Arterial spin labeling and dynamic susceptibility contrast CBF MRI in postischemic hyperperfusion,hypercapnia,and after mannitol injection

Abstract

Arterial spin labeling (ASL) and dynamic susceptibility contrast (DSC) magnetic resonance imaging (MRI) are widely used to image cerebral blood flow (CBF) in stroke. This study examined how changes in tissue spin-lattice relaxation-time constant (T₁), blood-brain barrier (BBB) permeability, and transit time affect CBF quantification by ASL and DSC in postischemic hyperperfusion in the same animals. In Group I (n = 6), embolic stroke rats imaged 48hours after stroke showed regional hyperperfusion. In normal pixels, ASL- and DSC-CBF linearly correlated pixel-by-pixel. In hyperperfusion pixels, ASL-CBF was significantly higher than DSC-CBF pixel-by-pixel (by 25%). T₁ increased from 1.76 ± 0.14 seconds in normal pixels to 1.93 ± 0.17 seconds in hyperperfusion pixels. Arterial transit time decreased from 300 milliseconds in normal pixels to 200 milliseconds in hyperperfusion pixels. ΔR₂* profiles showed contrast-agent leakages in the hyperperfusion regions. In Group II (n = 3) in which hypercapnic inhalation was used to increase CBF without BBB disruption, CBF increased overall but ASL- and DSC-CBF remained linearly correlated. In Group III (n = 3) in which mannitol was used to break the BBB, ASL-CBF was significantly higher than DSC-CBF. We concluded that in normal tissue, ASL and DSC provide comparable quantitative CBF, whereas in postischemic hyperperfusion, ASL-CBF and DSC-CBF differed significantly because ischemia-induced changes in T₁ and BBB permeability affected the two methods differently.

Keywords

arterial spin labeling arterial transit time cerebral blood flow dynamic susceptibility contrast permeability spin-lattice relaxation-time constant

Introduction

Accurate quantification of cerebral blood flow (CBF) could lead to improved clinical diagnosis of many neurologic diseases, including stroke. Magnetic resonance imaging (MRI) of CBF offers many advantages compared with other techniques. Cerebral blood flow is widely measured using arterial spin labeling (ASL) or dynamic susceptibility contrast (DSC) technique based on MRI. Arterial spin labeling is a completely noninvasive method and repeated measurements can be made to increase signal-to-noise ratio and spatial resolution because of the favorable short half-life (spin-lattice relaxation-time constant T₁) of endogenous water (Chalela et al, 2000; Detre et al, 1992; Kim, 1995; Williams et al, 1992). Dynamic susceptibility contrast imaging is time efficient and provides various cerebral hemodynamic parameters, such as CBF, cerebral blood volume, and arterial transit time (Ostergaard et al, 1996b; Smith et al, 2000). Absolute quantification of both ASL- and DSC-CBF methods, however, remains an active area of research. Specifically, the effects of spin-lattice relaxation-time constant, arterial transit time, and blood-brain barrier (BBB) permeability on ASL- and DSC-CBF quantification (Alsop and Detre, 1996; Donahue et al, 2000; Ostergaard et al, 1996b; Parkes and Tofts, 2002; Smith et al, 2000; Williams et al, 1992; Wu et al, 2003a; Zaharchuk et al, 2010), especially under perturbed conditions, remain poorly understood. Postischemic hyperperfusion is a common phenomenon associated with cerebral infarction. Early postischemic hyperperfusion has been reported to be both beneficial (i.e., prevent infarct growth) and harmful (i.e., aggravate edema and hemorrhage, and neuronal damage from reperfusion injury) (Pan et al, 2007; Schaller and Graf, 2004). Postischemic hyperperfusion in the subacute stage (48 hours after onset) has often been associated with tissue necrosis. The mechanisms of hyperperfusion are unknown and likely multifactorial (Marchal et al, 1999). In addition to many CBF regulating factors that are released during ischemia (Macfarlane et al, 1991), ischemia also affects tissue T₁, T₂, delayed arterial transit time, vascular resistance, and BBB disruption (Schaller and Graf, 2004) that could affect ASL- and DSC-CBF quantification. To improve understanding of postischemic hyperperfusion, it may be helpful to accurately measure CBF and understand how various biophysical and physiological parameters affect CBF under perturbed conditions.

The goal of this study was to examine the underlying factors that might have contributed to the differences in ASL- and DSC-CBF measurements in postischemic hyperperfusion at acute (2 hours after ischemia) and subacute (48 hours) phase in the same animals. Specifically, we examined how changes in tissue spin-lattice relaxation-time constant, BBB permeability, and arterial transit time affect ASL-and DSC-CBF quantification in three experimental groups: (1) postischemic hyperperfusion in embolic stroke rats, (2) normal animals breathing 5% CO₂ to increase CBF by dilating vessels without BBB disruption, and (3) normal animals injected with mannitol to modulate BBB permeability by osmotic shock (Rapoport, 2000). Arterial spin labeling CBF at different postlabeling delays (PLDs) and DSC-CBF, spin-lattice relaxation-rate constant and magnetic resonance angiography (MRA) were measured in the same animals.

