This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nabavi, D. G. Articles by Rosenberg, G. A. Search for Related Content PubMed PubMed Citation Articles by Nabavi, D. G. Articles by Rosenberg, G. A. Pubmed/NCBI databases Medline Plus Health Information CT Scans Related Collections Animal models of human disease Imaging Ischemic biology - basic studies Acute Cerebral Infarction Computerized tomography and Magnetic Resonance Imaging

(Stroke. 2001;32:175.)
© 2001 American Heart Association, Inc.

Original Contributions

Perfusion Mapping Using Computed Tomography Allows Accurate Prediction of Cerebral Infarction in Experimental Brain Ischemia

Presented in part at the 19th International Symposium on Cerebral Blood Flow, Metabolism, and Function, June 13–17, 1999, Copenhagen, Denmark, and published in abstract form (J Cereb Blood Flow Metab. 1999;19[suppl 1]:S586).

Darius G. Nabavi, MD; Aleksa Cenic, MSc; Sarah Henderson, BSc; Adrian W. Gelb, MB, ChB Ting-Yim Lee, PhD

From the Imaging Research Laboratories, John P. Robarts Research Institute, London, Ontario, Canada (D.G.N., A.C., S.H., T-Y.L.); Department of Neurology, Westfälische Wilhelms-Universität, Münster, Germany (D.G.N.); Department of Anesthesia, London Health Sciences Center, University of Western Ontario, London, Ontario, Canada (S.H., A.W.G.); and Imaging Division, Lawson Research Institute, London, Ontario, Canada (T-Y.L.).

Correspondence to Ting-Yim Lee, PhD, Imaging Research Laboratories, John P. Robarts Research Institute, PO Box 5015, London, Ontario, N6A 5K8. E-mail tlee{at}irus.rri.on.ca

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References Introduction References

 
Background and Purpose—We have developed a dynamic CT method to measure absolute cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit time (MTT). In this study we evaluated the ability of CT-derived functional maps to detect infarction in a rabbit model of focal cerebral ischemia.

Methods—Sequential dynamic CT studies were performed at 2 different slices in 5 control rabbits and another 8 after induction of focal cerebral ischemia. The size of critically ischemic tissue was correlated to size of infarction measured by postmortem 2,3,5-triphenyltetrazolium chloride staining. In the control rabbits, short-term variability of the parameters was assessed by ANOVA analysis.

Results—In 7 of 8 animals of the ischemia group, cerebral infarction was visible on 2,3,5-triphenyltetrazolium chloride staining, constituting 16.7±10.6% of the ipsilateral hemisphere. Good agreement of CBF functional maps with tissue specimens was found with respect to size and location of infarction. Best prediction of infarction was found for thresholds of CBF <10 mL/100 g per minute (mean size, 17.5±13.4%; r=0.95) and MTT >6 seconds (mean size, 15.6±13.5%; r=0.85), with regression slopes close to unity. CBV maps were less predictive of occurrence of infarction, especially in cases of small infarction. The short-term variability of CBF, CBV, and MTT in the control group was 10.9%, 15.2%, and 19.9%, respectively.

Conclusions—Functional CT measurements of absolute CBF and MTT early after onset of ischemia allow prediction of the size and location of cerebral infarction with good accuracy.

Key Words: cerebral blood flow • cerebral ischemia, focal • computed tomography • stroke, experimental

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References Introduction References

 
The recent introduction of intravenous thrombolysis has improved acute stroke therapy considerably.¹ However, efficient use of this potentially effective treatment modality requires robust algorithms for patient selection based on thorough risk-benefit considerations.² It has been shown that measurement of cerebral blood flow (CBF) can predict both extent and severity of cerebral infarction^{3 4 5} as well as clinical outcome^{6 7} in the acute stage of ischemia. Thus, early assessment of cerebral hemodynamics in stroke patients may guide the appropriate selection of treatment options.

Several methods are known to provide relative or absolute measures of CBF, cerebral blood volume (CBV), and mean transit time (MTT). Positron emission tomography (PET) remains the in vivo "gold standard" of in vivo perfusion measurements,⁸ even though single-photon emission CT⁹ and xenon-enhanced CT¹⁰ have been demonstrated to be prognostically useful in acute cerebral ischemia. However, limited availability of these modalities has restricted their clinical use to specially equipped centers. With the advent of diffusion- and perfusion-weighted images, MRI has made the most exciting advances in stroke imaging within the last decade.¹¹ However, not all stroke patients are suitable to undergo MRI studies, and, to date, the availability of MRI scanners is not prevalent enough to provide rapid "around-the-clock" accessibility.

CT of the head, which is readily available in most hospitals, is still the most frequently used imaging modality to investigate acute stroke patients.¹² Axel^{13 14} has shown that quantification of CBF and CBV can be performed by bolus injection of x-ray contrast followed by rapid CT scanning of the head. Absolute measurements require the simultaneous measurement of contrast enhancement within an artery together with that in a tissue sample; by deconvolution of these 2 enhancement curves, absolute values of CBF, CBV, and MTT can be calculated.¹⁴ On the basis of these principles, we have developed a CT technique to measure absolute CBF, CBV, and MTT. We have validated our CT CBF values in various animal models against CBF measurements using fluorescent microspheres.^{15 16} With the use of a peripheral artery not supplying cerebral tissue (eg, the radial artery) for the deconvolution process, results similar to measurements using the "true" arterial input from the carotid artery were achieved.¹⁷

The aim of this study was to assess the diagnostic potential of CT-derived functional maps of CBF, CBV, and MTT in a rabbit model of acute focal cerebral ischemia. Since, for a given species, there exists a consistent CBF threshold below which tissue necrosis occurs, we postulated that early hemodynamic mapping allows the accurate localization of critically ischemic tissue at high risk of cerebral infarction.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References Introduction References

 
Animal Protocol
A total of 15 male New Zealand White rabbits (2.8 to 3.9 kg body wt) were used in the experiments. The study procedures were approved by the Ethics Committee for Animal Studies at the University of Western Ontario (London, Ontario, Canada). All animals were initially anesthetized by intramuscular injection of ketamine/xylazine. Catheters were then placed in a femoral artery, a femoral vein, and an ear vein for physiological monitoring, blood sampling, and injection of x-ray contrast, respectively. After tracheotomy, mechanical ventilation was provided, and anesthesia was continued with 1.5% isoflurane in a mixture of room air and oxygen (40%). Muscle relaxation was achieved by intravenous administration of vecuronium (0.2 mg/kg) at regular intervals. A temperature probe was placed in the rectum for continuous monitoring of core temperature. A heating pad and bubble wrap were used to maintain core temperature at approximately 39°C.

