REVIEW ARTICLE

CT-based Techniques for Brain Perfusion

Pradeep Krishnan, MD, y Amanda Murphy, MD, z and Richard I. Aviv, MD, FRCR z

factors in aiding rapid triage of patients with suspected acute stroke.

Abstract: Recent rapid advances in endovascular treatment for acute ische-
Stroke management decisions, including endovascular procedures,
mic stroke highlight the crucial role of neuroimaging especially multimodal
are guided partly by the volume of ‘‘tissue at risk,’’ or extent of
computed tomography (CT) including CT perfusion in stroke triage and
the infarct core/penumbra mismatch.7,8 CTP may also facilitate
management decisions. With an increasing focus on changes in cerebral
improved patient selection for safe and effective treatment.9–11 By
physiology along with time-based matrices in clinical decisions for acute
providing a relative quantitative measure of brain perfusion, CTP
ischemic stroke, CT perfusion provides a rapid and practical modality for
may differentiate irreversibly damaged tissue (infarct ‘‘core’’) from
assessment and identification of salvageable tissue at risk and infarct core and
potentially salvageable tissue (‘‘penumbra’’)5,10,12 –14 The technical
provides a better understanding of the changes in cerebral physiology.
aspects of CTP image acquisition and processing are critical for
Although there are challenges with the lack of standardization and accuracy
ensuring good quality CTP images with accurate identification of
of quantitative assessment, CT perfusion is evolving as a cornerstone for
tissue at risk.
imaging-based strategies in the rapid management of acute ischemic stroke.
Key Words: brain, CT perfusion, CTP protocol, CTP technique, stroke
TECHNICAL CONSIDERATIONS: SCAN PROTOCOL
(Top Magn Reson Imaging 2017;26:113–119) AND CONTRAST INJECTION
The standard approach for multimodal CT evaluation of acute

A cute ischemic stroke is a leading cause of mortality and morbid-

ity in developed countries.1 The availability of thrombolytic
agents and emerging evidence of better outcomes related to rapidly
ischemic stroke usually entails a single session unenhanced CT of the
brain followed by CT angiogram (CTA) extending from the aortic
arch to the vertex, and CTP. CTA performed before CTP does not
improving endovascular procedures for recanalization emphasizes influence quantitative CTP parameters, but does reduce CTA-related
the need for continuous development of modern imaging techniques venous contamination.15 It also allows identification of any occlu-
as part of the diagnosis and management of acute stroke. sions and can be used to plan the CTP acquisition including timing
Multimodal computed tomography (CT) has emerged as the and coverage if whole brain scanning is not available.
widely available modality of choice for comprehensive assessment of The CTP scanning technique depends on the specific mode of
acute ischemic stroke, including defining the extent of salvageable or acquisition following a bolus of contrast injection. A continuous
‘‘at risk’’ tissue2 and understanding associated underlying cerebral scanning mode (cine-mode), designed to track the passage of the
hemodynamic disturbance.3 The primary objective of multimodal contrast column through the cerebral vasculature, is used in routine
CT use is improved stroke detection and increased the accuracy of practice.16 A volume set of images is acquired over the scan duration
parameters needed for the evaluation of acute ischemic stroke both such that a dataset of time-resolved images is obtained. Contrast
within the recommended and extended window period to ultimately enhancement of the tissue over time is depicted by the time density
facilitate better selection of patients who might be amenable to curve (TDC) (Fig. 2).3
endovascular intervention.4–6 CT perfusion (CTP) is increasingly Cine acquisitions are obtained following a bolus injection of
used as a part of the protocol for acute stroke evaluation. However, contrast and continuous mode scanning after a delay of approxi-
expanding applications of CTP necessitate the need for more stand- mately 5 to 7 seconds for up to 75 to 90 seconds or more. Continuous
ardized and optimized protocols and post-processing techniques. The scanning tracks the entire first passage of the contrast bolus through
following chapter reviews the technical considerations, principles, the intracranial vasculature from the arterial to the venous side within
challenges, and pitfalls for CTP elaborating on some of the more the brain volume defined by the user. Longer imaging acquisition
recent advances in stroke imaging. A brief review of Xenon CT is protocols avoid the potential pitfall of artificially attenuating the
also included. tissue concentration curve particularly in patients with cardiac
dysfunction or narrowing of an extra-cranial neck vessel.3,17
CT PERFUSION Extended, biphasic, acquisitions for usually >90 to 150 seconds
Recent advances in CT technology allow for faster acquisition, acquire 1 image per second in the arterial phase for 30 to 45 seconds,
higher resolution, and wider coverage. These advantages, in com- and 1 image every 2 to 3 seconds for the venous phase. Additional
bination with an easily accessible cost-effective modality, are crucial further delayed imaging may be undertaken for permeability map-
ping with images acquired every 10 to 15 seconds for about 2 to
From the Division of Neuroradiology, Department of Medical Imaging, Univer-
3 minutes.3,5,18
sity of Toronto and Sunnybrook Health Sciences Centre; yDiagnostic Imaging, Important factors related to contrast that contribute to high-
The Hospital for Sick Children; and zDivision of Neuroradiology, Department quality CTP images include a tight bolus injection of contrast
of Medical Imaging, University of Toronto and Sunnybrook Health Sciences material, the concentration of contrast agent, injection rate, and
Centre, Toronto, Ontario, Canada. the acquisition parameters, which will be discussed later. Forty to
Address correspondence to Richard I. Aviv, MD, FRCR, Sunnybrook Health
Sciences Centre, 2075 Bayview Avenue, Toronto, ON M4N 3M5, Canada. fifty milliliters of Iodinated contrast is commonly injected through a
(e-mail: [email protected]). large-bore venous access needle (18-gauge) at an injection rate of 4 to
Pradeep Krishnan, Amanda Murphy, and Richard I. Aviv contributed equally to the 7 mL/s, followed by 20-mL bolus saline (Table 1). Higher injection
manuscript. rates are recommended, as they produce a tight bolus with excellent
The authors report no conflicts of interest.
Copyright ß 2017 Wolters Kluwer Health, Inc. All rights reserved. peak opacification and minimal dispersion. A bolus of saline chaser
DOI: 10.1097/RMR.0000000000000129 behind the contrast column may further improve tissue attenuation