Materials and methods

Animal Preparation

Experiments were performed in accordance with a protocol approved by the institutional animal care and use committee. Male Sprague-Dawley rats (Charles River, Wilmington, MA, USA) weighing 200 to 250g were anesthetized with 1.5% isoflurane. They were orally intubated and mechanically ventilated. Rectal temperature was maintained at 37°C to 38°C throughout the experiment using an auto-feedback warm water pad. The left femoral artery and vein were cannulated with PE-50 catheters for continuous monitoring of blood pressure and drug injection, respectively. The inspired O₂ and expired CO₂ concentrations were monitored continuously, and tidal volume was adjusted to maintain an end-tidal CO₂ of 35 to 45 mm Hg throughout the experiments. Blood pressure, heart rate, and respiration rate were also maintained within normal physiologic ranges.

Three groups of experiments were performed as outlined in Figure 1. Magnetic resonance imaging consisted of ASL-CBF, DSC-CBF, and T₁ measurements. Time-of-flight (2D) MRA measurements were made in the stroke group to confirm recanalization.

Figure 1

Schematic drawings of the magnetic resonance imaging (MRI) protocols used in the three experimental groups.

Stroke group (n = 6): Embolic stroke was induced following procedures reported previously (Tanaka et al, 2007). Briefly, the right carotid artery was exposed. Retrograde cannulation was performed on the right external carotid artery with a PE-50 tubing, with its tip placed close to the carotid bifurcation for delivering clots toward the brain. Blood clots, obtained 24 hours before the stroke induction from the same animal, were injected to induce embolic stroke. One hour after clot embolism, animals received intravenous infusion of recombinant tissue plasminogen activator (10 mg/kg; Genetech, South San Francisco, CA, USA) dissolved in 2 mL of distilled water. Ten percent of the solution was given as a bolus, and the rest was given as a constant infusion over 1 hour. Magnetic resonance imaging was performed and the animals were allowed to recover and returned to standard animal housing. The animals were anesthetized again after 48 hours for a second set of MRI scans, and then euthanized. Note that these six rats herein (out of the seven rats) showed partial or total recanalization as confirmed by ASL-CBF map and MRA, and substantial CBF recovery within 1 hour after recombinant tissue plasminogen activator administration. All six rats survived for 48 hours after stroke.

Hypercapnia group (n = 3): Cerebral blood flow measurements were performed on normal rats associated with hypercapnic inhalation, which increased CBF without BBB disruption. As CO₂ inhalation affected global CBF, normalization with respect to contralateral hemisphere was not possible, and thus two sets of CBF measurement were made in the same animals. First, ASL-CBF and DSC-CBF measurements were made during room air. To avoid gadopentetate dimeglumine (Gd-DTPA) effect of the first DSC measurement on the second ASL measurement, we waited 2 hours before initiating the second set of CBF measurements during hypercapnic inhalation (5% CO₂, 21% O₂, and balance N₂). The 2-hour wait time was used to avoid Gd-DTPA effects on ASL contrast and was experimentally determined to be sufficient by monitoring ASL contrast returning to that of baseline (air).

Mannitol group (Group III, n = 3): Normal rats were anesthetized and their right carotid artery was exposed. Retrograde cannulation was performed on the right external carotid artery with a PE-50 tubing, with its tip placed close to the carotid bifurcation for delivering mannitol toward the brain. T₁ MRI measurements were made. Mannitol (0.25 mL 25% mannitol dissolved in phosphate buffer at the rate of 0.5 mL/min) was then infused. Cerebral blood flow measurements were made starting 1 minute after mannitol injection. T₁ MRI measurements were made again (about 20 minutes after mannitol injection). Mannitol was expected to break the BBB for 30 to 45 minutes (Duong et al, 2000).

Magnetic Resonance Imaging Measurements

All MRI was performed on a Bruker 7-T/30-cm scanner with a 40-G/cm gradient insert (ID = 12 cm, 120 milliseconds rise time) (Billerica, MA, USA). The rats were placed into a stereotaxic headset and then onto an animal holder, which consisted of an actively decoupled surface coil (2.3 cm ID) for brain imaging and a butterfly neck coil for ASL (Duong et al, 2000; Shen et al, 2003, 2004).

Arterial spin labeling CBF images were acquired using gradient-echo echo-planar imaging with matrix = 64 × 64, field of view = 2.56 × 2.56 cm², spectral width = 200 kHz, repetition time (TR) = 2.8 seconds (90° flip angle), echo time (TE) = 13.48 milliseconds, three 1.5 mm slices (Shen et al, 2003, 2004). Labeling duration was 2.3 seconds. Post-labeling delays of 100, 200, 300, 400, and 500 milliseconds were used in stroke group, and 100 milliseconds in hypercapnia and mannitol groups. Paired images were acquired alternately one with and the other without, ASL preparation. Sixty pairs of images (total time ∼ 5 minutes) were acquired for signal averaging.

Dynamic susceptibility contrast CBF measurements were performed using identical parameters as the ASL except with TR = 200 milliseconds (90° flip angle) and TE = 13.46-milliseconds. A total of 200 scans were obtained continuously in 40 seconds. A 0.2 mg/kg bolus of Gd-DTPA was injected 15 seconds from the start of the scan via the left femoral vein. Dynamic susceptibility contrast CBF measurements were made after ASL-CBF measurement to avoid contrast agent confounding ASL-CBF measurements.

