008, the ECASS 3 trial randomized 821 AIS patients to receive IV tPA versus placebo within 3 to 4.5 hour from symptom onset, demonstrating a modest but significantly higher rate of functional independence at 3 months (odds ratio 1.34, 95% confidence interval 1.02–1.76; P=0.04), while the risk of ICH was comparable with prior tPA trials.⁷⁶ Despite a subsequent robust meta-analysis supporting these findings,⁷⁷ the FDA failed to approve the extended therapeutic window beyond the 3-hour mark, leaving IV tPA for AIS patients presenting within 0 to 3 hours as the only FDA-approved pharmacological AIS treatment in the United States. The AHA/ASA guidelines, however, have recommended the extended time window ever since. The IST (International Stroke Trial)-3 randomized the largest cohort of patients (3035) to IV tPA versus placebo and showed that despite early hazards from an increased risk of ICH, thrombolysis within 6 hours improved functional outcomes even in patients aged >80 years.⁷⁸ Although some studies have suggested that genetic polymorphisms,^79,80 race, and sex may influence the efficacy of tPA,^81,82 only certain comorbidities such as hypertension and hyperglycemia have been found to significantly and negatively impact thrombolytic response in AIS patients.^83,84

Other Approaches to IVT

Although tPA is the mainstay of AIS therapy, it has several limitations, including a short half-life (4–8 minutes), low recanalization rate of 30% to 40% for proximal occlusions and of <5% for distal ICA occlusion,⁸⁵ effects on the blood–brain barrier and neurotoxicity, risk of ICH, time-sensitive administration, and a long list of absolute contraindications. The shortcomings of tPA have stimulated researchers to explore other thrombolytic agents. Clinical trials of tenecteplase began in 1999 after promising results in the acute myocardial infarction ASSENT-2 trial (Anglo-Scandinavian Study of Early Thrombolysis) emerged.⁸⁶ Early phase IIB^87,88 and a pilot studies’⁸⁹ data reported tenecteplase safety for the treatment of AIS within 3 hours at smaller doses compared with those used in cardiology, as well as its efficacy with increased early reperfusion rates measured by perfusion/diffusion scans and improved functional outcome (National Institute of Health Stroke Scale at 24 hours) when compared with tPA. These led to the ongoing phase III trials, TASTE (Tenecteplase Versus Alteplase for Stroke Thrombolysis Evaluation Trial) and NOR-TEST (The Norwegian Tenecteplase Stroke Trial).^90,91 Desmoteplase, a fibrin-specific thrombolytic drug with a longer half-life and less neurotoxicity compared with tPA, was initially tested in 2 phase II trials (DIAS [Desmoteplase in Acute Ischemic Stroke]⁹² and DEDAS [Dose Escalation of Desmoteplase for Acute Ischemic Stroke]⁹³) that documented reperfusion efficacy within a 9-hour window and safety at lower doses. Unfortunately, the subsequent phase III studies (DIAS-3⁹⁴ and DIAS-4⁹⁵) failed to show efficacy. Finally, the combination of tPA with other treatment modalities such as hypothermia, epifibatide, activated protein C, and extend alteplase therapeutic window ≤9 hours using novel imaging modalities are being studied with the hope of improving IV thrombolytic therapy in this dynamic, time-sensitive, and complex disease process.

Endovascular Therapy

EVT for AIS has undergone a major evolution since its origins back in the 1980s when it was purely based on the IA administration of thrombolytic agents directly into the clot. Potential benefits from this treatment were shown initially in nonrandomized studies, but it was not until the 1990s that the PROACT-I trial randomized patients with angiographically documented M1 and MCA–second segment occlusions to receive either IA recombinant prourokinase or placebo within 6 hours from stroke symptom onset.⁹⁶ IA recombinant prourokinase was associated with higher recanalization rates than placebo without a significantly increased incidence of ICH. A subsequent phase III study (PROACT-II) randomized 180 patients with AIS of <6 hours’ duration and angiographically proven proximal MCA occlusion to receive IA recombinant prourokinase plus heparin versus heparin alone.⁹⁷ Despite an increased incidence of early sICH (10.2%), patients treated with recombinant prourokinase had higher recanalization rates (66% versus 18%; P<0.001) and were significantly more likely to be independent at 90 days (40% versus 25%; P=0.04). The positive results of this trial stimulated the exploration of other EVT recanalization approaches. Some of the proposed approaches included microwire manipulation of the clot, balloon angioplasty, manual aspiration of the clot, and the combination of IV, IA tPA, and low energy ultrasound, but none of them succeeded.