Topics in Magnetic Resonance Imaging  Volume 26, Number 3, June 2017 www.topicsinmri.com | 113

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.

Krishnan et al Topics in Magnetic Resonance Imaging  Volume 26, Number 3, June 2017

TABLE 1. CT Perfusion: Contrast Injection

IV Contrast Injection rate
18–20 gauge IV in the ante-cubital vein >300 mg/mL iodinated contrast, low/iso-osmolar Maximum slope algorithms 8 mL/s
Dual-bore, 15–20 mL saline chase, injector >Volume 40–60 mL, 20–40 mL saline flush Deconvolution-based algorithms >4 mL/s
pump

values, thus reducing dispersion with a tight bolus. High concen- Current generation scanners (128 to 320 detector rows) provide near
tration of iodinated contrast is preferable, usually in the range of whole brain coverage ranging and the capability to assess multiple
300 to 370 mg/mL to achieve maximum peak enhancement and vascular territories in a single CTP acquisition. The volume of
attenuation density.18–20 coverage for CTP acquisitions has steadily increased with these
increasing scanner capabilities. Z-axis coverage has improved from
FIELD OF VIEW AND ADAPTIVE STRATEGIES 4 cm for 64-detector scanners to 16 cm for 320 detector rows.21,22,24
FOR EXTENDED COVERAGE Approximately 75 mm of coverage on CTP has been reported in the
CTP imaging performed on older generation (16 and 64 slices) literature as adequate for triaging patients for reperfusion therapy by
scanners is limited in Z-axis coverage, restricting the characteriz- a quantitative mismatch.33
ation and localization of the ischemic area within the affected Motion artifacts are often a major source of error in the para-
territory.21– 24 The restriction of coverage results in the limited metric measurements on CTP, seen in around 10% of the cases in our
quantification of the infarct core and penumbra as well as poor institutional experience and reported in up to 25% of patients in the
detection of lesions within the posterior fossa, which was rarely literature.34 Motion can readily be detected by reviewing the TDC for
targeted as a region of interest unless clinically directed. The a characteristic notching due to a drop in attenuation within the ROI
maximum Z-axis coverage is determined by the width of the scanner (Fig. 1) or by recognizing the characteristic rimming that is seen on
detector array. Several strategies can be employed to improve Z-axis parametric maps. Finally, sorting the perfusion images by the slice on
coverage with limited detector width. This includes dual bolus PACS and dynamically scrolling through the dataset will readily
techniques, which involve sequential imaging of 2 separate slabs, demonstrate the severity, timing, and direction of motion. Remedial
which together cover the entire brain but necessitate separate boluses measures are available to mitigate motion including the limited
of contrast injections for each. The table-toggle or shuttle techniques exclusion of motion degraded images from the scan volume before
are also employed where the table is moved back and forth between 2 generating CTP maps either directly at the scanner console or on
slabs during scanning after a single bolus injection. There is a slight most CTP software applications.
loss of temporal resolution due to the intervening movement between
the slabs. However, the drop in resolution does not cause any CT PERFUSION ACQUISITION PARAMETERS
significant impact on diagnostic accuracy or clinical decision- AND RADIATION DOSE
making.4,18,25,26 (Tables 2 and 3). Acquisition parameters routinely used for CTP are in the range
There is an associated risk of increased contrast and of around 100 to 200 mAs and 80 kV, producing photons with mean
radiation dosage27 with these adaptive methods. Several factors energy close to the K edge of Iodine to optimize contrast to noise
can modify radiation dose including current and tube voltage, ratio.3
collimation pitch, and iterative reconstruction techniques.28 – 30 There are several techniques used in practice to reduce the
The accuracy of quantitative perfusion information with these effective dose. Apart from toggle mode and reduced sampling
adaptive methods is maintained as long as the temporal resolution frequency, various iterative reconstructions and deconvolution
is less than 1 to 3 seconds31,32 with adequate sampling of the arterial models have been used. Increasing CTP utilization emphasizes
and tissue density curves. At temporal resolutions of 1 second, the need for a better understanding of the risk-benefit ratio of
although attenuation-to-noise ratios are better, there is a concom- CTP use and a good knowledge of radiation reduction strategies.
itant increase in the radiation dose compared with the lower Low radiation dose CTP does not adversely affect the quality and
sampling frequency. diagnostic utility of parametric perfusion maps. CTP studies with
The relationship between temporal resolution and accuracy of 70 kVp and tube currents in the range of 50 mAs have shown
perfusion parameter calculation is also dependent on the software comparable diagnostic evaluation in detecting acute infarction.28
algorithm. The different algorithms used to process and generate The effective radiation dosage with recommended protocol of
perfusion maps are broadly categorized into deconvolution and non- 100 mAs and 80 kVp for CTP has been estimated in the range of
deconvolution methods, which will be further discussed under 3 to 5 msv.35 The diagnostic accuracy of the CTP parametric maps
mathematical models in this chapter. generated depends on the SNR, independent of the specific algorithm
With the advent of wider detector coverage on 256 and 320-row applied. Several measures, which have been used to achieve reduced
detectors, the limitations of coverage are now becoming less relevant. radiation dose, while maximizing SNR include lower mA and kVp,

TABLE 2. CT Perfusion: Technical Considerations

Scan Onset and Duration Field of View
Scan Duration
Scan Onset Two Phase Single Phase 64-Slice Scanners Whole Brain Scanners
 5 s delay 1st phase: 35–37 s 1 image/s 60 s 1 image/s 4 cm centered on the basal ganglia 14–16 cm
At least 2 2nd phase: >33–37 s 1 image/3 s 8 cm if toggle table or shuttle mode Center on the area of
baseline images Total duration: 75–90 s 1 image/s clinical concern

114 | www.topicsinmri.com ß 2017 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.

