Optimal sequence timing of CT angiography and perfusion CT in patients with stroke

European Journal of Radiology 82 (2013) e286–e289

Contents lists available at SciVerse ScienceDirect

European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad

Optimal sequence timing of CT angiography and perfusion CT in patients with stroke D. Morhard a,b,∗ , C.D. Wirth a,c , M.F. Reiser a , G. Schulte-Altedorneburg b , B. Ertl-Wagner a a b c

Institute of Clinical Radiology, University of Munich, Munich, Germany Institute of Diagnostic and Interventional Radiology, Neuroradiology and Nuclear Medicine, Klinikum Muenchen-Harlaching, Munich, Germany Department of General Internal Medicine, University Hospital Inselspital, University of Bern, Bern, Switzerland

a r t i c l e

i n f o

Article history: Received 28 October 2012 Received in revised form 12 January 2013 Accepted 17 January 2013 Keywords: Cerebral angiography Computed tomography Perfusion Stroke

a b s t r a c t Objective: Standard stroke CT protocols start with non-enhanced CT followed by perfusion-CT (PCT) and end with CTA. We aimed to evaluate the influence of the sequence of PCT and CTA on quantitative perfusion parameters, venous contrast enhancement and examination time to save critical time in the therapeutic window in stroke patients. Methods and materials: Stroke CT data sets of 85 patients, 47 patients with CTA before PCT (group A) and 38 with CTA after PCT (group B) were retrospectively analyzed by two experienced neuroradiologists. Parameter maps of cerebral blood flow, cerebral blood volume, time to peak and mean transit time and contrast enhancements (arterial and venous) were compared. Results: Both readers rated contrast of brain-supplying arteries to be equal in both groups (p = 0.55 (intracranial) and p = 0.73 (extracranial)) although the extent of venous superimposition of the ICA was rated higher in group B (p = 0.04). Quantitative perfusion parameters did not significantly differ between the groups (all p > 0.18), while the extent of venous superimposition of the ICA was rated higher in group B (p = 0.04). The time to complete the diagnostic CT examination was significantly shorter for group A (p < 0.01). Conclusion: Performing CTA directly after NECT has no significant effect on PCT parameters and avoids venous preloading in CTA, while examination times were significantly shorter. © 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Multi-modal stroke imaging has become a diagnostic mainstay in the acute setting [1–6]. It generally includes non-enhanced computed tomography (NECT), perfusion CT (PCT) and CT angiography (CTA) of the supraaortic arteries. NECT is performed to rule-out hemorrhage as well as to detect early signs of cerebral ischemia. While PCT evaluates the hemodynamic relevance of an impaired cerebral blood flow, thereby differentiating between infarction and tissue-at-risk, CTA is able to identify the site and the extent of thrombosis or dissection as well as potential tandem stenoses and arterial collateralization [7–13]. Classically, NECT is performed first followed by PCT with CTA being the last examination. The reasoning behind this sequence of examinations was that a venous preloading could falsify the perfusion parameters to an extent that impairs diagnostic decision

∗ Corresponding author at: Institute of Diagnostic and Interventional Radiology, Neuroradiology and Nuclear Medicine, Klinikum Muenchen-Harlaching, Sanatoriumsplatz 2, D-81547 Munich, Germany. Tel.: +49 89 6201 2602. E-mail address: [email protected] (D. Morhard). 0720-048X/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2013.01.011

making. Recently, this sequence of examinations has been questioned, though [14,15]. As CTA has traditionally been performed after the contrast-load of PCT, it tended to suffer from a venous overlay thereby potentially interfering with the delineation of the supraaortic arteries, e.g. the cavernous segment of the internal carotid artery (ICA). In recent literature some authors used a different scanning sequence, where CTA was performed before PCT [16,17]. In clinical practice it remains unclear what scanning sequence is the best regarding image quality and acquisition time. As time is brain, the major aim of multimodal stroke imaging is to obtain a correct diagnosis regarding the presence or absence of hemorrhage as well as the site, extent and prognosis of vascular pathologies and cerebral ischemia in the shortest possible time. The critical time not only includes the imaging per se, but also the time for image interpretation. When analyzing a CTA of the supraaortic arteries, it is more straightforward to identify and delineate the respective vessels and to evaluate potential pathologies when there is little to no overlay by either venous contrast enhancement or bony structures [18,19]. It has been previously shown that thick maximum intensity projections (MIP) after the elimination of bone with either masked bone subtraction, i.e. the subtraction