Initial EVT Approaches

The era of clot retrievers was launched with the FDA approval of the MERCI (Mechanical Embolus Removal in Cerebral Ischemia) retriever in 2004. This approval was supported by the MERCI and multi-MERCI trials, a set of prospective, nonrandomized, multicenter, and single-arm trials designed to test the safety and efficacy of this device in large-vessel AIS patients with ≤8 hours of stroke symptoms.^98,99 Successful recanalization rates (Thrombolysis in Myocardial Infarction score 2 and 3) were lower than expected, with percentages ranging from 48 to 57 with MERCI only and 60 to 69 with MERCI in combination with adjunctive therapies. Although the proportion of favorable clinical outcomes tended to be higher and mortality rates lower with better degrees of recanalization, no significant difference was found when compared with controls from PROACT-2.⁹⁷ The FDA cleared a second first-generation device for AIS, the Penumbra Clot Aspiration system, in 2008 based on the results of the Penumbra Pivotal Stroke Trial.¹⁰⁰ This prospective multicenter single-arm study evaluated the safety and effectiveness of the penumbra system in the revascularization of large-vessel AIS patients presenting within 8 hours from symptom onset. Thrombolysis in Myocardial Infarction 2 or 3-reperfusion was achieved in 81.6% of patients without an increased risk of sICH (11.2%) when compared with historical controls, documenting both device efficacy and safety. However, despite the higher revascularization rates, only 27.7% were clinically independent at 90 days compared with 40% in the control arm of PROACT II.

Stent Retrievers

Uncertainty as to whether such recanalization could translate in better neurological outcomes in AIS led to the development of second-generation devices, the stent retrievers. The 2 main retrievable stent systems developed are the Solitaire (Medtronic) and Trevo (Stryker Neurovascular), both FDA approved in 2012. Unlike the detachable stents widely used in cardiology, the stent retrievers are characterized by being nondetachable, allowing for stent deployment within the clot for thrombus embedment and quick restoration of flow and subsequent clot removal. The safety and efficacy of this new technology was rapidly tested in several randomized trials that compared it directly to the first-generation devices.^101–105 The recanalization rates in these trials ranged from 61% to 87.5% for Solitaire and 91.7% to 92% for Trevo.¹⁰⁶ Both of these devices demonstrated high rates of good clinical outcomes as measured by the mRS at 90 days, 40% and 58% for Trevo and Solitaire, respectively.^101,104 Concomitantly, several landmark endovascular trials tested first- and, to a lesser extent, second-generation devices versus medical therapy in large populations of IV tPA–eligible patients. The IMS-III trial (Interventional Management of Stroke III) enrolled 656 participants to IV tPA within 3 hour from stroke onset with or without additional EVT ≤7 hours. As CTA and newer generation devices became more readily available in stroke centers, these technologies were implemented later in the course of the study to improve patient selection and treatment. Unfortunately, premature halting of the trial occurred, and no significant difference in clinical outcomes at 90 days was found across treatment groups (38.7% for IVT only and 40.8% IVT and EVT, P=0.25). SYNTHESIS (Synthesis Expansion: A Randomized Controlled Trial on IA Versus IV Thrombolysis in Acute Ischemic Stroke) randomized 362 AIS patients within 4.5 hours from symptom onset to receive IVT or EVT, which included IA tPA, mechanical clot disruption, or retrieval or a combination of these. At 3 months, only 30.4% of patients in the EVT group versus 34.8% treated medically had little or no disability (adjusted odds ratio 0.71; P=0.16). MR-RESCUE (Mechanical Retrieval and Recanalization of Stroke Clots using Embolectomy) randomized patients with anterior circulation LAO presenting within 8 hours from onset of symptoms to receive either EVT or IVT. Patients were also stratified based on favorable or nonpenumbral pattern. The recanalization rates were low, likely related to the use of first-generation devices, and no significant difference in functional outcomes at 90 days was appreciated, even after stratification. Although valuable preliminary data for the assessment of EVT in AIS was obtained, these trials had several limitations, including treatment delays, unstructured workflow, minimal use of stent retrievers, and lack of imaging-based selection to prove LAO, as well as to exclude patients with large areas of irreversible brain damage. As a consequence of these trials, pessimism spread in the stroke community.

