Original Research
Neuroradiology/Head and Neck Imaging
June 2009

Feasibility of Superficial Temporal Artery as the Input Artery for Cerebral Perfusion CT

Abstract

OBJECTIVE. The purpose of this study was to determine whether the superficial temporal artery as a surrogate arterial input function, compared with the anterior cerebral artery as the arterial input function, generates accurate perfusion CT maps with significant correlates for cerebral blood flow, cerebral blood volume, and mean transit time.
MATERIALS AND METHODS. One hundred perfusion CT examinations performed on 90 patients (51 women and girls, 39 men and boys) were retrospectively reviewed and postprocessed by one investigator using CT perfusion software at a workstation. Color-coded cerebral blood flow, cerebral blood volume, and mean transit time maps were constructed with the superficial temporal artery as a surrogate arterial input function and the anterior cerebral artery as the arterial input function. The effect of input artery choice on mean cerebral blood flow, cerebral blood volume, and mean transit time values in six regions of interest (one region of interest in each anterior cerebral artery, middle cerebral artery, and posterior cerebral artery territory) was assessed.
RESULTS. All graphs of correlation between the anterior cerebral artery as the arterial input function and the superficial temporal artery as a surrogate arterial input function produced significant results (p < 0.0001). Excellent correlation existed between the cerebral blood flow (r = 0.96 [Pearson correlation coefficient]; ρc = 0.96 [concordance correlation coefficient]), cerebral blood volume (r = 0.97; ρc = 0.97), and mean transit time (r = 0.97; ρc = 0.97) values obtained with the anterior cerebral artery and the values obtained with the superficial temporal artery. Linear regression lines produced strong agreement between use of the anterior cerebral artery and use of the superficial temporal artery (cerebral blood flow, y = 1.03x + 0.65; cerebral blood volume, y = 1.05x – 0.09; mean transit time, y = 0.92x + 0.21).
CONCLUSION. The preliminary results show that using an extracranial vessel such as the superficial temporal artery as a surrogate input artery can generate perfusion maps comparable with those generated with an intracranial vessel such as the anterior cerebral artery. This knowledge can be useful if the proximal intracranial vessels typically used for arterial input are not visible owing to diffuse disease, such as diffuse vasospasm and atherosclerosis, or are not included in the field of view of perfusion CT.

Introduction

Imaging is essential in the evaluation of stroke symptoms to help exclude intracranial hemorrhage and to detect signs of brain ischemia. The advent of thrombolytic therapy for acute nonhemorrhagic stroke has intensified the need for rapid imaging to help identify and quantify the presence and extent of a perfusion deficit. Evaluation of brain perfusion can help in the selection of patients for thrombolytic therapy by providing information that helps differentiate patients with potentially salvageable tissue at risk of infarction (ischemic penumbra) from those with extensive infarcts [1]. Perfusion imaging to assess perfusion and cerebrovascular reserves also may be helpful to patients with chronic cerebrovascular disease. Perfusion MRI, xenon CT, PET, and SPECT have been used to evaluate cerebral perfusion, but all are limited by availability, cost, patient tolerance, and lack of quantitative results. Perfusion CT was introduced as a means of rapid evaluation of cerebral perfusion in patients with acute stroke symptoms, most of whom would already be undergoing unenhanced head CT to exclude acute hemorrhage [2].
Perfusion CT is used to generate quantitative maps of cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit time (MTT). Calculation of CBF is based on the central volume principle, which states that the CBF is equal to the ratio of CBV and MTT (CBF = CBV / MTT) [3]. Perfusion CT requires the selection of input variables. The arterial input function (AIF) is usually a branch of the anterior cerebral artery (ACA) or the middle cerebral artery (MCA). Ideally, the input artery should be a vessel seen in cross section. The venous input function is usually a large dural venous sinus, such as the torcular herophili or transverse sinus (Fig. 1A, 1B).
Fig. 1A Venous input function. Anteroposterior venous phase angiogram of right internal carotid artery shows torcular herophili (arrowhead) and transverse (arrows) sinuses.
Fig. 1B Venous input function. Axial perfusion CT image shows torcular herophili (arrow) in cross section.
Fig. 2A Surrogate arterial input function. Lateral angiogram of external carotid artery shows branches (arrows) of superficial temporal artery.
Fig. 2B Surrogate arterial input function. Axial perfusion CT image shows superficial temporal artery (arrow) in cross section overlying calvarium.
Results of some studies have suggested that a diseased vessel should not be used as the arterial input, particularly with software that is sensitive to the tracer-delay effect [4]. Avoidance of a diseased vessel can be difficult, however, if it is not known at the time of perfusion CT that the vessel is diseased or whether multifocal cerebrovascular disease is present [5, 6]. Extracranial arteries have been used to generate perfusion maps in studies with animal and with human subjects [79]. Results of a more recent study [10] have indicated that the ACA is an appropriate AIF for perfusion CT studies regardless of disease status but that the ACA might not be visible or might not be included in the field of view of the perfusion examination. Because its course allows it to be seen in cross section on perfusion CT images, we chose the superficial temporal artery (STA) as an alternative to an intracranial vessel (Fig. 2A, 2B). The purpose of our study was to determine the feasibility of identifying and using the STA as a surrogate AIF for generation of perfusion CT maps and to determine the accuracy of perfusion CT maps generated with the STA as a surrogate input artery compared with those made with the ACA as the AIF.