D. Morhard et al. / European Journal of Radiology 82 (2013) e286–e289

of the non-enhanced scan [20], or with dual-energy based bone removal improves the delineation of supraaortic and/or intracranial arteries and reduces the reading time of the CTA [21]. The initial acquisition of CTA directly after NECT and prior to PCT could therefore be advantageous for CTA image interpretation. It may moreover enable an optimal placement of the PCT imaging slab depending on the CT-angiographic results. It would be desirable to avoid contrast preload by the CTA interfering with perfusion parameters. We therefore aimed to evaluate the impact of the sequence of CTA and PCT on the image quality of both modalities as well as on the contrast and delineation of the supraaortic arteries and the calculated perfusion parameters. 2. Materials and methods Institutional Review Board (IRB) approval was obtained prior to the commencement of the study. The datasets of 85 consecutive patients (46 female and 39 male, mean age: 60.6 ± 15 years; median age: 62.8; range 22–89 years) were retrospectively identified in a database query of our institutional radiology information system. Search criteria were: (1) suspected ischemic stroke; (2) a completed multimodal stroke-CT exam with a standardized protocol as described below, consisting of NECT, CTA of the supraaortic arteries and dynamic PCT of the brain; (3) examination on a modern CT scanner (40- or 64-slice multi detector CT); and (4) complete DICOM data sets of the PCT raw data and 1.0 mm or thinner axial slices of the CTA data. Exclusion criteria were known contraindications for the administration of contrast media, hemorrhage in the non-enhanced CT (NECT), age below 18 years and an incomplete dataset. All patient identifying information was pseudonymized prior to image analysis. The identified data sets were divided into two groups: group A: “CTA before PCT”, in which the CTA scan was performed prior to the PCT scan; and group B: “PCT before CTA”, in which the PCT scan was acquired prior to CTA. Group A consisted of 47 patient studies (21 male; 26 female; age: 60.8 ± 14 years; median age: 63.8 years; range 31–89 years). Group B consisted of 38 patient studies (18 male; 20 female; age: 60.4 ± 16 years; median age: 60.7 years; range 21–89 years). 75 examinations were acquired on a Somatom Sensation 64 scanner, 5 on a Somatom Definition and 5 on a Sensation Open Sliding Gantry (all Siemens Healthcare, Forchheim, Germany). Acquisition parameters were as follows: • CTA: 120 kV at 220 mAs reference tube current-time product with attenuation-based tube current modulation (CareDose, Siemens Healthcare, Forchheim, Germany), 40 or 64 mm × 0.6 mm collimation, scan start with 4 s delay after bolus-tracking at the level of the ascending aorta, pitch 0.8, caudocranial scan direction. A weight-adapted contrast media and saline chaser injection protocol was applied with 1.1 ml contrast material per kg body weight (Ultravist 370, Bayer Healthcare, Berlin, Germany, i.e. 0.42 g iodine per kg bodyweight) at an injection rate of (0.056 × bodyweight) ml/s followed by 100 ml saline at the identical flow rate. Axial images were reconstructed with a slice thickness of 0.75 or 1.0 mm and an increment of 0.5 or 0.8 mm using a standard CTA reconstruction kernel. • PCT: 80 kV, 250 mAs, 40 or 64 mm × 0.6 mm collimation, 2.4 or 2.8 cm coverage (2 slices of 1.2 resp. 1.4 cm width) at the level of the basal ganglia, 40 scans every 1.0 s after injection of 14–15 g iodine at 2.0–2.1 g/s followed by a saline flush of 40 ml NaCl at 5 ml/s flow. All CTA and PCT data were transferred to a multimodality workstation (Syngo MMWP, VA 22, Siemens Healthcare, Forchheim,

