- Email: [email protected]
Arterial spin tagging perfusion imaging of rat brain
Magnetic Resonance Imaging 18 (2000) 1109 –1113
Arterial spin tagging perfusion imaging of rat brain Dependency on magnetic field strength C. Franke,* F. A. van Dorsten, L. Olah, W. Schwindt, M. Hoehn Max-Planck-Institute for Neurological Research, Department of Experimental Neurology, Cologne, Germany Received 18 July 2000; accepted 2 September 2000
Abstract Perfusion-weighted imaging (PWI), using the method of arterial spin tagging, is strongly T1-dependent. This translates into a high field dependency of the perfusion signal intensity. In order to determine the expected signal improvement at higher magnetic fields we compared perfusion-weighted images in rat brain at 4.7 T and 7 T. Application of PWI to focal ischemia and functional activation of the brain and the use of two different anesthetics allowed the observation of a wide range of flow values. For all these (patho-)physiological conditions switching from 4.7 T to 7 T resulted in a significant increase of mean perfusion signal intensity by a factor of 2.96. The ratio of signal intensities of homotopic regions in the ipsi- and contralateral hemisphere was field-independent. The relative contribution of a) T1 relaxation time, b) net magnetization, c) the Q-value of the receiver coils and d) the degree of adiabatic inversion to the signal improvement at higher field strength were discussed. It was shown that the main parameters contributing to the higher signal intensity are the lengthening of T1 and the higher magnetization at the higher magnetic field. Keywords: Perfusion-weighted imaging; Arterial spin tagging; Magnetic field strength; Rat brain
1. Introduction The measurement of brain perfusion is important for the assessment of tissue viability. Several methods are available for measuring cerebral perfusion and related hemodynamic parameters using magnetic resonance imaging (MRI). One type of sequences utilizes magnetically labeled blood spins as a noninvasive diffusible tracer for blood flow measurements [1]. Blood flowing to the brain is saturated in the neck region with a slice-selective saturation preparation followed by the imaging sequence after a short delay. The magnetically labeled tracer has a decay rate of T1, which must be sufficiently long to allow the detection of tissue perfusion in the imaging plane. Using this method the regional concentration of saturated spins is defined by the regional blood flow and the regional relaxation time T1 [2,3]. This T1 dependence results in a high sensitivity of the perfusion signal intensity on magnetic field strength. Therefore, we investigated whether switching from 4.7 T to the higher magnetic field * Corresponding author. Tel.: ⫹49 (69) 305 27258; fax: ⫹49 (69) 308 236. E-mail address: [email protected] (C. Franke).
strength of 7 T leads to an increase of the signal-to-noiseratio (SNR) in perfusion-weighted images (PWI) using arterial spin labeling. PWI based on arterial spin labeling is successfully used to describe changes in local cerebral blood flow during early focal cerebral ischemia [4,5]. A rapid decrease of signal intensity can be detected in the ischemic hemisphere reflecting the area of perfusion deficit. Switching from halothane to ␣-chloralose anesthesia results in a reduction of cerebral blood flow (CBF) by approximately 50 –70%, as shown by Lindauer et al. [6]. This anesthetic-induced reduction of CBF leads to a drop in signal intensity of the perfusion-weighted image. On the other hand, under ␣-chloralose anesthesia it is possible to study the increased local cerebral blood flow with PWI during functional brain activation [7,8]. In the present investigation we compare the PWI sensitivity at two different magnetic field strengths, 4.7 T and 7 T, using the above-described animal models and physiological conditions to induce local or global changes in perfusion signal intensity. We demonstrate that switching to higher fields leads to a significant increase of perfusion signal intensity.
0730-725X/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 0 - 7 2 5 X ( 0 0 ) 0 0 2 1 1 - 3
1110
C. Franke et al. / Magnetic Resonance Imaging 18 (2000) 1109 –1113
2. Materials and methods 2.1. Animal preparation All animal handling and surgery was performed in accordance with animal protection guidelines and approved by local authorities. Male Wistar rats (n ⫽ 8; body weight: 350 – 400 g) were anesthetized with 1.5% halothane in a 70%/30% mixture of N2O/O2. The animals were tracheotomized, immobilized with pancuronium bromide (0.3 mg/kg/h intravenously) and mechanically ventilated for the whole duration of the experiment. Catheters were inserted into the femoral artery to monitor blood pressure and to obtain arterial blood samples for blood gas analysis. Body temperature, measured with a rectal thermometer, was maintained at 37°C using a feedback-controlled warm water blanket. 2.2. Functional activation Halothane was discontinued 45 min prior to electrical forepaw stimulation and replaced by i.v. anesthesia with ␣-chloralose (80 mg/kg). Supplemental doses (40 mg/kg) were given at 90-min intervals. For electrical stimulation, two small needle-electrodes were inserted into the skin of both forepaws. Rectangular pulses of 0.3 ms duration and 0.5 mA were applied at 3 Hz for 1 min using an electrical stimulator (FMI, Germany).