Materials and Methods

Data sets from 100 perfusion CT examinations of 90 patients were retrospectively analyzed. Because all links to subject identifiers were removed, institutional review board waiver of consent was obtained. All patients had undergone evaluation for known cerebrovascular disease: 51 patients with extracranial stenosis, occlusion, bypass, or dissection; 18 with intracranial stenosis, occlusion, or bypass; 10 with moyamoya disease; nine with subarachnoid hemorrhage, and two with internal carotid arterial and intracranial stenosis or occlusion (Table 1). Among the 100 perfusion CT examinations, 14 were performed for acute cerebrovascular symptoms (symptom onset within 24 hours preceding CT perfusion examination), 22 for subacute cerebrovascular disease (symptom onset within 25 hours–2 months preceding CT perfusion examination), and 62 for chronic cerebrovascular disease (symptom onset more than 2 months preceding CT perfusion examination). For two patients, the acuity of symptoms could not be determined other than that they were not acute (Table 2). Fifty-one women and girls (age range, 5–85 years) and 39 men and boys (age range, 11–86 years) were included in the study.
TABLE 1: Patient Demographics in 100 Examinations
SexAge (y)Disease Characteristics
M46RMCA occlusion
M73LVA occlusion, RVA stenosis, LICA stenosis
F38LICA occlusion
M34RMCA lacunar infarct
F49Fibromuscular dysplasia, dynamic occlusion
M66RICA occlusion
F70LICA occlusion, meningioma
M39RICA stenosis, occlusion of M2 segment of MCA
F63RICA severe stenosis, LICA moderate stenosis
F19Moyamoya disease
M79RMCA ischemia
M67RVA occlusion, LVA stenosis, RICA stenosis
F29LMCA aneurysm clip
M58RICA stenosis
F36LICA dissection
M6Moyamoya disease
M81LMCA stenosis
F37Moyamoya disease
F80LICA occlusion
M77Left common carotid and bilateral vertebral arterial occlusion
M72Bilateral carotid arterial occlusion
F69Left cavernous ICA stenosis
F11Moyamoya disease
M62Recurrent left skull base meningioma, LICA occlusion
F57RICA occlusion
F51Left stenosis of M1 segment of MCA after stent insertion
F42Left stenosis of A2 segment of ACA
M54Bilateral CEA
M51Right M1 occlusion
F59Innominate, right carotid stenosis, left subclavian stenosis with steal
F63SAH
F58Takayasu arteritis, left carotid stenosis
M53LICA occlusion after CVA, 46% stenosis of RICA
M54LICA occlusion, 90% stenosis of RICA
F54Moyamoya disease
M5Right M1 stenosis
M65Left carotid occlusion
M45LMCA and M1 stenosis after watershed CVA
F63LCCA to left subclavian artery bypass, right cerebellar CVA
F62Bilateral subclavian artery stenosis, left > right, 50% LCCA stenosis, cerebrovascular obstructive disease
M69Right carotid stenosis, left carotid occlusion, right vertebral stenosis
M85Right CVA
F47RMCA stenosis after CVA, two aneurysm clips
M69LICA occlusion, MCA aneurysm clip
F16Moyamoya disease after revascularization
F14Moyamoya disease after revascularization
M80RICA occlusion
M46Aftermath of right STA to MCA bypass
M46RICA occlusion
F50SAH after RICA aneurysm clip
F7090% stenosis of RICA, hypoplastic left vertebral artery
M37RICA stenosis, M2 occlusion
M36RICA stenosis, M2 occlusion
F31Moyamoya disease, 80% stenosis of RICA, >90% stenosis of LICA
F47Ruptured left PCOM clip
F44Ruptured ACOM clip
M55LICA occlusion, RICA stenosis, aftermath of CVA
F52Left ophthalmic artery aneurysm clip
F52Left ophthalmic artery aneurysm clip
M17RICA stenosis
F48LICA sacrifice, paraganglioma
F48LICA sacrifice, paraganglioma
F35Bilateral ICA occlusion, 60% stenosis of LVA
F64LICA stenosis
F35Moyamoya disease
F55Bilateral ICA occlusion, 60% stenosis of LVA
F33RICA occlusion after CVA
F32RICA occlusion after CVA
F29LMCA aneurysm clip, LMCA occlusion
M29Moyamoya disease, LICA occlusion after bypass
M29Moyamoya disease, LICA occlusion after bypass
F6790% stenosis of basilar artery
F46RICA occlusion
F37Moyamoya disease
F32Moyamoya disease
F61LVA occlusion
F13Moyamoya disease after left encephalodural synangiosis
F76RICA occlusion
F69Left PCOM aneurysm after SAH
M56ACOM aneurysm after SAH
M51Aftermath of aortic surgery
F53Right carotid stenosis
F61RICA occlusion
F55Cardiomyopathy, encephalopathy
M59Stenosis of M2 segment of RMCA
F57Left subclavian steal
M56RICA occlusion after right CEA, RCCA to ECA bypass
M56RICA occlusion after right CEA, RCCA to ECA bypass
M52LMCA stenosis, LICA fusiform aneurysm
M52LMCA stenosis, LICA fusiform aneurysm