Topics in Magnetic Resonance Imaging  Volume 26, Number 3, June 2017 CT-based Techniques for Brain Perfusion

TABLE 3. CT Perfusion: Technical Parameters time to peak (TTP), parameter of time to the maximum of impulse
residue function; (Tmax) and permeability surface produce (PS); can
Temporal Resolution kV mA be derived from the TDC information.36
Cine 1 image/s, 70–80 50–100 The arterial input function (AIF) is usually chosen within a large
Combined cine mode 1 image/s artery orthogonal to the scanning plane, such as the anterior cerebral
Axial mode 1 image/3 s artery or supra-clinoid ICA, as partial volume effects are minimized.
Several studies have failed to demonstrate a significant variation in
the quantitative values with respect to a specific AIF location.
lower temporal sampling, iterative reconstructions, and increased Although it has been previously reported that the ipsilateral AIF
pitch. in the affected hemisphere may underestimate the perfusion
parameters due to delay in arrival and dispersion, newer software
PRINCIPLE OF CT PERFUSION with delay-insensitive deconvolution modeling accounts for this.
The basic principle for CTP imaging is a first pass approach, However, placement of the AIF distal to occlusion location will
which represents the continuous serial sampling of the defined tissue lead to considerable variability and inaccuracy of quantitative CTP
volume while tracking the passage of the intravenously injected parameters, due to the marked delay and adequacy of contrast
contrast bolus through the cerebral parenchyma. Time-concentration opacification distal to the occlusion. The venous region of interest
curves are generated from each voxel based on the first pass principle is usually defined in a large vessel, for example, the superior sagittal
after selection of arterial and venous inputs. Several perfusion sinus, to minimize partial volume effects. The AIF and venous input
parameters, including but not limited, to cerebral blood flow functions can be defined manually by the user or with automated or
(CBF), cerebral blood volume (CBV), mean transit time (MTT), semi-automated CT Perfusion software Fig. 2.3,36

FIGURE 1. Time density curve and corresponding CTP maps that show artifact secondary to movement during the scan. Movement is identified
with a sharp spike at the peak of the TDC.

ß 2017 Wolters Kluwer Health, Inc. All rights reserved. www.topicsinmri.com | 115

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.

Krishnan et al Topics in Magnetic Resonance Imaging  Volume 26, Number 3, June 2017

FIGURE 2. CT arterial input function plot.

FIGURE 3. CBV/CBF/MTT and Tmax parametric maps in a left MCA stroke with mismatch.

116 | www.topicsinmri.com ß 2017 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.

Topics in Magnetic Resonance Imaging  Volume 26, Number 3, June 2017 CT-based Techniques for Brain Perfusion

