The authors address the use of computed tomography angiography (CTA), CT perfusion (CTP), and noncontrast CT imaging in acute stroke patients. With the enhanced clinical and anatomic details available, clinicians are increasingly able to assess a patient's hemorrhagic risk and determine prognosis. The use of CTA/CTP can facilitate triage of acute stroke patients to appropriate treatment.
Mr. Ussery is a medical student, Dr. Gonzalez is Division
Chief of Neuroradiology, and Dr. Lev is Director of the
Neurovascular Laboratory and Emergency Neuroradiology at
Massachusetts General Hospital, Harvard Medical School, Boston,
MA; and Dr. Sanelli is a Staff Neuroradiologist at New York
Presbyterian Hospital, Weill Cornell Medical School, New York,
The introduction of computed tomography angiography (CTA) and CT
perfusion (CTP) imaging in acute stroke patients provides useful
information for therapeutic decision making in the early stages of
These modalities are used in conjunction with non-contrast CT
(NCCT) of the brain.
Non-contrast CT is performed initially to exclude hemorrhage; an
absolute contraindication to thrombolytic therapy. Large
parenchymal hypodensity, typically indicating irreversible "core"
of infarction, is a relative contraindication to thrombolysis.
Computed tomography angiography demonstrates the anatomical
details of the neurovasculature, from the great vessel origins at
the aortic arch to their intracranial termination. This technique
has been shown to be highly accurate in the identification of
proximal large vessel circle of Willis occlusions, and, therefore,
in the rapid triage of patients to intra-arterial (IA) or
intravenous (IV) thrombolytic therapy.
An additional modality, CTP, can produce qualitative or
quantitative information used to predict the stroke end state.
Whole-brain perfused blood volume imaging data is acquired
simultaneously with the CTA data, without the need for additional
contrast administration or scanning.
Indeed, coregistration and subtraction of the NCCT images from the
post-contrast whole-brain source images provides quantitative maps
of cerebral blood volume (CBV).
The whole-brain perfused blood volume source images alone, however,
without subtraction, can provide important prognostic information
regarding final infarct size and clinical outcome in acute (<6
hours post onset) stroke patients.
Moreover, using cine imaging and dynamic contrast-enhanced bolus
tracking techniques, quantitative "first-pass" CT perfusion maps of
cerebral blood flow (CBF), CBV, and mean transit time (MTT) can be
Performing NCCT, CTA, and CTP in acute stroke patients upon arrival
to the emergency department can be a quick and useful method in
guiding the use of IV and/or IA thrombolytic therapy (figure
Computed tomography angiography provides anatomical assessment
of the vessel lumen regarding the presence of an acute embolus,
occlusion, or severe stenosis. The entire course of the carotid and
vertebral arteries can be imaged using multislice CT technology
(MSCT), including the intracranial and extracranial segments. This
data can be used to determine if a patient having an acute stroke
has an arterial blockage (embolus) requiring thrombolytic therapy.
Most acute emboli occur in the proximal segment of the middle
Quantitative dynamic cine CTP is a contrast-enhanced study
performed at selected image levels. Coverage is typically limited
by the maximum detector length in the z-axis
direction--approximately 2 cm for MSCT scanners. The selected
levels are imaged repeatedly during contrast bolus administration,
which reveals serial changes in attenuation of the brain
parenchyma. The relationship between tissue attenuation, measured
in Hounsfield units (HU), and the tissue concentration of iodinated
contrast material, at constant mAs and kV settings, is assumed to
Therefore, models for quantitative analysis of this perfusion
dataset can be applied conveniently. Such analysis requires
computer postprocessing to generate CBF, CBV, and MTT maps,
typically using a mathematical process known as deconvolution.
The actual quantification of flow values is based on the arterial
input and venous outflow curves, which can also be reconstructed
from the cine CTP dataset. According to the central volume
principle, CBF (in mL/100g/min), CBV (in mL/100g), and MTT (in
minutes) are related according to the formula CBF = CBV/MTT.
To date, the only FDA-approved treatment for acute stroke is IV
thrombolysis with recombinant tissue plasminogen activator (IV
rt-PA), administered within 3 hours of stroke onset.
If thrombolysis is applied beyond this time window, the increased
probability of intracranial hemorrhage is considered
The time window for treatment with IA agents is twice as long
for the anterior circulation, and indefinite for the posterior
circulation (depending on risk-to-benefit ratios); however, IA
thrombolysis--although of proven benefit in preliminary trials--has
not yet received FDA approval.
For thrombosis localized to the posterior circulation, the time
window for treatment may be extended beyond 6 hours due to the
extreme consequences of loss of blood flow to the brainstem,
despite the risk of hemorrhage.
Because of these narrow windows of opportunity, a goal of
advanced imaging is therefore to select patients for treatment more
rationally, potentially expanding the time window for thrombolysis
by identifying patients for whom rt-PA will be both safe and
Advanced CTA/CTP imaging of acute stroke has the potential to
not only help exclude from thrombolysis patients at high risk for
hemorrhage, but to identify those patients most likely to benefit
from thrombolysis. Even without hemorrhage, treatment failure with
thrombolytics is not uncommon. Many hospitals do not have the
facilities to use IA rt-PA, and it is important to be able to
pinpoint the state of evolution and cause of stroke in order to
select the most effective treatment. It may sometimes be necessary
to transfer patients to a tertiary care center with specialized
expertise in the clinical assessment, imaging, and treatment of
At most institutions the current standard of care is to perform
an NCCT exam to exclude hemorrhage; at this point, the imaging
evaluation of acute stroke patients typically stops. In some
centers, advanced imaging of stroke patients is obtained with
diffusion-weighted magnetic resonance imaging (MRI) and MR
perfusion transit time maps, but these techniques are often
time-consuming, expensive, or unavailable.