e287

Germany). The time between CTA and PCT or between PCT and CTA was measured in minutes using the DICOM time stamp of the first image of the corresponding series. Quantitative measurements of the contrast attenuation of the axial CTA “raw” data were performed by manually drawing a region on interest (ROI) at the left and right proximal extracranial ICA and the jugular veins (JV) of both sides. Mean attenuation was calculated in the ICA and JV. MIP reformations in axial, coronal and sagittal orientations were created for CTA reading with a slice thickness and increment of 3/3 mm for the axial images and 5/5 for coronal and sagittal images. From the PCT datasets, color-coded parameter maps (2 slices of 1.2 or 1.4 cm width) of cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT) and time to peak (TPP) were reconstructed using a standard deconvolution model. Quantitative measurements of the contrast enhancement (in Hounsfield units) were performed on the PCT data sets by manually drawing a ROI at the superior sagittal sinus (SSS) and in the white matter of the frontal lobe of an unaffected hemisphere. These measurements were done on the first image of the dynamic perfusion series and at the level of the peak enhancement. The difference of these two values was also calculated. For all four precalculated perfusion parameters a manually drawn ROI measurement was performed at the white matter of the frontal lobe of an unaffected hemisphere to assess quantitative values of CBF, CBF, MTT and TTP. Two board certified neuroradiologists, who were masked to all clinical data and to the sequence of CTA and PCT acquisition, evaluated the CTA data of all examinations in consensus in a randomized fashion. For each dataset, readers had to fill in a case report form, subjectively assessing contrast enhancement quality of the intracranial and extracranial arteries on a visual analogous scale (VAS with a range from 1 = “excellent” to 5 = “very poor”), venous contamination compromising the view of the extracranial ICA (on a VAS ranging from 1 = “none” to 5 = “extreme contamination”) and diagnostic quality of the cavernous segments of the ICA (on a VAS ranging from 1 = “excellent, no venous contamination” to 5 = “very poor, massive overlay due to venous enhancement”). For statistical analysis, all data were collected in a custom designed database using standard software (Access 2003, Microsoft, Redmond, USA). Standard statistical software (SPSS, IBM Inc., Chicago, USA) was used for statistical testing. Levene’s test for equality of variances was applied to all criteria of the two groups. The Wilcoxon matched-pairs signed-rank test was applied to test the differences of “venous contamination” of the samples for significance. Student’s t-test was used to test the other criteria of the samples for significance. A p-value of less than 0.05 was considered to indicate statistical significance.

3. Results There were no significant differences between the groups regarding the age and gender distribution. When evaluating the CTA parameters, the mean density of the internal jugular vein was significantly higher in group B, i.e. when PCT was performed prior to CTA (p < 0.01), while there were no significant differences between the groups regarding the mean density of the ICA (p = 0.90). The readers rated both groups to be equal regarding the subjective contrast of the intracranial arteries (p = 0.55) as well as the extracranial arteries (p = 0.73). However, the extent of venous superimposition of the ICA was rated to be higher (p = 0.04) and the diagnostic image quality of the cavernous segments of the ICA to be worse in group B (p = 0.03), see Table 1. Regarding the PCT raw data, the base line attenuation and the attenuation peak in the superior sagittal sinus were significantly higher in group A, i.e. in the group, in which CTA was performed prior to PCT (both p < 0.01). However, the delta of the attenuation in

e288

D. Morhard et al. / European Journal of Radiology 82 (2013) e286–e289

Table 1 Comparison of attenuation, image quality and venous contamination compromising the assessment of the extracranial ICA between CTA of groups A and B.