DEFINING TISSUE AT RISK CT PERFUSION: MATHEMATICAL MODELS

Definitions of common perfusion indices include CBV; the total Dynamic first pass contrast-enhanced CTP quantitative models
volume of blood in a given volume of brain tissue (mL/100 g), CBF; are based on the assumption that the tracer material is not diffusible.
the volume of blood perfusing the given volume of brain per unit time The 2 major types of models used in performing these calculations
(mL/100 g/min), MTT; the average transit time for the blood to pass are deconvolution-based and non-deconvolution based models.
through the given volume of the brain (seconds), Tmax; represents Deconvolution methods account for physiological variability in
the TTP of residual function, TTP; index of time between baseline the parenchymal delivery of contrast for example arterial delay,
and peak of enhancement, PS; the rate of contrast extravasation collateral vessels, and injection rates. These models also account
from intra-arterial to extravascular compartment through leaky for the dispersion of contrast before arriving at the level of the
blood-brain barrier. targeted tissue microcirculation. Deconvolution models calculate
CTP images are displayed in the different formats including quantitative values at the capillary level using a residual impulse
time density maps. These maps provide overall quick improved function, which simulates the tissue attenuation curves, which would
insight into altered hemodynamics and information related to the be obtained under ideal conditions. This is achieved using a variety of
vasculature. These images are usually displayed as thick sections to techniques including Fourier transformation and singular value
account for the noise from a source image. In clinical practice, CTP decomposition methods. Singular value decomposition and block
maps are most frequently qualitatively evaluated to assess potentially circular deconvolution are among the delay insensitive techniques
salvageable areas from infarct core. Although qualitative visual used to avoid CBF underestimation. These methods are slightly
assessment is reasonably effective, it is limited by inclusion of variable depending upon vascular anatomy and the degree of delay
oligemia in the penumbral assessment and relatively poor estimation or dispersal of contrast bolus. Potential pitfalls of CBF using the
of infarct core using CBV maps. The quantitative evaluation may deconvolution method include patient motion and partial volume
help to avoid these potential pitfalls and reduce interobserver differ- averaging. Singular value deconvolution appears to be the more
ences and variability in interpretation criteria. Although there is robust method among all deconvolution methods used for CBF
increased convergence of agreement on optimal thresholds and maps.3,5,36
parameters for infarct core and penumbra a determination, these Non-deconvolution models are based on Fick principle within a
quantitative thresholds are not yet clearly defined or standardized. region of interest in the brain with the underlying assumption of no
MTT, CBF, TTP, or Tmax commonly defines the ischemic leakage of contrast. Fick principle relates to the calculation of
penumbra, or tissue at risk. Unlike TTP, Tmax is not heavily perfusion in a particular distribution, which is proportional to the
influenced by the AIF and reflects local capillary level hemodynamic total number of extracted particles in the territory and the rate of
changes. In the evaluation of an acute stroke, time-based parameters accumulation. The simplification of this assumption yields relative
such as MTT or TTP are prolonged, and CBF is reduced both within perfusion values and makes cross comparisons difficult. The
the core and penumbra. CBV may be used to differentiate between maximal slope method calculates cerebral blood flow as a ratio of
