Is All Perfusion-Weighted Magnetic Resonance Imaging for Stroke Equal? The Temporal Evolution of Multiple Hemodynamic Parameters After Focal Ischemia in Rats Correlated With Evidence of Infarction

Abstract

Although perfusion-weighted imaging techniques are increasingly used to study stroke, no particular hemodynamic variable has emerged as a standard marker for accumulated ischemic damage. To better characterize the hemodynamic signature of infarction, the authors have assessed the severity and temporal evolution of ischemic hemodynamics in a middle cerebral artery occlusion model in the rat. Cerebral blood flow (CBF) and total and microvascular cerebral blood volume (CBV) changes were measured with arterial spin labeling and steady-state susceptibility contrast magnetic resonance imaging (MRI), respectively, and analyzed in regions corresponding to infarcted and spared ipsilateral tissue, based on 2,3,5-triphenyltetrazolium chloride histology sections after 24 hours ischemia. Spin echo susceptibility contrast was used to measure microvascular-weighted CBV, which had a maximum sensitivity for vessels with radii between 4 and 30 μm. Serial measurements between 1 and 3 hours after occlusion showed no change in CBF (22 ± 20% of contralateral, mean ± SD) or in total CBV (78 ± 13% of contralateral) in regions destined to infarct. However, microvascular CBV progressively declined from 72 ± 5% to 64 ± 11% (P < 0.01) during this same period. Microvascular CBV changes with time were entirely due to decreases in subcortical infarcted zones (from 73 ± 9% to 57±14%, P < 0.001) without changes in the cortical infarcted territory. The hemodynamic variables showed differences in magnitude and temporal response, and these changes varied based on histologic outcome and brain architecture. Such factors should be considered when designing imaging studies for human stroke.

Keywords

Focal ischemia Cerebral blood volume CBF MRI Microvasculature

In acute stroke, the ability to image the hemodynamic state of the brain may be useful for determining tissue viability and guiding therapeutic options (Powers, 1991; Heiss et al., 1994). So-called perfusion-weighted imaging studies in humans have measured a wide variety of potential markers, including cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time, bolus peak-relaxivity, and bolus time-to-peak maps (Warach et al., 1992; Sorensen et al., 1996; Hunter et al., 1997; Tong et al., 1998; Neumann-Haefelin et al., 1999). It is still unclear whether these different markers yield inherently different information and which, if any, best reflect the tissue energy status, final infarct size, or potentially salvageable tissue in the brain after ischemic injury.

Previous focal ischemia studies have shown that CBF thresholds maintained beyond a certain period of time have a high likelihood of predicting infarction, and that CBF reductions present in the first minutes of ischemia remain essentially unchanged over the next several hours (Jones et al., 1981; Heiss, 1983; Dirnagl et al., 1989; Pappata et al., 1992; Hossmann, 1994). Thus, CBF in ischemic areas does not change significantly over the critical period when injury is still partially reversible with reperfusion therapy (Memezawa et al., 1992; Garcia et al., 1993; Jiang et al., 1998). In clinical settings, the ischemic duration is often unknown, making it difficult, if not impossible, to determine whether salvageable tissue exists even given the most accurate measurement of CBF.

Cerebral blood volume changes during ischemia have been less frequently studied, primarily because equilibrium methods such as microwave fixation autoradiography and positron emission tomography (PET) are either expensive, technically demanding, or incompatible with continuous in vivo measurements (Sette et al., 1989; Todd et al., 1993). Cerebral blood volume alterations may reflect changes in cerebrovascular caliber mediated by the endothelium and vascular smooth muscle that are a reaction to sustained, low CBF, and as such may be uniquely sensitive to cumulative tissue damage. Preserved CBV in the setting of low CBF is believed to represent an appropriate vasodilatory response to maximize oxygen extraction (Gibbs et al., 1984; Powers, 1991; Derdeyn et al., 1998), and such CBV-CBF mismatched areas have been observed during early ischemia (Dijkhuizen et al., 1997; Schulz et al., 1997; Caramia et al., 1998; Hatazawa et al., 1999).

There are compelling reasons to focus on microvascular changes during stroke. Measurement of microvascular CBV changes would permit analysis of the response of small vessels to ischemia in isolation from changes in other vascular compartments. Small arterioles and capillaries are the site of reperfusion-mediated parenchymal hemorrhage, and their integrity may play an important role in the safety and efficacy of reperfusion therapy. Inflammatory mediators, both cytokine and cell-mediated, may cause microvascular plugging and attendant perfusion compromise (Ames et al., 1968; del Zoppo et al., 1991). Microvascular CBV measurement is possible with magnetic resonance imaging (MRI) susceptibility contrast because spin echo images may be made highly sensitive only to blood in small vessels; gradient echo images are insensitive to vessel size and therefore reflect total CBV changes (Fisel et al., 1991; Weisskoff et al., 1994). Experimental studies comparing quantitative MRI-based microvascular CBV measurements with PET are consistent with known vessel size distributions in brain (Østergaard et al., 1998). Previous studies have shown that microvascular and total CBV behave differently during global hypoperfusion and focal ischemia in rats (Jones et al., 1996; Zaharchuk et al., 1999).

In the current article, the authors have used MRI to assess the hemodynamic response of the rat brain after 1 to 3 hours of focal ischemia, and have correlated this MRI data by both brain region and by using 2,3,5-triphenyltetrazolium chloride (TTC) staining as evidence of infarction after 24 hours ischemia. Cerebral blood flow changes were measured with an arterial spin labeling (ASL) method, CAPTIVE imaging (Zaharchuk et al., 1998), that is also capable of simultaneous measurement of total and microvascular CBV using steady-state susceptibility contrast methods.