Mean attenuation of ICA [HU] Mean attenuation of JV [HU] Subjective contrast attenuation Intracranial arteries Extracranial arteries [1 = excellent; 5 = very poor] Venous contamination compromising the view of extracranial ICA [1 = none; 5 = extreme contamination] Diagnostic image quality of cavernous segments of ICA [1 = excellent; 5 = very poor]

the superior sagittal sinus did not significantly differ between the groups (p = 0.31). The baseline attenuation as well as the enhancement peak in the brain parenchyma did not significantly differ (p = 0.35 and p = 0.15 respectively), while the delta of the attenuation in the brain parenchyma was significantly higher in group B (p = 0.02), see Table 2. When analyzing the dynamic perfusion parameters, neither CBF, nor CBV, TTP or MTT differed between the groups (p = 0.07, 0.08, 0.14, 0.27 respectively). In all cases quality of CTA and PCT was good enough for diagnostic evaluation. The time between the two scans was significantly lower in group A compared to group B (group A mean: 3.2 ± 2.2 min, median: 2 min, range: 1–9 min; group B mean: 5.0 ± 3.3 min, median: 4 min, range: 2–11 min; p < 0.01). 4. Discussion In acute stroke evaluation it is crucial to optimize comprehensive stroke CT protocols in order to minimize examination and image interpretation times and to thereby reduce the delay to the initiation of therapy. Our results demonstrate that the dynamic perfusion parameters CBF, CBV, TTP and MTT were not significantly influenced when CTA was performed prior to PCT. These data are in close agreement with previous publications in MR imaging and CT [14,15]. When PCT is performed before CTA, it is customary to wait for several minutes in order to reduce venous overlay by the contrast preload. This prolongs the imaging protocol for the respective time period. In our study, prolongation was about 2 min, but did not influence subjective contrast attenuation (see Table 1). Thus, we believe that a delay of 2–3 min is a sufficient compromise between fast image acquisition as well as rapid therapeutic decision and image quality when acquiring PCT prior to CTA. In future, it might be

Group A: CTA before PCT

Group B: PCT before CTA

Level of significance

420 ± 108 185 ± 64

423 ± 84 233 ± 63

p = 0.90 p < 0.01

1.7 ± 0.8 1.7 ± 0.8

1.6 ± 0.8 1.7 ± 0.8

p = 0.55 p = 0.73

2.7 ± 1.1

3.1 ± 0.9

p = 0.04

2.8 ± 1.2

3.3 ± 1.2

p = 0.03

possible to decrease the amount of venous superimposition of the arterial vessels due to contrast preload by PCT (group B): By using iterative image reconstructions to obtain a better signal-to noise ratio or by modification of PCT-deconvolution algorithm, reduction of contrast agent can be achieved [22,23]. When acquiring CTA before PCT it is to be expected that venous overlay is absent and image quality of vessel segments suffering from contrast preload should be improved. Accordingly, our data demonstrated that the extent of venous superimposition of the ICA was significantly lower and the diagnostic image quality of the cavernous segments of the ICA was better when CTA was performed first. However, a contrast preload by CTA could also influence perfusion parameters in PCT. It is yet unclear, which delay times between CTA and PCT are optimal when reversing the classical order. In our study, the time delay between CTA and PCT was significantly shorter than between PCT and CTA. Even though the delay between CTA and PCT was short in our study and the effect of the contrast preload is expected to still be present, perfusion parameters were not significantly influenced (see Table 2). Most likely, delay times between CTA and PCT could be even further reduced. However, prospective studies will be needed to clarify this point. About 20% of ischemic strokes are found in the posterior circulation [3], with acute occlusion of the basilar artery potentially being the most dangerous [24,25]. Unlike the anterior circulation there is no set therapeutic window in basilar artery occlusion, as the spontaneous outcome without revascularization therapy carries a mortality rate of 70–80%. PCT, however, is very limited in the posterior circulation, due to pronounced artifacts from the posterior fossa and small vascular territories. Therefore, PCT could be omitted when CTA shows a clear thrombembolic occlusion of the vertebrobasilar system. In this setting, it is advantageous to perform CTA first in order to obtain a fast initial diagnosis and to initiate therapy without further delay.