M46Aftermath of embolization of RCCA for tumor
M65SAH, vermian bleed, embolization
F86SAH after coiling
F63SAH
F63SAH
F51LICA occlusion
M45HIV infection, cryptococcal meningitis
M53Temporary balloon occlusion of left infratemporal fossa mass
F48RICA occlusion, history of aneurysm
M
62
Transient ischemic attack with slurred speech and ataxia
Note—RMCA = right middle cerebral artery, LVA = left vertebral artery, RVA = right vertebral artery, LICA = left internal carotid artery, RICA = right internal carotid artery, MCA = middle cerebral artery, LMCA = left middle cerebral artery, ICA = internal carotid artery, ACA = anterior cerebral artery, CEA = carotid endarterectomy, SAH = subarachnoid hemorrhage, CVA = cerebrovascular accident, LCCA = left common carotid artery, STA = superficial temporal artery, PCOM = posterior communicating artery, ACOM = anterior communicating artery, RCCA = right common carotid artery, ECA = external carotid artery.
TABLE 2: Symptom Acuity at Perfusion CT (n = 100)
Disease AcuityNo. of Patients
Acute14
Subacute22
Chronic62
Unknown
2
The data sets were acquired with an MDCT scanner (LightSpeed, GE Healthcare). Unenhanced head CT and perfusion CT were performed. For perfusion CT, 50 mL of nonionic contrast material (300 mg I/mL) was injected at a rate of 4 mL/s. Five seconds into the injection, a cine (continuous) scan was initiated with the following parameters: 80 kVp, 190–200 mA, 4 × 5 mm slices, and 1 rotation/s for 50 seconds. The 4 × 5 mm slice selection was made at the level of the basal ganglia to include portions of the ACA, MCA, and posterior cerebral artery (PCA) territories. The 1-second images were reformatted at 0.5-second intervals, and the 5-mm slices were reformatted into two 10-mm-thick slices. The scans were obtained at 5 mm, rather than 10 mm, to lessen beam-hardening artifacts in the brain. The reformatted 10-mm-thick slices had a better signal-to-noise ratio than the 5-mm-thick slices and better temporal resolution (0.5-second cine interval). The perfusion CT images acquired were transferred to an imaging workstation (Advantage Windows, GE Healthcare) for postprocessing of CBF, CBV, and MTT maps with commercially available software (CT Perfusion, GE Healthcare).
A single investigator constructed color-coded CBF, CBV, and MTT maps for all 90 patients. Sets of these maps were constructed for each patient by varying a single user-defined parameter, AIF, while the other parameters were held constant at each of the two slices. Thus, a total of four sets of CBF, CBV, and MTT maps per patient were generated. Either the STA was chosen as the surrogate AIF or the ACA was chosen as the AIF. A large dural venous sinus, such as the torcular herophili or the transverse sinus, was chosen as the venous input function. Time–attenuation curves were generated for each input artery and vein (Fig. 3).
Six circular regions of interest (ROIs) with a diameter of approximately 2.5 cm each were placed along the cortical mantle in the bilateral ACA, MCA, and PCA territories (Fig. 4A, 4B). ROIs were identical in both size and placement for each patient to eliminate a potential source of error in comparison of the different maps produced per set. Quantitative perfusion values of CBF, CBV, and MTT were obtained in all six ROIs in each slice (Figs. 5A, 5B, 5C, 5D, 5E, 5F and 6A, 6B, 6C). The perfusion values were averaged over the two slice locations to obtain mean values of CBF, CBV, and MTT with the ACA as the input artery and the with the STA as a surrogate input artery.
We used linear regression analysis to assess the differences between the perfusion parameters obtained with the ACA and the STA. The Pearson linear correlation coefficient (r) was calculated to estimate the correlation of the perfusion parameters obtained with the ACA and the STA. The concordance correlation coefficient (ρc) was calculated to assess the reproducibility of the results. These correlations were estimated separately for each ROI and slice and overall. Ninety-five percent CIs for the correlation coefficients also were estimated. We used linear regression analysis to further evaluate the nature of any disagreement between ACA and STA data in calculating perfusion parameters. Specifically, we regressed the STA perfusion parameters on the ACA perfusion parameters and then estimated slopes and intercepts.