MATERIALS AND METHODS

Animal model

The following protocol was approved by the Office of Laboratory Animal Research at Massachusetts General Hospital, Charlestown, MA, U.S.A. Male Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA, U.S.A.; n = 12; 355 ± 30 g, mean ± SD) were anesthetized free-breathing with 0.7% to 1.0% halothane in a 1.5:1 nitrous oxide:oxygen mixture. Permanent middle cerebral artery occlusion (MCAO) was accomplished by inserting a 4-0 nylon monofilament (Ethicon; Johnson & Johnson, Edison, NJ, U.S.A.) coated with a mixture of silicone resin (Xantopren; Bayer Dental) and a hardener (Elastomer Activator; Bayer, Osaka, Japan) through the common carotid, internal carotid, and past the origin of the MCA. The surgery was performed with the animal positioned in a plastic stereotactic frame within the RF surface coil. A femoral intraarterial catheter was placed to monitor mean arterial blood pressure and arterial blood gases (pH, Paco₂, and Pao₂). A femoral intravenous catheter was placed to administer contrast agents. Core temperature was maintained with a heated water blanket.

After imaging, the rats were allowed free access to food and water. After death after 24 hours ischemia, the rats' brains were cooled with ice and the bregma point was identified visually. The skulls were carefully removed and the brains were positioned into a custom brain slicing device. Then, 2-mm coronal slices were taken, with the center slice aligned at the previously identified bregma level. Histologic staining for infarcted tissue was performed using 2,3,5-triphenyltetrazolium chloride. These stained sections were then photographed in a horizontal position using a photomicroscope integrated with an IBM-PC for comparison with the MR images.

Magnetic resonance imaging methods

General. Immediately after occlusion, the rats were placed in the 4.7 T magnet (Omega, CSI, Fremont, CA, U.S.A.). The rats were first aligned in the MRI scanner such that the long axis of their brain was parallel to the floor making the image appear upright and level on the MRI images. They were then shifted horizontally in the magnet until the bregma suture (identifiable on a coronal scout image) was at the isocenter of the magnet. Baseline conventional gradient echo (repetition time/echo time: 4000/15 milliseconds) and spin echo (TR/TE: 4000/60 milliseconds) images were acquired before injection of the contrast agent. Imaging parameters were as follows: field of view = 25 mm, slice thickness = 2 mm, matrix size = 64 × 64. Eight contiguous coronal slices were collected, covering the brain from the level of the olfactory bulb to the anterior thalamus.

CBV imaging. In 6 rats, a steady-state contrast agent, monocrystalline iron oxide nanoparticles (MION), at a concentration of 15 mg Fe/kg, was injected immediately after acquisition of baseline images (Mandeville et al., 1998). In the remaining 6 rats, a different contrast agent, MPEG-PL-DyDTPA, at a concentration of 210 μmol Dy/kg, was used to facilitate the measurement of CBF with CAPTIVE-ASL (Zaharchuk et al., 1998). Postcontrast gradient and spin echo images with the same parameters as the baseline images were taken approximately every 20 minutes between 1 and 3 hours after occlusion to measure total and microvascular CBV, respectively. Cerebral blood volume maps were created by mapping transverse relaxivity increases, by applying the formula -ln(S_post/S_pre)/TE to each voxel—S_post is the signal intensity postcontrast, S_pre is the signal intensity precontrast, and TE is the echo time. Based on simulations, the authors estimated that greater than 95% of the relaxivity increase in the spin echo images used to measure microvascular CBV arose from blood in vessels with radii between 4 and 30 μm, comprising capillaries and small arterioles and venules (Weisskoff et al., 1994).

CBF imaging. In the 6 animals given MPEG-PL-DyDTPA, CAPTIVE imaging was used to measure absolute CBF in a single slice located at bregma using a conventional proton-density imaging (repetition time/echo time: 4000/5 milliseconds) with the following parameters: field of view = 25 mm, slice thickness = 2 mm, matrix size = 32 × 32, labeling time = 3700 milliseconds, postlabeling delay = 200 milliseconds, cycling radiofrequency offset (Δf = ±3 kHz), and labeling gradient (G_z = ±0.5 G/cm). Cerebral blood flow was measured approximately every 20 minutes interleaved with the CBV measurements. Cerebral blood flow was calculated using the equation

CBF = \frac{λ}{T_{1 app}} \frac{S_{control} - S_{label}}{S_{label} + (2 α - 1) S_{control}}

where S_label and S_control represent signal intensity during proximal and distal spin labeling, T_1app is the gray matter T₁ (taken to be 1.5 seconds), λ is the brain:blood partition coefficient (taken to be 0.88), and α is the degree of spin labeling at the imaging plane, measured in previous experiments to be 0.7 (precontrast) and 0.6 (postcontrast;Zaharchuk et al., 1998).

Data analysis

Regions of interest corresponding to infarcted and spared ipsilateral tissue were determined by a coauthor (M. Yamada) blinded to the MR data by comparing the TTC-stained photomicrographs to the acquired MR images using anatomical landmarks for guidance. Spared tissue was defined as ipsilateral brain tissue that was not infarcted after 24 hours ischemia. Subcortical and cortical regions of interest were outlined, allowing examination of hemodynamic changes in a matrix of the following four areas: subcortical infarcted, subcortical spared, cortical infarcted, and cortical spared. Tissue of the olfactory bulb, substantia nigra, and preoptic regions were not included in these measurements. Measurements were also made in five anatomically defined regions in the single slice located at bregma. In one rat there was no infarction in this slice, therefore the CBF and CBV data presented for the infarcted and spared regions in Table 1 represent the mean of 5 rats and 11 rats, respectively.

TABLE 1.