using a previously described method [7]. Briefly, an arterial spin tagging technique [3] was used in combination with snapshot FLASH imaging. Measurement parameters were: echo time TE ⫽ 3.5 ms at 4.7 T and 4.4 ms at 7 T, repetition time TR ⫽ 7.4 ms (4.7 T) and 7.6 ms (7 T), field of view ⫽ 40 mm, slice thickness ⫽ 2 mm, matrix ⫽ 128 ⫻ 64, averages ⫽ 8, total scan time ⫽ 1.3 min. The sequence consisted of two similar image acquisition phases separated by a recovery period. During the first phase, blood flowing through the neck towards the brain was adiabatically inverted for three seconds, in the second phase inflowing spins were left undisturbed. The time delay between labeling and imaging was 40 ms. Perfusion-weighted images were obtained by subtraction of the acquisitions with and without prior adiabatic spin inversion. After positioning of the animal in the magnet, pilot scans in the sagittal plane were performed for correct positioning of the imaging plane for perfusion recording. The imaging plane for activation studies was 4.5 mm posterior to the rhinal fissure. In all other studies the coronal plane 5.9 mm posterior to the rhinal fissure was placed in the isocenter, thus focusing on the center of the ischemic territory in the MCAO model. Perfusion-weighted images were performed under control conditions in each animal, corresponding with PWI measurements under halothane anesthesia before MCA occlusion and under both anesthetics. Furthermore, perfusionweighted images were recorded 30 min after MCA occlusion and during functional activation.
2.3. Focal cerebral ischemia 2.5. Data processing and statistics Focal ischemia was produced by suture occlusion of the right middle cerebral artery using a previously described remotely controlled occlusion device [9]. Briefly, a monofilament nylon thread (4-0 Prolene, Ethicon Co., Nordstedt, Germany) with its distal end thickened to 0.28 – 0.30 mm in diameter was connected to an extension catheter and passed through a guide sheath fixed to the neck of the animal. The right common carotid artery was ligated and the filament was introduced into the right internal carotid artery via the base of the skull. 2.4. NMR experiments MR experiments were performed on Bruker Biospec systems (Bruker Medical, Ettlingen, Germany) operating at 4.7 T (MSL X-11) and 7 T (DBX), respectively. Both magnets are equipped with a 30 cm bore and actively shielded gradient coils (rise time ⬍250 s; 100 and 200 mT/m, respectively). At both systems a 12 cm Helmholtz coil was used for rf-transmission; signal detection followed via an inductively coupled surface coil placed over the head of the animal (16 and 24 mm diameter, respectively). The rf-coils were decoupled from each other, the transmitter coil actively, and the receiver coil passively. Single slice perfusion-weighted images were recorded
Perfusion-weighted images were transferred to a Macintosh Power PC 7200/66 (Apple, Cupertino, CA, USA) and image analysis was carried out using the image processing software IMAGE (NIH, Bethesda, MD, USA). After interpolation of images onto a 128 ⫻ 128 matrix SNRs were determined. In all animals the following regions of interest (ROI) were defined for SNR determination: one was placed in the area of the somatosensory cortex, another included the MCA territory. In all animals homotopic regions of the contralateral hemisphere and regions outside the brain were investigated, too. In order to define signal improvement after switching to the higher magnetic field strength the contrast to noise ratio was determined as the ratio between the signal of the perfusion image within an ROI inside the brain and the standard deviation of the background of the untagged image corrected for the different noise statistics proposed by Weisskoff [10]. ROIs were located in the central part of the basal ganglia of the brain (MCA territory). For the functional investigations the position was in the somatosensory cortex. To compare signal changes after MCA occlusion and during functional activation with the corresponding control situation we calculated the relative change of signal intensity between ipsi- and contralateral hemisphere and ex-
C. Franke et al. / Magnetic Resonance Imaging 18 (2000) 1109 –1113 Table 1. Signal-to-noise-ratio and contrast for different anesthesia, after occlusion of the MCA, and during functional activation as a function of magnetic field strength Signal-to noise-ratio
7T
4.7 T
Ratio
Halothane ␣-chloralose
30.89 ⫾ 6.67** 9.22 ⫾ 0.38
10.03 ⫾ 4.37 3.25 ⫾ 2.2
3.08 2.84
57 ⫾ 21 42 ⫾ 15
1.12 1.08
Signal change in percent of control Functional activation 77 ⫾ 12 MCA occlusion 46 ⫾ 5 * p ⬍ .005; ** p ⬍ 0.005.
pressed as percent of the signal intensity of the contralateral hemisphere. All values are given as mean ⫾ standard deviation (SD). Statistical analysis was performed between both field strengths using Student’s t-test (unpaired, unequal variances). p ⬍ 0.05 was accepted as significant. Symbols: *: p ⬍ 0.05; **: p ⬍ 0.005. 2.6. Determination of T1 and Q-values In order to determine the contribution of several factors to the signal-to-noise-ratio at both field strengths the T1 relaxation time at 7 T was determined using a multislice inversion recovery sequence (TR ⫽ 8000 ms, TE ⫽ 20 ms, ten varying inversion delays: 15–3000 ms). T1 was determined by monoexponential fitting. The Q-values of the receiver coils (200 and 300 MHz) were determined using a Network Analyzer (Hewlett Packard, USA) under loaded conditions.