Computed tomography angiography with CTP can offer a more
convenient advanced imaging alternative.
Clearly, given the potential severe consequences of stroke, better
methods of identifying its presence and etiology are required, and
deriving reliable predictors of the extent and degree of tissue
damage is desirable.
In the acute stroke setting CTA with CTP has the potential to be
useful for: 1) fast and reliable identification of stroke
signatures; 2) improved choice of treatment modality, including
exclusion of patients from thrombolytic therapy; 3) pinpointing the
vascular origin of the ischemic insult; and
4) determination of neurological consequences of the stroke,
including final infarct size, clinical outcome, and hemorrhagic
Available data suggests that CTA, when combined with NCCT, is
significantly more sensitive than NCCT alone in delineation of
The choice between IA or IV thrombolysis depends on a variety of
factors, including the time post-ictus, the clinical status of the
patient, and whether the clot is proximal (IA) or distal (IV). When
typical findings of occlusive thrombus on CTA and decreased tissue
enhancement on CTP are not present, the differential diagnoses
include lacunar infarct, early small distal embolic infarct,
transient ischemic attack, complex migraine headaches, and seizure.
At our institution, immediately following data acquisition for
CTA/CTP, rapid "collapsed" maximum intensity projection (MIP)
reconstructions (superior view) of the intracranial circulation are
created, for visualization of acute circle of Willis embolus.
Typically, these can be obtained in under a minute. Additional
reconstructions, not typically obtained urgently but prior to
endovascular treatment, are obtained displaying the aortic arch,
great vessels origins, and carotid bifurcations. The cine CTP
dataset, when obtained, can also be manipulated to generate
quantitative CBF, CBV, and MTT maps at selected levels (figure
The CBF, CBV, and MTT values are dynamic--they vary as the
stroke evolves. "Core" values for irreversible infarction are
typically considered to be in the CBF <10 to 12 mL/100g/min
range. Areas of less severe CBF reduction, with preserved CBV
values, are considered to represent "ischemic penumbra"--tissue
that is at high risk for infarction if circulation is not soon
restored, but not yet irreversibly infarcted. The larger the
ischemic penumbra relative to the core, the more likely the patient
would benefit from early thrombolytic therapy. If both CBV and CBF
are already reduced dramatically, however, the tissue is considered
irreversibly infarcted, and successful recanalization may have a
higher likelihood of causing hemorrhage.
At our institution, CTA/CTP imaging is performed on a multislice
scanner in the emergency department (GE LightSpeed, GE Medical
Systems, Milwaukee, WI), which enables us to acquire imaging data
from entire vascular territories in <1 minute. The initial NCCT
scan is performed at 140 kV, 170 mA, pitch = "high quality" (3:1),
and a table speed of 7.5 mm/sec. Images are obtained from the skull
base to the vertex. Slice thickness can be set at 2.5 mm, at 5 mm,
or at both. We typically use the thicker, 5-mm slices for visual
image review, and the thinner, 2.5-mm slices for 3D-reconstruction.
Initial image review is in "real-time," directly at the CT console.
The use of narrow window-width settings, with a center level of
about 30 HU (width of 5 to 30 HU), facilitates the detection of
early, subtle, ischemic changes contiguous with normal parenchyma.
Immediately following NCCT, CTA with simultaneous whole brain
perfused blood volume CTP is performed without repositioning the
patient. Approximately 90 to 120 mL of nonionic, iso-osmolar
contrast is used for CTA from the skull base to the vertex. The
injection rate is 3 mL/sec, with a 25-second scan delay. Initial
scan parameters are as per the NCCT scan described above. A second
phase of scanning is performed immediately, with minimal possible
delay, from the aortic arch to the skull base, with similar scan
parameters except for an increase in the table speed to 15 mm/sec.
Major advantages of first scanning the intracranial circulation
include: 1) obtaining the most important data first, which can be
reviewed during subsequent acquisition; and 2) allowing time for
clearance of dense, IV contrast from the subclavian, axillary, and
other veins at the thoracic inlet, reducing streak artifact.
Immediate assessment of the vasculature occurs at the scanner
console, where the source images are reviewed while scanning is
The axial source images from the CTA/CTP exam are typically
networked to a freestanding image analysis workstation. MIP
reconstructions of the intracranial circulation are performed to
identify acute embolus and the extent of the clot burden.
Reconstructions of the aortic arch, great vessels, and carotid
bifurcations are also performed to assess for additional stenoses
or occlusions of the neurovasculature prior to endovascular
Depending on the clinical circumstances--or as an entirely
separate exam--quantitative cine dynamic CT perfusion maps can also
be acquired. The level to be scanned is determined from the initial
NCCT findings, as well as from the involved vascular distribution
seen on CTA. Our MSCT scanner allows for the selection of four 5-mm
or two 10-mm thick slices in axial cine mode. Approximately 45 mL
of nonionic contrast is injected at 4 to 5 mL/sec, with a 5-second
preparation delay. In order to generate CTP maps, the operator must
select the arterial input and venous outflow parameters for the
Potential current and future benefits of CTA/CTP include
convenient, reliable, acute stroke imaging, which may improve
clinicians' ability to assess hemorrhagic risk and determine
prognosis, therefore facilitating triage of acute stroke patients
to appropriate treatment.
Quantitative CTP has shown further potential to characterize the
evolving status of stroke through the measurement of CBV, CBF, and
MTT. We encourage the use of CTA/CTP imaging for the emergency
evaluation and management of suspected stroke. A