Regional hemodynamic changes after middle cerebral artery occlusion in a 2-mm-thick coronal slide located at bregma*

		CBF(% conlralaieral)			Total CBV(% contralaieral)			Microvascular CBV(% contra lateral)
	Mean(mL/g/min)	Mean	Initial	Final	Mean	Initial	Final	Initial	Final
Time (min)			59 ± 17	169 ± 9		69 ± 16	178 ± 8	72 ± 18	178 ± 9
Contralateral	1.87 ± 0.19
Region 1	1.43 ± 0.25	69 ± 10	81 ± 26	75 ±7	75 ± 18	76 ± 17	74 ± 23	83 ± 6	81 ± 13
Region 2	0.42 ± 0.13	19 ± 9	28 ± 15	23 ± 22	98 ± 3 1	89 ± 28	103 ± 38	71 ± 18	71 ± 21
Region 3	0.39 ± 0.22	22 ± 18	36 ± 18	25 ± 23	74 ± 23	72 ± 21	73 ± 24	69 ± 20	65 ± 36
Region 4	0.23 ± 0.12	16 ± 14	21 ± 17	27 ± 17	68 ± 15	72 ± 16	64 ± 19	60 ± 18	44 ± 20†
Region 5	0.81 ± 0.37	45 ± 22	52 ± 17	45 ± 24	88 ± 18	91 ± 17	83 ± 19	82 ± 15	72 ± 14†
Infarcted	0.41 ± 0.33	22 ± 20	29 ± 20	25 ± 15	65 ± 13	67 ± 14	64 ± 15	62 ± 11	52 ± I8†
Spared	1.18 ± 0.54	64 ± 30	69 ± 25	57 ± 28	97 ± 19	95 ± 21	97 ± 21	93 ± 20	95 ± 26

Mean values represent the average of all measurements made within the I- to 3-hour period after middle cerebral artery occlusion.

Indicates % contra lateral, mean ± SD.

†

Initial measurements significantly different from final measurement (P > 0.01 paired Student's t-test).

Hemodynamic measurements were expressed as percent of the mean contralateral level at the same time point. All measurements in each rat were binned into groups based on the time from occlusion using 20-minute intervals (that is, <80 minutes, 80 to 100 minutes, 100 to 120 min, and so on). As CBV changes in the MION and MPEG-PL-DyDTPA groups were similar, all CBV results are shown as the pooled data for 12 animals. Paired two-tailed Student's t-tests using the Bonferroni correction for multiple comparisons were used to compare the different hemodynamic parameters. Repeated measures analysis of variance was used to compare the physiologic parameters. To determine the significance of changes in hemodynamic parameters over time, the authors used a test for nonzero slope on each parameter. All data is expressed as mean ± SD.

RESULTS

Physiologic monitoring was stable and reflected normal values for nonventilated rats (Table 2). For all rats, infarction volume as determined by TTC staining after 24 hours ischemia comprised 35 ± 15% of the ipsilateral hemisphere.

TABLE 2.

Physiologic parameters in rats subject to MCAO

Time(postocclusion)	MABP(mm Hg)	Temperature (C)	pH	PaCO₂(mm Hg)	PaO₂(mm Hg)
1 hour	103 ± 19	37.6 ± 0.7	7.33 ± 0.08	46 ± 9	177 ± 33
2 hours	108 ± 23	37.5 ± 0.6	7.32 ± 0.11	45 ± 9	149 ± 32
3 hours	108 ± 20	37.5 ± 0.5	7.33 ± 0.10	45 ± 8	157 ± 21

Values are expressed as mean ± SD in 12 rats, except for mean arterial blood pressure (MABP) and arterial blood gases (n = 10). There were no significant changes over time. MCAO, middle cerebral artery occlusion.

Fig. 1 is an example of total and microvascular CBV images after 1 hour of MCAO with representative TTC-stained sections demonstrating infarction after 24 hours occlusion. Differences in image contrast were evident—the total CBV maps outlined major vascular structures, whereas the microvascular CBV maps demonstrated primarily parenchymal (that is, gray–white) contrast. In addition, the relative reduction in microvascular CBV was greater than that for total CBV in regions that later infarcted. Fig. 2 is representative of the images used to examine the changes in several different hemodynamic parameters over time. Cerebral blood flow was decreased in a region that corresponded to infarcted tissue and did not change between 1 and 3 hours after occlusion. Total CBV in this rat was minimally affected and was also unchanged between 1 and 3 hours after occlusion. In contrast, a focal area of microvascular CBV decrease was seen to expand with time in the ischemic regions, particularly in striatum.

FIG. 1.

Spin echo anatomic images, total cerebral blood volume (CBV) maps, microvascular CBV maps, and 2,3,5-triphenyltetrazolium chloride (TTC) stained photomicrographs after middle cerebral artery occlusion using the steady-state contrast agent monocrystalline iron oxide nanoparticles (MION). Total and microvascular CBV images were taken after 1 hour, whereas TTC staining was performed after 24 hours occlusion. Total CBV images were dominated by large vessels, whereas microvascular CBV images emphasized parenchymal (that is, gray–white matter) differences. Note that microvascular CBV was more severely affected in the infarcted regions than was total CBV.

FIG. 2.

(A) 2,3,5-triphenyltetrazolium chloride staining after 24 hours ischemia, (B) cerebral blood flow (CBF), (C) total cerebral blood volume (CBV), and (D) microvascular CBV after 1, 2, and 3 hours middle cerebral artery occlusion using arterial spin labeling and the steady-state contrast agent MPEG-PL-DyDTPA. Cerebral blood flow was markedly reduced, whereas total CBV changes were minor; neither changed with the length of ischemia. Note the progressive microvascular CBV decrease in regions that infarct, which is particularly evident in the striatum (arrows).

Fig. 3 shows the temporal evolution of CBF in infarcted and spared tissue. Static CBF reduction wasobserved in infarcted tissue between 1 and 3 hours of occlusion, with a mean CBF of 0.41 ± 0.33 mL/g/min, corresponding to 22 ± 20% of contralateral. In spared tissue, CBF was higher but still not normal (1.18 ± 0.54 mL/g/min, or 64 ± 30% of contralateral). Contralateral CBF was 1.87 ± 0.19 mL/g/min. Cerebral blood flow in different anatomical regions is shown in Table 1.

FIG. 3.

Cerebral blood flow (CBF) versus time in ipsilateral tissue that was either infarcted (•) or spared (○) after 24 hours of middle cerebral artery occlusion in a single 2-mm-thick slice located at bregma. Cerebral blood flow was sharply decreased in the infarcted zone and did not change with time. Significant hypoperfusion was also seen in ipsilateral spared regions. The difference between CBF in the spared and infarcted tissue was significant at all time points (P < 0.05).