Two sets of T₁ maps were obtained to investigate the effect of BBB disruption in all groups, one before DSC-CBF (pre-Gd-DTPA) and another 20 minutes after (post-Gd-DTPA). T₁ map was calculated using multiple inversion-recovery echo-planar imaging sequence acquired with identical field of view and imaging matrix as the ASL and DSC images, but with TR = 12,000 milliseconds, TE = 31 milliseconds, and number of average = 1. Twenty-four inversion delays, ranging from 25 to 4,625 milliseconds in 200 milliseconds increments, were used for the calculation.

2D time-of-flight angiography was acquired to identify large vessels using TR = 20 milliseconds, flip angle = 80°, TE = 6.3 milliseconds, slice thickness = 0.9 mm, field of view = 2.56 × 2.56 cm², matrix = 128 × 128, 20 slices, and 4 averages.

Data Analysis

Data analysis was performed using codes written in Matlab (MathWorks, Natick, MA, USA) software. All data were reported as mean ± s.d. Unpaired t-test was used to compare among groups and between normal pixels and hyperperfusion pixels and paired t-test was used for the rest of the analysis with P < 0.05 considered to be statistically significant.

Quantitative ASL-CBF (mL/100g per minute) was calculated pixel-by-pixel using: CBF = λ/T₁(S_C–S_L)/(S_L + (2α–1)S_C), where S_C and S_L are the signal intensities of the control and labeled images, respectively; λ, the water brain-blood partition coefficient, was taken to be 0.9 (Herscovitch and Raichle, 1985); T₁ was the water spin-lattice relaxation time of tissue; and α, the ASL efficiency (Williams et al, 1992), was measured to be 0.75 (Shen et al, 2003, 2004). The signal difference between the control image and labeled image (ΔS/S) was calculated as (S_C–S_L)/S_C.

For DSC-CBF calculation, the transverse relaxation rate (ΔR₂*) was calculated using the equation, ΔR₂*(t) = –ln (S(t)/S_o)/TE, where S(t) is the signal intensity at time t, S_o is the average pre-Gd-DTPA baseline signal intensity, and TE is the echo time. Cerebral blood flow map was then generated by deconvolving the change in tissue concentration of Gd-DTPA over the first pass with an arterial input function using singular value decomposition (Ostergaard et al, 1996a,b).

Comparison of CBF data between the two techniques used normalization with respect to the unaffected hemisphere to reduce intersubject variations, instead of using quantitative CBF values. In the stroke group, ASL-CBF map obtained for each PLD was normalized with respect to the average CBF value of pixels in the middle cerebral artery territory of unaffected left hemisphere (normal pixels) obtained with the PLD of 100 milliseconds. Dynamic susceptibility contrast CBF maps were also normalized by the mean DSC-CBF value of the pixels in the same area. In the hypercapnia group, mean CBF of the pixels in the middle cerebral artery area of the left hemisphere under air breathing status was used as a reference for each rat.

Hyperperfusion pixels in the stroke group were defined as pixels that had both ASL-CBF and DSC-CBF values greater than mean + 2 s.d. of normal pixels. The ASL:DSC ratio map was obtained by dividing the normalized ASL-CBF by the normalized DSC-CBF. T₁ ratio map was generated by dividing the T₁ map obtained before the injection of Gd-DTPA to the T₁ map obtained post-Gd-DTPA. Regions of increased permeability were defined as regions that showed T₁ ratio increase greater than (mean + 2 s.d.) of the normal region, indicating a leakage of Gd-DTPA.

Results

Blood pressures were not statistically different among groups and between before and after mannitol injection (stroke group: 117 ± 8 mm Hg; hypercapnia group: air breathing 125 ± 8 mm Hg, 5% CO₂ inhalation 116 ± 9 mm Hg; mannitol group: before injection 115 ± 12 mm Hg, after injection 133 ± 15 mm Hg), except for the transient increase immediately after mannitol injection. Normalization with contralateral middle cerebral artery territory was used to minimize residual systemic effects of mannitol.

Average CBF values of normal pixels by ASL and DSC were, respectively (1) 84 ± 3 and 93 ± 10 mL/100g per minute (mean ± s.d., n = 6) in the stroke group, (2) 76 ± 14 and 106 ± 13mL/100g per minute (n = 3) in the hypercapnia group, and (3) 93 ± 9 and 109 ± 10 mL/100g per minute (n = 3) in the mannitol group. There were statistically significant differences between ASL and DSC values (paired t-test), a result of systematic errors between the two methods. The ASL data are consistent with those reported previously in our laboratory under essentially identical conditions (Liu et al, 2004; Shen et al, 2003, 2004, 2005; Sicard et al, 2003; Sicard and Duong, 2005).

The mean T₁ values from regions of interest of the cerebral cortex, caudoputamen, corpus callosum, infarct core, and hyperperfusion pixels from the stroke group were 1.76 ± 0.14, 1.74 ± 0.09, 1.35 ± 0.11, 2.27 ± 0.29, and 1.93 ± 0.17 (mean ± s.d., n = 6), respectively. T₁ values of the infarct core and hyperperfusion pixels were statistically different from that of the cortex (P < 0.05). T₁ differences were taken into account in calculating ASL-CBF. The hypercapnia group showed no significant changes in T₁ value, indicating the absence of Gd-DTPA leakage. The mannitol group showed significant T₁ decrease indicating Gd-DTPA leakage (before mannitol 1.76 ± 0.14, after mannitol 1.51 ± 0.24).

