Arterial spin labeling magnetic resonance imaging at short post-labeling delay reflects cerebral perfusion pressure verified by oxygen-15-positron emission tomography in cerebrovascular steno-occlusive disease

Abstract

Background

Arterial transit time correction by data acquisition with multiple post-labeling delays (PLDs) or relatively long PLDs is expected to obtain more accurate imaging in cases of the cerebrovascular steno-occlusive disease. However, there have so far been no reports describing the significance of arterial spin labeling (ASL) images at short PLDs regarding the evaluation of cerebral circulation in ischemic cerebrovascular disease.

Purpose

To clarify the role of short-PLD ASL in cerebrovascular steno-occlusive disease.

Material and Methods

Fifty-three patients with cerebrovascular steno-occlusive disease were included in this study. All patients underwent ASL magnetic resonance imaging and ¹⁵O-PET within two days of each modality. To compare the ASL findings with each parameter of PET, the right-to-left (R/L) ratio, defined as the right middle cerebral artery (MCA) value/left MCA value, was calculated.

Results

There is a significant correlation between the ASL images at a short PLD and the ratio of cerebral blood flow and cerebral blood volume by ¹⁵O-PET, which may accurately reflect the cerebral perfusion pressure. A receiver operating characteristic curve analysis indicated that ASL images at PLD 1000 and 1500 ms were more accurate than at PLD 2000–3000 ms for the detection of a ≥10% change in the PET cerebral blood flow.

Conclusion

ASL images at shorter PLDs may be useful at least as a screening modality to detect the changes in the cerebral circulation in cerebrovascular steno-occlusive disease. We must evaluate ASL images at multiple PLDs while considering the arterial transit time of each case at present.

Keywords

Arterial spin labeling post-labeling delay 15O-PET cerebrovascular steno-occlusive disease cerebral perfusion pressure

Introduction

Since arterial spin labeling (ASL) magnetic resonance imaging (MRI) utilizes water in blood as an intrinsic tracer, no radioactive tracer or contrast agent is needed with this approach (1,2). We are therefore exploring the utility of this imaging technique for the non-invasive and convenient measurement of regional cerebral perfusion. Although there have been reports (3 –5) describing a good correlation between ASL and oxygen-15-positron emission tomography (¹⁵O-PET) or single-photon emission computed tomography (SPECT) in healthy volunteers, ASL imaging is known to be strongly influenced by the arterial transit time (ATT), especially under conditions of cerebral ischemia (6 –9). The recruitment of the blood flow through collateral pathways may lead to the delayed arrival of the labeled blood, so ASL has the potential to overestimate the cerebral blood flow (CBF) reduction in such situations, as previously reported (8,9). Therefore, we must consider the post-labeling delay (PLD), which is the interval between the labeling of blood in the neck and acquiring the ASL imaging.

Generally, PLD can be in the range of 1000–3000 ms due to technical limitations at present. Although there have been reports indicating the optimal PLD to be ≥1500 ms under healthy conditions (3 –5), this is not applicable in cases of the cerebrovascular steno-occlusive disease. ATT correction by data acquisition with multiple PLDs or relatively long PLDs, such as 4000 ms, is expected to obtain more accurate imaging in cases of the cerebrovascular steno-occlusive disease, such as moyamoya disease (10). However, there have so far been no reports describing the significance of ASL images at short PLDs, such as 1000 or 1500 ms regarding the evaluation of cerebral circulation in ischemic cerebrovascular disease. We therefore focused on cerebral perfusion pressure (CPP), which would be initially reduced even in patients with mild cerebrovascular steno-occlusive disease. We assumed that ASL with a short PLD, which might have the potential to overestimate the CBF reduction, would correlate with the CBF/CBV that reflects the CPP, based on the findings of previous reports (11,12).

In the present study, we performed ASL imaging with multiple PLDs and compared our findings with those of ¹⁵O-PET to clarify the role of short-PLD ASL in cerebrovascular steno-occlusive disease.

Material and Methods

Participants

The present study was approved by the Ethics Committee based on ethical guidelines for epidemiological research. Informed consent was obtained from all participants.

Fifty-three patients (29 male, 24 female; mean age = 50.3 ± 23.6 years; age range = 7–83 years) with cerebrovascular steno-occlusive disease were included in this study from April 2013 to March 2017. Twenty-six patients had moyamoya disease, 25 had cervical internal carotid artery (ICA) stenosis/occlusion, and two had intracranial ICA/middle cerebral artery (MCA) stenosis. All patients underwent ASL MRI and ¹⁵O-PET within two days of each modality.

