Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Nabavi, D. G.
Right arrow Articles by Lee, T.-Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Nabavi, D. G.
Right arrow Articles by Lee, T.-Y.
(Radiology. 1999;213:141-149.)
© RSNA, 1999

Neuroradiology

CT Assessment of Cerebral Perfusion: Experimental Validation and Initial Clinical Experience1

Darius G. Nabavi, MD, Aleksa Cenic, MSc, Rosemary A. Craen, MB, BS, Adrian W. Gelb, MB, ChB, John D. Bennett, MDCM, FRCPC, Roman Kozak, MD, FRCPC and Ting-Yim Lee, PhD

1 From the Imaging Research Laboratories, John P Robarts Research Institute, PO Box 5015, 100 Perth Drive, London, Ontario, N6A 5K8 Canada (D.G.N., A.C., T.Y.L.); Dept of Radiology and Lawson Research Institute, St Joseph's Health Centre, London, Ontario (A.C., J.D.B., R.K., T.Y.L.); Dept of Anaesthesia (R.A.C., A.W.G.), London Health Sciences Centre, University Campus, Ontario; and Dept of Neurology, Westfälische Wilhelms-Universität, Munster, Germany (D.G.N.). Received Aug 26, 1998; revision requested Oct 22; final revision received Jan 14, 1999; accepted Feb 15. Supported in part by Medical Research Council of Canada, Heart and Stroke Foundation of Canada, and GE Medical Systems. D.G.N. supported by a research grant of the Deutsche Forschungsgemeinschaft. Address reprint requests to T.Y.L. (e-mail: tlee@irus.rri.on.ca).


   Abstract
TOP
 Abstract
Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Appendix 1
 References
 
PURPOSE: To validate a dynamic single-section computed tomographic (CT) method to measure cerebral blood volume (CBV) and cerebral blood flow (CBF) by using a noncarotid artery as the input and to demonstrate the feasibility of this method in a pilot series of patients.

MATERIALS AND METHODS: Twelve dynamic contrast material–enhanced CT studies were performed in beagles. CBV, CBF, and mean transit time (MTT) values were calculated by using an internal carotid artery (ICA) and a noncarotid artery as the input artery to the brain. Patient studies with use of the radial artery as the input were performed (a) repetitively in two patients after subarachnoid hemorrhage, (b) in a patient with a symptomatic ICA occlusion before and after the intravenous injection of 1 g of acetazolamide, and (c) in a patient with a malignant brain tumor.

RESULTS: Linear regression analyses revealed highly significant correlations (P < .001) between CBV (r, 0.98; slope, 0.96), CBF (r, 0.89; slope, 0.87), and MTT (r, 0.80; slope, 0.76) values calculated with the ICA and the noncarotid inputs. The CT-derived patient data correlated well with ancillary clinical and neuroradiologic findings.

CONCLUSION: Dynamic single-section CT scanning to measure CBV and CBF on the basis of a noncarotid input is a highly accessible and cost-effective blood flow measurement technique.

Index terms: Blood vessels, stenosis or obstruction, 17.72 • Brain, blood flow, 10.919 • Brain, CT, 10.12113, 10.12115 • Brain, hemorrhage, 10.2521 • Brain, perfusion, 10.919 • Cerebral blood vessels, flow dynamics, 17.12112, 17.919 • Computed tomography (CT), perfusion study, 10.919


   Introduction
TOP
 Abstract
 Introduction
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Appendix 1
 References
 
Adequate brain perfusion is fundamental for the integrity of the central nervous system. Cerebral blood flow (CBF) below a certain threshold leads to dysfunction (<20 mL/100 g/min) or death (<10–12 mL/100 g/min) of neuronal cells (1). For normal cerebral function, adjustment of the CBF according to the actual demand, determined by the activation level of the neurons, and maintenance of CBF above the ischemic threshold during changes of the systemic blood pressure are crucial. This autoregulation of CBF is controlled by complex metabolic and cellular mechanisms (2). Thus, assessment of CBF and its autoregulation can be used to investigate the normal physiology and the nature of various diseases of the brain.

Several in vivo methods, such as positron emission tomography (PET) (3), single photon emission computed tomography (SPECT) (4), and magnetic resonance (MR) imaging (5,6) are presently available to measure CBF and cerebral blood volume (CBV). However, high costs together with the limited accessibility of and the difficulty in obtaining quantitative values by using SPECT and MR imaging (7,8) have restricted the widespread clinical use of these imaging modalities. Because computed tomographic (CT) scanners are available in most clinical facilities, a CT-based measurement technique has the advantage of easy accessibility for clinical and experimental studies.

As described nearly 20 years ago (9), by using injection of a bolus of contrast material and subsequent dynamic CT imaging to follow the contrast enhancement in an input artery and the brain itself, CBF and CBV values can be obtained by application of the "central volume principle" and the technique of deconvolution. However, due to the limited scanning frequency of CT scanners at the time, this method was not used (9,10).

Since the publication of Axel's work (9,10), several alternative approaches have been proposed to evaluate the mean transit time (MTT) with CT, with promising experimental and clinical results (11–17). Compared with the deconvolution method, however, these techniques require additional assumptions and curve fitting procedures, each representing a potential source of variability. In a previous study (18), we validated our CT technique to measure CBF in a rabbit model against the reference standard of microspheres. When cerebral perfusion in humans at the level of the basal ganglia or higher is measured, however, the true supplying arteries (ie, the internal carotid artery [ICA]) usually are not depicted on the CT image. Therefore, other arteries must be selected to obtain the arterial contrast enhancement curve required for the deconvolution process. One suggestion was to use intraparenchymal cerebral arteries for this purpose (9,10), but they are very small and of variable appearance on the image, which reduces the reliability and reproducibility of this approach.

Alternatively, larger extracranial noncarotid arteries, such as the radial artery, can be scanned simultaneously together with the brain tissue, which provides a more consistent arterial input curve without partial volume averaging. It is known, however, that the vascular resistance of peripheral tissues is much higher than that of the brain parenchyma (19,20). Thus, differences in the shape of the contrast enhancement curves for normal ICA and normal noncarotid artery can be expected. To our knowledge, the effect of these differences on the calculation of CBF and CBV values is yet unknown.

We undertook this study to validate experimentally our CT method to measure CBF and CBV on the basis of a noncarotid input in beagles under normal physiologic conditions. In the second part of the study, the clinical feasibility of this technique and the spectrum of potential diagnostic applications is illustrated.

