Multidelay arterial spin-labeled perfusion magnetic resonance imaging in healthy individuals: A single-center experience
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.263246
Source of Support: None, Conflict of Interest: None
Purpose: To explore the optimal postlabeling delay (PLD) of arterial spin labeling (ASL) in different age groups and the correlation between cerebral blood flow (CBF) and age in adults.
Keywords: Arterial spin labeling, cerebral blood flow, postlabeling delay
Recently, arterial spin labeling (ASL) has been used in various fields, including the normal population,,, all kinds of cerebrovascular diseases,,,, and neuropsychiatric diseases,,, as a noninvasive perfusion technique. Postlabeling delay (PLD) is an important parameter of ASL and most of the researches applied single-phase PLD, except for a few studies that used multiphase PLDs. Wang et al., acquired four phases of perfusion data in patients with acute stroke. Wu et al., studied the reliability and repeatability of pseudo-continuous arterial spin labeling (pCASL) using PLDs of 1.5 and 2.5 s. Liu et al., investigated the effect of different delay times on cerebral blood flow (CBF) in patients with Alzheimer's disease using the same two PLDs as the former. Existing studies related to multiphase ASL were generally applied in patients, and the effect of different PLDs on regional CBF of normal adults has not been reported. Biagi et al., divided the subjects into children, adolescents, and adults and studied the CBF change from children to adults. The adults were not further grouped in their research; therefore, our study divided the adults into three age groups. Our study aimed to investigate the effects of different PLDs on the brain regional CBF in healthy individuals, to explore the optimal PLD of each age group, and to study the correlation between CBF and age in adults.
This research was undertaken at the Department of Radiology, the First Affiliated Hospital of Chongqing Medical University from June 2015 to November 2015. Eighty-four healthy volunteers (age range: 20–80 years, 36 male and 48 female subjects) were recruited and were divided into the following groups: youth group (27 cases, mean age 29.8 ± 4.3 years), middle-aged group (31 cases, mean age 52.2 ± 7.2 years), elderly group (26 cases, mean age 66.7 ± 7.8 years) according to the World Health Organization standards. Subjects were included in our study if they had (1) no history of trauma and stroke, (2) no history of hypertension and diabetes, (3) no anxiety, depression, and other mental disorders. Subjects were excluded if they had intracranial lesions shown in conventional magnetic resonance imaging (MRI), metal implants in vivo, claustrophobia, or other MRI contraindications. This study was approved by the Ethics Committee of our institution. All volunteers signed an informed consent form before the MRI examination. No smoking, no drinking coffee or caffeinated beverages, and no strenuous exercise within 24 h before examination was permitted.
Magnetic resonance imaging protocol
All subjects underwent a conventional MRI and perfusion scanning on a MR GE 3.0 T MR scanners (GE SignaHDxt 3.0 T) using an eight-channel head coil reception. The conventional sequences included axial T2 weighted image (T2-WI), T2 FLAIR, and three-dimensional (3D) T1-WI. The perfusion scanning was performed three times for each subject by 3D pCASL using different PLDs. The PLDs for the youth group were 1025, 1525, and 2525 ms. The PLDs for the middle-aged group were 1525, 2525, and 3025 ms. The PLDs for the elderly group were 1525, 2525, and 3025 ms. The 3D pCASL perfusion images were acquired with the following parameters: TR = 4279 ms (PLD = 1025 ms)/4521 ms (PLD = 1525 ms)/5216 ms (PLD = 2525 ms)/5451 ms (PLD = 3025 ms), TE = 9.8 ms, slice thickness = 4 mm, number of slices = 30, number of excitations = 3, field of view = 24 cm × 24 cm, matrix = 512 × 8 (3D spiral filling), acquisition time = 4:08 min (PLD = 1025 ms)/4:22 min (PLD = 1525 ms)/5:08 min (PLD = 2525 ms)/5:16 min (PLD = 3025 ms).
Postprocessing and data analysis
Data postprocessing was completed by using ADW4.5 (GE workstation), MATLAB2010b (MathWorks, Natick, MA), SPM8 (http://www.fil.ion.ucl.ac.uk/spm/software/spm8), and WFU Pick Atlas More Details (Wake Forest University, http://fmri.wfubmc.edu/cms/software).
The raw data were imported into Functool workstation (ADW4.5) to generate the CBF maps. The CBF maps were preprocessed by SPM, which is based on MATLAB, including image realignment and normalization. The 3D T1-WI were used to registrate and normalize the CBF into a standardized space (Montreal Neurological Institute template, MNI space). A predefined set of masks, including whole brain gray matter, frontal lobe, parietal lobe, temporal lobe, occipital lobe, and limbic lobe, were acquired by WFU Pickatlas [Figure 1]. Finally, the regional CBF values were extracted using the masks mentioned before.