these regions with increased CBV secondary to vasodilatation of the maximum initial slope of tissue to the peak height of arterial
the collateral vessels in the penumbra region, and reduced CBV tissue attenuation concentration. There is relative underestimation
within the core of the infarction with failure of compensatory of global CBF, although the qualitative assessment is similar to the
vasodilation.3,20,36 deconvolution method.
There is a wide variability in the threshold values for CTP used The model used to analyze CTP data may significantly impact
to define core and penumbra in the literature. These variabilities the CTP protocol used, for example, deconvolution models allow
relate to many factors, including differences in software/ deconvo- considerably slower injection rates compared to algorithms that
lution methods, variability in modality type, and duration to follow- assume no venous outflow.3,5,36 Deconvolution-based algorithms
up imaging used to define regions and scanning protocols/ are more robust, and temporal resolutions of up to 3 seconds do not
parameters.12 The threshold values used to determine the core of significantly alter perfusion parameter measurement. Non-decon-
the infarction are CBVabsolute values of <2.0 to 2.2/100 g, and CBF volution models such as maximum-slope algorithms may require
reduction of 70% relative to the contralateral side (rCBF <30%) and shorter sampling intervals to accurately identify the point of maxi-
for ischemic ‘‘at risk’’ tissue with MTT threshold of 145% compared mum slope on the time-density curve for quantification. Reduced
with the contralateral side, and Tmax >6 seconds for penumbra and sampling frequency in pre-enhancement and post-enhancement
>10 seconds for core are also widely used Fig. 3.12,37– 41 portions of the time-density curve and shorter sampling rate in
Standardization of perfusion assessment remains a priority due the arterial portion of the curve may be performed avoiding the
to a wide variability in CT-derived core volumes reported with potential risks of tissue density curve truncation due to delayed
different software applications. The accuracy of CTP data set contrast arrival.3
processing by the automated software for the estimation of the
volumes of the ischemic core is an important determinant for guiding CHALLENGES AND FUTURE DIRECTION: CTP IN
management and prognostication as well as for patient selection in ACUTE ISCHEMIC STROKE
endovascular procedures.3,4 Automatic rapid post-processing will Recent randomized control trials of endovascular therapy have
improve the speed of infarct core and penumbra determination, and is emphasized the critical role of neuroimaging in a rapid and efficient
achievable in minutes. Perfusion selection parameters were included management of acute stroke. The strategic goals of neuroimaging in
in 2 recent successful IA studies, SWIFT PRIME and EXTEND-IA acute ischemic stroke for which there is a high level of evidence are
demonstrating the feasibility of physiological patient selection. the exclusion of hemorrhage and the identification of a proximal
However, currently, AHA guidelines do not strongly support the vessel occlusion. Additional data for which there is growing evidence
use of CTP for the selection of patients for endovascular therapy. include evaluation of the core and ischemic penumbra, collateral
There remains significant potential for CTP to extend conventional status and risk of hemorrhagic transformation. However, in the
time windows of treatment based on physiology with potentially absence of any standardization of threshold values used for
salvageable tissue demonstrated up to 12 hours from ictus in some CTP parametric measurements or the method of quantification,
cases.42 the physiological selection of patients for thrombolysis and outcome

ß 2017 Wolters Kluwer Health, Inc. All rights reserved. www.topicsinmri.com | 117

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.