For the 5 control rabbits, no further surgical preparation was done. The following procedures were only performed for the 10 animals of the focal ischemia group. Needle electrodes were inserted on both sides 5 mm lateral to the bregma for continuous recording of the electroencephalogram during induction of focal ischemia. The midline neck incision for tracheotomy was extended to expose the bifurcation of the left distal common carotid artery (CCA). After identification of the ipsilateral internal carotid artery (ICA),¹⁸ the external carotid artery (ECA) was ligated and transected. The occipital artery was also ligated if it originated from the ICA. The isolated ECA stem was then retracted to introduce a 5-0 nylon filament via a 22-gauge catheter, while the ipsilateral CCA and the ICA were temporarily clipped. The 5-0 nylon suture, with its distal end coated with silicone up to a diameter of 0.3 to 0.4 mm, was introduced via ECA into the ICA. The filament was gently advanced intracranially until elastic resistance was felt (usually after 50 to 60 mm, as indicated by markers on the filament). Proper position of the filament was accompanied by prompt ipsilateral suppression (for duration of 1 to 5 minutes), followed by persistent slowing of the electroencephalographic activity. After its final position was reached, the lower end of the filament was attached to the ECA stem at the site of the insertion to prevent accidental dislocation before the neck incision was closed. One animal was excluded from the analysis because the filament could not be advanced >35 mm into the ICA.

The technical procedure described above is similar to the suture model in the rat,¹⁹ which has been shown to be more reliable with coating of the tip of the filament.²⁰ Our results agree with angiographic studies by Jeppson and Olin²¹ as well as the extensive work by Molnar et al.²² The latter group injected silver balls of various diameters (0.2 to 0.45 mm) attached to thin filaments into the extracranial ICA of rabbits and identified their intracranial location by x-ray and postmortem microscopy to evaluated the occurrence of cerebral infarction. Blockage of the middle cerebral artery (MCA) sufficient to cause infarction was only observed with balls <0.35 to 0.4 mm in diameter, while larger ones lodged more proximally within the ICA. In a pilot series of experiments, we had confirmed this previous work by showing that occlusion of the MCA stem was achieved only with sutures having a distal diameter of 0.3 to 0.35 mm. Use of larger filaments or thicker coating had consistently led to lodging of the suture within the distal ICA, which prevented it from entering the proximal MCA.

The animals were then placed in the prone position on the CT couch with the head secured in a conventional CT head holder supported laterally by cushions. Both forelimbs were turned forward and placed underneath the head so that the radial arteries could be simultaneously scanned together with the brain to provide the input function for our calculations.¹⁵ Vecuronium (0.2 mg/kg IV) was regularly given to suppress spontaneous breathing and to avoid movement artifacts during the scanning procedures. Throughout the experiment, mean arterial blood pressure and heart rate were continuously monitored. Immediately before each CT study, arterial blood was drawn to measure hematocrit, pH, PaO₂, PaCO₂, and blood glucose.

Six hours after onset of focal cerebral ischemia, the animals were euthanatized by barbiturate overdose (intravenous pentobarbital sodium). After the skull was removed, the brain was carefully lifted, and the location of the tip of the filament was documented. The brain was frozen, and then 4-mm slices were cut at distances 10 and 14 mm from the tip of the brain corresponding to the 2 CT slices studied (see Dynamic CT Imaging Protocol). The tissue specimens were incubated in 1% 2,3,5-triphenyltetrazolium chloride (TTC) for 30 minutes. Subsequently, the stained slices were immersed in 4% buffered formalin and photographed with a digital camera. Analysis of the extent of cerebral infarction was independently performed by 2 investigators (D.G.N., A.C.) blinded to the results of the perfusion measurements. Tissue areas without staining were manually outlined on the digital pictures of the tissue slices. Only clearly white areas were considered the infarcted area, excluding pink areas with incomplete infarction and white matter areas with insufficient staining due to the low concentration of mitochondria.^{23 24} The percent area of infarction was calculated on the basis of the pixel numbers of infarcted tissue divided by the pixel numbers of the entire hemisphere.

Dynamic CT Imaging Protocol
All imaging studies were performed with a HiSpeed CT scanner (General Electric Medical Systems) with the following technique: 80 kilovolts (peak) [kV(p)], 80 mA, 512x512 matrix size, 10-cm field of view, and 3-mm slice thickness. After the initial scout scan, nonenhanced coronal CT scans were performed at 3-mm spacing, followed by imaging at finer 1-mm spacing to identify the frontal end of the brain. The 2 target slices for dynamic CT measurements were then defined at distances of 10 mm (frontal slice) and 14 mm (temporal slice) from the frontal tip of the brain (Figure 1A). These measurements served as a guide for postmortem slicing of the frozen brain at sites identical to those of the CT measurements (see Animal Protocol). A total of 6 dynamic (cine) CT studies were performed in each animal at 45-minute intervals on the basis of the following schedule: 3 studies at the temporal slice 1.5, 3, and 4.5 hours and 3 studies at the frontal slice 2.25, 3.75, and 5.25 hours after onset of ischemia. The selected time interval between 2 studies was sufficient for the contrast material from the previous injection(s) to be cleared from the blood circulation. If the PaCO₂ was outside the target range of 36 to 44 mm Hg, the ventilation rate was adjusted and the CT study was postponed until normocapnic levels were reestablished (usually 5 to 10 minutes later). To have a comparable time schedule among all animals, scanning was always started at the temporal slice. Since the first CT study in ischemic rabbits was usually performed 3 hours after onset of anesthesia, the scanning schedule for the control animals was started accordingly 3 hours after onset of anesthesia.