Results

The mean CBF, CBV, and MTT values for the ACA, MCA, and PCA territories obtained with the ACA as the AIF and the STA as a surrogate AIF are summarized in Figure 6A, 6B, 6C. All graphs of correlation between ACA and STA input arteries showed significant results (p < 0.0001) (Fig. 7A, 7B, 7C). Linear regression lines showed strong agreement between ACA and STA values (Table 3).
TABLE 3: Linear Regression Results Showing Close Agreement Between Values Obtained with Anterior Cerebral Artery and Superficial Temporal Artery
Perfusion ParameterSlopey-Intercept
Cerebral blood flow1.030.65
Cerebral blood volume1.050.09
Mean transit time
0.92
0.21
With the ACA as the AIF, mean CBF in the ACA territory was 48.59 mL/100 g/min, in the MCA territory was 66.03 mL/100g/min, and in the PCA territory was 77.48 mL/100 g/min. Overall mean CBF was 64.04 ± 13.53 mL/100 g/min. With the STA as a surrogate AIF, mean CBF in the ACA territory was 50.83 mL/100 g/min, in the MCA territory was 68.57 mL/100 g/min, and in the PCA territory was 81.29 mL/100 g/min. Overall mean CBF was 66.9 ± 14.34 mL/100 g/min.
With the ACA as the AIF, mean CBV in the ACA territory was 2.38 mL/100 g, in the MCA territory was 2.95 mL/100 g, and in the PCA territory was 3.76 mL/100 g. Overall mean CBV was 3.03 ± 0.65 mL/100 g. With the STA as a surrogate AIF, mean CBV in the ACA territory was 2.4 mL/100 g, in the MCA territory was 2.99 mL/100 g, and in the PCA territory was 3.86 mL/100 g. Overall mean CBV was 3.08 ± 0.68 mL/100 g.
With the ACA as the AIF, the mean MTT in the ACA territory was 5.65 seconds, in the MCA territory was 5.06 seconds, and in the PCA territory was 5.68 seconds. The overall mean MTT was 5.46 ± 0.4 seconds. With the STA as a surrogate AIF, the mean MTT in the ACA territory was 5.54 seconds, in the MCA territory was 4.95 seconds, and in the PCA territory was 5.56 seconds. The overall mean MTT was 5.35 ± 0.37 seconds.
Regression analysis of the pooled ROIs (both hemispheres) for all patients revealed close agreement between the perfusion parameter values measured with the ACA as the arterial input and the perfusion parameter values measured with the STA as a surrogate arterial input. Perfect agreement occurs when the intercept equals 0 and the slope equals 1. Table 3 shows estimated values for intercepts and slopes to be close to 0 and 1 for CBF, CBV, and MTT. The linear (Pearson) correlation coefficients for CBF, CBV, and MTT were 0.96, 0.97, and 0.97, respectively, all significant at p < 0.001. The concordance correlation coefficients for CBF, CBV, and MTT were 0.96, 0.97, and 0.97, respectively, all significant at p < 0.001. The lower 95% confidence bounds for the concordance correlation coefficients were 0.95, 0.97, and 0.96, respectively. The lower 95% confidence bounds for the concordance correlation coefficients based on separate analysis of each slice and region ranged from 0.92 to 0.95 for CBF, 0.92 to 0.95 for CBV, and 0.91 to 0.96 for MTT.
Fig. 3 Artery versus vein time–attenuation curves. Curve for input artery has earlier rise (black arrow) and earlier peak attenuation (green arrow) compared with that of known venous structure, such as torcular herophili (yellow arrows). ↓