ASL MRI

ASL imaging was performed using a 3.0-T MR unit (Discovery MR 750W; GE Healthcare, Milwaukee, WI, USA) equipped with an eight-channel receive-only head coil for signal reception. ASL was prepared using a three-dimensional (3D) spiral fast-spin echo sequence with background suppression for perfusion imaging covering the entire brain, as previously described (9,13). A pseudo-continuous protocol was employed without vessel suppression. The labeling plane at the neck was perpendicular to the ICA located around the foramen magnum. Other acquisition parameters were as follows: six arms with 512 points in each spiral arm; phase encoding in the z direction = 36; section thickness = 4 mm; repetition time (TR) = 4728 ms; number of excitations (NEX) = 2. The labeling duration was 1.5 s. Five PLDs of 1000, 1500, 2000, 2500, and 3000 ms were chosen, as described elsewhere. The average acquisition time of ASL at each PLD was 2 min 39 s.

¹⁵O-PET

The Siemens Biograph micro-computed tomography (mCT) apparatus (software version syngo VG60A; Siemens Healthcare, Erlangen, Germany) consists of a PET detector with four rings, 48 detector blocks in each ring, and lutetium oxyorthosilicate (LSO) crystals of 4 × 4 × 20 mm in a 13 × 13 array coupled to a 2 × 2 PMT array in each detector block. This gives an axial PET field of view (FOV) of 22.1 cm. The transaxial FOV is 70 cm. The detector ring diameter is 84.2 cm. The time coincidence window is 4.1 ns and the energy window 435–650 keV. Integrated 64-slice CT is used for the attenuation correction of PET data (14). The CBF was determined while the participant continuously inhaled C¹⁵O₂ through a mask. Measurements of cerebral metabolic rate for oxygen (CMRO₂) and cerebral oxygen extraction fraction (OEF) were obtained during continuous inhalation of ¹⁵O₂. Data were collected for 5 min. The CBF, CMRO₂, and OEF were calculated using the steady-state method. According to Lammertsma and Jones (15), the CMRO₂ and OEF tend to be overestimated compared with values measured using arteriovenous difference techniques. This overestimation is theoretically due to the signal from non-extracted intravascular ¹⁵O. To correct for this intravascular component, patients took a single breath of C¹⁵O to measure the cerebral blood volume (CBV), and CMRO₂ and OEF were corrected according to the CBV (16).

Image analyses

The ASL data were transformed into a Montreal Neurological Institute (MNI) space. Because the distribution of ASL in the brain differs from that of ¹⁵O-PET, which is used as the PET template in the SPM program, we created an ASL template for spatial normalization. The ASL template was made by averaging ASL images from nine normal volunteers that had been spatially normalized using parameters from co-registered T1 images. Co-registration of ASL and T1 images and the spatial normalization of the ASL template were also performed with SPM2 (17). A total of 318 constant regions of interest (ROIs) were automatically placed in both the cerebral and cerebellar hemispheres using a 3D stereotaxic ROI template (3DSRT) with SPM2 (Fujifilm RI Pharma Co., Ltd., Tokyo, Japan) (18). 3DSRT is a fully automated regional CBF (rCBF) quantification software program developed in the MNI space. Therefore, the same regions can be compared between ASL and PET using 3DSRT. The ROIs were grouped into 10 segments (callosomarginal, pericallosal, precentral, central, parietal, angular, temporal, posterior, hippocampus, and cerebellar) in each hemisphere according to the arterial supply. Five of these segments (B: precentral, C: central, D: parietal, E: angular, and F: temporal) were combined and defined as an ROI perfused by the MCA (Fig. 1). The mean values of the ASL intensity signal and of each parameter on PET images were evaluated in the ROIs of the bilateral MCA territory. To compare the ASL findings with each parameter of PET, the right-to-left (R/L) ratio, defined as the right MCA value/left MCA value, was calculated.

Fig. 1.

Diagrams showing the regions of interest of three-dimensional stereotaxic ROI templates. B, middle and inferior frontal; C, primary sensorimotor; D, parietal; E, angular; F, temporal. B, C, D, E, and F are defined as MCA territory. MCA, middle cerebral artery; ROI, region of interest.