View larger version (100K): [in this window] [in a new window]   Figure 1. A, Definition of the 2 brain slices for dynamic CT studies and TTC staining by the distance to the tip of the brain. At each slice, 3 sequential dynamic CT studies were performed (see text). B, Standardized method for generation of 12 ROIs for analysis of short-term variability of CBF, CBV, and MTT in control rabbits. First, a large circular ROI was manually drawn to encompass the entire brain slice. Then a smaller copy of this ROI was placed within this ROI. The image center was automatically localized by the software, from which radial lines at equal angular increments of 60° were automatically generated.

All cine (continuous) CT scannings (at a rate of 1 scan per second) were initiated 5 seconds before a bolus of Ultravist 300 (1.0 mL/kg body wt) was injected into an ear vein at a rate of 1.0 mL/s (Medrad Injector). The short delay of contrast injection allowed for the acquisition of nonenhanced, baseline images (ie, background data for image analysis). Scanning duration was 1 minute; therefore, 60 continuous rotations of the x-ray tube were made as the couch remained stationary. The acquired images were then reconstructed at 0.5-second intervals with the use of a back-projection filter with a cutoff spatial frequency of 10 line pairs per centimeter.

CT Data Analysis
The arterial contrast concentration curve, C_a(t), was determined by drawing a 5-pixel radius circular region of interest (ROI) over both radial arteries. In the first CT study, the radial artery providing the higher contrast enhancement curve (after baseline subtraction) was defined as the input artery, and the same artery was used for all subsequent studies of the experiment. The C_a(t) was estimated because of partial volume averaging and was corrected by the method described previously.¹⁵ Throughout the study, a high consistency of the partial volume averaging scaling factors was found (range, 1.2 to 1.4) with a low intraindividual (<10%) variability.

For deconvolution, both the artery and the tissue curves were interpolated to a time interval of 0.25 seconds. The brain tissue enhancement was assessed in approximately 1000 pixel blocks, each with a size of 3x3 pixels, covering the entire brain area. The tissue curves for each pixel block were then deconvolved with the artery curve, on the basis of our previously described deconvolution algorithm, to determine CBF, CBV, and MTT.^{15 16 17} For better visualization, the raw maps were then smoothed with the use of a gaussian filter involving 8 immediately adjacent pixel blocks. For data analysis, ROIs were drawn covering the entire map and separating the 2 hemispheres. Figure 1B illustrates the generation of the 12 ROIs for the short-term variability studies in the normal control animals. In the ischemia group, the left-sided ischemic hemisphere will be further referred to as ipsilateral and the unaffected hemisphere as contralateral within this report. We defined critically ischemic tissue for cerebral infarction by different a priori defined thresholds for CBF (<7, <10, <13 mL/100 g per minute), CBV (<0.5, <1.0 mL/100 g), and MTT (>6, >8 seconds). The percent area of infarction was calculated on the basis of the pixel numbers of infarcted tissue identified by the aforementioned thresholds divided by the pixel numbers of the entire hemisphere.

Statistical Analysis
Statistical analysis was performed with the SPSS 9.0 for Windows Software Package (SPSS Inc). Repeated intragroup measurements were analyzed with a paired t test and repeated-measures ANOVA, respectively. Intergroup differences were compared with the 2-tailed t test and ANOVA for 2 groups, respectively. For nonnormally distributed data, Mann-Whitney and Kruskal-Wallis tests were applied. Agreement between the CT measurements and the size of infarction on TTC staining was assessed by means of linear regression analysis and the Pearson product moment test. Statistical significance was declared at the P<0.05 level.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References Introduction References

 
Results of the physiological data are given in the Table. No significant differences were found among the 3 sequential studies at each slice and between the control and the ischemia groups. In 7 of 10 animals of the ischemia group, the filament was successfully placed within the MCA stem. In another animal, the filament lodged within the distal part of the ICA (this animal was included in this series). One animal was excluded from the analysis because the filament could not be advanced >35 mm into the ICA. Another animal was excluded from the study because of vessel perforation of the filament with brain hemorrhage. Thus, 8 rabbits of the ischemia group (containing 16 brain slices) and the 5 control rabbits were included in the analysis.

View this table: [in this window] [in a new window]   Table 1. Survey of Physiological Measurements During 6 Sequential CT Perfusion Measurements

In the control rabbits, no cerebral infarction was found on postmortem investigations. In the ischemia group, TTC staining showed cerebral infarction in 14 of the 16 slices. The mean (±SD) area of infarction constituted 16.7±10.6% of the ipsilateral hemisphere (range, 7.3% to 40.1%). In contrast to the variable size, the location of infarction was consistent, affecting laterocaudal parts of the cortical and adjacent subcortical tissue of the temporal lobe. This was in agreement with TTC staining (Figures 2B and 3B). The frontal slice showed consistently larger infarcts (22.1±14.1%) than the temporal slice (11.3±6.7%), with wedge-shaped involvement of deep brain structures. In the animal with the filament lodging in the distal ICA, no infarction was visible on TTC staining.

View larger version (85K): [in this window] [in a new window]   Figure 2. A, Survey of CBF, CBV, and MTT maps in extensive focal cerebral ischemia. The sequential studies were performed 2.25 hours (study 1), 3.75 hours (study 2), and 5.25 hours after onset of ischemia. Note the slight improvement of tissue perfusion, indicated by slight increase in CBF and CBV and slight decrease of MTT. CBF, CBV, and MTT are in units of milliliters per minute per 100 g, milliliters per 100 g, and seconds, respectively. B, Corresponding postmortem TTC staining shows extended focal brain infarction (arrows).