Krishnan et al Topics in Magnetic Resonance Imaging  Volume 26, Number 3, June 2017

prediction remains challenging. However, CTP remains a powerful 11. Fox AJ, Symons SP, Howard P, et al. Acute stroke imaging: CT with CT
potential tool for the physiological assessment of acute ischemic angiography and CT perfusion before management decisions. AJNR Am J
evolution and will undoubtedly mature as the primary means of Neuroradiol. 2012;33:792–794.
patient selection both within and outside of conventional time 12. Bivard A, Levi C, Spratt N, Parsons M. Perfusion CT in acute stroke: a
windows of thrombolytic treatment.5,43–45 comprehensive analysis of infarct and penumbra. Radiology. 2013;267:
543–550.
XENON CT PERFUSION 13. Hopyan J, Ciarallo A, Dowlatshahi D, et al. Certainty of stroke diagnosis:
Xenon CTP is performed using an inert isotope Xenon (Xe), incremental benefit with CT perfusion over non-contrast CT and CT
which in its gaseous form when inhaled freely diffuses across blood- angiography. Radiology. 2010;255:142–153.
brain barrier to simulate a diffusible tracer or contrast agent for CTP. 14. Vagal A, Meganathan K, Kleindorfer DO, et al. Increasing use of computed
Xenon accumulation within the brain causes X-ray attenuation hence tomographic perfusion and computed tomographic angiograms in acute
easily identified. On Xenon CTP, the Xe tracer distribution is ischemic stroke from 2006 to 2010. Stroke. 2014;45:1029–1034.
interpreted as areas of increased density in the affected areas when 15. Morhard D, Wirth CD, Reiser MF, et al. Optimal sequence timing of CT
in equilibrium with blood gas levels, achieved due to its ability to angiography and perfusion CT in patients with stroke. Eur J Radiol.
diffuse across blood-brain barrier and lungs. A plain CT is obtained 2013;82:e286–289.
followed by inhalation of an inert mixture of xenon and oxygen for 16. Shetty SK, Lev MH. CT perfusion in acute stroke. Neuroimaging Clin N Am.
few minutes. Multiple serial images of 4 to 8 contiguous slices are 2005;15:481–501. ix.
obtained during inhalation with each level scanned 6 to 8 times. The
17. Sasaki M, Kudo K, Ogasawara K, Fujiwara S. Tracer delay-insensitive
scanner subtracts the baseline attenuation coefficients from that algorithm can improve the reliability of CT perfusion imaging for
obtained during inhalation obtained following xenon accumulation. cerebrovascular steno-occlusive disease: comparison with quantitative single-
The resultant data set provides a more quantitative and functional photon emission CT. AJNR Am J Neuroradiol. 2009;30:188–193.
measure of CBF and helps to delineate core and penumbra.46,47
18. Konstas AA, Goldmakher GV, Lee TY, Lev MH. Theoretic basis and technical
The major limitations include artifacts related to motion, mild
implementations of CT perfusion in acute ischemic stroke, part 2: technical
sedative effects, which may be a challenge in patients presenting with
implementations. AJNR Am J Neuroradiol. 2009;30:885–892.
acute stroke and limited availability due to cost and logistical
considerations. The only measurable parameter for Xe CT is CBF 19. Bae KT. Intravenous contrast medium administration and scan timing at CT:
unlike contrast CTP and also prolonged scans times that may limit considerations and approaches. Radiology. 2010;256:32–61.
the utility of the modality in acute stroke settings.46 20. Wintermark M, Sincic R, Sridhar D, Chien JD. Cerebral perfusion CT:
technique and clinical applications. J Neuroradiol. 2008;35:253–260.
REFERENCES 21. Thierfelder KM, Sommer WH, Baumann AB, et al. Whole-brain CT
1. Yoo AJ, Hu R, Hakimelahi R, et al. CT angiography source images acquired perfusion: reliability and reproducibility of volumetric perfusion deficit
with a fast-acquisition protocol overestimate infarct core on diffusion- assessment in patients with acute ischemic stroke. Neuroradiology.
weighted images in acute ischemic stroke. J Neuroimaging. 2012;22:329– 2013;55:827–835.
335. 22. Dorn F, Muenzel D, Meier R, et al. Brain perfusion CT for acute stroke using a
2. Bivard A, Levi C, Krishnamurthy V, et al. Perfusion computed tomography 256-slice CT: improvement of diagnostic information by large volume
to assist decision making for stroke thrombolysis. Brain. 2015;138(Pt 7): coverage. Eur Radiol. 2011;21:1803–1810.
1919–1931. 23. Breuer L, Knott M, Struffert T, et al. Limited versus whole-brain perfusion for
3. d’Esterre CD, Fainardi E, Aviv RI, Lee TY. Improving acute stroke the indication of thrombolysis in the extended time window of acute cerebral
management with computed tomography perfusion: a review of imaging ischemia. J Stroke Cerebrovasc Dis. 2015;24:2491–2496.
basics and applications. Transl Stroke Res. 2012;3:205–220. 24. Emmer BJ, Rijkee M, Niesten JM, et al. Whole brain CT perfusion in acute
4. Jia B, Scalzo F, Agbayani E, et al. Multimodal CT techniques for anterior circulation ischemia: coverage size matters. Neuroradiology.
cerebrovascular and hemodynamic evaluation of ischemic stroke: occlusion, 2014;56:1121–1126.
collaterals, and perfusion. Expert Rev Neurother. 2016;16:515–525. 25. Youn SW, Kim JH, Weon YC, et al. Perfusion CT of the brain using 40-mm-
5. Donahue J, Wintermark M. Perfusion CTand acute stroke imaging: foundations, wide detector and toggling table technique for initial imaging of acute stroke.
applications, and literature review. J Neuroradiol. 2015;42:21–29. AJR Am J Roentgenol. 2008;191:W120–W126.