1111
At 7 T we observed a T1 relaxation time of the brain tissue of 1382 ⫾ 96 ms, given as mean value averaged over the whole hemisphere of a healthy rat brain. The Q-values of the receiver coils were determined as 62 and 55 for 200 and 300 MHz, respectively. The systematic error of these measurements was estimated to approximately 8%. 3.2. Observations under different anesthetics Under halothane anesthesia, perfusion signal intensity was higher at 7 T than at 4.7 T by a factor of 3.08 (Table 1, Fig. 1 A, B). When changing from halothane to ␣-chloralose, the PWI signal intensity at 4.7 T was decreased by 70% throughout the brain due to a decrease of baseline cerebral blood flow (CBF) by the anesthetic [6]. This resulted in perfusion images with signal intensity remaining mostly in the cortical areas (Fig. 2 A). At 7 T a similar amount of signal decrease (68%) was observed when switching from halothane to ␣-chloralose. Due to the overall higher PWI signal intensity (factor 2.84) at 7 T, the brain hemispheres remained completely visible under ␣-chloralose, showing also signal intensity in the subcortical regions (Fig. 2 B). 3.3. Electrical forepaw stimulation Stimulation of the left forepaw caused a circumscribed increase in perfusion signal intensity in the right somatosensory cortex relative to the contralateral hemisphere, as shown in Fig. 2 A and B. Activation induced perfusion signal increase was determined and found to be independent of field strength: 57 ⫾ 21% at 4.7 T and 77 ⫾ 12% at 7 T (p ⫽ 0.3).
3. Results 3.4. MCA occlusion 3.1. General observations All physiological parameters of the animals were within normal range during the whole observation periods.
At 4.7 T, 30 min after occlusion of the right middle cerebral artery, a large area with reduced perfusion signal intensity was detectable, showing the perfusion deficit in the
Fig. 1. Perfusion-weighted images under halothane anesthesia at 4.7 T (A) and 7 T (B). Increasing the magnetic field strength from 4.7 T to 7 T resulted in an increase of perfusion-weighted signal-to-noise by a factor of 3.08. After MCA occlusion a large decrease of the perfusion signal intensity is detectable showing the perfusion deficit in the ischemic (right) hemisphere. The ischemia-related, relative signal change is similar at 4.7 T and 7 T.
1112
C. Franke et al. / Magnetic Resonance Imaging 18 (2000) 1109 –1113
Fig. 2. Perfusion-weighted images under ␣-chloralose anesthesia at 4.7 T (A) and 7 T (B). After switching from halothane to ␣-chloralose anesthesia the perfusion signal decreased by about 70% over the whole brain. At 4.7 T this resulted in a loss of perfusion signal in subcortical regions, due to a lack of sensitivity. At 7 T, the signal decrease under ␣-chloralose anesthesia was the same, but due to the 2.84 times higher signal-to-noise the whole cross-section of the brain including subcortical structures remained visible. Stimulation of the left forepaw caused a circumscribed increase in signal intensity in the right somatosensory cortex relative to the contralateral hemisphere.
ischemic hemisphere (0.42 ⫾ 0.15% of contralateral signal intensity, Fig. 1 A). This relative signal intensity change was similar at the higher field strength of 7 T (0.46 ⫾ 0.05%, p ⫽ 0.7; Fig. 1 B). 4. Discussion Switching from a magnetic field strength of 4.7 T to 7 T leads to a significant increase of PWI signal intensity. Our study shows that this improvement is independent of physiological conditions. Furthermore, we demonstrated that under different anesthetics the relative PWI signal intensity changes associated with changes in blood flow during functional activation or during occlusion of the MCA remained field-independent. The percentual change of the signal intensity was shown constant for the applied field strengths. This fact and the much higher PWI signal intensity at 7 T results in a higher sensitivity at higher magnetic field strengths when using arterial spin labeling PWI. Several factors may be expected to contribute to the observed signal intensity increase in arterial spin labeling PWI with higher field strength. These include: 1) the longer T1 relaxation time at 7 T in comparison to 4.7 T including the relaxation time of tissue as well as blood, 2) the higher net magnetization at the higher field strength, 3) the Q-value of the receiver rf-coils and 4) the degree of adiabatic inversion. To investigate the influence of all four parameters on the demonstrated signal improvement the T1 relaxation time at 7 T and the Q-values of the coils used at both systems were determined. At 7 T we observed a T1 relaxation time of 1382 ⫾ 96 ms. In comparison to already reported values at 4.7 T (1057 ⫾ 77 ms [11]) T1 is increased by a factor of 1.31. On the other hand, a further contribution in arterial spin tagging PWI comes from the lengthening of T1 of arterial blood. Nevertheless, the influence of this factor was neglected because the maximal systematic error without such correction is less than 2% [12].