View larger version (64K): [in this window] [in a new window]   Figure 3. A, Critically ischemic tissue in a temporal slice defined by CT perfusion measurements. Ischemic tissue is indicated by lowest values of CBF and CBV (blue) and highest MTT values (red). The sequential studies were performed 1.5 hours (study 1), 3 hours (study 2), and 4.5 hours after onset of ischemia. The CBF, CBV, and MTT maps were partitioned into 4 types of tissue by using 4 different thresholds for each parameter. The threshold values for CBF, CBV, and MTT are in units of milliliters per minute per 100 g, milliliters per 100 g, and seconds, respectively. B, Corresponding postmortem TTC staining shows large focal brain infarction (arrows). Note the blue 5-0 filament still positioned within the MCA attached to the base of the brain (arrowhead).

Control Group
The mean (±SD) values for all control studies were 68±19 mL/100 g per minute (CBF), 2.5±0.8 mL/100 g (CBV), and 2.2±0.8 seconds (MTT). In the temporal slice, significantly higher CBF (76 versus 61 mL/100 g per minute; P=0.01) and CBV (3.1 versus 2.5 mL/100 g; P=0.02) were found compared with the frontal slice because of the higher amount of gray matter in the former. The MTT values for the temporal and the frontal slices were similar (2.1 versus 2.2 seconds; P=0.8). No interhemispheric differences existed for either CBF, CBV, or MTT (all P>0.5). Over time, a significant drop of CBF occurred in study 3 compared with both study 1 and study 2 (P<0.05, ANOVA). No significant changes in CBV and MTT were found over time (P>0.1) despite a trend toward a slight increase of MTT with time. Comparing the 12 standardized ROIs per study (Figure 1B) using regression analysis, we found a close association between values of study 1 and study 2 for CBF (r=0.94, slope=1.01), CBV (r=0.94, slope=0.86), and MTT (r=0.76, slope=0.92). The short-term variabilities of CBF, CBV, and MTT in the control group were satisfactory. Mean (±SD) values of studies 1 and 2 were 72.8±29.1 and 74.1±31.8 mL/100 g per minute for CBF (variability 10.9%), 2.89±1.42 and 2.97±1.66 mL/100 g for CBV (variability 15.2%), and 2.09±0.79 and 2.02±0.58 seconds for MTT (variability 19.9%).

Ischemia Group
As observed in the control studies, there was a trend toward a reduction of CBF over time, which did not reach statistical significance either for the entire slice or when ipsilateral and contralateral hemispheres were assessed separately (all P>0.1). An overall significant reduction in CBV was found for studies 2 and 3 (1.8±0.3 and 1.7±0.3 mL/100 g per minute) compared with study 1 (2.1±0.6 mL/100 g; P<0.05), while no significant changes of MTT were found (P>0.5). When interhemispheric differences were evaluated, the ipsilateral hemisphere showed significantly lower CBF (33.9 versus 49.3 mL/100 g per minute; P<0.001) and higher MTT (3.60 versus 2.13 seconds; P<0.001), with a trend toward lower ipsilateral CBV (1.73 versus 1.93 mL/100 g; P=0.06).

In 14 of the 16 CT maps, an ischemic area could be visualized that corresponded well with the TTC slices (Figures 2 and 3). The ischemic core was located predominantly in the cortical rim of the frontotemporal lobe. According to the TTC staining, more extensive ischemia was found in the frontal slice, including deep brain structures such as the basal ganglia and the internal capsule, than in the temporal ones. In slices with large focal ischemia, all CT maps allowed depiction of the ischemic area by means of reduced CBF and CBV as well as increased MTT (Figures 2 and 3). In cases with minor cerebral ischemia, CBV maps did not consistently show a severe reduction, especially at earlier stages of the experiment. In these cases, the MTT maps were more sensitive in depicting the location and extent of ischemia.

As discussed in Materials and Methods, we defined critically ischemic tissue for cerebral infarction by different a priori defined thresholds for CBF, CBV, and MTT. The best correlation with respect to size and location of infarction was found for a threshold of CBF <10 mL/100 g per minute (r=0.95, P<10^-7) with regression slopes close to unity (slope [m]=1.05; Figure 4). When we assessed each of the 3 sequential CT studies for CBF <10 mL/100 g per minute, all 3 studies showed a good correlation with TTC staining (r=0.89 to 0.93, respectively). Notably, the number of pixels with critically reduced CBF was negligible in the 2 slices without subsequent infarction. Threshold of CBF <7 and <13 mL/100 g per minute likewise revealed good linear correlations (r=0.90 and 0.94, respectively) and regression slopes (m=0.96 and 0.94, respectively) with size of infarction.

View larger version (18K): [in this window] [in a new window]   Figure 4. Linear regression analysis between the size of tissue with critical ischemia defined by CBF <10 mL/100 g per minute (top), CBV <0.5 mL/100 g (middle), and MTT >6 seconds (bottom) and the size of infarction based on TTC staining. CT measurements are expressed as mean±SD based on the 3 subsequent studies. Sizes are given in percent area of the ipsilateral hemisphere. The lines represent the regression lines. The correlation coefficient (R) and slope (m) for each regression are given.

MTT maps likewise showed a good agreement with the TTC-stained slice, with the best correlation found for a threshold of >6 seconds (R=0.85, slope [m]=0.95; Figure 4), which is approximately 2.5 times above the average control MTT. A threshold of MTT >8 seconds revealed a correlation of 0.83 and a regression slope of 0.59 with TTC staining. In cases with a small area of ischemia, MTT maps were more sensitive than CBF and CBV in depicting ischemic tissue. This was especially true for perfusion measurement early after induction of ischemia.

In contrast, the CBV-based definition of critically ischemic tissue showed a weaker but still significant correlation to size of infarction (R=0.8, slope [m]=0.49, P<0.001; Figure 4). Critically ischemic tissue was strongly underestimated with the use of a low threshold of CBV <0.5 mL/100 g and was severely overestimated with the use of a threshold of CBV <1.0 mL/100 g compared with TTC staining. As a result, among the regressions between size of infarction defined with TTC staining and either CBF, CBF, or MTT thresholds, that using CBV is the weakest.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References Introduction References

 
Functional CT imaging after bolus injection of contrast material was evaluated 2 decades ago, but the poor scanning frequency of CT scanners at the time precluded its entry into clinical practice.^{13 14} With the advent of fast helical CT scanners, this method has renewed interest as a means of measuring CBF and CBV in patients with acute cerebral ischemia.^{25 26 27 28} A variety of mathematical algorithms are being applied to extract the perfusion-related information from the tissue enhancement curves. Whereas most of the studies are performed in clinical studies, only a few techniques were "calibrated" in experimental stroke models. We investigated the value of functional maps of absolute CBF, CBV, and MTT to define critically ischemic tissue for subsequent occurrence of cerebral infarction.