6. Wintermark M, Sanelli PC, Albers GW, et al. Imaging recommendations for 26. Wintermark M, Albers GW, Alexandrov AV, et al. Acute stroke imaging
acute stroke and transient ischemic attack patients: a joint statement by the research roadmap. Stroke. 2008;39:1621–1628.
American Society of Neuroradiology, the American College of Radiology 27. Yang CY, Chen YF, Lee CW, et al. Multiphase CT angiography versus single-
and the Society of NeuroInterventional Surgery. J Am Coll Radiol. 2013;10: phase CT angiography: comparison of image quality and radiation dose. AJNR
828–832. Am J Neuroradiol. 2008;29:1288–1295.
7. Campbell BC, Mitchell PJ, Kleinig TJ, et al. Endovascular therapy for 28. Murphy A, So A, Lee TY, et al. Low-dose CT perfusion in acute ischemic
ischemic stroke with the perfusion-imaging selection. N Engl J Med. stroke. Neuroradiology. 2014;56:1055–1062.
2015;372:1009–1018. 29. Fang XK, Ni QQ, Schoepf UJ, et al. Image quality, radiation dose and
8. Saver JL, Goyal M, Bonafe A, et al. Stent-retriever thrombectomy after diagnostic accuracy of 70 kVp whole brain volumetric CT perfusion imaging:
intravenous t-PA vs. t-PA alone in stroke. N Engl J Med. 2015;372:2285–2295. a preliminary study. Eur Radiol. 2016;26:4184–4193.
9. Burton KR, Dhanoa D, Aviv RI, et al. Perfusion CT for selecting patients with 30. Lin CJ, Wu TH, Lin CH, et al. Can iterative reconstruction improve imaging
acute ischemic stroke for intravenous thrombolytic therapy. Radiology. quality for lower radiation CT perfusion? Initial experience. AJNR Am J
2015;274:103–114. Neuroradiol. 2013;34:1516–1521.
10. van Seeters T, Biessels GJ, Kappelle LJ, et al. CT angiography and CT 31. Wintermark M, Smith WS, Ko NU, et al. Dynamic perfusion CT: optimizing
perfusion improve prediction of infarct volume in patients with anterior the temporal resolution and contrast volume for calculation of perfusion CT
circulation stroke. Neuroradiology. 2016;58:327–337. parameters in stroke patients. AJNR Am J Neuroradiol. 2004;25:720–729.