Furthermore, the received signal due to the higher field strength can be estimated to increase nearly linearly. The nuclear magnetization is proportional to the magnetic field and the induced voltage in the receiver coil is proportional to the product of the nuclear magnetization and the frequency resulting in an increase as the square of the field. The increase of the noise is more difficult to estimate. At low frequencies the increase is nonlinear and the noise is dominated by skin-depth-limited conduction losses and losses in the radiofrequency receiver coil. At higher fields it is more dominated by conduction losses in the animal and the noise voltage varies linearly with the frequency [13]. For all other factors we found no significant difference between our both scanner. Loaded Q-values of the receiver coils were indistinguishable and the degree of adiabatic inversion in the neck of the animal was found to be comparable (approx. 90%; Burke, private communication; [14]) for both systems. Magnetization transfer effects are not expected to contribute significantly to the perfusionweighted signal intensity, as this is properly accounted for by reversal of the labeling pulse in the control phase of the arterial spin tagging experiment. In conclusion, our study showed a strong field dependent effect in perfusion-weighted imaging based on the arterial spin tagging technique. It was shown for a wide range of flow values, that the PWI signal-to-noise-ratio increased with higher magnetic field strength. The main factors contributing to that higher signal intensity are the increase of the T1 relaxation time and the higher net magnetization at the higher field, whereas parameters concerning the NMR hardware or experimental conditions were excluded or are negligible. Acknowledgment This work was supported in part by the Deutsche Forschungsgemeinschaft (SFB194/B1) and the Bundesministerium fu¨r Bildung und Forschung (BMBF; Kompetenznetzwerk-Schlaganfall B5).
C. Franke et al. / Magnetic Resonance Imaging 18 (2000) 1109 –1113
References [1] Detre JA, Leigh JS, Williams DS, Koretsky AP. Perfusion Imaging. Magn Reson Med 1992;23:37– 45. [2] Williams DS, Detre JA, Leigh JS, Koretsky AP. Magnetic resonance imaging of perfusion using spin inversion of arterial water. Proc Natl Acad Sci USA 1992;89:212– 6. [3] Detre JA, Zhang WG, Roberts DA, Silva AC, Williams DS, Grandis DJ, Koretsky AP, Leigh JS. Tissue specific perfusion imaging using arterial spin labeling. NMR Biomed 1994;7:75– 82. [4] Calamante F, Lythgoe MF, Pell GS, Thoams DL, King MD, Busza AL, Sotak CH, Williams SR, Ordidge RJ, Gadian DG. Early changes in water diffusion, perfusion, T1, and T2 during focal cerebral ischemia in rat studied at 8.5 T. Magn Reson Med 1999;41:479 – 85. [5] van Dorsten FA, Hata R, Maeda K, Franke C, Eis M, Hossmann K-A, Hoehn M. Diffusion- and perfusion-weighted MR imaging of transient focal cerebral ischemia. NMR Biomed 1999;12:525–34. [6] Lindauer U, Villringer A, Dirnagl U. Characterization of CBF response to somatosensory stimulation: model and influence of anesthetics. Am J Physiol 1993;264:H1223– 8. [7] Kerskens CM, Hoehn-Berlage M, Schmitz B, Busch E, Bock C, Gyngell ML, Hossmann K-A. Ultrafast perfusion-weighted MRI of functional brain activation in rats during forepaw stimulation: comparison with T2*-weighted MRI. NMR Biomed 1996;8:20 –3.
1113
[8] Talagala SL, Noll DC. Functional MRI using steady-state arterial water labeling. Magn Reson Med 1998;39:179 – 83. [9] Kohno K, Back T, Hoehn-Berlage M, Hossmann K-A. A modified rat model of middle cerebral artery thread occlusion under electrophysiological control for magnetic resonance investigations. Magn Reson Imaging 1995;13:65–71. [10] Weisskoff RM. Simple measurement of scanner stability for functional NMR imaging of activation in the brain. Magn Reson Med 1996;36:643–5. [11] Hoehn-Berlage M, Bockhorst K. Quantitative magnetic resonance imaging of rat brain tumors: In vivo NMR relaxometry for the discrimination of normal and pathological tissues. Technology and Health Care 1994;2:247–54. [12] Schwarzbauer C, Morrissey SP, Haase A. Quantitative magnetic resonance imaging of perfusion using magnetic labeling of water proton spins within the detection slice. Magn Reson Med 1996;35: 540 – 6. [13] Hart HR, Bottomley PA, Edelstein WA, Karr SG, Leue WM, Mueller O, Redington RW, Schenck JF, Smith LS, Vatis D. Nuclear magnetic resonance imaging: Contrast-to-noise ratio as a function of strength of magnetic field. Am J Roentgenol 1983;141:1195–1201. [14] Kerskens CM. Entwicklung theoretischer und experimenteller Methoden zur Messung der Hirndurchblutung mit Hilfe der Kernspintomographie. Berlin: Logos-Verlag; 1998.
