ORIGINAL RESEARCH
Feasibility of ASL Spinal Bone Marrow Perfusion Imaging With Optimized Inversion Time Dong Xing, MD,1 Yunfei Zha, MD, PhD,1* Liyong Yan, MD,1 Kejun Wang, MD,1 Wei Gong, MD,1 and Hui Lin, MS2 Purpose: To assess the correlation between flow-sensitive alternating inversion recovery (FAIR) and dynamic contrastenhanced magnetic resonance imaging (DCE-MRI) in the measurement of spinal bone marrow (SBM) perfusion; in addition, to assess for an optimized inversion time (TI) as well as the reproducibility of SBM FAIR perfusion. Materials and Methods: The optimized TI of a FAIR SBM perfusion experiment was carried out on 14 volunteers; two adjacent vertebral bodies were selected from each volunteer to measure the change of signal intensity (DM) and the signal-to-noise ratio (SNR) of FAIR perfusion MRI with five different TIs. Then, reproducibility of FAIR data from 10 volunteers was assessed by the reposition SBM FAIR experiments. Finally, FAIR and DCE-MRI were performed on 27 subjects. The correlation between the blood flow on FAIR (BFASL) and perfusion-related parameters on DCE-MRI was evaluated. Results: The maximum value of DM and SNR were 36.39 6 12.53 and 2.38 6 0.97, respectively; both were obtained when TI was near 1200 msec. There were no significant difference between the two successive measurements of SBM BFASL perfusion (P 5 0.879), and the within-subject coefficients of variation (wCV) of the measurements was 3.28%. The BFASL showed a close correlation with Ktrans (P < 0.001) and Kep (P 5 0.004), and no correlation with Ve (P 5 0.082) was found. Conclusion: 1200 msec was the optimal TI for the SBM ASL perfusion image, which led to the maximum DM and a good quality perfusion image. The SBM FAIR perfusion scan protocol has good reproducibility, and as blood flow measurement on FAIR is reliable and closely related with the parameters on DCE-MRI, FAIR is feasible for measuring SBM blood flow. J. MAGN. RESON. IMAGING 2015;42:1314–1320.
V
ertebrae have a high frequency of predilection sites for degeneration, metastases, bone marrow infiltrates, and other diseases 1,2; those diseases can lead to serious harm to the health of patients. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), which can quantitatively evaluate microvascular structure and function utilizing the information about blood flow, permeability, and perfusion of the microcirculation, has been widely applied to investigate spinal bone marrow perfusion with degeneration and to discriminate between benign and malignant lesions of the musculoskeletal system, and has proven to be a powerful tool in monitoring tumor response to chemotherapy and antiangiogenic therapy.1–5 Arterial spin labeling (ASL) uses magnetically labeled blood as an endogenous tracer, and is another technique
for quantification of tissue perfusion.6,7 This simple and reproducible technique requires no contrast administration, has a preliminary application in the musculoskeletal system, and its research scope covers assessing the perfusion of skeletal muscle, bone, and joint, and monitoring the response to antiangiogenic therapy in patients with multiple myeloma.8–13 Thus, we hypothesized that ASL can be used to assess the perfusion of spinal bone marrow (SBM) as well.
MATERIALS AND METHODS Subjects This study was approved by the Institutional Review Board. Subjects with BMI (body mass index) >27 kg/m2, with hematologic, diabetes mellitus, or other systemic disorders, underlying malignant
View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.24891 Received Dec 24, 2014, Accepted for publication Mar 8, 2015. *Address reprint requests to: Y.Z., 99, Zhangzhidong Rd, Wuhan, 430060, P.R. China. E-mail: [email protected] From the 1Department of Radiology, Renmin Hospital of Wuhan University, Wuhan, Hubei, China; and 2MR Research, GE Healthcare China, Shanghai, China
1314
C 2015 Wiley Periodicals, Inc. V
Xing et al.: ASL Perfusion of Spinal Bone Marrow
FIGURE 1: ASL perfusion of spinal bone marrow. The signal intensity–acquisitions curves (A–E) and the perfusion images (F–J) of five inversion times (TIs).
disease, radiographic findings of fractures or degeneration of the lumbar spine, or recent trauma were excluded. Fifty-six volunteers were recruited in this work. Among them, 14 healthy volunteers (mean age, 26.28 years 6 3.63 [SD]; age range, 25–34 years) including 12 men and 2 women, were enrolled for the inversion time (TI) optimization study. Ten healthy volunteers (men; mean age, 41.67 years 6 11.88 [SD]; age range, 26–52 years) were recruited for ASL a test–retest reproducibility study. For the comparison study, 32 volunteers were recruited at first, among which five volunteers were excluded due to the motion artifacts. Therefore, 27 healthy volunteers (mean age, 52.22 years 6 17.28 [SD]; age range, 25–75 years) including 12 men and 15 women underwent back-to-back ASL and DCE-MRI successfully. All subjects signed written informed consent.