The theoretic basis of our measurement technique is the central volume principle (21). This principle relates CBF, CBV, and MTT values in the simple relationship

When contrast material is injected into a peripheral vein, the mass of the injected contrast material that resides in the brain (tissue residue function, Q[t]) and the arterial concentration (Ca[t]), as a function of time, t, can be measured with a CT scanner as an increase in CT number (Hounsfield units) as shown in Figure 1a. Meier and Zieler (21) have shown that

where {otimes} denotes the convolution operator and R(t) is the impulse residue function (Appendix, Fig 1b). R(t) is calculated by means of deconvolution (Appendix), and MTT is calculated by means of the area-over-height formula (10):

As shown by Axel (9), CBV in a capillary network is the ratio of areas,

and Equation (4) is valid only when the blood-brain barrier is intact and there is no recirculation of contrast material. In cases in which the blood-brain barrier is compromised, we have developed a correction method for the extravasation of contrast material (22,23) (Appendix).



View larger version (22K):
[in this window]
[in a new window]
  		Figure 1a. (a) Graph shows illustrative examples of CT contrast enhancement curves obtained from the ICA (•), a noncarotid artery of the neck ({bigcirc}), gray matter ({blacklozenge}), and white matter ({diamond}) brain tissue in a dog. For a better illustration, the tissue curves are displayed by using a separate scale of CT numbers (HU) (right y axis). Note the higher contrast enhancement of the gray matter as compared with white matter, which reflects differences in CBV. (b) Graph shows that after deconvolution of the arterial and tissue enhancement curves, the input-independent impulse residue functions for gray ({blacklozenge}) and white ({diamond}) matter were obtained. According to Equation (3), deconvolution of Q(t) and Ca(t) gives the product of CBF and R(t). Because R(t) is dimensionless, the scaled R(t) has the same units as CBF.

 


View larger version (14K):
[in this window]
[in a new window]
  		Figure 1b. (a) Graph shows illustrative examples of CT contrast enhancement curves obtained from the ICA (•), a noncarotid artery of the neck ({bigcirc}), gray matter ({blacklozenge}), and white matter ({diamond}) brain tissue in a dog. For a better illustration, the tissue curves are displayed by using a separate scale of CT numbers (HU) (right y axis). Note the higher contrast enhancement of the gray matter as compared with white matter, which reflects differences in CBV. (b) Graph shows that after deconvolution of the arterial and tissue enhancement curves, the input-independent impulse residue functions for gray ({blacklozenge}) and white ({diamond}) matter were obtained. According to Equation (3), deconvolution of Q(t) and Ca(t) gives the product of CBF and R(t). Because R(t) is dimensionless, the scaled R(t) has the same units as CBF.

 


View larger version (7K):
[in this window]
[in a new window]
  		Figure a1. Schematic shows the impulse residue function in the case when extravasation of contrast material occurs. The gray area denotes the intravascular portion, and the black area denotes the extravascular portion of the contrast material.

 
To eliminate the effect of recirculation from both Ca(t) and Q(t), the back slope of Ca(t) was extrapolated with a monoexponential function (24). The extrapolated Ca(t) was convolved with the calculated impulse residue function to generate the recirculation corrected Q(t). The CBV value was then calculated as the ratio of the area underneath the recirculation corrected Q(t) to that of the recirculation corrected Ca(t). Because the impulse residue function is known, the tissue curve corresponding to any arterial curve can be calculated by reconvolving the impulse residue function with Ca(t). Thus, an advantage of our method is that the exact functional form used for extrapolation does not affect the accuracy of the calculation of the CBV value. The correction factor applied to adjust for the difference in hematocrit values between the large arteries and the tissue capillaries was 0.9 according to Sakai et al (4).


   MATERIALS AND METHODS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 Appendix 1
 References
 
Animal Protocol
Five male adult beagles (R and V Kennels, North Java, NY) weighing 10–15 kg were used in the experiments, which were approved by the animal use ethics subcommittee of the University of Western Ontario, Canada. Each dog received anesthesia for the entire duration of the experiment, and a local anesthetic (lidocaine 1.0%, Xylocaine; Astra Pharmaceutical Products, Westborough, Mass) was administered prior to all surgical procedures. To minimize anesthetic complications, the dogs were not fed but had full access to water for 8 hours prior to administration of anesthesia. After intubation, the animals received mechanical ventilation to normocapnic (target range, 40 mm Hg ± 3 [mean ± SD]), hypocapnic (30 mm Hg ± 3), or hypercapnic (50 mm Hg ± 3) PaCO2 levels by using a mixture of air and oxygen (fraction of inspired O2, 0.4). Anesthesia was maintained throughout the experiments by using isoflurane ([1.5%] Isoflurane; Abbott Laboratories, Saint-Laurent, Quebec, Canada). The right femoral artery was cannulated to allow arterial blood sampling for hematocrit values and blood gas determination (ie, PaO2 and PaCO2) and the measurement of mean arterial pressure. A femoral vein was catheterized for drug administration if required. The core temperature was maintained at 37.5°C–38.5°C by using a water blanket.

After the surgical procedures were completed, each dog was placed in the prone position on the couch of the CT scanner, and the head was fixed in a conventional CT head holder to prevent movement during the subsequent experiments. Vecuronium bromide ([0.2 mg per kilogram of body weight] Norcuron; Organon Canada, Scarborough, Ontario, Canada) was administered intravenously every hour to stop spontaneous breathing. At completion of the experiments, the animals were sacrificed by means of barbiturate overdose, which was intravenously administered sodium pentobarbital ([162 mg/kg] Euthanyl Forte; MTC Pharmaceuticals, Cambridge, Ontario, Canada).