Arterial Spin Labeling Versus Dynamic Susceptibility Contrast Cerebral Blood Flow

Figure 2A shows representative ASL-CBF, DSC-CBF, and ASL:DSC ratio maps from each of the three groups. In the stroke group, both ASL- and DSC-CBF maps showed hyperperfusion in the lesion hemisphere. The ASL:DSC CBF ratio was higher in the hyperperfusion pixels than the contralateral normal hemisphere. In the hypercapnia group, both the ASL-and DSC-CBF maps showed CBF increases in both hemispheres but the ASL:DSC CBF ratio map did not increase relative to baseline. In the mannitol group, both ASL- and DSC-CBF maps showed CBF increase predominantly in the hemisphere to which mannitol was injected. Similar to the stroke group, the mannitol group also showed a higher ASL-CBF compared with DSC-CBF, resulting in increased ASL:DSC CBF ratio.

Figure 2

(A) Arterial spin labeling (ASL) cerebral blood flow (CBF) image, dynamic susceptibility contrast (DSC) CBF image, and ASL:DSC ratio maps from one animal of each of the three experimental groups. In the stroke animal (top row), the stroke lesion shows hyperperfusion. ASL yields a higher CBF than DSC. In the hypercapnia animal (middle row), CBF increases globally. ASL- and DSC-CBF maps show similar increases. In normal animal injected with mannitol (bottom row), CBF increases in the affected hemisphere. ASL yields a higher CBF increase than DSC. (B) Normalized ASL versus DSC CBF scatterplots for all animals in each of the three experimental groups. The slopes of the normal hemispheres in all three groups were close to unity. The slopes of the stroke, hypercapnia, and mannitol group were 1.21, 0.86, and 1.46, respectively.

Figure 2B shows the scatterplots of normalized ASL-CBF versus DSC-CBF group for a PLD of 100 milliseconds for each of the three groups. In the normal hemisphere or conditions, regression analysis showed strong correlations and slopes close to unity for all three groups (stroke group: y = 0.96x, R² = 0.44; hypercapnia group: y = 0.97x, R² = 0.37; and mannitol group: y = 0.96x, R² = 0.38). In contrast, the stroke and the mannitol-injected hemisphere showed greater-than-unity slopes (lesion hemisphere: y = 1.21x, R² = 0.57%, 26% increase; mannitol-affected hemisphere: y = 1.46x, R² = 0.43%, 52% increase), whereas the hypercapnia group showed a less-than-unity slope (y = 0.86x, R² = 0.42%, 11% decrease).

Cerebral Blood Flow Values and Gd-DTPA Leakage in Postischemic Hyperperfusion

Figure 3A shows the T₁ maps before and after Gd-DTPA injection, the subtraction image (T_1post-Gd–T_1pre-Gd), and the ASL-CBF map of a representative rat from the stroke group. Hyperintensity on the subtraction image corresponded to the area with Gd-DTPA leakage and matched well with the area of higher ASL-CBF. Scatterplots between the normalized ASL-CBF (PLD = 100 milliseconds) and DSC-CBF from all rats in the stroke group 48 hours after cerebral ischemia are shown in Figure 3B. The majority (748/962 pixels, 78%) of the hyperperfusion pixels showed increased T₁ ratio (blue dots). Moreover, the hyperperfusion pixels had higher ASL-CBF than DSC-CBF. Figure 3C shows the group-averaged ASL:DSC CBF ratio from normal and hyperperfusion pixels from the stroke group obtained at 2 and 48 hours after stroke (PLD = 100 milliseconds). The ASL:DSC CBF ratio increased significantly in the hyperperfusion area than that in the normal hemisphere at each time points. Similar conclusion was reached for the T₁ ratio data as depicted in Figure 3D.

Figure 3

(A) T₁ maps before and after Gd-DTPA intravenous injection, subtraction image, and arterial spin labeling cerebral blood flow (ASL-CBF) map of a representative stroke rat. Hyperintensity on the subtraction image indicates Gd-DTPA leakage, which corresponds well to the area of increased CBF. (B) Normalized ASL (postlabeling delay (PLD) = 100) versus dynamic susceptibility contrast (DSC) CBF scatterplots of stroke group obtained at 48 hours after stroke. Pixels with ratio of T₁ pre- to post-Gd-DTPA greater than mean T₁ + 2 s.d. of contralateral normal pixels are displayed as blue dots. Normal T₁ pixels of the stroke hemisphere are shown as red dots. (C) Group-averaged ASL:DSC CBF ratios of the normal and the hyperperfusion pixels at 2 and 48 hours after stroke onset (PLD = 100). The ratio was larger in the hyperperfusion territory. (D) Group-averaged ratios of T₁ before and after Gd-DTPA from the normal and the hyperperfusion pixels at 2 and 48 hours after stroke. The ratio was also larger in the hyperperfusion territory.