The CBF value of each age group was expressed as x± s. Shapiro-Wilk and Levene's tests were used to test the normality and homogeneity of variance, respectively. One-way analysis of variance test was used to compare the CBF difference among different PLDs and the CBF difference among different age groups. Further multiple comparisons were completed by Scheffe. The Pearson correlation analysis was performed to investigate the correlation between CBF and age. All the statistical analysis was performed using Statistical Package for the Social Sciences (SPSS) version 20.0 (SPSS, Chicago, IL, USA).
The CBF value of each age group
The CBF values of different PLDs in the three age groups are illustrated in [Table 1]. In the youth group, global gray matter and most brain lobes (frontal lobe, parietal lobe, temporal lobe, and limbic lobe) had a higher mean CBF value when PLDs were 1025 and 1525 ms, than that for 2525 ms. In the middle-aged group and elderly group, the CBFs of global gray matter and most brain lobes (frontal lobe, parietal lobe, temporal lobe, and limbic lobe) were higher with PLD 1525 and 3025 ms than with PLD 2525 ms. In addition, the standard deviation of CBF became lower as the PLD increased in all age groups.
The effect of PLD on CBF in each age group
Considering the effect on CBF of global gray matter of different PLDs, all patients had a statistically significant difference (P < 0.05) in the three age groups. On the contrary, the CBF was different in the temporal lobe (F = 10.476, P < 0.001), occipital lobe (F = 4.578, P = 0.013), and limbic lobe (F = 11.851, P < 0.001) in the youth group [Table 2]. In the middle-aged group, the CBF was different in the frontal lobe (F = 5.250, P = 0.008), parietal lobe (F = 4.123, P = 0.019), temporal lobe (F = 6.862, P = 0.002), occipital lobe (F = 7.368, P = 0.001), and limbic lobe (F = 5.076, P = 0.008) [Table 3]. The CBF was different in the frontal lobe (F = 4.108, P = 0.023), temporal lobe (F = 5.782, P = 0.006), and occipital lobe (F = 9.910, P < 0.001) in the elderly group [Table 4]. Further, multiple comparisons showed that the CBF of global gray matter of the temporal lobe and limbic lobe had no difference when the PLDs were 1025 and 1525 ms in the youth group [Table 2]. The CBF of global gray matter and all brain lobes had no difference when the PLDs were 1525 and 3025 ms in the middle-aged group [Table 3]. The CBF of global gray matter, frontal lobe and temporal lobe had no difference when the PLDs were 1525 and 3025 ms in the elderly group [Table 4].
The difference of CBF among the three age groups
The CBF of global gray matter, frontal lobe, parietal lobe, temporal lobe, occipital lobe, and limbic lobe had statistically significant difference (P < 0.05). Further, multiple comparisons showed that the global gray matter and all brain lobes had a statistically significant difference between the youth group and the middle-aged group, as well as the youth group and the elderly group. The global gray matter and all brain lobes had no statistically significant difference between the middle-aged group and the elderly group [Table 5].
Correlation between CBF and age
The Pearson correlation analysis showed that the CBF of global gray matter (r = − 0.440, P < 0.001), frontal lobe (r= − 0.425, P < 0.001), parietal lobe (r = − 0.412, P < 0.001), temporal lobe (r =− 0.553, P < 0.001), occipital lobe (r = − 0.464, P < 0.001), and limbic lobe (r = − 0.450, P < 0.001) all had negative correlation with age [Table 6].
Three-dimensional pseudo-continuous ASL (PCASL) has a higher labeling efficiency and signal-to-noise ratio, fewer artifacts, and a higher reliability compared with the traditional continuous ASL and pulsed ASL.,, The data postprocessing was completed by SPM and WFU. This method avoided errors while manually sketching the area of interest, and the CBF obtained from this study has a high consistency compared with previous researches.,,, In a study by Hales et al., the CBF of whole brain gray matter was 50 ± 6 (mL/100 g/min) in normal people of 32 years old. In our study, the CBF of whole gray matter was 51.27 ± 6.41, 50.73 ± 5.95, 46.95 ± 4.22 (mL/100 g/min) in the youth group when the PLD was 1025, 1525, and 2525 ms, respectively.