Discussion

Since the late 1990s, perfusion CT has been used increasingly to measure cerebral perfusion in a variety of clinical settings. This technique can be performed quickly with any standard helical CT scanner, and the perfusion maps can be generated in a short time at a workstation equipped with the appropriate software. Perfusion CT can be used to assess not only acute stroke but also a wide range of other cerebrovascular diseases. It also may be helpful in the diagnosis of a variety of tumors and in evaluation of the treatment response.
Deconvolution is one of the techniques used for calculation of perfusion parameters and allows generation of color-coded quantitative maps of CBF, CBV, and MTT from axial perfusion CT source images. This model requires selection of an ROI in an artery and a vein to represent the arterial and venous time–attenuation curves, which reflect the arterial input and venous outflow from the parenchymal capillary network. The AIF is required mathematically to perform the deconvolution, and the venous curve is required to correct for volume-averaging effects in the arterial input. Theoretic guidelines exist about the optimal selection of these user-defined input variables for the generation of CBV, CBF, and MTT maps [7]. A branch of the carotid artery, either the ACA or the MCA, is usually selected as the AIF, and a large dural venous sinus, such as the torcular herophili, superior sagittal sinus, or transverse sinus, is chosen as the venous input function.
Fig. 4A 65-year-old man with right internal carotid artery occlusion. Perfusion CT images show anterior cerebral artery (ACA) (A) and superficial temporal artery (STA) (B) as input arteries. Two-centimeter circular regions of interest are present in each anterior cerebral artery, middle cerebral artery, and posterior cerebral artery territory.
Fig. 4B 65-year-old man with right internal carotid artery occlusion. Perfusion CT images show anterior cerebral artery (ACA) (A) and superficial temporal artery (STA) (B) as input arteries. Two-centimeter circular regions of interest are present in each anterior cerebral artery, middle cerebral artery, and posterior cerebral artery territory.
Fig. 5A 65-year-old man with right internal carotid artery occlusion (same patient as in Fig. 4A, 4B). Cerebral blood flow (A and D), cerebral blood volume (B and E), and mean transit time (C and F) perfusion maps show results with anterior cerebral artery (ACA) (A–C) and superficial temporal artery (D–F) as input arteries.
Fig. 5B 65-year-old man with right internal carotid artery occlusion (same patient as in Fig. 4A, 4B). Cerebral blood flow (A and D), cerebral blood volume (B and E), and mean transit time (C and F) perfusion maps show results with anterior cerebral artery (ACA) (A–C) and superficial temporal artery (D–F) as input arteries.
Fig. 5C 65-year-old man with right internal carotid artery occlusion (same patient as in Fig. 4A, 4B). Cerebral blood flow (A and D), cerebral blood volume (B and E), and mean transit time (C and F) perfusion maps show results with anterior cerebral artery (ACA) (A–C) and superficial temporal artery (D–F) as input arteries.
Fig. 5D 65-year-old man with right internal carotid artery occlusion (same patient as in Fig. 4A, 4B). Cerebral blood flow (A and D), cerebral blood volume (B and E), and mean transit time (C and F) perfusion maps show results with anterior cerebral artery (ACA) (A–C) and superficial temporal artery (D–F) as input arteries.
Fig. 5E 65-year-old man with right internal carotid artery occlusion (same patient as in Fig. 4A, 4B). Cerebral blood flow (A and D), cerebral blood volume (B and E), and mean transit time (C and F) perfusion maps show results with anterior cerebral artery (ACA) (A–C) and superficial temporal artery (D–F) as input arteries.
Fig. 5F 65-year-old man with right internal carotid artery occlusion (same patient as in Fig. 4A, 4B). Cerebral blood flow (A and D), cerebral blood volume (B and E), and mean transit time (C and F) perfusion maps show results with anterior cerebral artery (ACA) (A–C) and superficial temporal artery (D–F) as input arteries.