To avoid Gd-DTPA effect of the first DSC measurement, the second ASL measurement was made 2 hours after the first. Arterial spin labeling was measured and confirmed ΔS/S fully recovered to baseline 90 minutes after Gd-DTPA injection in each rat.

ΔR₂* Time Courses of Dynamic Susceptibility Contrast Magnetic Resonance Imaging

Figure 4 shows representative normalized ΔR₂* time courses. ΔR₂* trace of the normal pixels returned close to baseline after Gd-DTPA injection (i.e., time points > 25seconds, normalized ΔR₂* between 25 and 30 seconds = 0.06 ± 0.003). Similarly, pixels from the hypercapnia group also showed the ΔR₂* trace returned close to baseline after Gd-DTPA injection (normalized ΔR₂* between 25 and 30seconds = 0.07 ± 0.005). In marked contrast, ΔR₂* trace of the hyperperfusion pixels in the stroke group remained significantly elevated from the baseline by comparison (normalized ΔR₂* between 25 and 30seconds = 0.16 ± 0.008). In the mannitol group, ΔR₂* trace of the pixels with T₁:T₁ with Gd-DTPA ratio larger than mean + 2 s.d. of normal values were slightly elevated compared with normal (normalized ΔR₂* between 25 and 30seconds = 0.09 ± 0.006). The timing of the first passes of the three groups differed, with the hyperperfusion pixels in the stroke group showed the most delayed first pass.

Figure 4

Normalized ΔR₂* time courses for three experimental groups. The traces from the normal and hypercapnia group returned close to baseline after bolus gadopentetate dimeglumine (Gd-DTPA) injection. The mannitol group shows a slight elevation post-Gd-DTPA. In contrast, trace from the hyperperfusion pixels remains significantly elevated above baseline.

Arterial Spin Labeling Cerebral Blood Flow Versus Arterial Transit Time in Postischemic Hyperperfusion

Figure 5A shows the representative MRA of a stroke rat imaged at different time points. The MRA acquired at 2 hours after recombinant tissue plasminogen activator injection revealed recanalization of the right middle cerebral artery (arrows). The MRA at 48hours after stroke also showed many dilated vessels in the affected hemisphere (arrow heads). The ASL:DSC CBF ratio from hyperperfusion and normal pixels were analyzed for multiple PLDs (Figure 5B). The ASL:DSC CBF ratio decreased significantly with increasing PLD in both normal and hyperperfusion pixels. However, the ratio of the hyperperfusion pixels dropped to a larger extent than that of the normal pixels (P < 0.05). The relationship between ΔS/S and PLD of the normal pixels and the hyperperfusion pixels is depicted in Figure 5C. The ΔS/S reached a maximum at PLD of 200 milliseconds in hyperperfusion pixels and 300 milliseconds in normal pixels.

Figure 5

(A) Magnetic resonance angiography (MRA) of a stroke rat imaged before, 2 and 48hours after recanalization. Recanalization was detected at 2hours (arrows). Additional vessels were highly perfused at 48hours (arrow heads). (B) Group-averaged arterial spin labeling to dynamic susceptibility contrast (ASL:DSC) cerebral blood flow (CBF) ratio of the hyperperfusion pixels and contralateral normal pixels of middle cerebral artery area at postlabeling delays (PLDs) of 100 and 500 milliseconds. ASL:DSC ratio decreases with increasing PLD. (C) ΔS/S versus PLD in the normal and hyperperfusion area. The ΔS/S of the hyperperfusion pixels peaked at PLD of 200 milliseconds and the ΔS/S of normal pixels peaked at PLD of 300 milliseconds. *P < 0.05.

Discussion

This study examined how changes in tissue spin-lattice relaxation-time constant, BBB permeability, and arterial transit time affect CBF quantification by ASL and DSC in postischemic hyperperfusion in same rats. Embolic stroke rats imaged 48hours after reperfusion showed reliable regional hyperperfusion. Arterial spin labeling and DSC-CBF of normal pixels linearly correlated, whereas ASL-CBF of hyperperfusion pixels were higher than DSC-CBF. T₁ of hyperperfusion pixels were higher, transit time was shortened, and ΔR₂* time courses showed Gd-DTPA leakages in hyperperfusion regions. Hypercapnic inhalation, which does not change BBB permeability, showed overall CBF increase but ASL- and DSC-CBF remain linearly correlated. Mannitol injection, which increases BBB permeability, showed ASL-CBF to be higher than DSC-CBF. We concluded that (1) under normal conditions the commonly used ASL and DSC provide comparable quantitative CBF values and (2) in ischemic hyperperfusion, T₁ and BBB disruption were responsible for discrepancy in CBF measured by ASL and DSC. These findings could have important implications in stroke MRI.