Statistical analyses

A linear regression analysis determined correlations between ASL images at multiple PLDs and each parameter of ¹⁵O-PET (CBF, CBV, the ratio of CBF and CBV [CBF/CBV], CMRO₂). A receiver operating characteristic (ROC) curve analysis was also performed to verify the accuracy of ASL imaging compared with the PET CBF. The data were analyzed by the SPSS software program (Ver. 20; SPSS Inc., Chicago, IL, USA). P < 0.01 was considered statistically significant.

Results

Figs. 2 and 3 show the results of the linear regression analysis of the R/L ratio between ASL images at multiple PLDs and each parameter of ¹⁵O-PET (A = CBF, B = CBF/CBV, C = CBV, D = CMRO₂). The correlation and regression coefficients between the parameters of PET and each PLD are shown in Table 1. ASL showed significant correlation with the PET CBF except for at a PLD of 3000 ms. The correlation coefficient of PLD1000 ms was the highest during all PLD. The regression coefficient of PLD 2000 ms was 1.25 which was the closest to 1. ASL showed significant correlation with the PET CBF/CBV except for at a PLD of 2500 and 3000 ms. The correlation coefficient of PLD 1000 ms was the highest among all PLDs and the regression coefficient of PLD 1500 ms was 0.95, which was the closest to 1.

Fig. 2.

The results of the linear regression analysis between MRI ASL and ¹⁵O-PET (CBF, CBF/CBV). (a) The correlation between PET CBF and ASL with each PLD. (b) The correlation between PET CBF/CBV and ASL with each PLD. ASL, arterial spin labeling; CBF, cerebral blood flow; CBV, cerebral blood volume; MRI, magnetic resonance imaging; PET, positron emission tomography; PLD, post-labeling delay.

Fig. 3.

The results of the linear regression analysis between MRI ASL and ¹⁵O-PET (CBV, CMRO₂). (a) The correlation between PET CBV and ASL with each PLD. (b) The correlation between PET CMRO₂ and ASL with each PLD. ASL, arterial spin labeling; CBV, cerebral blood volume; CMRO₂, cerebral metabolic rate for oxygen; MRI, magnetic resonance imaging; PET, positron emission tomography; PLD, post-labeling delay.

Table 1.

The correlation and regression coefficients between parameters of PET and each PLD.

		PLD (ms)
		1000	1500	2000	2500	3000
CBF	CC	0.77	0.66	0.58	0.36	0.06
	P value	<0.001	<0.001	<0.001	0.008	0.66
	RC	2.86	2.02	1.25	0.43	–0.053
CBF/CBV	CC	0.79	0.64	0.55	0.27	0.17
	P value	<0.001	<0.001	<0.001	0.048	0.22
	RC	1.42	0.95	0.57	0.16	–0.072
CBV	CC	0.6	0.47	0.39	0.13	0.23
	P value	<0.001	<0.001	0.004	0.37	0.1
	RC	–2.28	–1.49	–0.85	1.57	0.2
CMRO₂	CC	0.73	0.69	0.6	0.37	0.081
	P value	<0.001	<0.001	<0.001	0.006	0.56
	RC	2.94	2.32	1.41	0.49	–0.077

CBF, cerebral blood flow; CBV, cerebral blood volume; CC, correlation coefficient; CMRO₂, cerebral metabolic rate for oxygen; PET, positron emission tomography; PLD, post-labeling delay; RC, regression coefficient.

Representative cases

Case 1: moyamoya disease, a case of circulation delay

A 47-year-old woman underwent MRI and ¹⁵O-PET due to a headache. MRI fluid-attenuated inversion recovery (FLAIR) showed no ischemic lesions in the bilateral hemisphere (Fig. 4a). Magnetic resonance angiography (MRA) showed stage III on the right and stage I on the left (Fig. 4b). ¹⁵O-PET showed a decreased CBF/CBV in the right hemisphere. The CBV, CMRO₂, and OEF were not markedly changed (Fig. 4c). ASL images, particularly those at PLDs of 1.0 and 1500 ms, showed a markedly decreased ASL CBF in the right hemisphere. There was no laterality in ASL images of PLD ≥2000 ms. The R/L ratios of ASL images at 1000, 1500, 2000, 2500, and 3000 ms PLD were 0.71, 0.85, 0.93, 1.02, and 0.99, respectively. The R/L ratios of CBF, CBF/CBV, CBV, CMRO₂, and OEF were 0.93, 0.84, 1.11, 0.97, and 0.91, respectively.