Second-Generation Stent Retriever Trials

Several new trials were initiated, and the first to be completed in 2014 was MR CLEAN.¹⁰⁷ In this study, 500 patients with documented anterior circulation LAO on CTA presenting within 6 hours from stroke onset were enrolled to receive EVT or usual care alone. Stent retrievers were used in 81.5% patients, and tPA was administered in 90.6% of patients assigned to the EVT group and control group, respectively. An absolute difference of 13.5% points (adjusted odds ratio 1.67) in the rate of mRS score 0 to 1 at 90 days in favor of intervention was found, whereas mortality and sICH rates did not significantly differ between groups. The positive results precipitated an early interim analysis of the remaining 4 EVT trials: ESCAPE,⁴⁴ REVASCAT,¹⁰⁸ EXTEND-IA trial (Extending the Time for Thrombolysis in Emergency Neurological Deficits–Intra-Arterial),¹⁰⁹ and SWIFT-PRIME,¹¹⁰ which were subsequently prematurely stopped because of positive outcomes. Uniquely, the ESCAPE and EXTEND-IA trials used pre-enrollment imaging to identify a favorable penumbral pattern. Although the ESCAPE investigators used multimodal CT to estimate the degree of adequate collateral circulation as a marker of salvageable penumbra, the EXTEND-IA group used CTP analyzed by the Rapid Processing of Perfusion and Diffusion software. Overall, median onset to groin puncture ranged from 200 to 260 minutes, with a successful revascularization rate of 59% to 88%. A 50% reduction in mortality was noted in the EXTEND-IA and ESCAPE trials, although this finding was not universal (Table 2). The HERMES trial (Highly Effective Reperfusion evaluated in Multiple Endovascular Stroke), a recently published patient-level meta-analysis including 1287 individuals from the 5 EVT landmark trials, confirmed the stunning results. Disability at 90 days was significantly reduced compared with the control group (cOR 2.49, 95% confidence interval 1.76–3.53; P<0.0001). The number needed to treat to reduce disability by at least one level on mRS for each patient undergoing EVT is 2.6, one of the largest effect sizes across all disciplines in medicine. Benefit was irrespective of patients characteristics (elderly, time from onset to randomization >300 min, individuals not receiving recombinant tPA) and mortality, and sICH did not differ between groups.¹¹¹ Two additional EVT trials (THERAPY [The Randomized, Concurrent Controlled Trial to Assess the Penumbra System’s Safety and Effectiveness in the Treatment of Acute Stroke] and THRACE [Trial and Cost Effectiveness Evaluation of Intra-arterial Thrombectomy in Acute Ischemic Stroke])^112,113 with promising results presented at the European Stroke Conference in 2015 remain to be published, and others (POSITIVE [Perfusion Imaging Selection of Ischemic Stroke Patients for Endovascular Therapy] and DAWN [Diffusion Weighted Imaging or Computerized Tomography Perfusion Assessment With Clinical Mismatch in the Triage of Wake Up and Late Presenting Strokes Undergoing Neurointervention]) are ongoing with the hope to extend the therapeutic time window ≤12 and 24 hours, respectively. On the basis of available trials data, IV tPA remains the initial treatment for AIS patients presenting within 4.5 hours from symptoms onset. If an LAO is documented by expedited imaging while the ischemic core remains small, immediate EVT should be pursued ≤6 hours from stroke onset. Overall, <10% of patients with AIS receive stroke recanalization therapies in the United States, even in communities with highly organized stroke centers.¹¹⁴ Some challenges to treatment include small-vessel strokes, strokes in evolution, severe strokes from large-vessel occlusion, and the so-called wake-up strokes. Patients who wake up with stroke symptoms or wake-up strokes represent an undertreated subgroup of AIS patients (≈20%) who are generally excluded for recanalization therapies because of the unknown time of symptom onset.¹¹⁵ A series of DEFUSE trials (Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution)^116,117 have explored the role of perfusion imaging in assessing reversibility of brain ischemia in patients who present at a later time window. DEFUSE-3 is currently enrolling selected AIS patients ≤16 hours from symptom onset. For those stroke patients who are unable to receive any acute IV or IA therapies, a pivotal element for their care are the so-called stroke units. These organized inpatient stroke wards are staffed by a multidisciplinary team of experts specialized in stroke care and are dedicated to care for AIS patients. Individuals who receive care in stroke units are more likely to be independent, alive, and living at home at 1 year when compared with those who are admitted to alternative forms of care while the duration of hospital stays remain unchanged.¹¹⁸ Although AIS therapies had lagged behind those in cardiology, AIS treatment has finally gained momentum.

New AIS Therapy Guidelines

The recent influx of randomized trials has forced a recent AHA/ASA AIS guideline update, which took place in December 2015 for IV thrombolysis and in June 2015 for EVT.^21,24 The EVT AHA/ASA Guidelines recommend pursuing EVT with stent retrievers in all AIS patients who present within 6 hours from symptom onset, are ≥18 years of age, have minimal or no disability at baseline (prestroke mRS score 0–1), have a National Institute of Health Stroke Scale score and ASPECTS ≥6 on arrival, and a documented LVO on admission angiogram (Class I). The guidelines highlight the need for expedited treatment to ensure EVT benefit but recommend against either withholding IV tPA while EVT is being considered or prolonged observation of potential EVT candidates after IV tPA infusion to assess for clinical improvement. As with IV tPA, the EVT guideline is likely to become more inclusive with time, in particular with the advancement of perfusion imaging and devices. The main recommendations for EVT in AIS are summarized in Table 4.

Organizing and Implementing Systems of Care

The recently proven efficacy of IA therapy with stent retrievers and other devices is the most important advance for AIS treatment in 20 years. To best deliver IA therapy to as many AIS patients as possible will require careful planning, and the best approach will vary among cities, regions, and states depending on many factors. For the foreseeable future, the availability of IA device therapy will remain restricted to larger tertiary medical centers because the number of interventionalists capable of performing the procedure is limited, the establishment of an endovascular treatment center requires resources that many smaller hospitals will not have available, and as with many other procedures, there seems to be a relationship between successful performance and case volume.¹¹⁹ Regional AIS care delivery paradigms will need to be established based on available resources, including the number of endovascular centers, the number and distribution of primary stroke care centers where the initial evaluation and treatment of AIS patients with IV tPA can be done, and the local emergency medical transport system. A key question is what should be the preferred destination for AIS patients with an LAO appropriate for IA device therapy when both primary stroke centers and tertiary centers can both be reached relatively quickly.¹²⁰ In such a setting, it would be useful for ambulance crew to assess the severity of the AIS and the LAO likelihood by using currently available stroke rating scales and if it is determined that the stroke is likely caused by LAO to proceed directly to the endovascular center.¹²¹ The ambulance crew can be aided by developing technologies, such as telemedicine, that would allow them to interact with a physician remotely who can see the patient with them and help to determine stroke severity by guiding them through an examination such as the National Institute of Health stroke scale.¹²² Another developing technology is mobile stroke units that contain a CT scan and sophisticated personnel who interact with a hospital base station.¹²³ Currently, head CT scans can be performed and IV tPA can be started quickly in such units. It is unlikely that they will be widely available. A balance must be considered between choosing to route the patient to a primary stroke center and starting therapy with IV tPA with an inherent delay in reaching an endovascular center for patients with LAO versus bypassing the primary stroke center and proceeding directly to the endovascular center.¹²⁴ Factors that will need to be considered by the EMS system beyond distance and transport time include the rapidity of clinical and imaging evaluation at the primary stroke center, the door-to-needle time, and the door in door out time. If these time metrics are excessive, then routing patients with suspected LAO to an endovascular center and not to a closer primary stroke center may be the appropriate course of action. The development of regional AIS care plans will need to be done individually in different locales, but the basic concept will be to get the patients to the most appropriate treatment center as quickly as possible based on the likelihood of LAO and the other factors that will influence routing decisions.

For centers performing IA device therapy, it be incumbent on them to try to emulate the work flow paradigms as performed in the clinical trials so that appropriate AIS patients are treated as quickly as possible.²⁵ The best approach to imaging remains to be established but should at least include a head CT scan, CTA, and in some centers CTP.¹²⁵ A large ischemic core can be identified by the CT ASPECTS score, excluding patients with scores of ≤5 as was done in the clinical trials or a large ischemic core volume on CTP. Important time metrics that endovascular treatment centers should monitor will include time from emergency department arrival to imaging, time from imaging to start of the endovascular procedure, and time to reperfusion. The trials with the best outcome results also performed the best regarding these treatment metrics. Clinical centers will not replicate the trial results if they do not include patients similar to those in the trials and fail to establish reperfusion in a timely manner.

In other situations, the distance to an endovascular center is much greater, and all AIS patients will initially have to be taken to a primary stroke center or smaller hospital for initial assessment and treatment. Again the use of telemedicine technology will be useful because stroke expertise at a large center can be used to help evaluate the patient and make treatment decisions at the outlying medical center. Currently, most smaller hospitals do not routinely perform CT angiography, but now that EVT is of proven value for proximal vessel occlusions, the availability of CTA will need to increase because a vessel occlusion potentially amenable to endovascular treatment should initiate the rapid transfer of such patients to a center capable of performing the procedure. The telemedicine consultant can help to evaluate the CTA and the head CT scan performed at the outlying hospital to help local personnel decide if there is a proximal vessel occlusion amenable to IA therapy if the ASPECTS score on the head CT scan is compatible with a small to moderate ischemic core, supporting the suitability of the patient for IA therapy. For patients who are not candidates for IA therapy, transfer to a tertiary medical center may not be necessary in many cases, and the patient can then be managed locally with IV tPA in some cases. A transfer if deemed to be appropriate can be done either by helicopter or ambulance, but a major concern is that if the transfer will take several hours and that by the time the patient arrives at the endovascular center, the ischemic core will have enlarged to such an extent that the patient will no longer benefit from IA therapy. From a system-of-care perspective, patients who may be appropriate for IA therapy will need to be rapidly identified and transport times minimized as much as possible so that as many patients as possible will remain candidates for IA therapy. As will be discussed, the development of therapies to impede ischemic core growth that could be used during transfer is an exciting possibility that may substantially increase the percentage of AIS patients subject to long transfer times who might still benefit from IA therapy.