Effect of Permeability Change to Arterial Spin Labeling and Dynamic Susceptibility Contrast Measurements

Arterial spin labeling and DSC-CBF values were comparable in normal brain, in good agreement with previous reports (Lia et al, 2000; Weber et al, 2003). The ASL-CBF and DSC-CBF were higher in the hyperperfusion pixels in all three experimental groups. The ASL/DSC ratios of the hyperperfusion pixels were higher in the stroke group and mannitol group, but not in the hypercapnia group. The majority of hyperperfusion pixels in the stroke group showed Gd-DTPA leakage, whereas those in the hypercapnia group did not show significant T₁ change. A possible explanation of these observations is that BBB permeability differed among these groups. Water is not freely diffusible across the BBB (Herscovitch et al, 1987; Schwarzbauer et al, 1997; Silva et al, 1997). However, most published ASL-CBF calculations (including this study) assumed labeled water in the blood is freely diffusible endogenous tracer (Detre et al, 1992; Kim, 1995; Kwong et al, 1992). In postischemic hyperperfusion, BBB is disrupted by ischemia and water in the blood can extravasate more freely, resulting in CBF overestimation by the ASL approach herein. By contrast, CBF is likely underestimated in hypercapnia group because the water extraction fraction decreases with increased CBF in intact BBB (Silva et al, 1997).

Similarly, DSC MRI is also confounded by the compromised BBB in hyperperfusion regions. Leaked Gd-DTPA could shorten both T₁ and T₂* values of tissue. Whether DSC-CBF is underestimated or overestimated depends on whether T₁ or T₂* effects by leaked Gd-DTPA dominated. T₁ increase of accumulated Gd-DTPA in the interstitial space overwhelmed T₂* decrease when short TR spin echo echo-planar imaging was used (Lim et al, 2003). Our results showed that ΔR₂* of most of the hyperperfusion pixels remained higher than that in normal tissue after bolus passage, suggesting that T₂* effect was dominant. Although T₁ shortening effect could be minimized by using a double-echo technique (Heiland et al, 1999; Uematsu et al, 2000), T₂* effect by leaked Gd-DTPA cannot be distinguished from DSC-CBF effect.

It is interesting to note that some hyperperfusion pixels did not show significant Gd-DTPA leakage. It may be because of the difference between the permeability of water and Gd-DTPA. It may also be because of the sensitivity of T₁ measurement per se and measuring the time course of the leakage may be more sensitive (Ewing et al, 2003; Jiang et al, 2005; Tofts and Kermode, 1991).

Effect of T1 and Arterial Transit Time Change on Arterial Spin Labeling and Dynamic Susceptibility Contrast Measurement

T₁ value of hyperperfusion region was 1.93 seconds, compared with that of normal cortex of 1.76 seconds. The T₁ increase resulted in 8% decrease in ASL-CBF value. Although DSC-CBF quantification is not significantly affected by T₁ difference in a typical measurement, tissue T₂* changes due to Gd-DTPA leakage contributed to CBF accuracy (Ostergaard et al, 1996b).

In the hyperperfusion area, the arterial transit time was reduced. Thus, CBF accuracy with a PLD typically used for normal brain was not an issue for the hyperperfusion pixel. Although it was not investigated herein, ASL- and DSC-CBF accuracies are also affected in region of perfusion deficit associated with stroke. Thomas et al (2006) reported that longer PLD (e.g., 800 to 1,000 milliseconds) is required for spin echo CASL sequence to maintain CBF analysis insensitive to transit time because of the transit time delay. Delayed arterial transit time also affects DSC-CBF quantification, especially in ischemic hypoperfused area, resulting in underestimation of CBF (Calamante et al, 2000; Ostergaard et al, 1996b). Fourier-based or delay invariant block-circulant singular value decomposition methods have been proposed to minimize this effect (Ostergaard et al, 1996b; Smith et al, 2000; Wu et al, 2003b). Another problem for DSC-CBF quantification is dispersion (Calamante et al, 2000). Increasing dispersion underestimates DSC-CBF because DSC model cannot distinguish feeding vessels dispersion from tissue microvasculature dispersion. Our data show that DSC-CBF values in each group were affected by dispersion differently. Although correction for dispersion has been proposed (Ostergaard et al, 1999; Willats et al, 2008), dispersion correction is less straightforward.

Conclusion

Under normal conditions, the commonly used ASL and DSC approach provide comparable quantitative values. In ischemic hyperperfusion, changes in T₁ and BBB disruption that result in increased water permeability and Gd-DTPA leakage are likely responsible for discrepancy between CBF measured by ASL and DSC technique. The ASL model overestimates CBF due to increased permeability in postischemic hyperperfusion and DSC overestimates CBF due to ΔR₂* effects from Gd-DTPA leakage. Arterial spin labeling CBF is higher than DSC-CBF by ∼25% in the hyperperfusion pixels. Caution must be exercised when using various CBF methods in stroke imaging. Future studies will incorporate T₁, arterial transit time, and permeability into the ASL and DSC analysis, and compare with positron emission tomography in the same animals.

Footnotes

The authors declare no conflict of interest.

References

Alsop

Detre

(1996) Reduced transit-time sensitivity in noninvasive magnetic resonance imaging of human cerebral blood flow. J Cereb Blood Flow Metab 16: 1236–49.

Calamante

Gadian

Connelly

(2000) Delay and dispersion effects in dynamic susceptibility contrast MRI: simulations using singular value decomposition. Magn Reson Med 44:466–73.

Chalela

Alsop

Gonzalez-Atavales

Maldjian

Kasner

Detre

(2000) Magnetic resonance perfusion imaging in acute ischemic stroke using continuous arterial spin labeling. Stroke 31:680–7.