Arterial spin tagging perfusion imaging of rat brain Dependency on magnetic field strength C. Franke,* F. A. van Dorsten, L. Olah, W. Schwindt, M. Hoehn Max-Planck-Institute for Neurological Research, Department of Experimental Neurology, Cologne, Germany Received 18 July 2000; accepted 2 September 2000
Abstract Perfusion-weighted imaging (PWI), using the method of arterial spin tagging, is strongly T1-dependent. This translates into a high field dependency of the perfusion signal intensity. In order to determine the expected signal improvement at higher magnetic fields we compared perfusion-weighted images in rat brain at 4.7 T and 7 T. Application of PWI to focal ischemia and functional activation of the brain and the use of two different anesthetics allowed the observation of a wide range of flow values. For all these (patho-)physiological conditions switching from 4.7 T to 7 T resulted in a significant increase of mean perfusion signal intensity by a factor of 2.96. The ratio of signal intensities of homotopic regions in the ipsi- and contralateral hemisphere was field-independent. The relative contribution of a) T1 relaxation time, b) net magnetization, c) the Q-value of the receiver coils and d) the degree of adiabatic inversion to the signal improvement at higher field strength were discussed. It was shown that the main parameters contributing to the higher signal intensity are the lengthening of T1 and the higher magnetization at the higher magnetic field. Keywords: Perfusion-weighted imaging; Arterial spin tagging; Magnetic field strength; Rat brain
1. Introduction The measurement of brain perfusion is important for the assessment of tissue viability. Several methods are available for measuring cerebral perfusion and related hemodynamic parameters using magnetic resonance imaging (MRI). One type of sequences utilizes magnetically labeled blood spins as a noninvasive diffusible tracer for blood flow measurements [1]. Blood flowing to the brain is saturated in the neck region with a slice-selective saturation preparation followed by the imaging sequence after a short delay. The magnetically labeled tracer has a decay rate of T1, which must be sufficiently long to allow the detection of tissue perfusion in the imaging plane. Using this method the regional concentration of saturated spins is defined by the regional blood flow and the regional relaxation time T1 [2,3]. This T1 dependence results in a high sensitivity of the perfusion signal intensity on magnetic field strength. Therefore, we investigated whether switching from 4.7 T to the higher magnetic field * Corresponding author. Tel.: ⫹49 (69) 305 27258; fax: ⫹49 (69) 308 236. E-mail address: [email protected] (C. Franke).
strength of 7 T leads to an increase of the signal-to-noiseratio (SNR) in perfusion-weighted images (PWI) using arterial spin labeling. PWI based on arterial spin labeling is successfully used to describe changes in local cerebral blood flow during early focal cerebral ischemia [4,5]. A rapid decrease of signal intensity can be detected in the ischemic hemisphere reflecting the area of perfusion deficit. Switching from halothane to ␣-chloralose anesthesia results in a reduction of cerebral blood flow (CBF) by approximately 50 –70%, as shown by Lindauer et al. [6]. This anesthetic-induced reduction of CBF leads to a drop in signal intensity of the perfusion-weighted image. On the other hand, under ␣-chloralose anesthesia it is possible to study the increased local cerebral blood flow with PWI during functional brain activation [7,8]. In the present investigation we compare the PWI sensitivity at two different magnetic field strengths, 4.7 T and 7 T, using the above-described animal models and physiological conditions to induce local or global changes in perfusion signal intensity. We demonstrate that switching to higher fields leads to a significant increase of perfusion signal intensity.
0730-725X/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 0 - 7 2 5 X ( 0 0 ) 0 0 2 1 1 - 3
1110
C. Franke et al. / Magnetic Resonance Imaging 18 (2000) 1109 –1113
2. Materials and methods 2.1. Animal preparation All animal handling and surgery was performed in accordance with animal protection guidelines and approved by local authorities. Male Wistar rats (n ⫽ 8; body weight: 350 – 400 g) were anesthetized with 1.5% halothane in a 70%/30% mixture of N2O/O2. The animals were tracheotomized, immobilized with pancuronium bromide (0.3 mg/kg/h intravenously) and mechanically ventilated for the whole duration of the experiment. Catheters were inserted into the femoral artery to monitor blood pressure and to obtain arterial blood samples for blood gas analysis. Body temperature, measured with a rectal thermometer, was maintained at 37°C using a feedback-controlled warm water blanket. 2.2. Functional activation Halothane was discontinued 45 min prior to electrical forepaw stimulation and replaced by i.v. anesthesia with ␣-chloralose (80 mg/kg). Supplemental doses (40 mg/kg) were given at 90-min intervals. For electrical stimulation, two small needle-electrodes were inserted into the skin of both forepaws. Rectangular pulses of 0.3 ms duration and 0.5 mA were applied at 3 Hz for 1 min using an electrical stimulator (FMI, Germany).