118 | www.topicsinmri.com ß 2017 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.

Topics in Magnetic Resonance Imaging  Volume 26, Number 3, June 2017 CT-based Techniques for Brain Perfusion

32. Kämena A, Streitparth F, Grieser C, et al. Dynamic perfusion CT: optimizing 39. Wintermark M, Flanders AE, Velthuis B, et al. Perfusion-CT assessment of
the temporal resolution for the calculation of perfusion CT parameters in infarct core and penumbra: receiver operating characteristic curve analysis in
stroke patients. Eur J Radiol. 2007;64:111–118. 130 patients suspected of acute hemispheric stroke. Stroke. 2006;37:979–985.
33. Furtado AD, Lau BC, Vittinghoff E, et al. Optimal brain perfusion CT 40. McVerry F, Dani KA, MacDougall NJ, et al. Derivation and evaluation of
coverage in patients with acute middle cerebral artery stroke. AJNR Am J thresholds for core and tissue at risk of infarction using CT perfusion.
Neuroradiol. 2010;31:691–695. J Neuroimaging. 2014;24:562–568.
34. Fahmi F, Beenen LF, Streekstra GJ, et al. Head movement during CT brain 41. Lin L, Bivard A, Krishnamurthy V, et al. Whole-brain CT perfusion to quantify
perfusion acquisition of patients with suspected acute ischemic stroke. Eur J acute ischemic penumbra and core. Radiology. 2016;279:876–887.
Radiol. 2013;82:2334–2341.
42. Thomalla G, Gerloff C. Treatment concepts for wake-up stroke and stroke with
35. Diekmann S, Siebert E, Juran R, et al. Dose exposure of patients undergoing unknown time of symptom onset. Stroke. 2015;46:2707–2713.
comprehensive stroke imaging by multidetector-row CT: comparison of 320-
detector row and 64-detector row CT scanners. AJNR Am J Neuroradiol. 43. Lev MH. Perfusion imaging of acute stroke: its role in current and future
2010;31:1003–1009. clinical practice. Radiology. 2013;266:22–27.

36. Konstas AA, Goldmakher GV, Lee TY, Lev MH. Theoretic basis and technical 44. Turowski B, Schramm P. An appeal to standardize CT- and MR-perfusion.
implementations of CT perfusion in acute ischemic stroke, part 1: theoretic Clin Neuroradiol. 2015;25(Suppl 2):205–210.
basis. AJNR Am J Neuroradiol. 2009;30:662–668. 45. Liebeskind DS, Parsons MW, Wintermark M, et al. Computed tomography
37. Konstas AA, Wintermark M, Lev MH. CT perfusion imaging in acute stroke. perfusion in acute ischemic stroke: is it ready for prime time? Stroke.
Neuroimaging Clin N Am. 2011;21:215–238. ix. 2015;46:2364–2367.

38. Schaefer PW, Barak ER, Kamalian S, et al. Quantitative assessment of core/ 46. Mullins ME. Stroke imaging with xenon-CT. Semin Ultrasound CT MR.
penumbra mismatch in acute stroke: CT and MR perfusion imaging are 2006;27:219–220.
strongly correlated when sufficient brain volume is imaged. Stroke. 47. Hage ZA, Alaraj A, Arnone GD, Charbel FT. Novel imaging approaches to
2008;39:2986–2992. cerebrovascular disease. Transl Res. 2016;175:54–75.

ß 2017 Wolters Kluwer Health, Inc. All rights reserved. www.topicsinmri.com | 119

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.