Fig. 4.

A 47-year-old woman underwent MRI and ¹⁵O-PET due to a headache. (a) MRI FLAIR images showed no ischemic lesions in the bilateral hemisphere. (b) MRA showed stages III in the right and stage I in the left. (c) ¹⁵O-PET showed a decreased CBF/CBV in the right hemisphere. The CBV, CMRO₂, and OEF were not changed. (d) The R/L ratios of ASL images at 1000, 1500, 2000, 2500, and 3000 ms PLD were 0.71, 0.85, 0.93, 1.02, and 0.99, respectively. The R/L ratios of CBF, CBF/CBV, CBV, CMRO₂, and OEF were 0.93, 0.84, 1.11, 0.97, and 0.91, respectively. ASL images with PLD, especially 1000 and 1500 ms, showed markedly decreased ASL CBF in the right hemisphere. There was no laterality after PLD 2000 ms. It is obvious that ASL images with short PLD overestimated the CBF reduction in comparison to PET CBF. ASL, arterial spin labeling; CBF, cerebral blood flow; CBV, cerebral blood volume; CMRO₂, cerebral metabolic rate for oxygen; MRA, magnetic resonance angiography; MRI, magnetic resonance imaging; OEF, oxygen ejection fraction; PET, positron emission tomography; PLD, post-labeling delay.

ASL images at shorter PLDs obviously overestimated the CBF reduction compared with the PET CBF (Fig. 4d).

Case 2: right cervical ICA severe stenosis, a case of misery perfusion

A 76-year-old man was admitted to our hospital due to transient ischemic attack (TIA). MRI FLAIR showed only mild white matter lesions at the bilateral hemispheres (Fig. 5a). MRA showed right cervical ICA severe stenosis and signal reduction of the right intracranial vessel (Fig. 5b). ¹⁵O-PET showed a markedly decreased CBF and CBF/CBV in the right hemisphere. CBV and OEF were increased (Fig. 5c). The R/L ratios of ASL images at 1000, 1500, 2000, 2500, and 3000 ms PLD were 0.50, 0.53, 0.55, 0.74, and 0.87, respectively. The R/L ratios of CBF, CBF/CBV, CBV, CMRO_2, and OEF were 0.76, 0.60, 1.17, 0.93, and 1.27, respectively. ASL images at PLDs of 1000, 1500, and 2000 ms showed a markedly decreased ASL CBF in the right hemisphere. ASL images at PLD 3000 ms showed no reduction in the ASL CBF in the right hemisphere (Fig. 5d). ASL images at PLD 1000 and 1500 ms more closely resembled the CBF/CBV images than those images obtained at a longer PLD.

Fig. 5.

A 76-year-old man admitted to our hospital with TIA. (a) MRI FLAIR images showed only a mild white matter lesion of the bilateral hemisphere. (b) MRA showed right cervical ICA severe stenosis and signal reduction of the right intracranial vessel. (c) ¹⁵O-PET showed markedly decreased CBF, CBF/CBV in the right hemisphere. CBV and OEF were increased. (d) The R/L ratios of ASL images at 1000, 1500, 2000, 2500, and 30 ms PLD were 0.50, 0.53, 0.55, 0.74, and 0.87, respectively. The R/L ratios of CBF, CBF/CBV, CBV, CMRO₂ and OEF were 0.76, 0.60, 1.17, 0.93, and 1.27, respectively. ASL images with PLD 1000, 1500, and 2000 ms showed markedly decreased ASL CBF in the right hemisphere. ASL images with PLD 3000 ms showed no reduction of ASL CBF in the right hemisphere. ASL, arterial spin labeling; CBF, cerebral blood flow; CBV, cerebral blood volume; CMRO₂, cerebral metabolic rate for oxygen; ICA, internal carotid artery; MRA, magnetic resonance angiography; MRI, magnetic resonance imaging; OEF, oxygen ejection fraction; PET, positron emission tomography; PLD, post-labeling delay; TIA, transient ischemic attack.

ROC curve analyses

Figure 6 shows the results of the ROC analysis of the ASL images at each PLD, which identified a change of ≥10% in the R/L ratio of ¹⁵O-PET CBF. Table 2 shows the area under the curve (AUC) of each PLD. These results indicate that ASL images at PLD 1000 and 1500 ms are more accurate than those at PLD 2000–3000 ms for the detection of a ≥10% change in the PET CBF. The cut-off value of the R/L ratio of the ASL images at PLD 1000 ms for ≥10% of the R/L ratio of a 15O PET CBF was 10.7%. The sensitivity and specificity were 75% and 87.8%, respectively.