using a previously described method [7]. Briefly, an arterial spin tagging technique [3] was used in combination with snapshot FLASH imaging. Measurement parameters were: echo time TE ⫽ 3.5 ms at 4.7 T and 4.4 ms at 7 T, repetition time TR ⫽ 7.4 ms (4.7 T) and 7.6 ms (7 T), field of view ⫽ 40 mm, slice thickness ⫽ 2 mm, matrix ⫽ 128 ⫻ 64, averages ⫽ 8, total scan time ⫽ 1.3 min. The sequence consisted of two similar image acquisition phases separated by a recovery period. During the first phase, blood flowing through the neck towards the brain was adiabatically inverted for three seconds, in the second phase inflowing spins were left undisturbed. The time delay between labeling and imaging was 40 ms. Perfusion-weighted images were obtained by subtraction of the acquisitions with and without prior adiabatic spin inversion. After positioning of the animal in the magnet, pilot scans in the sagittal plane were performed for correct positioning of the imaging plane for perfusion recording. The imaging plane for activation studies was 4.5 mm posterior to the rhinal fissure. In all other studies the coronal plane 5.9 mm posterior to the rhinal fissure was placed in the isocenter, thus focusing on the center of the ischemic territory in the MCAO model. Perfusion-weighted images were performed under control conditions in each animal, corresponding with PWI measurements under halothane anesthesia before MCA occlusion and under both anesthetics. Furthermore, perfusionweighted images were recorded 30 min after MCA occlusion and during functional activation.
2.3. Focal cerebral ischemia 2.5. Data processing and statistics Focal ischemia was produced by suture occlusion of the right middle cerebral artery using a previously described remotely controlled occlusion device [9]. Briefly, a monofilament nylon thread (4-0 Prolene, Ethicon Co., Nordstedt, Germany) with its distal end thickened to 0.28 – 0.30 mm in diameter was connected to an extension catheter and passed through a guide sheath fixed to the neck of the animal. The right common carotid artery was ligated and the filament was introduced into the right internal carotid artery via the base of the skull. 2.4. NMR experiments MR experiments were performed on Bruker Biospec systems (Bruker Medical, Ettlingen, Germany) operating at 4.7 T (MSL X-11) and 7 T (DBX), respectively. Both magnets are equipped with a 30 cm bore and actively shielded gradient coils (rise time ⬍250 s; 100 and 200 mT/m, respectively). At both systems a 12 cm Helmholtz coil was used for rf-transmission; signal detection followed via an inductively coupled surface coil placed over the head of the animal (16 and 24 mm diameter, respectively). The rf-coils were decoupled from each other, the transmitter coil actively, and the receiver coil passively. Single slice perfusion-weighted images were recorded
Perfusion-weighted images were transferred to a Macintosh Power PC 7200/66 (Apple, Cupertino, CA, USA) and image analysis was carried out using the image processing software IMAGE (NIH, Bethesda, MD, USA). After interpolation of images onto a 128 ⫻ 128 matrix SNRs were determined. In all animals the following regions of interest (ROI) were defined for SNR determination: one was placed in the area of the somatosensory cortex, another included the MCA territory. In all animals homotopic regions of the contralateral hemisphere and regions outside the brain were investigated, too. In order to define signal improvement after switching to the higher magnetic field strength the contrast to noise ratio was determined as the ratio between the signal of the perfusion image within an ROI inside the brain and the standard deviation of the background of the untagged image corrected for the different noise statistics proposed by Weisskoff [10]. ROIs were located in the central part of the basal ganglia of the brain (MCA territory). For the functional investigations the position was in the somatosensory cortex. To compare signal changes after MCA occlusion and during functional activation with the corresponding control situation we calculated the relative change of signal intensity between ipsi- and contralateral hemisphere and ex-
C. Franke et al. / Magnetic Resonance Imaging 18 (2000) 1109 –1113 Table 1. Signal-to-noise-ratio and contrast for different anesthesia, after occlusion of the MCA, and during functional activation as a function of magnetic field strength Signal-to noise-ratio
7T
4.7 T
Ratio
Halothane ␣-chloralose
30.89 ⫾ 6.67** 9.22 ⫾ 0.38
10.03 ⫾ 4.37 3.25 ⫾ 2.2
3.08 2.84
57 ⫾ 21 42 ⫾ 15
1.12 1.08
Signal change in percent of control Functional activation 77 ⫾ 12 MCA occlusion 46 ⫾ 5 * p ⬍ .005; ** p ⬍ 0.005.
pressed as percent of the signal intensity of the contralateral hemisphere. All values are given as mean ⫾ standard deviation (SD). Statistical analysis was performed between both field strengths using Student’s t-test (unpaired, unequal variances). p ⬍ 0.05 was accepted as significant. Symbols: *: p ⬍ 0.05; **: p ⬍ 0.005. 2.6. Determination of T1 and Q-values In order to determine the contribution of several factors to the signal-to-noise-ratio at both field strengths the T1 relaxation time at 7 T was determined using a multislice inversion recovery sequence (TR ⫽ 8000 ms, TE ⫽ 20 ms, ten varying inversion delays: 15–3000 ms). T1 was determined by monoexponential fitting. The Q-values of the receiver coils (200 and 300 MHz) were determined using a Network Analyzer (Hewlett Packard, USA) under loaded conditions.