We found an excellent agreement between size and location of critically ischemic tissue defined by acute CT perfusion mapping and size of infarction postmortem by TTC staining. In this respect, CBF maps showed the best correlation between critical ischemia and postmortem tissue staining, with regression slopes close to unity for thresholds of 10 mL/100 g per minute (m=1.05). Critically ischemic tissue was determined as the mean of the areas of tissue with CBF below the aforementioned thresholds for the 3 sequential CT studies. For a given species, CBF threshold for critical ischemia is expected to be constant among different animals. This, together with our short-term variability studies in the control animals, underscores the high consistency of our functional perfusion measurements. Furthermore, the optimum CBF threshold found for infarct prediction (ie, <7 to 10 mL/100 g per minute) is plausible with regard to most of the experimental studies reporting thresholds for cerebral infarction at CBF levels <5 to 12 mL/100 g per minute among the different species.^{29 30 31 32} This supports the accuracy of our CBF measurements, which has already been shown in our validation studies using fluorescent microspheres.^{15 16} Notably, in the control studies and in all contralateral slices of the ischemia group, mean (±SD) areas with CBF <10 mL/100 g per minute were as small as 0.8±0.5% and 1.1±1.2% of the hemisphere, respectively. This indicates the high specificity of our CBF measurements and the threshold applied to detect critical ischemia. The results of this study, together with previous experimental investigations,^{15 16 17} indicate that the accuracy of this method of CT perfusion mapping with the use of an extracranial artery as an input function is appropriate for the assessment of severe ischemia.

Mapping of MTT with a threshold of >6 seconds likewise showed a good predictability of later occurrence of infarction (Figure 4). In cases of limited severe ischemia, MTT maps seemed more sensitive than CBF maps in depicting the lesion site, especially at very early stages. Since MTT reflects the velocity of the contrast agent traveling through the capillaries of a tissue sample, it is a very sensitive marker of locally reduced perfusion pressure downstream of an occluded vessel. Since a threshold of 6 s is a 2.5-fold increase of MTT compared with the control studies, it corresponds to an MTT increase from a normal of 5.5 to 6 seconds to 14 to 15 seconds in humans. Notably, PET investigations in humans have shown that an MTT beyond 13 to 15 seconds indicates tissue at high risk of infarction because of exhaustion of autoregulation.³³ Thus, our optimum threshold of MTT for prediction of infarction agrees with this result. MTT maps are a useful complement to CBF maps, especially to detect minor areas of ischemia.

In contrast, the CBV maps were only of moderate diagnostic value to define critically ischemic tissue (Figure 4). Using a low CBV threshold of <0.5 mL/100 g led to an underestimation of subsequent infarction. Most likely, vasodilatory effects, likely to occur within ischemic tissue,³ make CBV maps less useful for the prediction of tissue necrosis. This observation is in accordance with previous experimental studies using well-established measurement techniques demonstrating a low sensitivity for CBV measurements to depict cerebral ischemia.^{34 35} In a PET investigation of acute stroke patients, Heiss et al⁶ found a significant reduction of CBF but not of CBV in the ischemic core and surrounding tissue compared with contralateral tissue. In a recent study by Hatazawa et al,³⁵ CBF was definitely superior to CBV in predicting occurrence of cerebral infarction. They hypothesized that a preserved CBV level reflects maintained compensatory mechanisms indicative of potentially salvageable brain tissue. Thus, the additional assessment of CBV may help to define the individual therapeutic window, which deserves further experimental consideration.

Only one previous study has correlated results of dynamic CT mapping with postmortem size of infarction. Lo et al²⁸ compared the results of CT-derived maps of relative bolus delay of contrast material and relative blood volume with TTC staining in a rabbit model of focal cerebral ischemia. They likewise observed an overall good correlation for their relative perfusion maps and postmortem findings despite differences of up to 50% in individual cases. They demonstrated that the final area of infarction included not only regions without any contrast enhancement (ie, without any flow) but also tissue with disturbed hemodynamics (relative bolus delays >6 seconds). For defining the core of ischemia with maximum CBF reduction, ie, containing noise only without any contrast enhancement, the relative and absolute approach should be of similar accuracy. In contrast to absolute calculations, however, the relative approach requires comparison of each ROI with a "normal" reference region (usually within the contralateral hemisphere). The occurrence of perfusion disturbances in the contralateral hemisphere^{36 37} makes the assumption that the contralateral hemisphere contains entirely normal tissue questionable. The latter constitutes a significant source of errors of relative perfusion measurements. Furthermore, interstudy comparisons are more difficult without normalization of the individual arterial input. In summary, more data are needed to better define the diagnostic potential and limitations of absolute versus relative measurements of the cerebral perfusion.

Further methodological considerations must be discussed. First, a prerequisite for comparison of in vivo measurements with an ex vivo gold standard is the appropriate registration of the respective slices. We selected 2 target slices by measuring the distance from the tip of the brain through CT scanning at 1-mm intervals. These distances were used for ex vivo tissue slicing, which revealed an overall good agreement. This procedure could not eliminate all residual mismatches between CT image and tissue slice. However, the standardization avoided bias from a purely visual selection. Second, the radiation dose of this dynamic CT technique must be considered. Although 60 cine images are required for a dynamic CT study, the low x-ray parameters [80 kV(p) and 80 mA per image] ensure that the effective dose equivalent for a dynamic CT study (2.0 to 2.5 mSv) is only slightly higher than that for a routine CT head scan (1.5 mSv).³⁸ With the use of a faster contrast injection,^{25 26} the number of images required could be reduced further, lowering the overall radiation dose. The duration to produce high-resolution maps has become negligible because recent advances of programs and computer systems have enabled us to generate maps of absolute cerebral perfusion within 3 to 5 minutes. This time is sufficiently short to allow the use of these hemodynamic data for clinical decisions even in hyperacute strokes. Finally, our CT measurement is currently restricted to a single CT slice. In the study by Koenig et al,²⁶ acute stroke patients were also investigated with the use of monoslice dynamic CT perfusion imaging. Although assessment of tissue ischemia was good overall, in 3 of 16 studies ischemia occurred at a site distant to the target slice and was missed by the single-slice measurements. With the advent of multislice CT scanners, this limitation will be reduced in the near future.³⁹