Detre

Leigh

Williams

Koretsky

(1992) Perfusion imaging. Magn Reson Med 23:37–45.

Donahue

Krouwer

Rand

Pathak

Marszalkowski

Censky

Prost

(2000) Utility of simultaneously acquired gradient-echo and spin-echo cerebral blood volume and morphology maps in brain tumor patients. Magn Reson Med 43:845–53.

Duong

Silva

Lee

Kim

(2000) Functional MRI of calcium-dependent synaptic activity: cross correlation with CBF and BOLD measurements. Magn Reson Med 43:383–92.

Ewing

Knight

Nagaraja

Yee

Nagesh

Whitton

Fenstermacher

(2003) Patlak plots of Gd-DTPA MRI data yield blood-brain transfer constants concordant with those of 14C-sucrose in areas of blood-brain opening. Magn Reson Med 50:283–92.

Heiland

Benner

Debus

Rempp

Reith

Sartor

(1999) Simultaneous assessment of cerebral hemodynamics and contrast agent uptake in lesions with disrupted blood-brain-barrier. Magn Reson Imaging 17:21–7.

Herscovitch

Raichle

(1985) What is the correct value for the brain–blood partition coefficient for water? J Cereb Blood Flow Metab 5:65–9.

10.

Herscovitch

Raichle

Kilbourn

Welch

(1987) Positron emission tomographic measurement of cerebral blood flow and permeability-surface area product of water using [15O]water and [11C]butanol. J Cereb Blood Flow Metab 7:527–42.

11.

Jiang

Ewing

Ding

Zhang

Whitton

Knight

Chopp

(2005) Quantitative evaluation of BBB permeability after embolic stroke in rat using MRI. J Cereb Blood Flow Metab 25:583–92.

12.

Kim

(1995) Quantification of relative cerebral blood flow change by flow-sensitive alternating inversion recovery (FAIR) technique: application to functional mapping. Magn Reson Med 34:293–301.

13.

Kwong

Belliveau

Chesler

Goldberg

Weisskoff

Poncelet

Kennedy

Hoppel

Cohen

Turner

(1992) Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. Proc Natl Acad Sci USA 89:5675–9.

14.

Lia

Guang Chen

Ostergaard

Hindmarsh

Moseley

(2000) Quantification of cerebral blood flow by bolus tracking and artery spin tagging methods. Mag Reson Imaging 18:503–12.

15.

Lim

Roberts

Sitoh

Hui

(2003) Rising signal intensity observed in extra-axial brain tumours—a potential pitfall in perfusion MR imaging. Singapore Med J 44:526–30.

16.

Liu

Schmidt

Sicard

Duong

(2004) Imaging oxygen consumption in forepaw somatosensory stimulation in rats under isoflurane anesthesia. Magn Reson Med 52:277–85.

17.

Macfarlane

Moskowitz

Sakas

Tasdemiroglu

Wei

Kontos

(1991) The role of Neuroeffector mechanisms in cerebral hyperperfusion syndromes. J Neurosurg 75:845–55.

18.

Marchal

Young

Baron

(1999) Early postischemic hyperperfusion: pathophysiologic insights from positron emission tomography. J Cereb Blood Flow Metab 19:467–82.

19.

Ostergaard

Chesler

Weisskoff

Sorensen

Rosen

(1999) Modeling cerebral blood flow and flow heterogeneity from magnetic resonance residue data. J Cereb Blood Flow Metab 19:690–9.

20.

Ostergaard

Sorensen

Kwong

Weisskoff

Gyldensted

Rosen

(1996a) High resolution measurement of cerebral blood flow using intravascular tracer bolus passages. Part II: Experimental comparison and preliminary results. Magn Reson Med 36:726–36.

21.

Ostergaard

Weisskoff

Chesler

Gyldensted

Rosen

(1996b) High resolution measurement of cerebral blood flow using intravascular tracer bolus passages. Part I: mathematical approach and statistical analysis. Magn Reson Med 36:715–25.

22.

Pan

Konstas

Bateman

Ortolano

Pile-Spellman

(2007) Reperfusion injury following cerebral ischemia: pathophysiology, MR imaging, and potential therapies. Neuroradiology 49:93–102.

23.

Parkes

Tofts

(2002) Improved accuracy of human cerebral blood perfusion measurements using arterial spin labeling: accounting for capillary water permeability. Magn Reson Med 48:27–41.

24.

Rapoport

(2000) Osmotic opening of the blood–brain barrier: principles, mechanism, and therapeutic applications. Cell Mol Neurobiol 20:217–30.

25.

Schaller

Graf

(2004) Cerebral ischemia and reperfusion: the pathophysiologic concept as a basis for clinical therapy. J Cereb Blood Flow Metab 24:351–71.

26.

Schwarzbauer

Morrissey

Deichmann

Hillenbrand

Syha

Adolf

Noth

Haase

(1997) Quantitative magnetic resonance imaging of capillary water permeability and regional blood volume with an intravascular MR contrast agent. Magn Reson Med 37:769–77.

27.

Shen

Fisher

Sotak

Duong

(2004) Effects of reperfusion on ADC and CBF pixel-by-pixel dynamics in stroke: characterizing tissue fates using quantitative diffusion and perfusion imaging. J Cereb Blood Flow Metab 24:280–90.