Fig. 6.

The results of a ROC curve analysis of the ASL images at each PLD to detect a change of ≥10% in the R/L ratio of a 15O-PET CBF. These results indicate that ASL images with PLD 1000 and 1500 ms are more accurate than PLD 2000–3000 ms for the detection of ≥10% change in PET CBF. ASL, arterial spin labeling; CBF, cerebral blood flow; PET, positron emission tomography; PLD, post-labeling delay; ROC, receiver operating characteristic.

Table 2.

The AUC of each PLD according to a ROC analysis.

PLD (ms)	AUC	P value
1000	0.74	0.01
1500	0.715	0.02
2000	0.62	0.21
2500	0.596	0.32
3000	0.567	0.48

AUC, area under the curve; PLD, post-labeling delay; ROC, receiver operating characteristic.

Discussion

Since MRI ASL was first reported by Williams et al. (1,2) in 1992, the quality of ASL imaging has been improving with the advent of various technologies. We, therefore, expect ASL imaging to become more universally applicable for the evaluation of cerebral circulation due to its noninvasiveness, convenience, and low cost compared with PET and SPECT. However, several issues remain to be resolved in order to obtain reliable images of as good a quality as PET and SPECT (19).

Several reports (3 –5) have indicated a good correlation between the ASL CBF at PLD ≥1500 ms and ¹⁵O-PET or SPECT in healthy volunteers. However, the ASL images with short PLDs, such as 1000 or 1500 ms, reportedly overestimated the CBF reduction compared with PET CBF in cases of cerebrovascular steno-occlusive disease (6 –9), as the recruitment of the blood flow through collateral pathways can lead to the delayed arrival of the labeled blood. Regarding this overestimation, we need to pay attention to the ATT and consider the PLD, which is the time delay between labeling blood in the neck and the detection of such labeled blood in the target lesion. Although there have been reports indicating the optimal PLD to be ≥1500 ms under healthy conditions (3 –5), ATT correction by data acquisition with multiple PLDs or relatively long PLDs, such as 4000 ms, would enabling more accurate ASL imaging (10,20).

On the other hand, there have been no reports describing the significance of ASL images at a short PLD, such as a PLD of 1000 or 1500 ms, for evaluating the cerebral circulation in ischemic cerebrovascular disease. The present findings demonstrated a more significant correlation between the ASL images at a short PLD and CBF/CBV of ¹⁵O-PET. In contrast, no significant correlation was noted between ASL images at PLD 3000 s and the parameters of PET (Figs. 2 and 3, Table 1). The signal-to-noise ratio (SNR) of ASL images at 3.0 s may have been too low to affect the evaluation of the cerebral circulation adversely.

Since the CBF/CBV measured by ¹⁵O-PET is considered to reflect the cerebral perfusion pressure (CPP) according to previous reports (11,12), our findings suggest that a short PLD may reflect the CPP, which is the one of the factors helping to maintain the cerebral circulation. Powers (21) reported that the staging of cerebral hemodynamic state was based on CPP. Thus, the reduction in CPP is considered an initial change in cerebrovascular steno-occlusive patients. Since the CBF/CBV is also equal to “1 divided by the mean transit time (MTT)” according to the central volume principle. As the MTT would be longer with a lower CPP, a longer PLD might be more appropriate in ASL images. However, we need to consider the possibility of arterial transit artifacts (ATAs) (22), such as stagnation of blood in the vessel, especially with a longer PLD. Short-PLD images may be less susceptible to ATAs than longer ones. Even a mild reduction in the CPP could then induce changes in ASL images with a short PLD. Since T max images in MR perfusion image indicates the delay on the arrival of contrast (23), ASL based on ATT might be more relevant with T max images. Thus, even though the principle of measurement is different between direct susceptibility contrast (DSC) MR perfusion using macro particle and ASL using water molecular, ASL images might have the possibility to behave like DSC MR perfusion images. However, further investigation of this point is needed. In brief, CBF is negatively correlated with CBV and positively correlated with CMRO₂ in the patients with cerebrovascular steno-occlusive disease except misery perfusion state. It is therefore believed that CBV or CMRO₂ correlates to some extent with ASL.