Patient Studies
The baseline characteristics of the four patients examined with the dynamic CT method are shown in Table 1. All patient studies were performed, after written informed consent was obtained, for diagnostic reasons to assess brain perfusion. In the two patients with subarachnoid hemorrhage, serial CT studies were performed twice per week within the first 3 weeks after subarachnoid hemorrhage and surgical clipping of the aneurysm. In the patient who experienced repetitive transient ischemic attacks owing to a known right ICA occlusion, digital subtraction angiograms revealed evidence of insufficient intracranial collateral flow with hypoplasia of the anterior communicating artery and very limited leptomeningeal collateral flow. One gram of acetazolamide (Diamox; Wyeth-Ayerst Canada, Saint-Laurent, Quebec, Canada) was injected intravenously to evaluate the reserve capacity of cerebral autoregulation. In patient 4, a primary malignant brain tumor within the left frontal cortical area was diagnosed by means of CT and MR imaging and was later confirmed by means of histologic findings.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Characteristics of Patients Undergoing Dynamic CT Studies
 
CT Imaging Protocol for the Beagle Studies
The imaging studies were performed by using a slip-ring, helical CT scanner (CT HiSpeed Advantage; GE Medical Systems, Milwaukee, Wis), which allows uninterrupted scanning at the same anatomic level, or cine scanning, at a rate of one scan per second. The CT imaging protocol involved two steps: the localization coronal scanning and the cine CT scanning. For localization, nonenhanced scans were first obtained at 3-mm spacing. Then scans were obtained at 1-mm intervals with identical radiologic parameters to select the target section containing the lateral and the largest cross section of the third ventricle, while including the carotid arteries (Fig 2). CT imaging after the injection of 5 mL of iopromide (Ultravist 300; Berlex Laboratories, Wayne, NJ) confirmed visualization of the carotid arteries in this section. Parameters for the localizing scans were as follows: 80 kVP, 80 mAs, 512 x 512 matrix, 15-cm field of view, and 5-mm section thickness. The back projection filter used in the reconstruction of CT images has a cutoff frequency of 10 line pairs per centimeter.



View larger version (155K):
[in this window]
[in a new window]
  		Figure 2. Axial CT image of the dog's head in the prone position. Arrows indicate small cortical gray matter ROIs, arrowheads indicate small white matter ROIs, and BS denotes brain stem ROI. At the bottom of the Figure, the circular ROIs over the two ICAs and four noncarotid arteries (NCA) are displayed.

 
After the target section was selected, cine scanning was performed by using the same radiologic parameters as those used for the localization scans. We chose an overall interval of 60 seconds for the cine scan to ensure that it covered the entire passage of the bolus of contrast material within the cerebral circulation. Thus, in our cine scanning protocol, 60 continuous rotations of the x-ray tube were made as the couch remained stationary. In addition, image reconstruction was performed at 0.5-second intervals. Therefore, each study was composed of 119 sequential images (60 prospective and 59 retrospective images) with a 0.5-second time resolution.

CT scanning was initiated 5 seconds prior to the intravenous injection of 1.5 mL/kg iopromide by means of an automated injector (Mark IV; Medrad, Pittsburgh, Pa). This delay in injection of contrast material allowed for the acquisition of nonenhanced baseline images. To obtain a large range of CBF values, CT studies were performed with PaCO2 levels between 30 and 55 mm Hg. An interstudy delay of at least 20 minutes was chosen to ensure that most of the contrast material from previous injections had left the circulatory system.

CT Imaging Protocol for the Human Studies
Owing to the longer circulation time in humans, cine scanning of 90-second duration was performed with the same radiologic parameters as in the experimental studies. In the patient with the brain tumor, the section with the largest tumor cross section was chosen. In the remaining patients, the section containing the upper parts of the lateral ventricles was selected (Fig 3). The right forearm was placed on a specially designed support above the head fixed with Velcro to the head holder (Fig 3). This enabled axial scanning of the arteries of the forearm simultaneously with the head. Image reconstruction was performed as described earlier. A dose of 1 mL/kg iopromide was injected at a rate of 1.5 mL/sec.



View larger version (97K):
[in this window]
[in a new window]
  		Figure 3. Axial dynamic CT scan obtained during patient studies. The patient's head was placed in the head holder, and one of the forearms was supported above the head with a platform we made. The space between the plastic support and the forearm is filled with foam (arrows). Note the enhancement of the radial (RA) and ulnar (UA) arteries. The arterial contrast enhancement curves were obtained from the radial artery by using a 4-pixel-diameter circular ROI. The set of standardized tissue ROIs were drawn by using the method described in Materials and Methods.

 
CT Data Analysis
The CT data were stored on digital audiotape (Sony, Tokyo, Japan) and then transferred to a workstation (Ultra I; Sun Microsystems, Palo Alto, Calif) for further computational analysis. An image of the first CT study was used as a template to register the images of subsequent studies of the same subject (SPM software package; University College, London, UK). By using a mouse-guided cursor, freehand regions of interest (ROIs) were drawn (D.G.N.) to obtain the tissue curve in the beagle studies. The following sets of tissue ROI were drawn: (a) small ROIs that contained cortical or deep gray matter and white matter, (b) large ROIs that contained the entire white and gray matters of each hemisphere, and (c) large ROIs that contained mixed tissue. Major blood vessels were excluded from the ROIs. A circular ROI was placed within both ICAs and within up to six additional noncarotid arteries visualized in the neck in the same image (Fig 2). In each beagle, identical ROIs were used in repetitive studies.

In the human studies, a standardized set of ROIs was used as illustrated in Figure 3. A circular ROI was drawn (D.G.N.) to encompass the entire brain. A smaller copy of this ROI was then automatically generated and placed at the border between white and gray matters. The brain area was then automatically partitioned by radial lines at equal angular increments from a starting radial line chosen by the operator. In the patient with the brain tumor, additional ROIs were drawn that covered the tumor and the surrounding tissue.

The contrast enhancement curves were obtained by subtracting the regional mean baseline CT number (Hounsfield units) in the images obtained before the administration of contrast material from the mean CT number in serial contrast-enhanced images. Correction for partial volume averaging for the ICA and noncarotid contrast enhancement curves was performed as described previously (18). In brief, Gaussian curves were fitted to either the horizontal or vertical profile through the center of each arterial ROI. As has been shown experimentally, the SD of the fitted Gaussian curve is a good estimate of the internal diameter of the artery (18). The partial volume averaging correction factor can then be obtained from an experimentally determined calibration curve relating the correction factor to the internal diameter.

Statistics
Statistical analysis was performed by using the SPSS Package (SPSS, Chicago, Ill). The Student t test and Mann-Whitney U test were used to compare normally and nonnormally distributed data, respectively. Linear regression analysis and the Pearson product moment test were used to analyze the correlation of data sets. Significance was declared at P values less than .05.