1111
At 7 T we observed a T1 relaxation time of the brain tissue of 1382 ⫾ 96 ms, given as mean value averaged over the whole hemisphere of a healthy rat brain. The Q-values of the receiver coils were determined as 62 and 55 for 200 and 300 MHz, respectively. The systematic error of these measurements was estimated to approximately 8%. 3.2. Observations under different anesthetics Under halothane anesthesia, perfusion signal intensity was higher at 7 T than at 4.7 T by a factor of 3.08 (Table 1, Fig. 1 A, B). When changing from halothane to ␣-chloralose, the PWI signal intensity at 4.7 T was decreased by 70% throughout the brain due to a decrease of baseline cerebral blood flow (CBF) by the anesthetic [6]. This resulted in perfusion images with signal intensity remaining mostly in the cortical areas (Fig. 2 A). At 7 T a similar amount of signal decrease (68%) was observed when switching from halothane to ␣-chloralose. Due to the overall higher PWI signal intensity (factor 2.84) at 7 T, the brain hemispheres remained completely visible under ␣-chloralose, showing also signal intensity in the subcortical regions (Fig. 2 B). 3.3. Electrical forepaw stimulation Stimulation of the left forepaw caused a circumscribed increase in perfusion signal intensity in the right somatosensory cortex relative to the contralateral hemisphere, as shown in Fig. 2 A and B. Activation induced perfusion signal increase was determined and found to be independent of field strength: 57 ⫾ 21% at 4.7 T and 77 ⫾ 12% at 7 T (p ⫽ 0.3).
3. Results 3.4. MCA occlusion 3.1. General observations All physiological parameters of the animals were within normal range during the whole observation periods.
At 4.7 T, 30 min after occlusion of the right middle cerebral artery, a large area with reduced perfusion signal intensity was detectable, showing the perfusion deficit in the
Fig. 1. Perfusion-weighted images under halothane anesthesia at 4.7 T (A) and 7 T (B). Increasing the magnetic field strength from 4.7 T to 7 T resulted in an increase of perfusion-weighted signal-to-noise by a factor of 3.08. After MCA occlusion a large decrease of the perfusion signal intensity is detectable showing the perfusion deficit in the ischemic (right) hemisphere. The ischemia-related, relative signal change is similar at 4.7 T and 7 T.
1112
C. Franke et al. / Magnetic Resonance Imaging 18 (2000) 1109 –1113
Fig. 2. Perfusion-weighted images under ␣-chloralose anesthesia at 4.7 T (A) and 7 T (B). After switching from halothane to ␣-chloralose anesthesia the perfusion signal decreased by about 70% over the whole brain. At 4.7 T this resulted in a loss of perfusion signal in subcortical regions, due to a lack of sensitivity. At 7 T, the signal decrease under ␣-chloralose anesthesia was the same, but due to the 2.84 times higher signal-to-noise the whole cross-section of the brain including subcortical structures remained visible. Stimulation of the left forepaw caused a circumscribed increase in signal intensity in the right somatosensory cortex relative to the contralateral hemisphere.
ischemic hemisphere (0.42 ⫾ 0.15% of contralateral signal intensity, Fig. 1 A). This relative signal intensity change was similar at the higher field strength of 7 T (0.46 ⫾ 0.05%, p ⫽ 0.7; Fig. 1 B). 4. Discussion Switching from a magnetic field strength of 4.7 T to 7 T leads to a significant increase of PWI signal intensity. Our study shows that this improvement is independent of physiological conditions. Furthermore, we demonstrated that under different anesthetics the relative PWI signal intensity changes associated with changes in blood flow during functional activation or during occlusion of the MCA remained field-independent. The percentual change of the signal intensity was shown constant for the applied field strengths. This fact and the much higher PWI signal intensity at 7 T results in a higher sensitivity at higher magnetic field strengths when using arterial spin labeling PWI. Several factors may be expected to contribute to the observed signal intensity increase in arterial spin labeling PWI with higher field strength. These include: 1) the longer T1 relaxation time at 7 T in comparison to 4.7 T including the relaxation time of tissue as well as blood, 2) the higher net magnetization at the higher field strength, 3) the Q-value of the receiver rf-coils and 4) the degree of adiabatic inversion. To investigate the influence of all four parameters on the demonstrated signal improvement the T1 relaxation time at 7 T and the Q-values of the coils used at both systems were determined. At 7 T we observed a T1 relaxation time of 1382 ⫾ 96 ms. In comparison to already reported values at 4.7 T (1057 ⫾ 77 ms [11]) T1 is increased by a factor of 1.31. On the other hand, a further contribution in arterial spin tagging PWI comes from the lengthening of T1 of arterial blood. Nevertheless, the influence of this factor was neglected because the maximal systematic error without such correction is less than 2% [12].