PLD refers to the time from the blood labeling to the image acquisition, whereas ATT (arterial transit time) refers to the time taken for arterial blood flowing from the labeling level to the acquisition level. When PLD equals to ATT, the perfusion value is accurate. When PLD is less than ATT, the CBF value will underestimate the true perfusion for labeled blood that has not totally reached the imaging level. When PLD is longer than ATT, the CBF value will underestimate the true perfusion for excessive relaxation after labeling. The inappropriate PLD may falsely suggest a reduced or increased CBF. Therefore, the use of multi-PLD ASL sequences may overcome this methodological shortcoming. Wang et al., used multidelay (1.5, 2, 2.5, and 3 s) ASL and found that multiphase PLDs can assess the collateral perfusion information in patients with acute ischemic stroke. Liu et al., studied the effect of delay time on CBF in patients with Alzheimer's disease using two PLDs (1.5 and 2.5 s), and found that there was no difference between the two PLDs. Wu et al., studied the CBF of eight young people with two PLDs (1.5 and 2.5 s) and recommended that the longer PLD be used to measure the CBF for a higher repeatability. The PLD used in our study included 1025, 1525, 2525, 3025 ms. Considering that blood flows fast in the youth, 1025 ms was used in the youth group and 3025 ms was used in the middle-aged and the elderly groups. The result showed that most of the brain lobes had a higher mean CBF value when PLDs were 1025 and 1525 ms in the youth group, and the CBF of global gray matter as well as the temporal lobe and limbic lobe showed no difference between these two PLDs. In the youth group, the reason of a lower CBF with PLD 2525 ms may be excessive arterial relaxation. In the middle-aged and elderly-aged group, the mean CBF of global gray matter and most of the brain lobes were higher when PLDs were 3025 and 1025 ms, and the CBF showed no difference between these two PLDs. The CBF with PLD 2525 ms was lower than with 3025 ms; it may be due to the shorter delay time that had underestimated the perfusion. The CBF values were higher when the PLD was 1025 ms compared with 2525 ms in those two age groups; it may due to lots of labeled blood in the cerebrovascular compartment that resulted in an overestimation of the perfusion. Our study also found that the standard deviation of CBF in global gray matter and all brain lobes decreased as the delay time increased. These results indicated that the shorter PLD is suitable for the youth group, and the longer PLD is suitable for the middle-aged group and elderly group. 1525 ms is the best PLD for the youth, and 3025 ms is the best PLD for the middle-aged group and the elderly group, considering the smaller variation of CBF with longer PLD.
Our results indicated that the CBF of global gray matter and all brain lobes had a negative correlation with age, and the CBF of the elderly was lower than that of the youth. This was roughly consistent with the findings of the previous studies. The multiple comparison of different age groups showed that all brain lobes had a statistically significant difference between the youth group and the middle-aged group, and all brain lobes had no statistically significant difference between the middle-aged group and the elderly group. The reason may be that the CBF decreased with age and gradually reached a plateau after middle age (>44 years). Biagi's studies showed that CBF continued to increase during 4–10 years, decreased rapidly during adolescence, and reached a plateau at 25–30 years. Our study revealed that the plateau was attained earlier in a younger population than that seen in their study. In Biagi's research, the numbers of adult patients were relatively small (21 cases) and these were not further grouped. These limitations may have led to CBF changes in the adults that were not observed. The plateau obtained in younger patients in their study may also have the same explanation. Our results on age-related changes in CBF were highly consistent with Soni's study. However, in our study, adults were further divided into the youth (20–44 years), the middle-aged (45–59 years) and the elderly groups (60-80 years). We found no significant difference between the middle-aged group and the elderly group while examining all brain lobes. In the Pearson correlation analysis, the absolute value of r in our study was relatively small compared with their study. This may have been due to the fact that our study did not include children and adolescents, and many previous researches have shown that CBF tends to decrease obviously with age in children and adolescents. At the same time, we used the three-phases of PLDs, and got the optimal delay time in different age groups. The longer PLD is suitable for the middle-aged group and the elderly group, and the shorter PLD may lead to an underestimation of CBF. In their study, only one PLD was used, perhaps resulting in a lower CBF value in the elderly than the actual value, which may have led to a greater absolute value of r in their study. In addition, our study showed that 1525 ms is the best PLD for youth, and 3025 ms is the best PLD for the middle-aged group and elderly group. We used the CBF of the corresponding delay time in each age group for analysis. The result of our study is perhaps more accurate compared with previous studies., At the same time, in our study, there were no gender differences and no differences between the right and left lobes of the same volunteer.
A limitation of our study was that children and adolescents were not included in the research. In order to obtain a more complete conclusion, the next part of our study is to explore the best delay time of children and adolescents.
Our research demonstrates that PLD was an important parameter of 3D-ASL and the most optimal PLD was different among each age group. The shorter PLD is suitable for the young people (20–44 years) and 1525 ms is the best PLD. The longer PLD is suitable for the middle-aged group and elderly people (≥45 years) and 3025 ms is the best PLD. In most regions of the brain, the regional CBF decreased with age; the trend gradually flattens with aging and reached a plateau after middle age (>44 years).
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]