Furthermore, the results of our ROC analysis indicate that ASL images at a short PLD are more accurate for detecting a ≥10% change in the PET CBF than those at longer PLDs. Therefore, ASL images at short PLDs may at least be useful as a screening modality for the detection of changes in the cerebral circulation in cases of the cerebrovascular steno-occlusive disease.

However, we must bear in mind that ASL images at short PLDs may overestimate the reduction in the CBF, as previously mentioned. Since the regression coefficient is higher at PLDs of 1000 and 1500 ms than at PDLs of 2000 and 2500 ms (Fig. 2, Table 1), we can easily see that the degree of change in the cerebral circulation is more substantial at shorter PLDs than at longer PLDs. To prevent this overestimation, we must judge the condition of the cerebral circulation based on ASL images at multiple PLDs (at least two values with a difference of 1000 ms or more). ASL images at short PLDs are thus suitable for the detection of slight changes in the cerebral circulation, while longer PLDs are suitable for assessing the degree of change in the cerebral circulation. Alternatively, we can consider ATT based on multiple PLDs. However, the commercial-based application does not allow us to do that at present.

Several limitations associated with the present study warrant mention. First, many of the participants of this study had moyamoya disease. The CPP of patients in this study might be lower than that of patients with cerebrovascular steno-occlusive disease in other studies. Thus, the ATT in this series may demonstrate more variation than other ischemic cerebrovascular diseases. Second, the CBF was evaluated not as an absolute value but as an R/L ratio. This is because the absolute value of ASL can be affected by various factors, such as a difference in the T1 of arterial blood, uncertainly in labeling efficiency, and variation in transport times paper). Indeed, we found insufficient correlation between the ASL absolute value and ¹⁵O-PET in our preliminary data (data not shown). Third, the average age of our participants was relatively young. Previous reports have suggested that the ATT may be delayed in older people compared with younger ones. The results of our study may, therefore, have been different had our participants been older (24,25). We need to interpret our results while taking into account those limitations.

Given these findings, it appears to still be difficult for ASL imaging to be used instead of SPECT and PET in the evaluation of cerebral circulation at present. In the future, further progress in ASL, such as the advent of multi-delay ASL (26) and progress in ATT correction (10), may allow us to obtain more accurate ASL images.

In conclusion, the findings of the present study show that ASL images with a short PLD closely reflect the ¹⁵O-PET CBF/CBV, which may accurately reflect the cerebral perfusion pressure and are therefore able to accurately show the cerebral perfusion pressure in patients with cerebrovascular steno-occlusive disease. Although ASL might have utility for screening cerebral circulation, similar to SPECT or PET, we must evaluate ASL images at multiple PLDs while considering the ATT of each case at present.

Footnotes

Acknowledgements

The authors thank Mr. Ooba, whose generous support and insightful comments were invaluable during the course of our study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Yasuaki Kokubo

References

Williams

Detre

Leigh

, et al. Magnetic resonance imaging of perfusion using spin inversion of arterial water. Proc Natl Acad Sci USA 1992; 89:212–216.

Detre

Leigh

Williams

, et al. Perfusion imaging. Magn Reson Med 1992; 23:37–45.

Berman

Ellmore

, et al. H₂¹⁵O PET validation of steady-state arterial spin tagging cerebral blood flow measurements in humans. Magn Reson Med 2000; 244:450–456.

Qiu

Paul Maguire

Arora

, et al. Arterial transit time effects in pulsed arterial spin labeling CBF mapping: insight from a PET and MR study in normal human subjects. Magn Reson Med 2010; 63:374–384.

Rowley

, et al. Reliability and precision of pseudo-continuous arterial spin labeling perfusion MRI on 3.0 T and comparison with 15O-water PET in elderly subjects at risk for Alzheimer’s disease. NMR Biomed 2010; 23:286–293.

Kimura

Kado

Koshimoto

, et al. Multislice continuous arterial spin-labeled perfusion MRI in patients with chronic occlusive cerebrovascular disease: a correlative study with CO2 PET validation. J Magn Reson Imaging 2005; 22:189–198.

Bokkers

Bremmer

van Berckel

, et al. Arterial spin labeling perfusion MRI at multiple delay times: a correlative study with H2 15O positron emission tomography in patients with symptomatic carotid artery occlusion. J Cereb Blood Flow Metab 2010; 30:222–229.