Table 2 Comparison of attenuation and PCT parameters between groups A and B. Group A: CTA before PCT

Group B: PCT before CTA

Level of significance

SSS Base line attenuation [HU] Attenuation peak [HU] Delta of attenuation [HU]

124 ± 39 366 ± 145 242 ± 122

48 ± 11 265 ± 108 216 ± 106

p < 0.01 p < 0.01 p = 0.31

Brain parenchyma Base line attenuation [HU] Attenuation peak [HU] Delta of attenuation [HU]

43 ± 5 58 ± 10 15 ± 7

42 ± 4 61 ± 10 19 ± 8

p = 0.35 p = 0.15 p = 0.02

CBF [ml/100 ml/min] CBV [ml/100 ml] TTP [s] MTT [s]

53 3.0 9.0 3.2

± ± ± ±

p = 0.07 p = 0.08 p = 0.14 p = 0.27

± ± ± ±

9 0.6 1.8 0.4

58 3.2 9.6 3.3

11 0.7 1.9 0.6

D. Morhard et al. / European Journal of Radiology 82 (2013) e286–e289

On clinical grounds alone, it can be very challenging to differentiate between severe ischemic events of the anterior and of the posterior circulation. Therefore, the PCT imaging slab cannot be individually placed, if PCT is performed prior to CTA. The detection of the site of occlusion, however, can aid in choosing the optimal scan level and in accelerating the diagnostic and therapeutic decision making. Previous studies have shown that automatic or semiautomatic bone removal significantly shortens the reading time of CTA images in patients with suspected cerebrovascular events [20,21]. Diagnostic decision-making on the basis of CTA images is known to be shorter and more efficient if there is little or no venous overlay and no interference of bone. Moreover, segmentation algorithms, e.g. dual energy bone removal, function better in the absence of venous overlay, especially in the skull base [26]. It therefore seems to be advantageous to eliminate venous overlay by performing CTA directly after NECT. There are some limitations to our study that need to be kept in mind when interpreting the results. As the approach was retrospective, there was no randomization of the scanning order. Moreover, there were no randomized subgroups with different interval lengths between CTA and PCT and between PCT and CTA respectively. In order to study the effects of potentially further reducing the interval between CTA and PCT, further studies with larger patient populations would be needed. Overall, the results of our study demonstrate that the reversal of the classical sequence of comprehensive CT stroke imaging does not interfere with quantitative PCT parameters and increases image quality of the CT angiographic images. Moreover, the time of the diagnostic CT scan can be significantly reduced and the PCT slab can be optimally placed on the basis of the CTA data. Disclosure There are no conflicts of interest, nor were there any sources of funding which could introduce bias to the scientific work of any of the authors. References [1] Wardlaw JM, Seymour J, Cairns J, Keir S, Lewis S, Sandercock P. Immediate computed tomography scanning of acute stroke is cost-effective and improves quality of life. Stroke 2004;35(11):2477–83. [2] Wardlaw JM, Stevenson MD, Chappell F, et al. Carotid artery imaging for secondary stroke prevention: both imaging modality and rapid access to imaging are important. Stroke 2009;40(11):3511–7. [3] Ringleb P, Schellinger PD, Hacke W. European Stroke Organisation 2008 guidelines for managing acute cerebral infarction or transient ischemic attack. Part 1. Der Nervenarzt 2008;79(8):936–57. [4] Wintermark M, Albers GW, Alexandrov AV, et al. Acute stroke imaging research roadmap. Stroke: A Journal of Cerebral Circulation 2008;39(5):1621–8. [5] Schaefer PW, Barak ER, Kamalian S, et al. Quantitative assessment of core/penumbra mismatch in acute stroke: CT and MR perfusion imaging are strongly correlated when sufficient brain volume is imaged. Stroke: A Journal of Cerebral Circulation 2008;39(11):2986–92.