28.

Shen

Meng

Fisher

Sotak

Duong

(2003) Pixel-by-pixel spatiotemporal progression of focal ischemia derived using quantitative perfusion and diffusion imaging. J Cereb Blood Flow Metab 23:1479–88.

29.

Shen

Ren

Cheng

Fisher

Duong

(2005) Functional, perfusion and diffusion MRI of acute focal ischemic brain injury. J Cereb Blood Flow Metab 25:1265–79.

30.

Sicard

Shen

Brevard

Sullivan

Ferris

King

Duong

(2003) Regional cerebral blood flow and BOLD responses in conscious and anesthetized rats under basal and hypercapnic conditions: implications for functional MRI studies. J Cereb Blood Flow Metab 23:472–81.

31.

Sicard

Duong

(2005) Effects of hypoxia, hyperoxia and hypercapnia on baseline and stimulus-evoked BOLD, CBF and CMRO2 in spontaneously breathing animals. Neuroimage 25:850–8.

32.

Silva

Zhang

Williams

Koretsky

(1997) Estimation of water extraction fractions in rat brain using magnetic resonance measurement of perfusion with arterial spin labeling. Magn Reson Med 37:58–68.

33.

Smith

Grandin

Duprez

Mataigne

Cosnard

(2000) Whole brain quantitative CBF and CBV measurements using MRI bolus tracking: comparison of methodologies. Magn Reson Med 43:559–64.

34.

Tanaka

Ishii

Hiraoka

Miyasaka

Kuroiwa

Hajjar

Nagaoka

Duong

Ohno

Yoshida

(2007) Efficacy of recombinant annexin 2 for fibrinolytic therapy in a rat embolic stroke model: a magnetic resonance imaging study. Brain Res 1165:135–43.

35.

Thomas

Lythgoe

Van der Weerd

Ordidge

Gadian

(2006) Regional variation of cerebral blood flow and arterial transit time in the normal and hypoperfused rat brain measured using continuous arterial spin labeling MRI. J Cereb Blood Flow Metab 26:274–82.

36.

Tofts

Kermode

(1991) Measurement of the blood–brain barrier permeability and leakage space using dynamic MR imaging. 1. Fundamental concepts. Magn Reson Med 17:357–67.

37.

Uematsu

Maeda

Sadato

Matsuda

Ishimori

Koshimoto

Yamada

Kimura

Kawamura

Matsuda

Hayashi

Yonekura

Ishii

(2000) Vascular permeability: quantitative measurement with double-echo dynamic MR imaging—theory and clinical application. Radiology 214:912–7.

38.

Weber

Gunther

Lichy

Delorme

Bongers

Thilmann

Essig

Zuna

Schad

Debus

Schlemmer

(2003) Comparison of arterial spin-labeling techniques and dynamic susceptibility-weighted contrast-enhanced MRI in perfusion imaging of normal brain tissue. Invest Radiol 38:712–8.

39.

Willats

Connelly

Calamante

(2008) Minimising the effects of bolus dispersion in bolus-tracking MRI. NMR Biomed 21:1126–37.

40.

Williams

Detre

Leigh

Koretsky

(1992) Magnetic resonance imaging of perfusion using spin inversion of arterial water. Proc Natl Acad Sci USA 89:212–6.

41.

Ostergaard

Koroshetz

Schwamm

O'Donnell

Schaefer

Rosen

Weisskoff

Sorensen

(2003a) Effects of tracer arrival time on flow estimates in MR perfusion-weighted imaging. Magn Reson Med 50:856–64.

42.

Ostergaard

Weisskoff

Benner

Rosen

Sorensen

(2003b) Tracer arrival timing-insensitive technique for estimating flow in MR perfusion-weighted imaging using singular value decomposition with a block-circulant deconvolution matrix. Magn Reson Med 50:164–74.

43.

Zaharchuk

Straka

Marks

Albers

Moseley

Bammer

(2010) Combined arterial spin label and dynamic susceptibility contrast measurement of cerebral blood flow. Magn Reson Med 63:1548–56.

Arterial spin labeling and dynamic susceptibility contrast CBF MRI in postischemic hyperperfusion,hypercapnia,and after mannitol injection

Abstract

Keywords

Introduction

Materials and methods

Animal Preparation

Magnetic Resonance Imaging Measurements

Data Analysis

Results

Arterial Spin Labeling Versus Dynamic Susceptibility Contrast Cerebral Blood Flow

Cerebral Blood Flow Values and Gd-DTPA Leakage in Postischemic Hyperperfusion

ΔR2* Time Courses of Dynamic Susceptibility Contrast Magnetic Resonance Imaging

Arterial Spin Labeling Cerebral Blood Flow Versus Arterial Transit Time in Postischemic Hyperperfusion

Discussion

Effect of Permeability Change to Arterial Spin Labeling and Dynamic Susceptibility Contrast Measurements

Effect of T1 and Arterial Transit Time Change on Arterial Spin Labeling and Dynamic Susceptibility Contrast Measurement

Conclusion

Footnotes

References

ΔR₂* Time Courses of Dynamic Susceptibility Contrast Magnetic Resonance Imaging