In summary, we have experimentally demonstrated that the location and extent of cerebral ischemia can accurately be assessed by deconvolution-based CT mapping of CBF and MTT. On the basis of our measurements, CBF threshold of <10 mL/100 g per minute and MTT threshold of >6 seconds were accurate predictors of cerebral infarction in a rabbit model of acute focal cerebral ischemia. Because of the low cost and broad availability of CT scanners, this method has the potential to improve early stroke management on a broad scale; however, further clinical studies are required to confirm the experimental results reported here before widespread application of the method can be recommended.

	Acknowledgments

 
This study was supported by the Medical Research Council of Canada (MT-13117). Dr Nabavi was supported by the Deutsche Forschungsgemeinschaft (Na 357/1-1). The authors are grateful to Jay Davis, MSc, for development of the Xstatpak program used in the analysis of CT images.

Received April 20, 2000; revision received July 24, 2000; accepted September 7, 2000.

	References

Top Abstract Introduction Materials and Methods Results Discussion References Introduction References

 

1. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 1995;333:1581–1587.[Abstract/Free Full Text]
2. Furlan A, Kanoti G. When is thrombolysis justified in patients with acute ischemic stroke? Stroke. 1997;28:214–218.[Abstract/Free Full Text]
3. Hossmann KA, Mies G, Paschen W, Csiba L, Bodsch W, Rapin JR, Poncin-Lafitte ML, Takahashi K. Multiparametric imaging of blood flow and metabolism after middle cerebral artery occlusion in cats. J Cereb Blood Flow Metab. 1985;5:97–107.[Medline] [Order article via Infotrieve]
4. Symon L, Pasztor E, Branston NM. The distribution and density of reduced cerebral blood flow following acute middle cerebral artery occlusion: an experimental study by the technique of hydrogen clearance in baboons. Stroke. 1974;5:355–364.[Abstract/Free Full Text]
5. Hakim AM, Hogan MJ, Carpenter S. Time course of cerebral blood flow and histological outcome after focal cerebral ischemia in rats. Stroke. 1992;23:1138–1144.[Abstract/Free Full Text]
6. Heiss WD, Huber M, Fink GR, Herholz K, Wagner R, Wienhard K. Progressive derangement of periinfarct viable tissue in ischemic stroke. J Cereb Blood Flow Metab. 1992;12:193–203.[Medline] [Order article via Infotrieve]
7. Hakim AM, Pokrupa RP, Villanueva J. The effect of spontaneous reperfusion on metabolic function in early human cerebral infarcts. Ann Neurol. 1987;21:279–289.[Medline] [Order article via Infotrieve]
8. Powers WJ, Grubb RLJ, Darriet D, Raichle ME. Cerebral blood flow and cerebral metabolic rate of oxygen requirements for cerebral function and viability in humans. J Cereb Blood Flow Metab. 1985;5:600–608.[Medline] [Order article via Infotrieve]
9. Weir CJ, Bolster AA, Tytler S, Murray GD, Corrigall RS, Adams FG, Lees KR. Prognostic value of single-photon emission tomography in acute ischaemic stroke. Eur J Nucl Med. 1997;24:21–26.[Medline] [Order article via Infotrieve]
10. Yonas H, Gur D, Claassen D, Wolfson SK, Moossy J. Stable xenon enhanced computed tomography in the study of clinical and pathologic correlates of focal ischemia in baboons. Stroke. 1988;19:228–239.[Abstract/Free Full Text]
11. Baird AE, Warach S. Magnetic resonance imaging of acute stroke. J Cereb Blood Flow Metab. 1998;18:583–609.[Medline] [Order article via Infotrieve]
12. Adams HP, Brott TG, Crowell RM, Furlan AJ, Gomez CR, Grotta J, Helgason CM, Marler JR, Woolsen RF, Mayberg M. Guidelines for the management of patients with acute ischemic stroke: a statement for healthcare professionals from a special writing group of the Stroke Council, American Heart Association. Stroke. 1994;25:1901–1914.[Medline] [Order article via Infotrieve]
13. Axel L. Cerebral blood flow determination by rapid-sequence computed tomography: a theoretical analysis. Radiology. 1980;137:679–686.[Abstract/Free Full Text]
14. Axel L. Tissue mean transit time from dynamic computed tomography by a simple deconvolution technique. Invest Radiol. 1983;18:94–99.[Medline] [Order article via Infotrieve]
15. Cenic A, Nabavi DG, Craen RA, Gelb AW, Lee TY. Dynamic CT measurement of cerebral blood flow: a validation study. AJNR Am J Neuroradiol. 1999;20:63–73.[Abstract/Free Full Text]
16. Nabavi DG, Cenic A, Dool J, Smith RM, Espinosa F, Craen RA, Gelb AW, Lee TY. Quantitative assessment of cerebral hemodynamics using CT: stability, accuracy, and precision studies in dogs. J Comput Assist Tomogr. 1999;23:506–515.[Medline] [Order article via Infotrieve]
17. Nabavi DG, Cenic A, Craen RA, Gelb AW, Bennett JD, Kozak JD, Lee TY. CT assessment of cerebral perfusion in man: experimental validation and initial clinical experience. Radiology. 1999;213:141–149.[Abstract/Free Full Text]
18. Scremin OU, Sonnenschein RR, Rubinstein EH. Cerebrovascular anatomy and blood flow measurements in the rabbit. J Cereb Blood Flow Metab. 1982;2:55–66.[Medline] [Order article via Infotrieve]
19. Zea Longa E, Weinstein PR, Carlson S, Cummins R. Reversible middle cerebral artery occlusion without craniectomy in rats. Stroke. 1989;20:84–91.[Abstract/Free Full Text]
20. Laing RJ, Jakubowski J, Laing RW. Middle cerebral artery occlusion without craniectomy in rats: which method works best? Stroke. 1993;24:294–298.[Abstract/Free Full Text]
21. Jeppsson PG, Olin T. Cerebral Angiography in the Rabbit: An Investigation of Vascular Anatomy and Variation in Circulation Pattern With Conditions of Injection [dissertation]. Lund, Sweden: Lund University, 1960. (Lunds Universitets Arsskrift. N.F., Avd 2, Bd 56, Nr 14.)>
22. Molnar L, Hegedüs K, Fekete I. A new model for inducing transient cerebral ischemia and subsequent reperfusion in rabbits without craniectomy. Stroke. 1988;19:1262–1266.[Abstract/Free Full Text]
23. Bederson J, Pittis LH, Germano SM, Nishimura MC, Davis RL, Bartkowski HM. Evaluation of 2,3,5-triphenyltetrazolium chloride as a stain for detection and quantification of experimental cerebral infraction in rats. Stroke. 1986;17:1304–1308.[Abstract/Free Full Text]
24. Isayama K, Pitts LH, Nishimura MC. Evaluation of 2,3,5-triphenyltetrazolium chloride staining to delineate rat brain infarcts. Stroke. 1991;22:1394–1398.[Abstract/Free Full Text]
25. Reichenbach JR, Röther J, Jonetz-Mentzel L, Herzau M, Fiala A, Weiller C, Kaiser W. Acute stroke evaluated by time-to-peak mapping during initial and early follow-up perfusion CT studies. AJNR Am J Neuroradiol. 1999;20:1842–1850.[Abstract/Free Full Text]
26. 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.[Abstract/Free Full Text]
27. Hamberg LM, Hunter GJ, Halpern EF, Hoop B, Gazelle GS, Wolf GL. Quantitative high-resolution measurement of cerebrovascular physiology with slip-ring CT. AJNR Am J Neuroradiol. 1996;17:639–650.[Abstract]
28. Lo EH, Rogowska J, Bogorodzki P, Matsumoto K, Saffran B, Wolf GL. Temporal correlation analysis of penumbral dynamics in focal cerebral ischemia. J Cereb Blood Flow Metab. 1996;16:60–68.[Medline] [Order article via Infotrieve]
29. Hossmann K. Viability thresholds and the penumbra of focal ischemia. Ann Neurol. 1994;36:557–565.[Medline] [Order article via Infotrieve]
30. Mies G, Ishimaru S, Xie Y, Seo K, Hossmann KA. Ischemic threshold of cerebral protein synthesis and energy state following middle cerebral artery occlusion in rat. J Cereb Blood Flow Metab. 1991;11:753–761.[Medline] [Order article via Infotrieve]
31. Symon L, Pasztor E, Branston NM. The distribution and density of reduced cerebral blood flow following acute middle cerebral artery occlusion: an experimental study by the technique of hydrogen clearance in baboons. Stroke. 1974;5:355–364.
32. Astrup J, Seymon L. Thresholds in cerebral ischemia: the ischemic penumbra. Stroke. 1981;12:723–725.[Free Full Text]
33. Sette G, Baron JC, Mazoyer B, Levasseur M, Pappata S, Crouzel C. Local brain haemodynamics and oxygen metabolism in cerebrovascular disease: positron emission tomography. Brain. 1989;112:931–951.[Abstract/Free Full Text]
34. Touzani O, Young AR, Derlon JM, Beaudouin V, Marchal G, Rioux P, Mezenge F, Baron JC, MacKenzie ET. Sequential studies of severely hypometabolic tissue volumes after permanent middle cerebral artery occlusion: a positron emission tomography investigation in anaesthetized baboons. Stroke. 1995;26:2112–2119.[Abstract/Free Full Text]
35. Hatazawa J, Shimosegawa E, Toyoshima H, Ardekani BA, Suzuki A, Okudera T, Miura Y. Cerebral blood volume in acute brain infarction: a combined study with dynamic susceptibility contrast MRI and 99 mTc-HMPAO-SPECT. Stroke. 1999;30:800–806.[Abstract/Free Full Text]
36. Lavy S, Melamed E, Portnoy Z. The effect of cerebral infarction on the regional cerebral blood flow of the contralateral hemisphere. Stroke. 1975;6:160–163.[Abstract/Free Full Text]
37. Feeney DM, Baron JC. Diaschisis. Stroke. 1986;17:817–830.[Free Full Text]
38. Atherton JV, Huda W. Energy imparted and effective doses in computed tomography. Med Phys. 1996;23:735–741.[Medline] [Order article via Infotrieve]
39. McCollough CH, Zink FE. Performance evaluation of a multi-slice CT system. Med Phys. 1999;26:2223–2230.[Medline] [Order article via Infotrieve]

Editorial Comment

Presented in part at the 19th International Symposium on Cerebral Blood Flow, Metabolism, and Function, June 13–17, 1999, Copenhagen, Denmark, and published in abstract form (J Cereb Blood Flow Metab. 1999;19[suppl 1]:S586).

Gary A. Rosenberg, MD, Guest Editor

Departments of Neurology, Neuroscience, and, Cell Biology and Physiology, University of New Mexico Health Sciences Center, Albuquerque, New Mexico

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References Introduction References

 
Rational stroke therapy will depend on accurate assessments of the extent of tissue injury in the very early stages of the ischemic event.^R1 The preceding report by Nabavi and colleagues demonstrates the usefulness of dynamic computed tomography (dCT) for this purpose. Algorithms to extract cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit times (MTT) from CT were developed over 20 years ago.^R2 However, it was not until the development of rapid spiral, or slip-ring, CT scanners that the information available from those algorithms could be obtained in a clinically relevant time frame.

This group has shown earlier, in animals, that the measurements obtained with the dCT scanners correlate well with standard methods of blood flow measurement.^R3 In this report, they extend those observations to correlate the CBF, CBV, and MTT with 2,3,5-triphenyltetrazolium chloride (TTC) measurements of infarct size.

Absolute measurements of CBF are possible with dCT with a rapid bolus injection of Ultravist, using an arterial measurement and a tissue measurement. If both are known, the CBF can be extracted mathematically.^R2 Similarly, CBV and MTT data are available. The authors found that dCT measurements of CBF were tightly correlated with infarct size as measured by TTC. MTT was also well correlated, but the relationship with CBV was less accurate. Because of the availability of dCT scanners in many clinical centers, human studies are being done in parallel with the animal studies. Early results look very promising for the use of dCT to accurately predict stroke size and to aid in the selection of patients for thrombolysis.^{R4 R5}

CT scanners are more widely available than MRI scanners. Use of CT contrast agents introduces a small risk of fatal reaction to the dye, and the radiation doses involved with dCT are the same as standard CT. Neither of these factors, therefore, should impede the use of these methods. Another problem with dCT is the ability to obtain only 1 slice at a time, which means that the region of the stroke may be missed. This may be overcome with the use of multislice scanners.

Technology advances are being made faster than our ability to fully evaluate their utility. Positron emission tomography, single-photon emission CT, and perfusion- or diffusion-weighted MRI all provide information on extent of tissue damage.^{R6 R7} Dynamic CT is an important advance, and in the next few years information should be available that will help guide our rational use of current thrombolytic agents and facilitate the testing of new agents.

Received April 20, 2000; revision received July 24, 2000; accepted September 7, 2000.

	References

Top Abstract Introduction Materials and Methods Results Discussion References Introduction References

 

1. Dirnagl U, Iadecola C, Moskowitz MA. Pathobiology of ischaemic stroke: an integrated view. Trends Neurosci. 1999;22:391–397.[Medline] [Order article via Infotrieve]
2. Axel L. Cerebral blood flow determination by rapid-sequence computed tomography: theoretical analysis. Radiology. 1980;137:679–686.
3. Cenic A, Nabavi DG, Craen RA, Gelb AW, Lee TY. Dynamic CT measurement of cerebral blood flow: a validation study. AJNR Am J Neuroradiol. 1999;20:63–73.
4. Rother J, Jonetz-Mentzel L, Fiala A, Reichenbach JR, Herzau M, Kaiser WA, Weiller C. Hemodynamic assessment of acute stroke using dynamic single-slice computed tomographic perfusion imaging. Arch Neurol. 2000;57:1161–1166.[Abstract/Free Full Text]
5. Lee KH, Lee SJ, Cho SJ, Na DG, Byun HS, Kim YB, Song HJ, Jin IS, Chung CS. Usefulness of triphasic perfusion computed tomography for intravenous thrombolysis with tissue-type plasminogen activator in acute ischemic stroke. Arch Neurol. 2000;57:1000–1008.[Abstract/Free Full Text]
6. Ostergaard L, Sorensen AG, Chesler DA, Weisskoff RM, Koroshetz WJ, Wu O, Gyldensted C, Rosen BR. Combined diffusion-weighted and perfusion-weighted flow heterogeneity magnetic resonance imaging in acute stroke. Stroke. 2000;31:1097–1103.[Abstract/Free Full Text]
7. Schellinger PD, Jansen O, Fiebach JB, Heiland S, Steiner T, Schwab S, Pohlers O, Ryssel H, Sartor K, Hacke W. Monitoring intravenous recombinant tissue plasminogen activator thrombolysis for acute ischemic stroke with diffusion and perfusion MRI. Stroke. 2000;31:1318–1328.[Abstract/Free Full Text]

This article has been cited by other articles:

				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, October 7, 2008; (2008) 2492080026. [Abstract] [Full Text]

				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]

				K. Lin, K.S. Kazmi, M. Law, J. Babb, N. Peccerelli, and B.K. Pramanik Measuring Elevated Microvascular Permeability and Predicting Hemorrhagic Transformation in Acute Ischemic Stroke Using First-Pass Dynamic Perfusion CT Imaging AJNR Am. J. Neuroradiol., August 1, 2007; 28(7): 1292 - 1298. [Abstract] [Full Text] [PDF]

				G TAN and T GODDARD Neuroimaging applications of multislice CT perfusion Imaging, June 1, 2007; 19(2): 142 - 152. [Abstract] [Full Text] [PDF]

				B.D. Murphy, X. Chen, and T.-Y. Lee Serial Changes in CT Cerebral Blood Volume and Flow after 4 Hours of Middle Cerebral Occlusion in an Animal Model of Embolic Cerebral Ischemia AJNR Am. J. Neuroradiol., April 1, 2007; 28(4): 743 - 749. [Abstract] [Full Text] [PDF]

				R.G. Gonzalez Imaging-guided acute ischemic stroke therapy: From "time is brain" to "physiology is brain". AJNR Am. J. Neuroradiol., April 1, 2006; 27(4): 728 - 735. [Abstract] [Full Text] [PDF]

				A.M. Laslo, J.D. Eastwood, F.-X. Chen, and T.-Y. Lee Dynamic CT Perfusion Imaging in Subarachnoid Hemorrhage-Related Vasospasm AJNR Am. J. Neuroradiol., March 1, 2006; 27(3): 624 - 631. [Abstract] [Full Text] [PDF]

				A. Sitek and R. G. Sheiman Small-Bowel Perfusion Measurement: Feasibility with Single-Compartment Kinetic Model Applied to Dynamic Contrast-enhanced CT Radiology, November 1, 2005; 237(2): 670 - 674. [Abstract] [Full Text] [PDF]

				S. P. Kloska, D. G. Nabavi, C. Gaus, E.-M. Nam, E. Klotz, E. B. Ringelstein, and W. Heindel Acute Stroke Assessment with CT: Do We Need Multimodal Evaluation? Radiology, October 1, 2004; 233(1): 79 - 86. [Abstract] [Full Text] [PDF]