Furthermore, the received signal due to the higher field strength can be estimated to increase nearly linearly. The nuclear magnetization is proportional to the magnetic field and the induced voltage in the receiver coil is proportional to the product of the nuclear magnetization and the frequency resulting in an increase as the square of the field. The increase of the noise is more difficult to estimate. At low frequencies the increase is nonlinear and the noise is dominated by skin-depth-limited conduction losses and losses in the radiofrequency receiver coil. At higher fields it is more dominated by conduction losses in the animal and the noise voltage varies linearly with the frequency [13]. For all other factors we found no significant difference between our both scanner. Loaded Q-values of the receiver coils were indistinguishable and the degree of adiabatic inversion in the neck of the animal was found to be comparable (approx. 90%; Burke, private communication; [14]) for both systems. Magnetization transfer effects are not expected to contribute significantly to the perfusionweighted signal intensity, as this is properly accounted for by reversal of the labeling pulse in the control phase of the arterial spin tagging experiment. In conclusion, our study showed a strong field dependent effect in perfusion-weighted imaging based on the arterial spin tagging technique. It was shown for a wide range of flow values, that the PWI signal-to-noise-ratio increased with higher magnetic field strength. The main factors contributing to that higher signal intensity are the increase of the T1 relaxation time and the higher net magnetization at the higher field, whereas parameters concerning the NMR hardware or experimental conditions were excluded or are negligible. Acknowledgment This work was supported in part by the Deutsche Forschungsgemeinschaft (SFB194/B1) and the Bundesministerium fu¨r Bildung und Forschung (BMBF; Kompetenznetzwerk-Schlaganfall B5).
C. Franke et al. / Magnetic Resonance Imaging 18 (2000) 1109 –1113
References [1] Detre JA, Leigh JS, Williams DS, Koretsky AP. Perfusion Imaging. Magn Reson Med 1992;23:37– 45. [2] Williams DS, Detre JA, Leigh JS, Koretsky AP. Magnetic resonance imaging of perfusion using spin inversion of arterial water. Proc Natl Acad Sci USA 1992;89:212– 6. [3] Detre JA, Zhang WG, Roberts DA, Silva AC, Williams DS, Grandis DJ, Koretsky AP, Leigh JS. Tissue specific perfusion imaging using arterial spin labeling. NMR Biomed 1994;7:75– 82. [4] Calamante F, Lythgoe MF, Pell GS, Thoams DL, King MD, Busza AL, Sotak CH, Williams SR, Ordidge RJ, Gadian DG. Early changes in water diffusion, perfusion, T1, and T2 during focal cerebral ischemia in rat studied at 8.5 T. Magn Reson Med 1999;41:479 – 85. [5] van Dorsten FA, Hata R, Maeda K, Franke C, Eis M, Hossmann K-A, Hoehn M. Diffusion- and perfusion-weighted MR imaging of transient focal cerebral ischemia. NMR Biomed 1999;12:525–34. [6] Lindauer U, Villringer A, Dirnagl U. Characterization of CBF response to somatosensory stimulation: model and influence of anesthetics. Am J Physiol 1993;264:H1223– 8. [7] Kerskens CM, Hoehn-Berlage M, Schmitz B, Busch E, Bock C, Gyngell ML, Hossmann K-A. Ultrafast perfusion-weighted MRI of functional brain activation in rats during forepaw stimulation: comparison with T2*-weighted MRI. NMR Biomed 1996;8:20 –3.
1113
[8] Talagala SL, Noll DC. Functional MRI using steady-state arterial water labeling. Magn Reson Med 1998;39:179 – 83. [9] Kohno K, Back T, Hoehn-Berlage M, Hossmann K-A. A modified rat model of middle cerebral artery thread occlusion under electrophysiological control for magnetic resonance investigations. Magn Reson Imaging 1995;13:65–71. [10] Weisskoff RM. Simple measurement of scanner stability for functional NMR imaging of activation in the brain. Magn Reson Med 1996;36:643–5. [11] Hoehn-Berlage M, Bockhorst K. Quantitative magnetic resonance imaging of rat brain tumors: In vivo NMR relaxometry for the discrimination of normal and pathological tissues. Technology and Health Care 1994;2:247–54. [12] Schwarzbauer C, Morrissey SP, Haase A. Quantitative magnetic resonance imaging of perfusion using magnetic labeling of water proton spins within the detection slice. Magn Reson Med 1996;35: 540 – 6. [13] Hart HR, Bottomley PA, Edelstein WA, Karr SG, Leue WM, Mueller O, Redington RW, Schenck JF, Smith LS, Vatis D. Nuclear magnetic resonance imaging: Contrast-to-noise ratio as a function of strength of magnetic field. Am J Roentgenol 1983;141:1195–1201. [14] Kerskens CM. Entwicklung theoretischer und experimenteller Methoden zur Messung der Hirndurchblutung mit Hilfe der Kernspintomographie. Berlin: Logos-Verlag; 1998.