Fig. 6A Graphs of region-of-interest (RO1) values for cerebral blood flow, cerebral blood volume, and mean transit time. Graphs show values obtained from cerebral blood flow (A), cerebral blood volume (B), and mean transit time (C) perfusion maps. R-ACA = right anterior cerebral artery, L-ACA = left anterior cerebral artery, R-MCA = right middle cerebral artery, L-MCA = left middle cerebral artery, R-PCA = right posterior cerebral artery, L-PCA = left posterior cerebral artery, STA = superficial temporal artery.
Fig. 6B Graphs of region-of-interest (RO1) values for cerebral blood flow, cerebral blood volume, and mean transit time. Graphs show values obtained from cerebral blood flow (A), cerebral blood volume (B), and mean transit time (C) perfusion maps. R-ACA = right anterior cerebral artery, L-ACA = left anterior cerebral artery, R-MCA = right middle cerebral artery, L-MCA = left middle cerebral artery, R-PCA = right posterior cerebral artery, L-PCA = left posterior cerebral artery, STA = superficial temporal artery.
Fig. 6C Graphs of region-of-interest (RO1) values for cerebral blood flow, cerebral blood volume, and mean transit time. Graphs show values obtained from cerebral blood flow (A), cerebral blood volume (B), and mean transit time (C) perfusion maps. R-ACA = right anterior cerebral artery, L-ACA = left anterior cerebral artery, R-MCA = right middle cerebral artery, L-MCA = left middle cerebral artery, R-PCA = right posterior cerebral artery, L-PCA = left posterior cerebral artery, STA = superficial temporal artery.
Fig. 7A Correlation between results obtained with anterior cerebral artery as arterial input function and superficial temporal artery as surrogate arterial input function. Graphs show excellent correlation between cerebral blood flow (A), cerebral blood volume (B), and mean transit time (C) values obtained with anterior cerebral artery and superficial temporal artery.
Fig. 7B Correlation between results obtained with anterior cerebral artery as arterial input function and superficial temporal artery as surrogate arterial input function. Graphs show excellent correlation between cerebral blood flow (A), cerebral blood volume (B), and mean transit time (C) values obtained with anterior cerebral artery and superficial temporal artery.
Fig. 7C Correlation between results obtained with anterior cerebral artery as arterial input function and superficial temporal artery as surrogate arterial input function. Graphs show excellent correlation between cerebral blood flow (A), cerebral blood volume (B), and mean transit time (C) values obtained with anterior cerebral artery and superficial temporal artery.
We chose the STA as an alternative vessel for extracranial input because it is less likely to be involved in cerebrovascular disease and its course allows it to be visualized in cross section on axial perfusion CT images. In our study, of the total 102 perfusion CT examinations studied, only two examinations were discarded because of inability to visualize the STA on axial CT images, an estimated visualization rate of at least 98.0%. Because we used the ACA as the validated standard, we did not include any patients in whom the ACA was not visible.
Fig. 8A Variation in time–attenuation curves for anterior cerebral artery (ACA) and superficial temporal artery (STA). 72-year-old man with left vertebral artery occlusion and stenoses of right vertebral and left internal carotid arteries. Time–attenuation curve for superficial temporal artery shows delayed contrast arrival time (curve 1) (black arrow, A) compared with that of anterior cerebral artery (black arrow, B) and greater dispersion of contrast material (curve 1) with superficial temporal artery (red arrow, A) than with anterior cerebral artery (green arrow, B).