   RESULTS
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 Appendix 1
 References
 
Comparison of ICA- and Noncarotid-derived Data in Beagles
No significant changes of the mean arterial blood pressure, hematocrit values, and other physiologic parameters occurred throughout the experimental studies. In 12 dynamic CT studies, 227 tissue ROIs of different sizes and tissue compositions and 72 arterial ROIs were analyzed. In 10 of the 12 studies, subtle partial volume averaging of the noncarotid contrast enhancement curve was disclosed by a correction factor of 1.05–1.1, while partial volume averaging of the ICA contrast enhancement curves was not found. Notwithstanding the existence of partial volume averaging, the shape of the noncarotid contrast enhancement curve was very close to that of the ICA as illustrated in Figure 1a.

For both calculation modes (ICA or noncarotid input), significantly higher CBV and CBF and lower MTT values were found for gray matter ROIs as compared with those for white matter ROIs (all P < .001, Mann-Whitney U test). A trend toward higher values for CBV and CBF was found when using the noncarotid artery as compared with the ICA to obtain the input curve (all P > .1, Mann-Whitney U test). This was found consistently for all subgroup analyses according to tissue type and size of ROI, although no significance was found (all P > .1, Mann-Whitney U test) (Table 2). Notably, the MTT showed an opposite trend toward lower values for the noncarotid- as compared with the ICA-based calculations (all P > .1), which reflects subtle differences in the shapes of the curves. The difference in the ICA- and noncarotid-derived MTT among the ROI subgroups approximated 0.2 second (Table 2).


View this table:
[in this window]
[in a new window]
  		TABLE 2. Survey of the Dynamic CT Measurements Based on the ICA Input and the Noncarotid Input Obtained from the Dog Studies
 
Linear regression analysis (Fig 4) revealed an overall strong correlation between CBV values (r, 0.98; CBVICA = 0.96 x CBVNCA), CBF (r, 0.89; CBFICA = 4.4 + [0.87 x CBFNCA]), and MTT (r, 0.80; MTTICA = 0.84 + [0.76 MTTNCA]) calculated with the ICA and the noncarotid, or NCA, enhancement curves. Similar good correlations were found for the subgroup regression analyses with respect to size and tissue composition of the ROI (all r > 0.75, all P < .01). The absolute differences (in mL/100 g/min) between the ICA- and the noncarotid-derived CBF values were calculated for all ROIs and for those ROIs with ICA-derived flow values less than 80 mL/100 g/min (Fig 5). More than 60% of all ROIs and nearly 70% of the ROI subgroups showed flow differences less than 10 mL/100 g/min. Larger differences (>20 mL/100 g/min) mostly occurred for ROIs with flow rates of 80–200 mL/100 g/min, and were very rare for lower flow values (Fig 5).



View larger version (11K):
[in this window]
[in a new window]
  		Figure 4a. Scatterplot shows linear regression of ICA-derived values versus noncarotid-derived (NCA) values of (a) CBV and (b) CBF in beagles. For both parameters, a strong correlation was found with slopes close to unity.

 


View larger version (18K):
[in this window]
[in a new window]
  		Figure 4b. Scatterplot shows linear regression of ICA-derived values versus noncarotid-derived (NCA) values of (a) CBV and (b) CBF in beagles. For both parameters, a strong correlation was found with slopes close to unity.

 


View larger version (16K):
[in this window]
[in a new window]
  		Figure 5. Bar graph shows absolute differences between ICA-derived and noncarotid-derived CBF values. On the y axis, the frequency of ROI is shown as a percentage. Note that for approximately two-thirds of the ROIs, the CBF difference was less than 10 mL/100 g/min. Differences greater than 20 mL/100 g/min were relatively rare for all ROIs (black bars) and nearly absent for ROIs with flow values less than 80 mL/100 g/min (gray bars).

 
Clinical Feasibility Studies
Follow-up after subarachnoid hemorrhage.—Neurovascular work-up, including digital subtraction angiography and transcranial Doppler ultrasonography, revealed evidence of moderate to severe vasospasm in patient 1 starting the 7th day after subarachnoid hemorrhage, while in patient 2 a normal vessel status was found. Compared with white matter ROIs, in all studies gray matter ROIs had significantly higher mean CBF (41.4 vs 22.9 mL/100 g/min; P < .01) and CBV (4.2 vs 2.6 mL/100 g; P < .01) values. Differences in the overall mean CBF values were found between the two patients, with much lower flow values for the patient with evidence of vasospasm beyond the third day after subarachnoid hemorrhage (Fig 6). Likewise, differences were found in the proportion of ROIs with moderate or severe ischemia defined as a CBF of 15–25 mL/100 g/min and less than 15 mL/100 g/min, respectively. In patient 1, who had vasospasm, nearly two-thirds of the ROI showed moderate to severe ischemia beyond the 9th day after subarachnoid hemorrhage, whereas in patient 2, who did not have vasospasm, there was absence of severe ischemia, and only relatively few ROIs (<20%) showed moderate ischemia.



View larger version (13K):
[in this window]
[in a new window]
  		Figure 6. Line graph shows the time course of mean CBF ± SD (error bars) in the two patients with subarachnoid hemorrhage (SAH). Much lower overall mean CBF values beyond day 9 were observed in the patient with vasospasm (•) compared with those in the patient free of vasospasm ({diamond}).

 
Symptomatic ICA occlusion.—As in the patients with subarachnoid hemorrhage, gray matter ROIs showed significantly higher CBF and CBV values than did white matter ROIs (all P < .01). Prior to injection of acetazolamide, differences were found between the mean hemispheric CBF and CBV values. Compared with the left hemisphere, the ipsilateral (right) affected hemisphere had a lower CBF value (28 mL/100 g/min ± 14 vs 33 mL/100 g/min ± 18) and a higher CBV value (2.5 mL/100 g ± 0.7 vs 2.1 mL/100 g ± 0.5), which indicates the activation of the autoregulation in the hemisphere. Twenty minutes after injection of acetazolamide, a pathologic vasomotor reaction was found for the affected hemisphere, with an overall slight decrease in the CBF value (-1%) and only a minor increase in the CBV value (+11%), in contrast to normal changes in the CBF value (+19.3%) and the CBV value (+22%) on the contralateral side. Differences in the vasoreactivity were most prominent for the cortical ROI of the middle cerebral artery, or MCA, territory. On the affected side, a considerable decrease in the CBF value (-9%) and almost no change in the CBV value (+3%) were found as compared with large increases in the CBF value (+19.3%) and the CBV value (+23%) on the contralateral side.