Total and microvascular CBV changes are shown in Fig. 4. There was no significant difference between total CBV in the infarcted zone depending on the contrast agent (MION: 78 ± 8% versus MPEG-PL-DyDTPA: 74 ± 16%, P > 0.10) used during the 2 hour period of the experiment. Based on this, all CBV data shown represents pooled data from both groups of animals. Total and microvascular CBV changes in different anatomical regions are shown in Table 1. While all regions except region 1 (which is in the anterior cerebral artery rather than the MCA territory) demonstrated some degree of CBV/CBF mismatch, the most striking mismatches were present in cortex. For example, in region 2, mean CBF was only 19 ± 9% of contralateral while the total CBV was unchanged (98 ± 30% of contralateral); microvascular CBV in this region showed an intermediate decline to 71 ± 18% of contralateral.

FIG. 4.

Total cerebral blood volume (CBV) (•) and microvascular CBV (▪) versus time in (A) ipsilateral spared tissue and (B) infarcted tissue. Note the different absolute ranges of CBV levels for the different graphs. Mild CBV decreases were seen in spared tissue. In infarcted tissue, total CBV reduction was slightly more severe but did not change with time. Microvascular CBV was significantly lower than total CBV at all time points measured in infarcted tissue (P < 0.05). Microvascular CBV in infarcted tissue decreased significantly between 1 and 3 hours after occlusion (P < 0.01; test for nonzero slope).

Small reductions in total CBV (97 ± 10% of contralateral, P > 0.05) and microvascular CBV (92 ± 9% of contralateral, P < 0.05) were observed in spared ipsilateral tissue (Fig. 4A). No changes with time were present. In the infarcted zones (Fig. 4B), total CBV was mildly but significantly reduced and did not change between 1 and 3 hours occlusion (78 ± 10% after 1 hour versus 76 ± 14% of contralateral after 3 hours ischemia). Mean total CBV during the 2-hour interval was 76 ± 12% of contralateral. After 1-hour occlusion, microvascular CBV in the infarcted zone was more severely affected than total CBV and declined steadily during the following 2 hours (Fig. 4B). The decrease of microvascular CBV in infarcted tissue with advancing time was the only hemodynamic parameter measured in which a linear fit showed a slope significantly different from zero (P < 0.01).

When the temporal changes were further broken down into subcortical and cortical regions (Table 3), it became clear that the progressive microvascular CBV decline in infarcted zones was specific only to subcortical areas, where a decrease from 73 ± 9% to 57 ± 14% (P < 0.001) of contralateral was observed between 1 and 3 hours of ischemia; cortical infarcted regions did not change significantly (from 72 ± 12% to 75 ± 20% of contralateral, respectively). Slight decreases in total CBV were also seen in subcortical infarcted areas (from 82 ± 10% to 76 ± 16%), but this was not statistically significant (P > 0.10).

TABLE 3.

Total and microvascular cerebral blood volume changes in cortical and subcortical regions of the full brain*

		Cortex		Subcortical
	Time (min)	Infarcted	Spared	Infarcted	Spared
Total CBV
Initial	69 ± 16	81 ± 16	87 ± 16	82 ± 10	105 ± 11
Final	178 ± 8	81 ± 12	82 ± 22	76 ± 16	105 ± 15
Average		82 ± 16	83 ± 19	78 ± 12	106 ± 13
Microvascular CBV
Initial	72 ± 18	72 ± 12	85 ± 7	73 ± 9	95 ± 8
Final	178 ± 9	75 ± 20	87 ± 15	57 ± 14**	90 ± 12

Indicates % contralateral, mean ± SD, n = 12.

P < 0.001; paired Student's t-test for initial versus final measurement.

DISCUSSION

Cerebral blood flow

The measurement of mean CBF in the contralateral hemisphere of 1.87 ± 0.19 mL/g/min is slightly higher than that reported in some other rodent studies (1.1 to 1.7 mL/g/min) (Bolander et al., 1989; Back et al., 1995), but the value is consistent with the slightly higher Paco₂ levels in the authors' nonventilated rats and the use of halothane anesthesia. In regions that infarcted, marked CBF deficits were observed, whereas for spared regions, a lesser but still significant degree of hypoperfusion was present. The finding of static flow reduction in tissue destined to infarct was consistent with previous results in different species, using autoradiographic and noninvasive MR perfusion measurements (Dirnagl et al., 1989; Pappata et al., 1992; Schulz et al., 1997; Calamante et al., 1999). Thus, the current study further demonstrates that CBF alone cannot distinguish reversible from irreversible tissue damage without knowledge of ischemic duration.

Although this is the first study to measure CBF with a high-resolution imaging modality during early ischemia and compare it with infarction at a later time point in the same animals, several previous studies attempted to address this issue using either early histologic markers or different groups of animals. Tyson et al. (1984) demonstrated that CBF levels below 0.25 mL/g/min after 4 hours ischemia produced consistent histologic signs of derangement, whereas for CBF levels between 0.25 and 0.35 mL/g/min such signs were inconsistently observed. By examining the spatial coincidence of CBF and infarction measured in a separate group of animals, a flow level of 0.23 mL/g/min was found to be consistent with infarction using a photochemically-induced distal MCA thrombosis model (Back et al., 1995). In two separate studies, Jacewicz et al. (1990 and 1992) subjected normal Fisher and spontaneously hypertensive rats to both MCA and common carotid occlusion and found that a CBF level of 0.50 mL/g/min was most consistent with that determined histologically in separate groups of animals. The results presented in the current article of absolute CBF of 0.41 ± 0.33 mL/g/min between 1 and 3 hours of ischemia in infarcted tissue in the same rats after 24 hours are more consistent with the studies of Jacewicz et al. (1990 and 1992). This may reflect the similar timing of CBF measurements and infarct determination. Although the blinded observer relied on anatomical landmarks to delineate infarction areas, it is possible that a small amount of misregistration due to edema could still be present, making precise correlation of early and late markers of infarction difficult.