Fig. 8B Variation in time–attenuation curves for anterior cerebral artery (ACA) and superficial temporal artery (STA). 72-year-old man with left vertebral artery occlusion and stenoses of right vertebral and left internal carotid arteries. Time–attenuation curve for superficial temporal artery shows delayed contrast arrival time (curve 1) (black arrow, A) compared with that of anterior cerebral artery (black arrow, B) and greater dispersion of contrast material (curve 1) with superficial temporal artery (red arrow, A) than with anterior cerebral artery (green arrow, B).
Fig. 8C Variation in time–attenuation curves for anterior cerebral artery (ACA) and superficial temporal artery (STA). 46-year-old man with right middle cerebral artery occlusion. Time–attenuation curves for superficial temporal artery (C) and anterior cerebral artery (D) show similar times for contrast arrival (curve 1) and dispersion of contrast material (curve 1).
Fig. 8D Variation in time–attenuation curves for anterior cerebral artery (ACA) and superficial temporal artery (STA). 46-year-old man with right middle cerebral artery occlusion. Time–attenuation curves for superficial temporal artery (C) and anterior cerebral artery (D) show similar times for contrast arrival (curve 1) and dispersion of contrast material (curve 1).
One possible limitation of using the STA as the surrogate AIF is the higher vascular resistance in this vessel compared with intracranial vessels such as the ACA [11]. In theory, differences in vascular resistance may have an effect on time–attenuation curves, causing greater delay in arrival of contrast material and greater dispersion of contrast material in the higher-resistance STA compared with the ACA. Although there was a difference in the shape of the time–attenuation curves for the STA compared with the ACA in some patients, this difference did not occur in all cases (Fig. 8A, 8B, 8C, 8D). Increased delay and dispersion can be expected to result in overestimation of MTT and underestimation of CBF [10, 12], which did not occur in our study. In addition, although the mean hemodynamic values are not statistically significantly different for the STA compared with the ACA, there is a trend toward greater CBV and CBF and shorter MTT when the STA is used. These results in our study may be related to the patient sample: Subacute to chronic disease was found in 84 of 100 perfusion CT examinations. In this group of patients, external carotid artery branches may provide collateral supply to the brain, resulting in vasodilation and lower vascular resistance in the extracranial vessels. This phenomenon may explain the similar appearance of the time–attenuation curves for the STA and ACA in some patients and the trend toward greater CBV and CBF and shorter MTT with use of the STA as a surrogate AIF.
Although external carotid artery stenosis is much less common than internal carotid artery stenosis, the assumption that no external carotid artery cerebrovascular disease is present is a limitation to the use of the STA as the arterial input. One of the other possible limitations of our study was the lack of normative data from healthy persons for perfusion CT with variation of the input artery. In addition, for quantification of CBF in perfusion CT, uniform hematocrit levels in both capillaries and large vessels were assumed; however, this condition does not hold true for the altered hematocrit levels that accompany chronic cerebrovascular disease [13].
The 2007 Acute Stroke Imaging Research Roadmap [14] suggests that the delay-compensated deconvolution method may be the most appropriate way to process perfusion CT data sets. However, without formal comparisons of the superiority of this method in terms of accurate representation of perfusion status and prediction of tissue outcome compared with previous deconvolution models, our results on the optimal selection of AIF are still applicable to those with current perfusion CT techniques. We found significant correlation between quantitative CBF, CBV, and MTT values obtained with the ACA as the AIF and the values obtained with the STA as a surrogate AIF, the strongest correlation occurring between CBV values. Thus, the STA is a viable alternative AIF for generation of perfusion CT maps if the ACA is not visible because of disease or is not included in the field of view.