Malignant brain tumor.—Within the tumor, a pathologic increase in both CBF (155 mL/100 g/min) and CBV (5.5 mL/100 g) values was found. These were markedly higher than both the mean CBF and mean CBV values (44.3 mL/100 g/min and 3.3 mL/100 g) of all the other ROIs, as well as the mean values (52.1 mL/100 g/min and 3.9 mL/100 g) for the gray matter ROIs. Owing to the mass effect of the tumor, the two cortical ROIs adjacent to the tumor had much lower CBF (17.5 mL/100 g/min) and CBV (2.0 mL/100 g) values than those of the average of the remaining cortical ROI (52.1 mL/100 g/min and 3.9 mL/100 g).


   DISCUSSION
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
Appendix 1
 References
 
In the first part of the study, we experimentally demonstrated that CBV and CBF values can be measured accurately by means of our dynamic contrast-enhanced CT technique by using a noncarotid artery as a surrogate input artery to the brain. A strong correlation was observed between the dynamic CT measurements in which the ICA was used as the input and those in which the noncarotid artery was used. The CBV values, calculated as the ratio of the areas under the (recirculation corrected) tissue and the arterial enhancement curves, had a correlation coefficient close to unity (r, 0.98; slope, 0.96). The smaller size of the noncarotid arteries and their inherent partial volume averaging led to slightly smaller areas under the noncarotid enhancement curves and, accordingly, higher CBV values (Table 2). Our partial volume averaging correction algorithm enabled us to approximately halve this error down to 5%, but we could not fully eliminate it. Due to the larger size of the human radial artery, which was used as the input artery in human studies, even smaller differences between the ICA- and the radial artery–derived CBV values could be expected.

The MTT and the CBF measurements likewise showed a strong correlation between the ICA- and the noncarotid-derived values (r, 0.80 and r, 0.89, respectively). However, the slopes of their regression lines, especially for the MTT, were not as good as the CBV results. In contrast to the CBV calculation, the initial slope of the arterial enhancement curve determines the results of the deconvolution-based MTT and CBF values. Our results suggest that minor differences between the shape of the ICA contrast enhancement curve and the noncarotid contrast enhancement curve existed, which affected the deconvolution process (Fig 1a). With lower MTT values (eg, in gray matter ROIs), an increasing effect on the flow measurements is expected, even with subtle differences in the ascending slopes of two different arterial curves. The latter is supported by the observation that the largest differences between the ICA and noncarotid CBF measurements (>20 mL/100 g/min) mostly occurred for tissue flow values higher than 80 mL/100 g/min (Figs 4b, 5). In human subjects, however, the range in which accurate CBF measurements are most critical is 0–80 mL/100 g/min. Within this range, the absolute measurement error with use of the surrogate noncarotid input was acceptable for both CBV and CBF values (Figs 4, 5).

In the second part of the study, the clinical feasibility of our dynamic single-section CT technique was demonstrated. By using a simple support for one of the patient's forearms, which can be easily attached to the scanner's head holder, the radial artery can be scanned simultaneously with the patient's head, which provides a consistent and reliable arterial input for the CT measurements. Similar to the extracranial neck arteries used in the beagle experiments, the radial artery also supplies muscle and connective tissue. Thus, similar vascular resistances in the vascular beds downstream from both arteries can be expected. However, in the beagle studies, both ICA and noncarotid enhancement occurred simultaneously in the CT images. In contrast, in the patient studies, radial arterial enhancement appeared with a short delay after the onset of brain enhancement owing to the longer travel time to the forearm.

To demonstrate the validity of our radial artery approach in humans, it has to be shown that this longer travel time has no substantial effect on the deconvolution-based calculations. The delay itself can be corrected for by shifting the arterial enhancement curve to the left of the brain enhancement curve and letting the deconvolution algorithm determine the optimal separation (in the least squares sense) between the two curves. However, it generally is believed that a longer arterial travel time increases the dispersion of a given tracer (25). A more pronounced dispersion, as compared with that of the ICA curve, would make the radial artery invalid as a surrogate for the arterial input for the assessment of cerebral hemodynamics.

We have shown recently by means of a theoretic approach, as well as in vitro experiments in which vessel phantoms were used, that arterial dispersion is not a substantial point of concern for normal flow conditions (23). Thus, with the normal hemodynamic conditions present in the circulation system, an intravenously introduced tracer will appear with nearly identical dispersion in all the major vessels of the arterial tree. In addition, pathologically increased dispersion due to a high-grade arterial stenosis along the path to the forearm is a potential source of errors. However, peripheral arterial disease mostly occurs in the legs, and high-grade arterial obstruction in the upper extremities represents less than 5% of all clinical cases (26). Furthermore, substantial arteriopathy of the arms can be excluded by routine clinical history and the presence of regular radial arterial pulses (27).

Thus, alterations of the shape of the radial artery curve, owing to either the longer path to the forearm, as compared with the path to the brain, or arterial stenoses are not substantial points of concern. Hence, the radial artery's contrast enhancement curve can serve as a surrogate of the arterial input to the brain to measure CBV and CBF values. We are now investigating the CBF in healthy human subjects to obtain normal values for CBV and CBF by using this method.

However, the case in which a normal radial artery is used as the input function for brain tissue supplied by a severely obstructed ICA has to be discussed. For the core of an ischemic region, it is reasonable to expect that the arterial input to this type of tissue will be subjected to more dispersion than regions with normal or slight decrease in blood flow owing to the increase in the length of the collateral pathways supplying that region. Østergaard et al (28) have shown preliminary evidence of this phenomenon in a patient with cerebrovascular disorder.

If we assume that the additional dispersion can be characterized by the function h(t), then instead of the true impulse residue function, R(t), our radial artery–based deconvolution algorithm will calculate the convolution of h(t) and R(t) (ie, h(t) {otimes} R(t)). In the case of insubstantial dispersion, the width of h(t) is equal to or less than the plateau width of R(t). Then the plateau height of R(t), and hence, CBF calculation, is not affected. In addition, because the area of both R(t) and h(t) {otimes} R(t) will be identical, MTT will not be affected. However, in case of substantial dispersion (ie, the width of h[t] is greater than the plateau width of R[t]), then the maximum height of h(t) {otimes} R(t) will be less than that of R(t). As a result, CBF will be underestimated. Even in this situation, the area of h(t) {otimes} R(t) will be the same as R(t), so that the CBV value calculated by using the radial arterial input will not be affected.