MRI All MRI examinations were performed at 3T (Signa HDxt, GE Healthcare, Milwaukee, WI) with a spine-array coil. Conventional MRI including sagittal T1-FLAIR (TR 5 2000 msec, TE 5 8.4/Ef, TI 5 1080 msec, ETL 5 8) and FRFSE T2WI (TR 5 2140 msec, TE 5 120 msec, ETL 5 25) sequence with thickness of 4.8 mm, gap of 2.4 mm, and field of view (FOV) of 320 3 320 mm were scanned. The FAIR-EPI (echo-planar imaging) sequence (TR 5 800 msec, TE mini, FOV 5 300 3 300 mm, matrix 5 128 3 96, NEX 5 1, average 5 32 [16 pairs of images], slice thickness 5 5 mm, gap 5 1.5 mm) was chosen for ASL SBM perfusion, and FAIR images were acquired in the axial plane with one slice placed in the middle portion of the each lumbar vertebral body. To ascertain the optimal TI value, FAIR with five different TI values (800, 1000, 1200, 1400, 1600 msec) were performed on two adjacent lumbar bodies orderly selected from L1–L5; 28 vertebral bodies were eventually included for TI comparison, and the results will be used in subsequent experiments. Reproducibility of FAIR data was assessed by a test–retest experiment. After the first scan, subjects left the scanner, then returned to it and were rescanned in the same position and location as far as possible, and one vertebral body (L4 or L5) was chosen from each subject. November 2015
In the correlation study, two adjacent lumbar vertebral bodies were selected for FAIR imaging as mentioned above, and DCE-MRI experiments were performed. Sagittal DCE-MR images were obtained with a LAVA-XV (3D-SGRE-T1WI) sequence (TR 5 2.7 msec, TE 5 1.2 msec, flip angle 5 10 , FOV 5 320 3 320 mm, matrix 5 256 3 160, slab thickness 5 5 mm, gap 5 2.4 mm) then a bolus injection of 0.2 mmol/kg of Gd-BOPTA (Multihance, BRACCOSINE) by means of an automatic injector at a rate of 3 mL/s followed by a 15-mL saline flush. Each slab acquisition took 7 seconds, and the whole DCE-MRI lasted for 105 seconds.
Data Processing The axial FAIR raw images of SBM were analyzed with Functool with an AW 4.4 workstation. Each vertebral body was covered by one region of interest (ROI) (sizes ranged form 30 to 95 pixels, an average of 63 pixels), and care was taken to make sure the ROI did not exceed the border of the lumbar body. The average signal intensity curve and perfusion map of SBM were drawn with Functool automatically (Fig. 1). For each vertebral body ROI, the signal intensities of both control (Mcon) and tagging images (Mtag) were taken from the average signal intensity curve, and the BFASL of SBM was taken from the perfusion map. The signal intensity change (DM) was calculated according to Eq. [1],14 and the SNRs of SBM FAIR images were calculated as in Eq. [2].15 SDMcon-noise is the standard deviation of the background noise, and the ROI of noise (sizes ranged from 190 to 197 pixels, an average of 192 pixels) was set in the lower right corner of the unorganized background area of the control image.15 The data were divided into five groups according to five different TIs. FAIR DATA PROCESSING.
DM 5 Mcon 2Mtag SNR 5 Mcon 2Mtag =SDMcon2noise
Eq. [1] Eq. [2]
Quantification of the BFASL was performed using Standard Kinetic Mode,14 as follows: 1315
Journal of Magnetic Resonance Imaging
FIGURE 2: DCE-MRI: The Ktrans, Kep, and Ve map of L2.
DM ðtÞ50 52aM0a f ðt2dtÞe
2t=T1a
0 < t < dt qðtÞ
dt < t < dt1s
52aM0a f se 2t=T1a qðtÞ
Eq. [3]
t > dt1s
where M0a is the steady state equilibrium value of arterial blood; f is blood flow; T1a 5 longitudinal relaxation time of arterial blood; a is the tagging efficiency, which means the fraction of the maximum change in the longitudinal magnetization that was achieved; transit time (dt) and trail time (s) are the times required for the head and tail arrival the imaging level, respectively. The above processes were repeated once, and the average value should be taken as the final data. We measured the DCE perfusion by drawing an ROI encompassing the entire vertebral bodies manually with Cine tool software at the same vertebral level with FAIR imaging (Fig. 2). Three quantitative parameters (Kep, Ktrans, Ve) were calculated with the bicompartmental Toft model,16,17 as follows: Kep 5 Ktrans / Ve. Ktrans is the volume transfer constant, kep is the flux rate constant between extravascular extracellular space (EES) and plasma, and Ve represents the volume of EES per unit volume of tissue. Those parameters were measured twice and the average values were taken. DCE DATA PROCESSING.