e289

[6] Lovblad KO, Baird AE. Computed tomography in acute ischemic stroke. Neuroradiology 2010;52(3):175–87. [7] Puetz V, Dzialowski I, Hill MD, et al. Intracranial thrombus extent predicts clinical outcome, final infarct size and hemorrhagic transformation in ischemic stroke: the clot burden score. International Journal of Stroke 2008;3(4): 230–6. [8] Tomandl BF, Klotz E, Handschu R, et al. Comprehensive imaging of ischemic stroke with multisection CT. Radiographics 2003;23(3):565–92. [9] Koenig M, Klotz E, Luka B, Venderink DJ, Spittler JF, Heuser L. Perfusion CT, of the brain: diagnostic approach for early detection of ischemic stroke. Radiology 1998;209(1):85–93. [10] Tan JC, Dillon WP, Liu S, Adler F, Smith WS, Wintermark M. Systematic comparison of perfusion-CT and CT-angiography in acute stroke patients. Annals of Neurology 2007;61(6):533–43. [11] Hopyan J, Ciarallo A, Dowlatshahi D, et al. Certainty of stroke diagnosis: incremental benefit with CT perfusion over noncontrast CT and CT angiography. Radiology 2010;255(1):142–53. [12] Wardlaw GM, Wong R, Boylan C, Rebello R, Noseworthy MD. Investigation of a logistic model for T2* dynamic susceptibility contrast magnetic resonance imaging (dscMRI) perfusion studies. Journal of Computer Assisted Tomography 2011;35(6):728–33. [13] Shinohara Y, Ibaraki M, Ohmura T, et al. Whole-brain perfusion measurement using 320-detector row computed tomography in patients with cerebrovascular steno-occlusive disease: comparison with 15O-positron emission tomography. Journal of Computer Assisted Tomography 2010;34(6): 830–5. [14] Dorn F, Liebig T, Muenzel D, et al. Order of CT stroke protocol (CTA before or after CTP): impact on image quality. Neuroradiology 2012. [15] Ryu CW, Lee DH, Kim HS, et al. Acquisition of MR perfusion images and contrastenhanced MR angiography in acute ischaemic stroke patients: which procedure should be done first? British Journal of Radiology 2006;79(948):962–7. [16] Konstas AA, Wintermark M, Lev MH. CT perfusion imaging in acute stroke. Neuroimaging Clinics of North America 2011;21(2):215–38, ix. [17] Johnson MM. Stroke and CT perfusion. Radiologic Technology 2012;83(5):467CT–86CT. [18] Lell MM, Ruehm SG, Kramer M, et al. Cranial computed tomography angiography with automated bone subtraction: a feasibility study. Investigative Radiology 2009;44(1):38–43. [19] Jayakrishnan VK, White PM, Aitken D, Crane P, McMahon AD, Teasdale EM. Subtraction helical CT angiography of intra- and extracranial vessels: technical considerations and preliminary experience. American Journal of Neuroradiology 2003;24(3):451–5. [20] Morhard D, Fink C, Becker C, Reiser MF, Nikolaou K. Value of automatic bone subtraction in cranial CT angiography: comparison of bone-subtracted vs. standard CT angiography in 100 patients. European Radiology 2008;18(5): 974–82. [21] Morhard D, Fink C, Graser A, Reiser MF, Becker C, Johnson TR. Cervical and cranial computed tomographic angiography with automated bone removal: dual energy computed tomography versus standard computed tomography. Investigative Radiology 2009;44(5):293–7. [22] Itatani R, Oda S, Utsunomiya D, et al. Reduction in radiation and contrast medium dose via optimization of low-kilovoltage CT protocols using a hybrid iterative reconstruction algorithm at 256-slice body CT: Phantom study and clinical correlation. Clinical Radiology 2012, http://dx.doi.org/10.1016/j.crad.2012.10.014, pii: S0009-9260(12)00545-4. [Epub ahead of print], PMID: 23245269. [23] Yamada Y, Jinzaki M, Hosokawa T, et al. Dose reduction in chest CT: comparison of the adaptive iterative dose reduction 3D, adaptive iterative dose reduction, and filtered back projection reconstruction techniques. European Journal of Radiology 2012;81(12):4185–95. [24] Pfefferkorn T, Holtmannspotter M, Schmidt C, et al. Drip, ship, and retrieve: cooperative recanalization therapy in acute basilar artery occlusion. Stroke 2010;41(4):722–6. [25] Schulte-Altedorneburg G, Mayer TE. Management of acute basilar artery occlusion. Radiologe 2007;47(4):355–8. [26] Johnson TR, Krauss B, Sedlmair M, et al. Material differentiation by dual energy CT: initial experience. European Radiology 2007;17(6):1510–7.