The magnitude of the underestimation in CBF and overestimation in MTT cannot be quantified without detailed knowledge of h(t). Greitz (29) has shown that intracranial dispersion in the brain is generally insubstantial both in healthy subjects and in patients with obstructed input pathways. However, this was an old study based on nondigital angiography with limited time resolution. Owing to the importance of this assumption for our method, confirmation of Greitz's results by using modern digital subtraction angiographic equipment capable of higher framing rates is desirable and being planned.

Results in our pilot series were in good agreement with the current understanding of pathophysiology. In all patients, gray matter CBF (~45–50 mL/100 g/min) and CBV (~4 mL/100 g) values were approximately two times higher than the respective white matter values (3,4). During the follow-up in two patients with subarachnoid hemorrhage, a marked reduction of the CBF was observed in the patient with vasospasm in contrast to normal flow rates in the other patient. In the patient with vasospasm, the time course and amount of flow changes for white and gray matters correlated well with observations from a large-scale study in which the xenon 133 inhalation method was used (30).

In the patient with symptomatic ICA occlusion, evidence of activated autoregulation at rest and exhausted vasomotor reactivity after injection of acetazolamide were found for the affected hemisphere, in accordance with clinical and angiographic findings. The contralateral increases in CBF and CBV values in this patient (~20%) were notably lower than the fractional increases of 30%–40% reported for healthy subjects after injection of acetazolamide (31). This could be explained by underlying arteriopathy on the clinically unaffected side or by a steal effect due to collateral flow toward the ischemic hemisphere. The striking increase in CBF (150 mL/100 g/min) and CBV (5.5 mL/100 g) values measured within the tumor, and substantial ischemia present in the adjacent cortical tissue, is in accordance with knowledge of the pathophysiology and hemodynamics of brain tumors (32). Thus, this pilot series demonstrates the feasibility of this technical approach and illustrates the broad clinical spectrum of potential applications of this CT method. Nevertheless, controlled clinical studies are necessary to delineate more accurately the clinical capability and limitations of this measurement technique.

Further methodological issues have to be discussed. First, the radiation dose of this CT technique has to be defined. Although the number of sections required for a dynamic CT study in humans is about six to seven times higher compared with that required for a regular diagnostic scan of the brain (90 vs 10–14 scans), much lower radiologic parameters are used (80 kVp and 80 mAs vs 120 kVp and 340 mAs). Thus, the overall effective dose equivalent required for a dynamic CT study (~2.0 mSv) is not much higher than that required for a routine CT scan (~1.5 mSv) (33). Moreover, this dose is below the dose equivalent delivered with other blood flow measurement techniques such as PET (34) and SPECT (35).

Second, our CT measurements are restricted to a single anatomic level of the brain. Although measurements at different brain levels can be obtained sequentially, this would lead to a very low time resolution of the measured contrast enhancement curves. The latter would result in less than optimal data interpolation and curve fitting procedures and lead to an unacceptably low accuracy of CBF measurements. The introduction of multisection CT scanners with the ability to scan simultaneously at different levels may overcome this limitation in the future.

In summary, we have validated a dynamic single-section CT technique to measure CBV and CBF values by using a noncarotid artery as a surrogate arterial input for deconvolution-based calculations and have demonstrated its feasibility in patient studies. Because the use of the radial artery in humans is easy to apply and provides consistent enhancement curves without substantial partial volume averaging, this CT approach enables accurate, readily available, and cost-effective measurements of CBV and CBF values in a variety of clinical and experimental settings.


   Appendix 1
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Appendix 1
References
 
Deconvolution Process
As discussed earlier, the process of calculating the impulse residue function, R(t) (36), given Q(t) and Ca(t), is the opposite (ie, inverse) of convolution and is called "deconvolution." The process of deconvolution is extremely sensitive to noise in the measured CT data of the arterial and tissue enhancement curves (37,38). Without reliable methods to limit the deleterious effects of noise, the calculated R(t) will be wildly oscillatory and make the calculation of MTT according to Equation (3) impossible. We have decreased the noise sensitivity of deconvolution considerably by using a method described previously (18,39) so that solutions of the general shape shown in Figure 1b are produced.

Correction for Extravasation of Contrast Material
When the capillary endothelium leaks contrast material, we have shown both by heuristic arguments (22) and experimental validation (23) that R(t), in this case, can be expressed as:

where E is the extracted fraction of the contrast material, Ri(t) is the impulse residue function for the fraction, (1 - E), of contrast material that remains intravascular, while Re(t) is the impulse residue function for the fraction, E, of contrast material that leaks out into the extravascular space. Figure A1 is a schematic representation of R(t) as expressed by Equation (A1). As shown, the first plateau is from Ri(t), while the second plateau is from Re(t). By extrapolating the beginning portion of the second plateau in Figure A1 to time zero and subtracting it from R(t), we have shown that this will give the impulse residue function, Ri(t), for the intravascular portion of the contrast material (22,23). The area-over-height formula (Eq [3]) when applied to Ri(t) gives the correct MTT, and if Ri(t) is convolved with the monoexponentially extrapolated Ca(t), this gives the tissue residue curve, Q(t), that has been corrected both for leakage of contrast material into the extravascular space and for recirculation.

Correction for Difference in Tissue and Large-Vessel Hematocrit Values
Contrast material is confined to the plasma phase of blood. As long as the hematocrit values of blood remain the same both in peripheral (large) blood vessels and in tissue capillaries, then the enhancement measured in tissue and blood are equivalent. To correct the tissue enhancement for the difference in large-vessel and small-vessel (tissue) hematocrit values, Q(t) has to be multiplied by the factor {phi} = (1 - H)/(1 - rH) = 0.88, where H (0.4) is the hematocrit value of blood in large vessels and r (0.8–0.85) is the ratio of small- to large-vessel hematocrit values (4).


   Acknowledgments
 
Berlex (Canada) provided the contrast material (Ultravist 300) used in the studies. The authors also thank the animal care technician, Sarah Henderson, AHT, for surgical preparation of the dogs. We acknowledge the help of D.H. Lee, MD, and A.J. Fox, MD, for support in the CT studies and Lisa LeBlanc, BSc, MSc, for editing the manuscript. We are grateful to members of the Department of Clinical Neurosciences at London Health Sciences Centre for help in patient recruitment.


   Footnotes
 
Abbreviations: CBF = cerebral blood flow CBV = cerebral blood volume ICA = internal carotid artery MTT = mean transit time ROI = region of interest

Author contributions: Guarantors of integrity of entire study, D.G.N., T.Y.L.; study concepts, T.Y.L., A.W.G.; study design, A.C., R.A.C.; definition of intellectual content, D.G.N., T.Y.L., A.W.G.; literature research, D.G.N., A.C.; clinical studies, J.D.B., R.K., T.Y.L.; experimental studies, J.D.B., R.K., D.G.N., A.C., R.A.C., T.Y.L.; data acquisition, A.C., R.A.C.; data analysis, D.G.N., A.C.; statistical analysis, D.G.N., T.Y.L., R.A.C.; manuscript preparation, D.G.N., T.Y.L.; manuscript editing and review, all authors


   References
TOP
 Abstract
 Introduction
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 Appendix 1
 References
 

  1. Powers WJ, Grubb RL, Darriet D, et al. Cerebral blood flow and cerebral metabolic rate of oxygen requirements for cerebral functions and viability in man. J Cereb Blood Flow Metab 1985; 5:600-608.[Medline]
  2. Paulson OB, Strandgaard S, Edvinsson L. Cerebral autoregulation. Cerebrovasc Brain Metab Rev 1990; 2:161-192.[Medline]
  3. Leenders KL, Perani D, Lammertsma , et al. Cerebral blood flow, blood volume, and oxygen utilization: normal values and effect of age. Brain 1990; 113:27-47.[Abstract/Free Full Text]
  4. Sakai F, Kakazawa K, Tazaki Y, et al. Regional cerebral blood volume and hematocrit measured in human volunteers by single-photon emission tomography. J Cereb Blood Flow Metab 1985; 5:207-213.[Medline]
  5. Rempp KA, Brix G, Wenz F, Becker CR, Gückel F, Lorenz WJ. Quantification of regional cerebral blood flow and volume with dynamic susceptibility contrast-enhanced MR imaging. Radiology 1994; 193:637-641.[Abstract/Free Full Text]
  6. Baird AE, Warach S. Magnetic resonance imaging of acute stroke. J Cereb Blood Flow Metab 1998; 18:583-609.[Medline]
  7. Goldman H. Techniques for measuring cerebral blood flow In: Phillis JW, ed. The regulation of cerebral blood flow. Boca Raton, Fla: CRC Press, 1993.
  8. Weisskorff RM, Chesler D, Boxerman JL, Rosen BR. Pitfalls in MR measurement of tissue blood flow with intravascular tracers: which mean transit time?. Magn Reson Med 1993; 29:553-559.[Medline]
  9. Axel L. Cerebral blood flow determination by rapid-sequence computed tomography: a theoretical analysis. Radiology 1980; 137:679-686.[Abstract/Free Full Text]
  10. Axel L. Tissue mean transit time from dynamic computed tomography by a simple deconvolution technique. Invest Radiol 1983; 18:94-99.[Medline]
  11. Fike JR, Cann CE, Berninger WH. Quantitative evaluation of the canine brain using computed tomography. J Comput Assist Tomogr 1982; 6:325-333.[Medline]
  12. Nagata K, Asano T. Functional image of dynamic computed tomography for the evaluation of cerebral hemodynamics. Stroke 1990; 21:882-889.[Abstract/Free Full Text]
  13. Norman D, Axel L, Berninger WH, et al. Dynamic computed tomography of the brain: techniques, data analysis, and applications. AJNR 1991; 2:1-12.
  14. Gobbel GT, Cann CE, Iwamoto HS, Fike JR. Measurement of regional cerebral blood flow using ultrafast computed tomography: experimental validation. Stroke 1991; 22:772-779.[Abstract/Free Full Text]
  15. Steiger HJ, Aaslid R, Stooss R. Dynamic computed tomography imaging of regional cerebral blood flow and blood volume: a clinical pilot study. Stroke 1993; 24:591-597.[Abstract/Free Full Text]
  16. Hamberg LM, Hunter GJ, Halpern EF, Hoop B, Gazelle GS, Wolf GL. Quantitative high-resolution measurement of cerebrovascular physiology with slip-ring CT. AJNR 1996; 17:639-650.[Abstract]
  17. Lo EH, Rogowska J, Batchelder KF, Wolf GL. Hemodynamical alterations in focal cerebral ischemia: temporal correlation analysis for functional imaging. Neurol Res 1996; 18:150-156.[Medline]
  18. Cenic A, Nabavi DG, Craen RA, Gelb AW, Lee TY. Dynamic CT measurement of cerebral blood flow: a validation study. AJNR 1999; 20:63-73.[Abstract/Free Full Text]
  19. Kroeker EJ, Wood EJ. Comparison of simultaneous recorded central and peripheral artery pressure pulses during rest, exercise and tilted position in man. Circ Res 1955; 3:623-632.[Abstract/Free Full Text]
  20. Carter SA. Hemodynamic considerations in peripheral and cerebrovascular disease. In: Zwiebel WJ, eds. Introduction to vascular ultrasonography. Philadelphia, Pa: Saunders, 1992.
  21. Meier P, Zieler KL. On the theory of the indicator-dilution method for measurement of blood flow and volume. J Appl Physiol 1954; 6:731-744.[Free Full Text]
  22. Singal RK, Lee TY, Razvi HA, et al. Evaluation of Doppler ultrasound and dynamic contrast-enhanced CT in acute and chronic renal obstruction. J Endourol 1997; 11:5-13.[Medline]
  23. Lawrence KS, Lee TY. An adiabatic approximation to the tissue homogeneity model for water exchange in the brain I. Theoretical derivation. J Cereb Blood Flow Metab 1998; 18:1365-1377.[Medline]
  24. Nilsson BW, Rikner G, Wolgast M. On the theory of an intravenous isotope method for cerebral blood flow measurements. Scand J Clin Lab Invest 1977; 37:195-200.[Medline]
  25. Meyer E. Simultaneous correction for tracer arrival delay and dispersion in CBF measurements by the H215O autoradiographic method and dynamic PET. J Nucl Med 1989; 30:1069-1078.[Abstract/Free Full Text]
  26. Welling RE, Cranley JJ, Krause RJ, Hafner CD. Obliterative disease of the upper extremity. Arch Surg 1981; 116:1593-1596.[Abstract]
  27. Bernstein EF, Stuart SH, Fronek A. The predictive value of noninvasive testing in peripheral artery disease. In: Bernstein EF, eds. Noninvasive diagnostic techniques in vascular disease. St Louis, Mo: Mosby, 1982.
  28. Østergaard L, Sorensen AG, Kwong KK, Weisskoff RM, Gyldensted C, Rosen BR. High resolution measurement of cerebral blood flow using intravascular tracer bolus passages II. Experimental comparison and preliminary results. Magn Reson Med 1996; 36:726-736.[Medline]
  29. Greitz T. A radiologic study of the brain circulation by rapid serial angiography of the carotid artery. Acta Radiol Suppl 1956; 140:1-123.[Medline]
  30. Meyer CH, Lowe D, Meyer M, Richardson RL, Neil-Dwyer G. Progressive change in cerebral blood flow during the first three weeks after subarachnoid hemorrhage. Neurosurgery 1983; 12:58-76.[Medline]
  31. Okudaira Y, Bandoh K, Arai H, Sato K. Evaluation of the acetazolamide test: vasoreactivity and cerebral blood volume. Stroke 1995; 26:1234-1239.[Abstract/Free Full Text]
  32. Kuroda K, Skyhoj OT, Lassen NA. Regional cerebral blood flow in various types of brain tumor: effect of the space-occupying lesion on blood flow in brain tissue close and remote from tumor site. Acta Neurol Scand 1982; 66:160-171.[Medline]
  33. Atherton JV, Huda W. Energy imparted and effective doses in computed tomography. Med Phys 1996; 23:735-741.[Medline]
  34. Huda W, Sandison GA. Estimates of the effective dose equivalent, HE, in positron emission tomography studies. Eur J Nucl Med 1990; 17:116-120.[Medline]
  35. Huda W, Sandison GA. The use of the effective dose equivalents, HE, for 99mTc labelled radiopharmaceuticals. Eur J Nucl Med 1989; 15:174-179.[Medline]
  36. Bassingthwaigthe JB, Chinard FP, Crone C, Lassen NA, Perl W. Definitions and terminology for indicator dilution methods. In: Crone C, Lassen NA, eds. Capillary permeability. Copenhagen, Denmark: Munksgaard, 1970.
  37. Gamel J, Rousseau WF, Katholi CR, Mesel E. Pitfalls in digital computation of the impulse response of vascular beds from indicator dilution curves. Circ Res 1973; 32:516-523.[Abstract/Free Full Text]
  38. Bronikowski TA, Dawson CA, Linehan JH. Model-free deconvolution techniques for estimating vascular transport functions. Int J Biomed Comput 1983; 14:411-429.[Medline]
  39. Yeung IWT, Lee TY, Del Maestro RF, Kozak R, Bennet JD, Brown T. An absorptiometry method for the determination of arterial blood concentration of injected iodinated contrast agent. Phys Med Biol 1992; 37:1741-1758.[Medline]



This article has been cited by other articles:


Home page
 RadiologyHome page
K. Murayama, K. Katada, M. Nakane, H. Toyama, H. Anno, M. Hayakawa, D. San Millan Ruiz, and K. J. Murphy
Whole-Brain Perfusion CT Performed with a Prototype 256-Detector Row CT System: Initial Experience
Radiology, January 1, 2009; 250(1): 202 - 211.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
R. G. Sheiman and A. Sitek
Feasibility of Measurement of Pancreatic Perfusion Parameters with Single-Compartment Kinetic Model Applied to Dynamic Contrast-enhanced CT Images
Radiology, December 1, 2008; 249(3): 878 - 882.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
S. Makkat, R. Luypaert, T. Stadnik, C. Bourgain, S. Sourbron, M. Dujardin, J. De Greve, and J. De Mey
Deconvolution-based Dynamic Contrast-enhanced MR Imaging of Breast Tumors: Correlation of Tumor Blood Flow with Human Epidermal Growth Factor Receptor 2 Status and Clinicopathologic Findings--Preliminary Results
Radiology, November 1, 2008; 249(2): 471 - 482.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
A. Sabir, R. Schor-Bardach, C. J. Wilcox, S. Rahmanuddin, M. B. Atkins, J. B. Kruskal, S. Signoretti, V. D. Raptopoulos, and S. N. Goldberg
Perfusion MDCT Enables Early Detection of Therapeutic Response to Antiangiogenic Therapy
Am. J. Roentgenol., July 1, 2008; 191(1): 133 - 139.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Neuroradiol.Home page
H.M. Silvennoinen, L.M. Hamberg, P.J. Lindsberg, L. Valanne, and G.J. Hunter
CT Perfusion Identifies Increased Salvage of Tissue in Patients Receiving Intravenous Recombinant Tissue Plasminogen Activator within 3 Hours of Stroke Onset
AJNR Am. J. Neuroradiol., June 1, 2008; 29(6): 1118 - 1123.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Neuroradiol.Home page
F. Gaudiello, V. Colangelo, F. Bolacchi, M. Melis, R. Gandini, F.G. Garaci, V. Cozzolino, R. Floris, and G. Simonetti
Sixty-Four-Section CT Cerebral Perfusion Evaluation in Patients with Carotid Artery Stenosis before and after Stenting with a Cerebral Protection Device
AJNR Am. J. Neuroradiol., May 1, 2008; 29(5): 919 - 923.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Neuroradiol.Home page
S.K. Ellika, R. Jain, S.C. Patel, L. Scarpace, L.R. Schultz, J.P. Rock, and T. Mikkelsen
Role of Perfusion CT in Glioma Grading and Comparison with Conventional MR Imaging Features
AJNR Am. J. Neuroradiol., November 1, 2007; 28(10): 1981 - 1987.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Neuroradiol.Home page
H.M. Silvennoinen, L.M. Hamberg, L. Valanne, and G.J. Hunter
Increasing Contrast Agent Concentration Improves Enhancement in First-Pass CT Perfusion
AJNR Am. J. Neuroradiol., August 1, 2007; 28(7): 1299 - 1303.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
S. C. Faria, C. S. Ng, K. R. Hess, S. Phongkitkarun, J. Szejnfeld, D. Daliani, and C. Charnsangavej
CT Quantification of Effects of Thalidomide in Patients with Metastatic Renal Cell Carcinoma
Am. J. Roentgenol., August 1, 2007; 189(2): 378 - 385.
[Abstract] [Full Text] [PDF]

Home page
 Br J AnaesthHome page
J. P. Coles
Imaging after brain injury
Br. J. Anaesth., July 1, 2007; 99(1): 49 - 60.
[Abstract] [Full Text] [PDF]