A previous study comparing a volume-based ASL method for measuring CBF with radiolabeled microspheres concluded that ASL underestimated flow by as much as 32% at low flow and suggested that this was likely due to increases in the arterial transit delay of labeled spins (Walsh et al., 1994). In this measurement, the authors aimed to decrease transit delay effects by using a postlabeling delay (Alsop and Detre, 1996), which has been demonstrated to be effective (Calamante et al., 1999). Arterial spin labeling is likely to have heightened sensitivity for CBF changes at low flow because direct flow effects and transit delay increases combine to affect the ASL signal; however, because of the overall variability in CBF levels measured, as demonstrated by the error bars in Fig. 3, small changes in CBF during this time period cannot be ruled out.

Total and microvascular CBV

Several experimental studies of susceptibility contrast MRI have demonstrated the accuracy of this tool for studying CBV changes in vivo.Boxerman et al. (1995) showed concordance between Monte Carlo and experimental susceptibility contrast studies with measurement of total CBV in rat brain using ¹²⁵I-labeled albumin autoradiography. Also, Ostergaard et al. (1998), using a quantitative approach to dynamic susceptibility contrast found that the spin echo susceptibility contrast measurement was sensitive to approximately 50% of the total CBV, which is consistent with estimates of small vessel volume in the pig brain.

Steady-state susceptibility contrast methods, in which the contrast agent reaches an equilibrium blood level during the imaging period, improve the sensitivity of CBV measurements by increasing the available imaging time (Kent et al., 1990). Equilibrium methods may also be required for accurate measurement of CBV because they are less sensitive to regional flow differences. Incomplete or delayed access of the bolus to low flow regions may cause significant underestimations in bolus-based CBV measurements (Lassen and Perl, 1979; Hunter et al., 1997). Such underestimation is suggested by previous reports during early focal ischemia, in which steady-state techniques such as PET have consistently reported higher CBV levels than bolus MRI and CT studies (Pappata et al., 1992; Warach et al., 1992; Heiss et al., 1994; Furlan et al., 1996; Sorensen et al., 1996; Hunter et al., 1997). Gamma-variate fitting of the agent's first pass should minimize these problems; however, fitting a broad gamma-variate function under noisy conditions can be unreliable, especially in the presence of heterogeneous flow within a voxel (Boxerman et al., 1997). In the current study, total and microvascular CBV in tissue destined to infarct were significantly reduced from their mean value in the contralateral hemisphere during the 1 to 3 hours after occlusion. The absolute level of total CBV decrease was mild, which was consistent with other steady-state techniques in animals and humans (Powers et al., 1985; Heiss et al., 1992; Pappata et al., 1992; Furlan et al., 1996; Quast et al., 1997).

Although coupling between CBF and CBV has been demonstrated during Paco₂ alterations (Grubb et al., 1974) and functional activation (Mandeville et al., 1998), it appears to break down during ischemia (Gibbs et al., 1984; Dijkhuizen et al., 1997; Quast et al., 1997; Schulz et al., 1997; Caramia et al., 1998). Microvascular CBV images are maximally sensitized to blood volume in capillaries, small arterioles, and small venules (Weisskoff et al., 1994). An important limitation of the technique is that it cannot distinguish between blood volume in these different vessels, which may behave differently during hypoperfusion. Nevertheless, by removing large vessel effects the technique offers a closer examination of microvascular hemodynamic changes, whose precise anatomical location and mechanism may be explored further with more invasive techniques. In this study, microvascular CBV reduction within the infarcted zones were larger than total CBV reduction, and both were less severely affected than CBF. No hyperemic regions surrounding infarcted tissue were observed; rather, small CBV decreases were noted in the hypoperfused regions that were spared. Such hemodynamic relations were in agreement with a previous study of global hypoperfusion (Zaharchuk et al., 1999).

Several studies have examined total and microvascular CBV changes after focal ischemia in rats. Jones et al. (1996) demonstrated microvascular CBV reduction greater than total CBV reduction in ipsilateral caudoputamen and cortex after 1 to 2 hours of MCAO. Mild microvascular CBV decreases in the MCA territory of rats using spin echo susceptibility contrast have also been shown after 4 hours of permanent MCAO (Quast et al., 1997). However, the current study differs from these previous studies in two key respects: the examination of temporal changes and the use of regions of interest based on histologic evidence of infarction in the same animals. Such an experimental protocol may be important given the known variability of rat stroke models.

In this study, the combination of constant CBF and total CBV reductions coupled with a progressive drop in microvascular CBV reflects a redistribution of blood volume between different compartments. One reasonable explanation is that flow is diverted from the microvasculature bed to thoroughfare channels, causing a decrease in capillary perfusion and a shift in blood volume from the microvasculature to larger vessels. Such behavior would be consistent with the concept of ineffective perfusion and venous pooling after focal ischemia (Tomita, 1988). Ultimately, such hemodynamic changes should lead to decreases in the oxygen extraction fraction, which appears to be a marker for irreversibly damaged tissue based on PET measurements (Sette et al., 1989; Powers, 1991).

Temporal evolution of microvascular cerebral blood volume

Progressively decreasing microvascular CBV levels during the period from 1 to 3 hours were seen only in the regions destined to infarct. This decrease could be attributed solely to changes in subcortical infarcted regions with static CBV reductions in cortical infarcted areas. Although the etiology of this decrease could not be elucidated in this study, several reports suggest that microvascular compromise is a significant feature of early ischemia, particularly in subcortical regions. Polymorphonuclear leukocytes have been found to occlude up to 40% of capillaries in the striatum of baboons after 2 hours focal ischemia (del Zoppo et al., 1991) and may be responsible for the “no-reflow” phenomenon after reperfusion (Ames et al., 1968). Polymorphonuclear leukocyte migration into tissue is maximal after 48 to 72 hours ischemia, but increases in polymorphonuclear leukocyte density that may cause vessel occlusion occur during the initial 4 hours of ischemia (Knight et al., 1994). Although a recent study suggested that all capillaries are perfused during the first hour after MCAO (Theilen et al., 1994), if ischemia is further prolonged, the ratio of perfused to nonperfused capillaries may change. Also, the swelling of astrocytic foot processes, which is known to occur with cytotoxic edema (Willard-Mack et al., 1996), may cause significant changes in microvascular CBV with minimal effects on total CBV.