Statistical Analysis Statistical analysis was performed with SPSS Statistics 17.0 software (Chicago, IL) with a significance level of 0.05.
First, the mean and SD of DM and SNR of each TI group were calculated. Second, in order to evaluate the test-retest reproducibility, paired t-test was used to describe the difference in BFASL between the first and the second experiments. Third, the withinsubject coefficient of variation was also calculated for the BFASL as follows 18,19: To begin with, three statistical assumptions should be satisfied, 1) the difference of BFASL between the repeated examinations are normally distributed (tested by boxplots and the Shapiro– Wilk W-test); 2) the difference value of BFASL between examinations are independent of the mean value of it (tested by Kendall’s rank correlation coefficient); 3) the distribution of BFASL between examinations are comparable (tested by Wilcoxon’s signed ranks test). Then several parameters were calculated: 1) the squared root of the mean squared difference (dSD), is the standard deviation of the difference between test–retest measurements; 2) the withinsubject SD (wSD) 5 dSD/冑2; 3) repeatability 5 2.77 3 wSD; 4) within-subject coefficient of variation (wCV) 19 5 wSD/subject mean BFASL. Finally, the correlation of the BFASL to the DCE perfusion parameters were evaluated using the two-tailed Pearson correlation coefficient.
RESULTS TI Optimization The values of DM and SNR of each TI group are summarized in Table 1 and the curves of the mean value of DM and SNR response to TIs are illustrated in Fig. 3. We learned from the curves that when TI (msec) was 800, 1000, 1200, 1400, 1600 successively, the mean value of DM and SNR increases first and then decreases; the peak value was reached with a TI near 1200 msec. Reproducibility Analysis The repeat measurements of BFASL (ml/100 g/min, mean 6 SD) of SBM were 108.936 6 4.607 and 109.186 6 4.617 respectively, and no significant difference between them was found (t 5 –0.157, P 5 0.879). For the reposition experiment, the BFASL of SBM demonstrated a wCV of 3.28% (Table 2). FAIR vs. DCE-MRI (Correlation Analysis) In this part, data collected in five subjects were excluded because of aortic motion artifacts, body motion artifacts, or
TABLE 1. Signal Intensity Change (DM) and SNR of Spinal Bone Marrow With Different Inversion Time (TI) (Mean 6 SD)
TI (msec) 800 28
1000 28
1200 28
1400 28
1600 28
DM
20.76 6 9.04
28.97 6 10.91
36.39 6 12.53
26.17 6 10.14
23.82 6 11.53
SNR
2.02 6 1.02
2.31 6 0.82
2.38 6 0.97
2.26 6 0.77
1.95 6 0.71
Sample size
1316
Volume 42, No. 5
Xing et al.: ASL Perfusion of Spinal Bone Marrow
FIGURE 3: The curves of the mean value of DM and SNR response to TIs: With the increment of TI from 800 msec to 1600 msec, DM and SNR both increased first and then decreased; the maximum value of both happened at TI approximately equal to 1200 msec.
both. The range and mean 6 SD values of BFASL, Ktrans, Kep, and Ve calculated from 54 vertebral bodies are represented in Table 3. Correlation analysis showed that there was a strong positive correlation between the BFASL and Ktrans (r 5 0.646, P < 0.001) and a weak positive correlation with Kep (r 5 0.387, P 5 0.004), but no correlation was observed with Ve (r 5 0.239, P 5 0.082) (Figs. 4, 5).