Kamano

Yoshiura

Hiwatashi

, et al. Arterial spin labeling in patients with chronic cerebral artery steno-occlusive disease: correlation with ¹⁵O-PET. Acta Radiol 2013; 54:99–106.

Itagaki H, Kokubo Y, Saso K, et al. Finding of MRI ALS compared with 15O-PET in two cases of moyamoya disease with a delay of cerebral circulation and post-operative hyperperfusion. Jpn J Stroke 2018;40:190–194.

10.

Fan

Guo

Khalighi

, et al. Long-delay arterial spin labeling provides more accurate cerebral blood flow measurements in moyamoya patients: a simultaneous positron emission tomography/MRI study. Stroke 2017; 48:2441–2449.

11.

Schumann

Touzani

Young

, et al. Evaluation of the ratio of cerebral blood flow to cerebral blood volume as an index of local cerebral perfusion pressure. Brain 1998; 121:1369–1379.

12.

Watabe

Shimosegawa

Kato

, et al. CBF/CBV maps in normal volunteers studied with ¹⁵O PET: a possible index of cerebral perfusion pressure. Neurosci Bull 2014; 30: 857–862.

13.

Haga

Morioka

Shimogawa

, et al. Arterial spin labeling perfusion magnetic resonance image with dual postlabeling delay: a correlative study with acetazolamide loading 123I-iodoamphetamine single-photon emission computed tomography. J Stroke Cerebrovasc Dis 2016; 25:1–6.

14.

Okazawa

Yamauchi

Sugimoto

, et al. Quantitative comparison of the bolus and steady-state methods for measurement of cerebral perfusion and oxygen metabolism: positron emission tomography study using 15O-gas and water. J Cereb Blood Flow Metab 2001; 21:793–803.

15.

Lammertsma

Jones

Correction for the presence of intravascular oxygen-15 in the steady-state technique for measuring regional oxygen extraction ratio in the brain: 1. Description of the method. J Cereb Blood Flow Metab 1983; 3:416–424.

16.

Frackowiak

Lenzi

Jones

, et al. Quantitative measurement of regional cerebral blood flow and oxygen metabolism in man using 15O and positron emission tomography: theory, procedure, and normal values. J Comput Assist Tomogr 1980; 4:727–736.

17.

Ohnishi

Matsuda

Hashimoto

, et al. Abnormal regional cerebral blood flow in childhood autism. Brain 2000; 23:1838–1844.

18.

Takeuchi

Matsuda

Yoshioka

, et al. Cerebral blood flow SPET in transient global amnesia with automated ROI analysis by 3DSRT. Eur J Nucl Med Mol Imaging 2004; 31:578–589.

19.

Fan

Jahanian

Holdsworth

, et al. Comparison of cerebral blood flow measurement with [15O]-water positron emission tomography and arterial spin labeling magnetic resonance imaging: A systematic review. J Cereb Blood Flow Metab 2016; 36:842–861.

20.

Tsujikawa

Kimura

Matsuda

, et al. Arterial transit time mapping obtained by pulsed continuous 3D ASL imaging with multiple post-label delay acquisitions: comparative study with PET-CBF in patients with chronic occlusive cerebrovascular disease. PLoS One 2016; 11:e0156005.

21.

Powers

WJ.

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

22.

Yoo

Yun

Rhim

, et al. Bright vessel appearance on arterial spin labeling MRI for localizing arterial occlusion in acute ischemic stroke. Stroke 2015; 46:564–567.

23.

Shih

Saver

Alger

, et al. Perfusion-weighted magnetic resonance imaging thresholds identifying core, irreversibly infarcted tissue. Stroke 2003; 34:1425–1430.

24.

Campbell

Beaulieu

Pulsed arterial spin labeling parameter optimization for an elderly population. J Magn Reson Imaging 2006; 23:398–403.

25.

Yoshiura

Hiwatashi

Yamashita

, et al. Simultaneous measurement of arterial transit time, arterial blood volume, and cerebral blood flow using arterial spin-labeling in patients with Alzheimer disease. Am J Neuroradiol 2009; 30:1388–1393.

26.

Guo

Holdsworth

Fan

, et al. Comparing accuracy and reproducibility of sequential and Hadamard-encoded multidelay pseudocontinuous arterial spin labeling for measuring cerebral blood flow and arterial transit time in healthy subjects: A simulation and in vivo study. J Magn Reson Imaging 2018; 47:1119–1132.