The different microvascular structure of the striatum may explain the regional differences in microvascular CBV behavior. Cortical regions may be more resistant to microvascular compromise based on their ability to form perfusion collaterals. As the striatum is known to be the most difficult tissue to salvage with neuroprotective strategies, it is possible that targeting strategies that preserve the microvasculature may prove effective. Although it could not be observed in this study, microvascular CBV changes may occur later in cortical ischemic regions, as suggested by a previous report using a bolus technique (Dijkhuizen et al., 1997).

Microvascular CBV lesions developed in regions with the most severe CBF deficits, suggesting that microvascular CBV status may reflect increasing severity of the ischemic insult with time. This decrease occurred over the time period reported by previous studies to represent the therapeutic window in reperfusion models of ischemia in rats (Memezawa et al., 1992; Garcia et al., 1993) and evolves over a somewhat longer time period than the changes in apparent diffusion coefficient thought to represent the onset of cytotoxic edema (Moseley et al., 1990). Furthermore, apparent diffusion coefficient changes also appear to be different in cortical and subcortical regions (Hoehn-Berlage et al., 1995). Studies of the relation between apparent diffusion coefficient changes and microvascular CBV may shed light on the relation between microvascular perfusion changes and their effect on tissue metabolism.

Potential errors in susceptibility contrast cerebral blood volume measurements

Cerebral blood volume measurement errors with steady-state MR susceptibility contrast can occur with changes in blood contrast agent concentration or in such MR parameters as proton density, T1, T2, or T2*. The contrast agents used in the current study have biological half-lives on the order of 24 hours (Bogdanov et al., 1993; Mandeville et al., 1998). During the time between 1 and 3 hours ischemia, previous studies have shown no changes in proton density and only minimal changes in T1 (Hoehn-Berlage et al., 1995; Calamante et al., 1999). Because T1 insensitive sequences were used in the current study, the authors estimate an error of less than 0.3% based on T1 increases during this time frame. The relatively large contrast agent doses used in this study mitigate the effects of T2 and T2* changes that may reflect either blood oxygenation level-dependent effects or the onset of vasogenic edema (Hoehn-Berlage et al., 1995; Jiang et al., 1998). Also, any such errors would affect the measurements in rats using MPEG-PL-DyDTPA more significantly than those that received MION (a maximum possible effect of 11% versus 4%, respectively), because the baseline relaxivity increase for MPEG-PL-DyDTPA treated rats was greater than for those rats receiving MION. In the current study, no differences were seen between the two groups, and there were no parallel changes between microvascular (T2) and total (T2*) CBV measurements. Based on this, the authors believe that the actual error is far less, particularly as contrast agents remove the intravascular blood oxygenation level-dependent effect. Changes in the permeability of the blood–brain barrier would not affect the CBV measurements presented in this article because the contrast agents used are macromolecules with effective radii greater than 20 nm.

Finally, several authors have suggested that changes in flow may lead to errors in measuring CBV using susceptibility contrast. White et al. (1992) suggest that decreased flow would lead to CBV underestimation, whereas others (Hardy and Henkelman, 1991; Boxerman et al., 1995) predict that there may be a slight CBV overestimation with a maximum possible effect on the order of 10% for total and microvascular CBV. It is unlikely that such effects are responsible for the progressive decline in microvascular CBV in subcortical infarcted areas because similar flow levels are present in the infarcted cortex in the absence of such effects.

Conclusion

The fundamental finding of this study was that several different hemodynamic parameters (CBF, total CBV, and microvascular CBV) were found to have significantly discordant behavior in the early ischemic period. Thus, great care is required when interpreting perfusion-weighted imaging studies in humans that use different markers or combinations of markers to assess ischemic damage after human stroke, as not all markers of ischemia are equal. Differing degrees of CBF and total CBV reduction were observed in infarcted versus spared tissue, which were independent of ischemic time during the first 3 hours after occlusion. The authors found that microvascular CBV, measured by spin echo susceptibility contrast MRI, progressively decreased with the duration of the ischemic insult as the tissue became irreversibly damaged. Measurement of microvascular CBV, particularly using steady-state paradigms, may prove valuable for examining stroke in humans, as it may be sensitive to accumulated ischemic damage.

Footnotes

Acknowledgements

The authors thank Dr. Eng H. Lo for valuable discussions and encouragement.

References

Alsop

Detre

(1996) Reduced transit time sensitivity in noninvasive magnetic resonance imaging of human cerebral blood flow. J Cereb Blood Flow Metab 16:1236–1249

Ames

Wright

Kowade

Thurston

Majors

(1968) Cerebral ischemia: II. the no-reflow phenomenon. Am J Pathol 52:437–453

Back

Zhao

Ginsberg

(1995) Three-dimensional image analysis of brain glucose metabolism-blood flow uncoupling and its electrophysiological correlates in the acute ischemic penumbra following middle cerebral artery occlusion. J Cereb Blood Flow Metab 15:566–577

Bogdanov

Jr Weissleder

Frank

Bogdanova

Nossif

Schaffer

Tsai

Papisov

Brady

(1993) A new macromolecule as a contrast agent for MR angiography: preparation, properties, and animal studies. Radiology 187:701–706

Bolander

Persson

Hillered

d'Argy

Ponten

Olsson

(1989) Regional cerebral blood flow and histopathologic changes after middle cerebral artery occlusion in rats. Stroke 20:930–937

Boxerman

Hamberg

Rosen

Weisskoff

(1995) MR contrast due to intravascular magnetic susceptibility perturbations. Magn Reson Med 34:555–566

Boxerman

Rosen

Weisskoff

(1997) Signal-to-noise analysis of cerebral blood volume maps from dynamic NMR imaging studies. J Magn Reson Imaging 7:528–537

Calamante

Lythgoe

Pell

Thomas

King

Busza

Sotak

Williams

Ordidge

Gadian

(1999) Early changes in water diffusion, perfusion, T1, and T2 during focal cerebral ischemia in the rat studied at 8.5 T. Magn Reson Med 41:479–485

Caramia

Huang

Hamberg

Weisskoff

Zaharchuk

Moskowitz

Cavagna

Rosen

(1998) Mismatch between cerebral blood volume and flow index during transient focal ischemia studied with MRI and GD-BOPTA. Magn Reson Imaging 16:97–103

10.

del Zoppo

Schmid-Schönbein

Mori

Copeland

Chang

(1991) Polymorphonuclear leukocytes occlude capillaries following middle cerebral artery occlusion and reperfusion in baboons. Stroke 22:1276–1283

11.