DISCUSSION The accuracy of the BFASL calculation is dependent on the appropriate selection of a TI value, which leads to not only
a better perfusion image quality, but also a perfect match of the transit time.20,21 Transit time, dt, is thought to be the major source of error in the quantitative estimate of perfusion due to an improper TI.21 Thus, considering that dt differs according to the histology and hemodynamics of the tested tissue,22–25 it is highly necessary to ascertain the optimal TI which matches the dt of SBM the best. Our results showed that the magnitude of DM and SNR of ASL SBM perfusion imaging changes along with the different value of TI, and that when a TI of 1200 msec was used the mean magnitude of both reached a peak. In our research, the five sequences with different TIs were performed consecutively in a short period of time; thus, the five sequences with different TIs roughly have the same noise intensity and the value of SNR differs mainly according to the change of ASL intensity (ie, signal intensity change DM). DM represents the signal difference between the labeling and control images in specified ROIs; its magnitude depends on the inflowing of labeled protons, which are restricted by three parameters: inversion time (TI), transit time (dt), and trail time (s). TI represents the delay between labeling and image acquisition, which allowed the labeled blood spins to reach the capillaries to exchange with tissue water and thereby bear the perfusion signal change.6 According to the general kinetic model proposed by Buxton et al, at the short TIs (TI < dt) no DM can be detected, as there is no arrival of labeled blood; when TI is between dt and dt1s, DM increases along with the tracer inflow; and
TABLE 2. Reproducibility of BF Evaluated In Lumbar Bone Marrow
Parameter
Lumbar bone marrow
BF Means 6 SD (ml100g21min21) The former experiment (N 5 10)
108.94 6 4.61
The later experiment (N 5 10)
109.19 6 4.62
Global measurements (N 5 20)
109.06 6 4.49
Assessment of model assumptions Shapiro-Wilk Kendall’s s
a
P 5 0.87
b
P 5 0.97
Wilcoxon’s signed rank test
c
P 5 0.65
Reproducibility analysis dSD (ml100g21min21) wSD (ml100g
21
min
5.05
21
)
3.57 21
Repeatability(a 5 0.05) (ml100g wsCV(%)
min
21
)
9.89 3.28
a
The original data demonstrated an abnormal distribution. A significant increase in difference between measurements as the magnitude of the CBF measurement increased. c A significant difference in distributions between the initial and 1-week measurements. b
November 2015
1317
Journal of Magnetic Resonance Imaging
TABLE 3. Perfusion Parameter of ASL and DCE-MRI Obtained in Lumbar Bone Marrow
Perfusion parameter BFASL (ml/100g/min) trans
K
(min
Kep (min Ve (%)
21
)
21
)
Sample size
Mean 6 SD
Range
54
124.457 6 8.772
102.420145.825
54
0.387 6 0.155
0.0770.782
54
1.662 6 0.494
0.2762.915
54
0.244 6 0.131
0.1000.621
the DM would decrease when TI exceeds dt1s as a result of washout of the labeled spins and T1 relaxation.14,20,22,23 Good reproducibility is an essential element to ensure measurement accuracy and is the prerequisite to judge the response of surgery, drugs, and other interventions.26 The reproducibility of ASL perfusion imaging is influenced by three aspects 18,19,27: physiologic fluctuation, scanning and postprocessing technique, and environmental variation. For test–retest experiments, the longer the time span between experiments, the greater the variation of them.18,28 Thus, Floyd et al,18 in a reproducibility analysis of CBF measurements using the CASL sequence, found that for the whole brain CBF case, the result of a 1-hour frame experiment indicated a CBF wCV of 5.8%, while the CBF wCV of a 1week frame increased to 13%. The paired t-test showed that there was no statistically significant difference between the former and latter measurement of BFASL of SBM and illustrate good agreement between them. Also, the BFASL of SBM demonstrated a wsCV of 3.28%, slightly less than the wCV of whole brain in a 1-hour frame (5.8%) measured by Floyd et al.18 This is mainly because the time interval of our experiments was shorter, which effectively reduced physiologic and environmental variation, and this also demonstrated that the scanning and postprocessing technique itself has good reproducibility.
Several studies had demonstrated the feasibility of ASL for measuring the perfusion of brain, lung, kidney, prostate, and skeletal muscle, etc., and proved the correlation between the BFASL and perfusion-related parameters of other perfusion technologies.7,15,29–32 Our results found correlations between ASL and DCE measurements in the SBM, and indicated the feasibility of ASL for measuring the SBM blood flow. The correlations of BFASL to Ktrans and Kep have different interpretations. The BFASL means the volume of blood flow per 100 g of tissue per minute, while the explanation of Ktrans depends on the physiologic situation of tissue.17 In high permeability situations, Ktrans means the blood flow per unit volume of tissue, which is determined by the flow; and in low limited situations, Ktrans means the surface area product between blood plasma and the EES, which is limited by capillary permeability. The anatomical basis of bone marrow blood flow is sinusoid, which only has a single layer of endothelial cells and high permeability. In addition to the high permeability of sinusoid, Cho et al 29 thought that when using low-molecular contrast material as the contrast agent, the contrast agent can freely perfuse through the vascular endothelial gaps, and the vascular permeability is mostly determined by the blood flow. Thus, Ktrans of SBM equal to the blood flow per unit volume of SBM, and for the same reason, Kep, which represents the
FIGURE 4: Graph shows a positive correlation between blood flow (BFASL) and Ktrans (Pearson r 5 0.646, P < 0.001).
FIGURE 5: Graph shows a positive correlation between blood flow (BFASL) and Kep (Pearson r 5 0.387, P 5 0.004).
1318
Volume 42, No. 5
Xing et al.: ASL Perfusion of Spinal Bone Marrow
permeability from EES to intravascular space, is restricted by the blood flow. Our study has two limitations. First, perfusion data were acquired in a different plane, the sagittal plane for DCE-MRI and the axial plane for ASL. For SBM DCEMRI perfusion imaging, both the axial and sagittal image acquisition methods are commonly used. