I m a g i n g o f Ac u t e S t rok e

Current State
Steven Hetts, MDa, Rajkamal Khangura, MDb,*

KEYWORDS
 Acute ischemic stroke  Endovascular therapy  Thrombectomy  Computer tomography perfusion
 Neurointervention  Large-vessel occlusion

KEY POINTS
 Endovascular therapy for large-vessel occlusions may now be performed up to 24 hours after time
last known well.
 Stroke systems of care should be protocolized to allow for rapid, efficient stroke imaging to allow for
faster revascularization.
 Efficient, rapid neuroimaging differentiates acute ischemic stroke from stroke mimics, identifies
intracranial and extracranial occlusions, and delineates cranio-cervical vascular anatomy, extent
of core infarction, volume of salvageable brain (ischemic penumbra), and status of leptomeningeal
collaterals.

INTRODUCTION treatment benefit in nearly all subgroups and iden-

tified time to treatment as the primary predictor of
Acute ischemic stroke (AIS) affects up to 1 in 6 good outcome (modified Rankin Scale score <2).5
adults in their lifetime and is the leading cause of Diagnostic neuroimaging is essential in triaging
long-term severe disability in North America, with stroke patients for intravenous thrombolysis ther-
nearly 800,000 cases of AIS in the United States apy inclusion and, if LVO is identified, proceeding
annually. Ischemic stroke represents most acute to endovascular therapy. Efficient, rapid neuroi-
stroke events (87%), whereas hemorrhagic stroke maging differentiates AIS from stroke mimics,
represents a minority of cases (13%).1 Approxi- identifies intracranial and extracranial occlusions,
mately 30% of ischemic strokes involve acute delineates cranio-cervical vascular anatomy,
large-vessel occlusion (LVO) of the intracranial ar- extent of core infarction, volume of salvageable
teries, predominantly involving the anterior circula- brain (ischemic penumbra), and status of leptome-
tion (internal carotid artery, middle cerebral artery, ningeal collaterals. This article discusses the cur-
or anterior cerebral artery).2 Overwhelmingly posi- rent state of acute stroke imaging, focusing on
tive results from 5 randomized controlled trials imaging findings that have an impact on endovas-
published in 2015 supported the use of endovas- cular stroke treatment.
cular therapy (EVT) in LVO within the first 6 to
8 hours of stroke onset. More recent trials using SYSTEMS OF STROKE CARE
advanced imaging to select patients with salvage-
able brain tissue have extended the LVO treatment The mantra “time is brain” encapsulates the goal
window to 18 to 24 hours.3,4 Meta-analysis of the of stroke intervention systems of care—stream-
major trials, the HERMES collaboration, confirmed lined diagnosis and rapid revascularization. Faster
radiologic.theclinics.com

No disclosures.
a
Division of Neurointerventional Radiology, University of California, San Francisco, 505 Parnassus Avenue #
M-380, San Francisco, CA 94143, USA; b Department of Radiology and Biomedical Imaging, Division of Neuro-
interventional Radiology, University of California, San Francisco, 505 Parnassus Avenue # M-380, San Francisco,
CA 94143, USA
* Corresponding author.
E-mail address: rajkamal.khangura@ucsf.edu

Radiol Clin N Am 57 (2019) 1083–1091

https://doi.org/10.1016/j.rcl.2019.07.009
0033-8389/19/Ó 2019 Elsevier Inc. All rights reserved.
1084 Hetts & Khangura

diagnosis and recanalization of LVOs are associ- time. In this scenario, a CT and potentially CTA/CTP
ated with greater likelihood of a disability-free are acquired using cone-bean CT technology within
outcome. Clinical suspicion of AIS, activated in the angiography suite. A pilot study in Barcelona
the field or the hospital, results in a cascade of demonstrated the feasibility and significant time
events often with imaging as the rate-limiting reduction using this strategy in patients with high
step. Once prehospital or in-hospital stroke activa- prehospital stroke scale scores.7
tion occurs, providers and staff involved in stroke
treatment should also be notified; this includes NEUROIMAGING
the neuroradiologist. Initial noncontrast computed Noncontrast Head Computed Tomography
tomography (CT) interpretation, either at the scan-
ner or in the reading room, begins the treatment AIS remains a clinical diagnosis although treat-
algorithm for intravenous thrombolytic therapy ment is centered around exclusionary criteria ob-
and subsequent endovascular therapy. After tained from patient history and diagnostic
noncontrast head CT, computed tomography imaging. CT imaging is the cornerstone of routine
angiography (CTA) covering from the left atrial ischemic stroke imaging, followed by CTA and
appendage through the cranial vertex should be advanced imaging such as CTP or multiphase
obtained. If the patient is beyond 6 hours from CTA in patients who present during the extended
time last seen normal, then computed tomography treatment window. The noncontrast head CT
perfusion (CTP) should be performed to determine should be obtained from the vertex through C2,
extended window eligibility. Initial imaging should provided in true axial reconstructions and best
be protocolized to prevent delays associated viewed in 35-45/35-45 window/level settings. The
with imaging. Protocolized imaging provides the primary objective of the head CT is to exclude pa-
greatest capture of patients presenting with tients from intravenous (IV) thrombolysis; these
LVOs. Additional components of a stroke protocol exclusionary criteria include evidence of hemor-
CT should include pre-set slice thickness, multi- rhage, mass, and completed large infarction. In
planar reformats, and true axial images. the absence of these findings, the patient remains
CT with CTA remains the clinical standard to a candidate for IV thrombolysis within the appro-
evaluate for LVO because it was the primary priate treatment time (currently within 4.5 hours af-
noninvasive vascular imaging in the major trials, ter time last seen well in most centers). Additional
although some sites used MR imaging and mag- findings suggestive of acute ischemia, although
netic resonance angiography (MRA) with similar specific, are not sensitive. These include the
results and speed. According to the 2018 Amer- hyperdense vessel sign, loss of insular ribbon
ican Heart Association guidelines, the data to sup- sign, territorial gray-white differentiation loss, and
port the routine use of MR imaging for in-window rarely, sulcal effacement (seen toward the end of
thrombectomy candidates are inadequate, and the acute ischemic phase and more commonly
CT is preferred. In select patients who present in as a subacute ischemic finding).
the extended time window, diffusion-weighted Major findings that should be evaluated include
MR imaging provides definitive evaluation of the 1. Evidence of hemorrhage, mass or mass effect,
core infarct volume and perfusion MR imaging and hydrocephalus
characterizes tissue mismatch/tissue at risk.6 2. Evidence of gray-white differentiation loss and
Use of MR imaging with MRA in this scenario assessment of Alberta Stroke Program Early
must be balanced with rapid imaging acquisition Computed Tomography Score (ASPECTS) to
and diagnosis to include patients for endovascular characterize core infarction
treatment. 3. Hyperdense vessel sign
In efforts to dramatically lower “door to tissue
plasminogen activator” time and treat a greater Following the exclusion of stroke mimics and
percentage of patients with early thrombolysis, hemorrhage, the scan should be assessed for ev-
mobile stroke units have been developed; these idence of core infarction, best assessed with AS-
are modified ambulances with an on-board CT PECTS. The total volume of the middle cerebral
scanner, technologist, critical care nurse, emer- artery territory is roughly 300 cm3, divided into 10
gency medical service provider/paramedic, and anatomic regions, with each area representing
neurologist. Patients are scanned using the mobile approximately 30 cm3 of brain tissue (Fig. 1). All
CT scanner to evaluate for exclusionary criteria for slices from the lower ganglionic level through the
thrombolysis. cephalad aspect of the supraganglionic level
Lastly, direct transfer to the angiography suite should be assessed. Each area of hypoattenua-
is being evaluated, bypassing the emergency tion/ischemic change/core infarction results in
room and CT scanner to reduce “door-to-puncture” subtraction of 1 point for each involved region
 Imaging of Acute Stroke 1085

Fig. 1. Ten regions with 4 for ganglionic

gray-matter structures: caudate (C); len-
tiform (L); internal capsule (IC); insular
ribbon (I); and 6 for cortically based
structures in the MCA territory: M1–
M6.8 (From Schröder, J., & Thomalla, G.
(2017). A Critical Review of Alberta
Stroke Program Early CT Score for
Evaluation of Acute Stroke Imaging .
Frontiers in Neurology . Retrieved from
https://www.frontiersin.org/article/10.
3389/fneur.2016.0024; with permission.)

from a normal score of 10. Level 1 evidence sup- LVO, because this is amenable to EVT. Again, this
ports treatment of patients with ASPECTS greater imaging should be acquired immediately after the
than 6. noncontrast CT head as part of the protocolized
The dense vessel sign is suggestive of higher Stroke CT with CTA study.
red blood cell (RBC) content within thrombus,9
which seems to be secondary to a cardioembolic Computed Tomography Angiography
source of embolism (Fig. 2). Less dense thrombi
may be suggestive of fibrin-rich, platelet-rich A CTA of both the head and neck should then
emboli, which typically arise from large-vessel be performed, extending from the left atrial
atherosclerotic sources.10 Lastly, a calcified appendage through the vertex, and reconstructed
embolus may be secondary to a chronically in multiplanar reconstruction and maximum inten-
formed clot from a cardioembolic source or sity projection (MIP) images for evaluation of the
from an artery to artery embolus. Because these origin of the great vessels, carotid disease or nar-
factors may affect the potential devices used rowing, and the intracranial vasculature.11 Given
during thrombectomy, this information should be that time is of the essence, imaging is obtained
promptly conveyed to the neurointerventionalist. without evaluating the patient’s renal status.
Patients who receive IV thrombolysis should then Treatment planning should be considered when
be evaluated with vessel imaging for evidence of evaluating the CTA. Characterization of the aortic
arch, patency of the independent origins of the
great vessels as well as common origins should
be reported (Fig. 3). The course and tortuosity of
vessels should be described. Evaluation of the
distal common carotid artery (CCA), internal ca-
rotid artery (ICA) bulb and proximal ICA should
be documented to establish the potential source
of stroke. Most strokes are cardioembolic, fol-
lowed by artery to artery emboli from the carotid
artery.
Intracranially, the length of occlusion should be
described, because it may aid in device selection.
In describing the middle cerebral artery (MCA) M1
segment, it should be noted that an early bifurca-
tion of the MCA is considered an M1 location as
long as it is within the horizontal segment of the
sylvian fissure. Thrombus involving the bifurcation
may involve the superior or inferior truck, repre-
senting M2 occlusions.
Occlusions are best seen on MIP slab images,
Fig. 2. Hyperdensity of the left middle cerebral artery and reconstructions should always be provided
M1 segment. as part of the standard stroke protocol.
1086 Hetts & Khangura

populations was that, based on the HERMES

pooled patient analysis, a subset of patients
benefit from treatment in the extended window
time period. This ability to maintain a small
core volume despite late presentation is
most likely explained by robust leptomeningeal
collaterals.
Advanced imaging combined with automated
processing may alert physicians to patients with
undiagnosed LVO in the early window. Systems
should be streamlined to reduce time to allow for
rapid detection, diagnosis, thrombolysis, and
transfer if LVO present. Protocolized studies for
CTP in the extended window should be initiated
according to the particular institution’s policy,
either up to 16 hours (DEFUSE 3) or up to 24 hours
(DAWN).
Multiphase acquisition of head CTA to charac-
terize leptomeningeal collaterals has been devel-
Fig. 3. Coronal CTA reconstruction demonstrates a
common origin of the innominate artery and left
oped and pioneered by the Calgary/Alberta
common carotid artery. group (ESCAPE Trial).13 Patients are characterized
as having poor, intermediate, or robust collaterals,
which can aid patient selection for the extended
Extending the Treatment Window for Patients
window patient.
with Large-Vessel Occlusion
Perfusion imaging is intended to identify
Analysis of the 5 major trials revealed that patients with salvageable brain tissue (ischemic
patients can be stratified into 2 categories: penumbra). The inclusion criteria of the 2
patients with rapid expansion of the ischemic randomized control trials, DEFUSE 3 and
core (so-called fast progressors) and patients DAWN, evaluating endovascular stroke treat-
with slow growth of their ischemic core (so- ment in the extended window are shown in
called slow progressors). Their outcome seems Table 1.
to be determined primarily by their collaterals Briefly, CTP is a time-resolved multiphase
(Fig. 4).12 acquisition following a contrast bolus injection.
Two recent studies evaluating the role of Reconstructed images are then post processed
thrombectomy in patients beyond the 6 hour to generate color maps of the mean transit time
time window demonstrated unequivocal benefit (MTT), time to peak (TTP), cerebral blood flow
of mechanical thrombectomy between 6 and (CBF), and cerebral blood volume (CBV) parame-
16 hours (DEFUSE 3) and up to 24 hours ters. The best estimation of the infarct core is an
(DAWN).3,4 The principle behind these study area of tissue with approximately 30% CBF

Fig. 4. The dashed lines reflect the

95% confidence interval range as
well as reflecting patients with robust
collaterals (upper dashed line) or low
collaterals (lower dashed line), sup-
porting both fast treatment and EVT
for extended window patients with
good collaterals and low core infarc-
tion. (From Saver JL, Goyal M, van
der Lugt A, et al. Time to treatment
with endovascular thrombectomy
and outcomes from ischemic stroke:
a meta-analysis endovascular throm-
bectomy and outcomes in ischemic
stroke endovascular thrombectomy
and outcomes in ischemic stroke.
JAMA 2016;316(12):1279-89; with
permission.)
 Imaging of Acute Stroke 1087

Table 1
The inclusion criteria of the 2 randomized control trials, DEFUSE 3 and DAWN

DEFUSE 3 DAWN
Age and baseline Age: 18–90 years Age: At least 18 years
mRS Baseline mRS: 0–1 Baseline mRS: 0–1
IV thrombolysis tPA if started within 4.5 h of onset tPA given with persistence of
LVO or outside of tPA window
Time since last 6–16 h 6–24 h
known well
Imaging ASPECTS 6 No intracerebral hemorrhage on
CT or MR imaging
No evidence of stroke in >1/3 of
MCA territory
Mismatch Initial core infarct volume <70 cm3 Group A: >80 years
inclusion criteria NIHSS 6 NIHSS  10, core infarct <21 cm3
Ratio of volume of ischemic tissue to Group B: <80 year, NIHSS 10, core
initial infarct volume of 1.8 or more infarct <31 cm3
Absolute volume of penumbra tissue of Group C: <80 years, NIHSS 20, core
15 cm3 or more (defined as Tmax >6 s infarct 31–50 cm3
using RAPID software)
Occlusion location ICA, MCA, or both ICA, MCA, or both

Abbreviations: ICA, intracerebral hemorrhage; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke
Scale.

compared with normal brain tissue.14 Tissue at PEARLS, PITFALLS, VARIANTS

risk or ischemic penumbra is characterized by Intracranial Atherosclerotic Disease
prolongation of MTT or delayed TTP/Tmax. The
Acute plaque rupture of intracranial atheroscle-
mismatch between areas of deceased CBF or
rosis may present similar to cardioembolic or ar-
CBV compared with prolongation of a time
tery to artery embolic stroke but typically does
parameter reflects salvageable brain tissue
not respond to mechanical thrombectomy or
(Fig. 5). This assessment forms the basis of
thromboaspiration.15 Treatment with a stentriever
treatment of patients in the extended treatment
results in further platelet aggregation of the
window.

Fig. 5. Generated CTP threshold images demonstrate suspected core infarct (purple) and ischemic tissue at risk
(green), representing a favorable mismatch profile. (Courtesy of Ischemaview RAPID.)
1088 Hetts & Khangura

Fig. 6. (A) Sagittal CTA of the neck shows decreased opacification with tapered appearance of the proximal left
ICA. (B) Post-contrast axial CT of the head in the same patient reveals a patent distal left cervical ICA, suggesting
the initial nonopacification of the left cervical ICA represented a “pseudo-occlusion.”

thrombus, and patients may ultimately require cervical vertebral level, tortuosity of the vessel,
intravenous antiplatelet treatment, balloon angio- length of stenosis, and degree of peripheral calci-
plasty, or intracranial stenting. The noncontrast fication should be reported to help determine
CT may suggest intracranial atherosclerotic eligibility for carotid endarterectomy versus
disease (ICAD) with evidence of calcium deposi- carotid stenting.
tion in the vessel wall or absence of a hyperdense
vessel sign. Common locations of intracranial Tandem Occlusions
atherothrombotic occlusions are in the mid to
Tandem occlusions consist of concomitant intra-
distal MCAs and mid basilar artery.16 Evidence
cranial LVO and ipsilateral extracranial ICA occlu-
of other sites of intracranial atherosclerosis
sion, occurring in approximately 10% to 20% of
(excluding the carotid siphons) may suggest
cases of AIS.18 Treatment is often complex, influ-
ICAD. CTP and multiphase CTA may reveal well-
enced by the ability to cross the proximal occlu-
developed collaterals and a favorable mismatch
sion, the volume of the core infarction, the
profile, likely reflective of the chronicity of the
degree of success of intracranial revascularization,
underlying lesion.
adequacy of circle of Willis collaterals, and extra-
cranial lesion morphology. Common causes of
Chronic Extracranial Occlusion/Near-Total
tandem occlusion include acute occlusion of a
Occlusion
previously nearly occluded common/internal ca-
Chronic occlusions of the extracranial circulation rotid artery, acute plaque rupture of an ICA athero-
pose clinical challenges, because there is no intra- sclerotic lesion with distal embolization, or carotid
cranial LVO but upstream narrowing or occlusion dissection with thrombus formation and emboliza-
with greater stress placed on the circle of Willis tion. First-pass CTA limits the evaluation of a
or leptomeningeal collaterals. In general, recent tandem lesion, but delayed imaging may help eval-
extracranial carotid occlusions are not typically uate vascular patency, vessel contour, distal
treated, although revascularization may be collapse of the ICA, string sign, or filling defect
appropriate in certain clinical scenarios (increasing suggestive of an ulcerated plaque. During the arte-
NIHSS, recurrent strokes). Symptomatic ICA rial phase of acquisition, there may be absence of
occlusion requires revascularization within opacification of the distal CCA or proximal ICA.
14 days of the stroke.17 The degree of stenosis, This is typically explained by 2 potential pro-
the location relative to both the mandible and cesses: “pseudo-occlusion” or tandem occlusion.
 Imaging of Acute Stroke 1089

Pseudo-occlusions of the cervical carotid occur

with occlusion of the ICA terminus or in an isolated
hemisphere with MCA M1 occlusion (Fig. 6). Poor
vascular opacification of the cervical ICA occurs
because there are no branches of the cervical
ICA; therefore poor antegrade flow results in
poor/incomplete opacification during the arterial
phase of imaging.
A post-contrast scan may be obtained to eval-
uate for delayed filling of the extracranial vascula-
ture, which may appear to be occluded, hence
the so-called pseudo-occlusion on the first-pass
angiographic phase of imaging.

Malignant Pattern of Collaterals

Patients with poor collaterals have poor opacifica-
tion of vessels in the occluded territory, which is
associated with more rapid core infarction
expansion and poorer prognosis (Fig. 7).19
Typically, some degree of arterial and venous
filling is seen as a result of collateral arterial filling
from anterior cerebral artery–MCA or posterior
cerebral artery–MCA leptomeningeal collaterals.
Patients with a so-called malignant pattern of
collaterals will have a rapidly expanding infarct
core unless the LVO is opened. For these patients,
time is of the utmost importance. This should
be promptly recognized and relayed to the
neurointerventionalist.

Basilar Artery Occlusions

Posterior circulation LVOs, specifically basilar ar- Fig. 7. Subtracted collapsed MIP axial images of the
tery occlusions, were not included in the major circle of Willis reveal a right ICA terminus occlusion
randomized studies. To date, there are no ran- with little contrast opacification of the right MCA ter-
domized studies evaluating modern era endovas- ritory, suggestive of poor collaterals, a malignant
cular therapy devices for treatment of basilar collateral pattern. (Courtesy of Ischemaview RAPID.)
occlusions. Retrospective data suggest that endo-
vascular therapy is safe and effective for basilar
occlusions.20 Occlusion of the proximal and mid- for thrombolysis, such as hemorrhage, mass, or
dle segments of the basilar artery are more likely hydrocephalus. In the absence of these exclu-
to reflect underlying ICAD.21 Distal occlusions sionary imaging findings and other patient factors
and evidence of thalamic infarctions are more determined by their medical history, intravenous
often seen with embolic stroke. Guideline recom- thrombolysis can be administered.
mendations from the Society of Neurointerven- A dense vessel sign is highly specific for LVO,
tional Surgery recommend imaging evaluation specifically, an RBC-rich thrombus. Measurement
initially with CT/CTA with possible diffusion- of the thrombus length provides some clinical
weighted imaging for late arrivals, as institutionally value to size the stentriever or rule out primary
defined, to exclude those with a poor pc- aspiration (which is less favorable for longer
ASPECTS (posterior circulation ASPECTS) or large thrombi). Ischemic change should be reported as
core infarction.22 part of the patient’s ASPECTS, which can be re-
ported as a range (eg, ASPECTS of approximately
WHAT THE REFERRING PHYSICIAN NEEDS TO 8–10). CTA establishes the LVO and in the
KNOW absence of extensive early ischemic change
(ASPECTS <6), patients are candidates for EVT.
Noncontrast CT imaging should provide exclu- Neurointerventionalists benefit from learning
sionary reasons why patients are not candidates about aortic arch anat6omic variants, cervical
1090 Hetts & Khangura

carotid disease status, and vascular tortuosity. from five randomised trials. Lancet 2016;
Tandem lesions secondary to carotid artery 387(10029):1723–31.
atherosclerosis or ICA dissection present a 6. Powers WJ, Rabinstein AA, Ackerson T, et al. 2018
dilemma for the neurointerventionalists; earlier Guidelines for the early management of patients
diagnosis provides the opportunity to plan inter- with acute ischemic stroke: a guideline for health-
vention. CTP/multiphase CTA best evaluates the care professionals from the American Heart Associ-
ischemic tissue at risk and potential collaterals, ation/American Stroke Association. Stroke 2018;
respectively. CBV and MTT/TTP should be re- 49(3):e46–99.
ported along with parameter ratios. The quality of 7. Ribo M, Boned S, Rubiera M, et al. Direct transfer to
collaterals should be reported in the impression. angiosuite to reduce door-to-puncture time in throm-
If present, malignant-type collaterals should be re- bectomy for acute stroke. J Neurointerv Surg 2018;
ported immediately to the neuerointerventionalist. 10(3):221–4.
8. Schröder J, Thomalla G. A critical review of Alberta
SUMMARY Stroke Program early CT score for evaluation of
acute stroke imaging. Front Neurol 2017;7:245.
Recent randomized control trials demonstrate the 9. Liebeskind DS, Sanossian N, Yong WH, et al. CT and
benefit of thrombectomy potentially up to 24 hours MRI early vessel signs reflect clot composition in
in selected patients. CT remains the standard and acute stroke. Stroke 2011;42(5):1237–43.
primary mode of imaging for acute stroke. Treat- 10. Benson JC, Fitzgerald ST, Kadirvel R, et al. Clot
ment benefit is seen in the absence of extensive permeability and histopathology: is a clot’s pervious-
early ischemic change. Diagnostic imaging plays ness on CT imaging correlated with its histologic
a central role in providing efficient exclusionary composition? J Neurointerv Surg 2019. https://doi.
criteria for thrombolysis, diagnosis of LVO, and org/10.1136/neurintsurg-2019-014979.
selecting appropriate patients in the extended 11. de Lucas EM, Sánchez E, Gutiérrez A, et al. CT pro-
window based on expedient automated advanced tocol for acute stroke: tips and tricks for general ra-
imaging software. Radiologists may aid neuroin- diologists. Radiographics 2008;28(6):1673–87.
terventionalists with the diagnosis of LVO on the 12. Rocha M, Jovin TG. Fast versus slow progressors of
noncontrast head CT, potentially bypassing nonin- infarct growth in large vessel occlusion stroke.
vasive vascular imaging. In addition, identification Stroke 2017;48(9):2621–7.
of potential ICAD or tandem lesions may change 13. Menon BK, d’Esterre CD, Qazi EM, et al. Multiphase
the course of treatment, benefiting the patient’s CT angiography: a new tool for the imaging triage of
outcome. patients with acute ischemic stroke. Radiology 2015;
275(2):510–20.
ACKNOWLEDGMENTS 14. Campbell BC, Christensen S, Levi CR, et al. Cerebral
Dr Steven Hetts acknowledges NIH grant blood flow is the optimal CT perfusion parameter for
R01EB012031 as a source of funding. assessing infarct core. Stroke 2011;42(12):3435–40.
15. Jang-Hyun B, Moon KB, Hoe HJ, et al. Outcomes of
endovascular treatment for acute intracranial
REFERENCES
atherosclerosis–related large vessel occlusion.
1. Benjamin EJ, Blaha MJ, Chiuve SE, et al, on behalf Stroke 2018;49(11):2699–705.
of the American Heart Association Statistics Com- 16. Arenillas JF. Intracranial atherosclerosis. Stroke
mittee and Stroke Statistics Subcommittee. Heart 2011;42(1_suppl_1):S20–3.
disease and stroke statistics—2017 update: a report 17. Sacco RL, Adams R, Albers G, et al. Guidelines for
from the American Heart Association. Circulation prevention of stroke in patients with ischemic stroke
2017;135:e229–445. or transient ischemic attack. Stroke 2006;37(2):
2. Go S. Posterior circulation ischemic stroke. Mo Med 577–617.
2015;112(3):192–6. 18. Marta R, Marc R, Raquel D-M, et al. Tandem internal
3. Albers GW, Marks MP, Kemp S, et al. Thrombectomy carotid artery/middle cerebral artery occlusion.
for stroke at 6 to 16 hours with selection by perfusion Stroke 2006;37(9):2301–5.
imaging. N Engl J Med 2018;378(8):708–18. 19. Souza LCS, Yoo AJ, Chaudhry ZA, et al. Malignant
4. Nogueira RG, Jadhav AP, Haussen DC, et al. Throm- CTA collateral profile is highly specific for large admis-
bectomy 6 to 24 hours after stroke with a mismatch sion DWI infarct core and poor outcome in acute
between deficit and infarct. N Engl J Med 2017; stroke. AJNR Am J Neuroradiol 2012;33(7):1331–6.
378(1):11–21. 20. Dong Hun K, Cheolkyu J, Woong Y, et al. Endovas-
5. Goyal M, Menon BK, van Zwam WH, et al. Endovas- cular thrombectomy for acute basilar artery occlu-
cular thrombectomy after large-vessel ischaemic sion: a multicenter retrospective observational
stroke: a meta-analysis of individual patient data study. J Am Heart Assoc 2018;7(14):e009419.
 Imaging of Acute Stroke 1091

21. Lee YY, Yoon W, Kim SK, et al. Acute basilar artery 22. Kayan Y, Meyers PM, Prestigiacomo CJ, et al. Cur-
occlusion: differences in characteristics and out- rent endovascular strategies for posterior circulation
comes after endovascular therapy between patients large vessel occlusion stroke: report of the Society
with and without underlying severe atherosclerotic of NeuroInterventional Surgery Standards and
stenosis. AJNR Am J Neuroradiol 2017;38(8): Guidelines Committee. J Neurointerv Surg 2019.
1600–4. https://doi.org/10.1136/neurintsurg-2019-014873.