Derdeyn

Powers

Grubb

Jr (1998) Hemodynamic effects of middle cerebral artery stenosis and occlusion. American Journal of Neuroradiology 19:1463–1469

12.

Dijkhuizen

Berkelbach van der Sprenkel

Tulleken

KAF

Nicolay

(1997) Regional assessment of tissue oxygenation and the temporal evolution of hemodynamic parameters and water diffusion during acute focal ischemia in rat brain. Brain Res 750:161–170

13.

Dirnagl

Kaplan

Jacewicz

Pulsinelli

(1989) Continuous measurement of cerebral cortical blood flow by laser Doppler flowmetry in a rat stroke model. J Cereb Blood Flow Metab 9:589–596

14.

Fisel

Ackerman

Buxton

Garrido

Belliveau

Rosen

Brady

(1991) MR contrast due to microscopically heterogeneous magnetic susceptibility: numerical simulations and applications to cerebral physiology. Magn Reson Med 17:336–347

15.

Furlan

Marchal

Viader

Derlon

J-M

Baron

J-C

(1996) Spontaneous neurological recovery after stroke and the fate of the ischemic penumbra. Ann Neurol 40:216–226

16.

Garcia

Yoshida

Chen

Zhang

Lian

Chen

Chopp

(1993) Progression from ischemic injury to infarct following middle cerebral artery occlusion in the rat. Am J Pathol 142:623–635

17.

Gibbs

Wise

RJS

Leenders

Jones

(1984) Evaluation of cerebral perfusion reserve in patients with carotid-artery occlusion. Lancet i:310–314

18.

Grubb

Raichle

Eichling

Ter-Pogossian

(1974) The effects of changes in PaCO₂ on cerebral blood volume, blood flow, and vascular mean transit time. Stroke 5:630–639

19.

Hardy

Henkelman

(1991) Transverse relaxation rate enhancement induced by diffusion of spins through inhomogeneous fields. Magn Reson Med 17:348–356

20.

Hatazawa

Shimosegawa

Toyoshima

Ardekani

Suzuki

Okudera

Miura

(1999) Cerebral blood volume in acute brain infarction: a combined study with dynamic susceptibility contrast MRI and 99m-Tc-HMPAO SPECT. Stroke 30:800–806

21.

Heiss

W-D

(1983) Flow thresholds for functional and morphological damage of brain tissue. Stroke 14:329–331

22.

Heiss

W-D

Graf

Wienhard

Lottgen

Saito

Fujita

Rosner

Wagner

(1994) Dynamic penumbra demonstrated by sequential multitracer PET after middle cerebral artery occlusion in cats. J Cereb Blood Flow Metab 14:892–902

23.

Heiss

W-D

Huber

Fink

Herholz

Pietrzyk

Wagner

Weinhard

(1992) Progressive derangement of peri-infarct viable tissue in ischemic stroke. J Cereb Blood Flow Metab 12:193–203

24.

Hoehn-Berlage

Eis

Back

Kohno

Yamashita

(1995) Changes of relaxation times (T1, T2) and apparent diffusion coefficient after permanent middle cerebral artery occlusion in the rat: temporal evolution, regional extent, and comparison with histology. Magn Reson Med 34:824–834

25.

Hossmann

K-A

(1994) Viability thresholds and the penumbra of focal ischemia. Ann Neurol 36:557–565

26.

Hunter

Hamberg

Ponzo

Huang-Hellinger

Morris

Rabinov

Farkas

Lev

Schaefer

Ogilvy

Schwamm

Buonanno

Koroshetz

Wolf

Gonzalez

(1997) Assessment of cerebral perfusion and arterial anatomy in hyperacute stroke with three-dimensional functional CT: early clinical results. American Journal of Neuroradiology 19:29–37

27.

Jacewicz

Tanabe

Pulsinelli

(1992) The CBF threshold and dynamics for focal cerebral infarction in spontaneously hypertensive rats. J Cereb Blood Flow Metab 12:359–370

28.

Jacewicz

Tanabe

Wang

Pulsinelli

(1990) Blood flow in the borderzone of focal cerebral infarcts in rats. In: Proceedings of the Society for Neuroscience, 20th Annual Meeting, St. Louis, MO, October 28–November 2, 1990, p 1278

29.

Jiang

Zhang

Ewing

Divine

Chopp

(1998) Diffusion, T2, and perfusion-weighted nuclear magnetic resonance imaging of middle cerebral artery embolic stroke and recombinant tissue plasminogen activator intervention in rat. J Cereb Blood Flow Metab 18:758–767

30.

Jones

Haraldseth

Baptista

Müller

Øksendal

(1996) A study of the contribution of changes in the cerebral blood volume to the haemodynamic response to anoxia in rat brain. NMR Biomed 9:233–240

31.

Jones

Morawetz

Crowell

Marcoux

FitzGibbon

DeGirolami

Ojemann

(1981) Thresholds of focal cerebral ischemia in awake monkeys. J Neurosurg 54:773–782

32.

Kent

Quast

Kaplan

Lifsey

Eisenberg

(1990) Assessment of a superparamagnetic iron oxide (AMI-225) as a brain contrast agent. Magn Reson Med 13:434–443

33.

Knight

Dereski

Helpern

Ordidge

Chopp

(1994) Magnetic resonance imaging assessment of evolving focal cerebral ischemia - comparison with histopathology in rats. Stroke 25:1252–1261

34.

Lassen

Perl

(1979) Tracer kinetic methods in medical physiology. New York, NY: Raven Press

35.

Mandeville

Marota

JJA

Kosofsky

Keltner

Weissleder

Rosen

Weisskoff

(1998) Dynamic functional imaging of relative cerebral blood volume during rat forepaw stimulation. Magn Reson Med 39:615–624

36.

Memezawa

Smith

Siesjo

(1992) Penumbral tissues salvaged by reperfusion following middle cerebral artery occlusion in rats. Stroke 23:552–559

37.

Moseley

Cohen

Mintorovitch

Chileuitt

Shimizu

Kucharczyk

Wendland

Weinstein

(1990) Early detection of regional cerebral ischemia in cats: comparison of diffusion and T2 weighted MRI and spectroscopy. Magn Reson Med 14:330–346

38.

Neumann-Haefelin

Wittsack

Wenserski

Siebler

Seitz

Modder

Freund

(1999) Diffusion- and perfusion-weighted MRI. The DWI/PWI mismatch region in acute stroke. Stroke 30:1591–1597

39.

Østergaard

Smith

Vestergaard-Poulsen

Hansen

Gee

Gjedde

Gyldensted

(1998) Absolute cerebral blood flow and blood volume measured by magnetic resonance imaging bolus tracking: comparison with positron emission tomography values. J Cereb Blood Flow Metab 18:425–432

40.

Pappata

Fiorelli

Rommel

Hartmann

Dettmers

Yamaguchi

Chabriat

Poline

Crouzel

di Giamberardino

Baron

(1992) PET study of changes in local brain hemodynamics and oxygen metabolism after unilateral middle cerebral artery occlusion in baboons. J Cereb Blood Flow Metab 13:416–424

41.

Powers

(1991) Cerebral hemodynamics in ischemic cerebrovascular disease. Ann Neurol 29:231–240

42.

Powers

Grubb

Jr Darriet

Raichle

(1985) Cerebral blood flow and cerebral metabolic rate of oxygen requirements for cerebral function and viability in humans. J Cereb Blood Flow Metab 5:600–608

43.

Quast

Wei

Huang

Brunder

Sell

Gonzalez

Hillman

Kent

(1997) Perfusion deficit parallels exacerbation of cerebral ischemia/reperfusion injury in hyperglycemic rats. J Cereb Blood Flow Metab 17:553–559

44.

Schulz

Panahian

Chen

Beal

Moskowitz

Rosen

Jenkins

(1997) Facilitation of post-ischemic cerebral reperfusion using alpha-PBN. Assessment using NMR and laser Doppler flow techniques. Am J Physiol 272(4 pt 2):H1986–1995

45.

Sette

Baron

Mazoyer

Levasseur

Pappata

Crouzel

(1989) Local brain haemodynamics and oxygen metabolism in cerebrovascular disease. Brain 112:931–951

46.

Sorensen

Buonanno

Gonzalez

Schwamm

Lev

Huang-Hellinger

Reese

Weisskoff

Davis

Suwanwela

Can

Moreira

Copen

Look

Finkelstein

Rosen

Koroshetz

(1996) Hyperacute stroke: evaluation with combined multisection diffusion-weighted and hemodynamically weighted echo-planar MR imaging. Radiology 199:391–401

47.

Theilen

Schröck

Kuschinsky

(1994) Gross persistence of capillary plasma perfusion after middle cerebral artery occlusion in the rat brain. J Cereb Blood Flow Metab 14:1055–1061

48.

Todd

Weeks

Warner

(1993) Microwave fixation for the determination of cerebral blood volume in rats. J Cereb Blood Flow Metab 13:328–336

49.

Tomita

(1988) Significance of cerebral blood volume. In: Cerebral hyperemia and ischemia: from the standpoint of cerebral blood volume ( Tomita

Sawata

Naritomi

Heiss

W-D

, eds), Amsterdam, The Netherlands: Excerpta Medica, pp. 3–31

50.

Tong

Yenari

Albers

O'Brien

Marks

Moseley

(1998) Correlation of perfusion- and diffusion-weighted MRI with NIHSS score in acute (<6.5 hour) ischemic stroke. Neurology 50:864–870

51.

Tyson

Teasdale

Graham

McCulloch

(1984) Focal cerebral ischemia in the rat: topography of hemodynamic and histopathological changes. Ann Neurol 15:559–567

52.

Walsh

Minematsu

Leppo

Moore

(1994) Radioactive microsphere validation of a volume localized continuous saturation perfusion measurement. Magn Reson Med 31:147–153

53.

Warach

Ronthal

Edelman

(1992) Acute cerebral ischemia: evaluation with dynamic contrast-enhanced MR imaging and MR angiography. Radiology 182:41–47

54.

Weisskoff

Zuo

Boxerman

Rosen

(1994) Microscopic susceptibility variation and transverse relaxation: theory and experiment. Magn Reson Med 31:601–610

55.

White

Aicher

Tzika

Kucharczyk

Engelstad

Moseley

(1992) Iron-dextran as a magnetic susceptibility contrast agent: flow-related contrast effects in T2-weighted spin-echo MRI of normal rat and cat brain. Magn Reson Med 24:14–28

56.

Willard-Mack

Koehler

Hirata

Cork

Takahashi

Traystman

Brusilow

(1996) Inhibition of glutamine synthetase reduces ammonia-induced astrocyte swelling in rat. Neuroscience 71:589–599

57.

Zaharchuk

Bogdanov

Jr Marota

JJA

Shimizu-Sasamata

Weisskoff

Kwong

Jenkins

Weissleder

Rosen

(1998) Continuous assessment of perfusion by tagging including volume and water extraction (CAPTIVE): a steady-state contrast agent techniques for measuring blood flow, relative blood volume fraction, and the water extraction fraction. Magn Reson Med 40:666–678

58.

Zaharchuk

Mandeville

Bogdanov

Jr Weissleder

Rosen

Marota

JJA

(1999) Cerebrovascular dynamics of autoregulation and hypoperfusion: an MRI study of CBF and changes in total and microvascular cerebral blood volume during hemorrhagic hypotension Stroke 30:2197–2205