Footnotes

K. Sheikh is supported by the 2006 Lita Annenberg Hazen Award for Neurobiology and the 2006 American Academy of Neurology Summer Research Scholarship.
Address correspondence to E. G. Hoeffner ([email protected]).
WEB
This is a Web exclusive article.

References

1.
Wintermark M, Reichhart M, Thiran JP, et al. Prognostic accuracy of cerebral blood flow measurement by perfusion computed tomography, at the time of emergency room admission, in acute stroke patients. Ann Neurol 2002; 51:417 –432
2.
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; 209:85–93
3.
Hoeffner EG, Case I, Jain R, et al. Cerebral perfusion CT: technique and clinical applications. Radiology 2004; 231:632 –644
4.
Ibaraki M, Shimosegawa E, Toyoshima H, et al. Effect of regional tracer delay on CBF in healthy subjects measured with dynamic susceptibility contrast-enhanced MRI: comparison with 15O-PET. Magn Reson Med Sci 2005; 4:27 –34
5.
Lee T, Lev MH, Eastwood JD, Provenzale JM, Azhari T, Herzau MA. Effect of choice of artery in the measurement of cerebral blood flow in stroke by CT perfusion. Radiology 2001; 221 (P):481
6.
Eastwood JD, Loving VA, DeLong DM. Assessment of the influence of variables related to arterial and venous input functions and CT perfusion image signal-to-noise value: importance of the venous outflow curve. Radiology 2002; 225 (P):280
7.
Nabavi DG, Cenic A, Craen RA, et al. CT assessment of cerebral perfusion: experimental validation and initial clinical experience. Radiology 1999; 213:141–149
8.
Nabavi DG, LeBlanc LM, Baxter B, et al. Monitoring cerebral perfusion after subarachnoid hemorrhage using CT. Neuroradiology 2001; 43:7–16
9.
Nabavi DG, Cenic A, Henderson S, Gelb AW, Lee TY. Perfusion mapping using computed tomography allows accurate prediction of cerebral infarction in experimental brain ischemia. Stroke 2001; 32:175–183
10.
Wintermark M, Lau BC, Chien J, Arora S. The anterior cerebral artery is an appropriate input function for perfusion: CT processing in patients with acute stroke. Neuroradiology 2008; 50:227 –236
11.
Sherriff SB, Barber DC. A simple quantitative screening test for the detection of extracranial carotid disease. Clin Phys Physiol Meas 1989; 10:23 –32
12.
Calamante F, Gadian DG, Connelly A. Delay and dispersion effects in dynamic susceptibility contrast MRI: simulations using singular value decomposition. Magn Reson Med 2000; 44:466–473
13.
Yamauchi H, Fukuyama H, Nagahama Y, Katsumi Y, Okazawa H. Cerebral hematocrit decreases with hemodynamic compromise in carotid artery occlusion: a PET study. Stroke 1998; 29:98–103
14.
Provenzale JM, Wintermark M. Optimization of perfusion imaging for acute cerebral ischemia: review of recent clinical trials and recommendations for future studies. AJR 2008; 191:1263 –1270

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: W321 - W329
PubMed: 19457797

History

Submitted: January 14, 2008
Accepted: December 11, 2008

Keywords

  1. anterior cerebral artery
  2. cerebrovascular disease
  3. input artery
  4. perfusion CT
  5. superficial temporal artery

Authors

Affiliations

Kiran Sheikh
Department of Radiology, University of Michigan Health System, B2A 209G University Hospital, 1500 E Medical Center Dr., Ann Arbor, MI 48109.
Present address: Department of Radiology, New York Hospital–Weil Cornell Medical Center, New York, NY.
Matthew J. Schipper
Department of Radiology, University of Michigan Health System, B2A 209G University Hospital, 1500 E Medical Center Dr., Ann Arbor, MI 48109.
Present address: Biostatistician, Innovative Analytics, Kalamazoo, MI.
Ellen G. Hoeffner
Department of Radiology, University of Michigan Health System, B2A 209G University Hospital, 1500 E Medical Center Dr., Ann Arbor, MI 48109.

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

View Options

View options

PDF

View PDF

PDF Download

Download PDF

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media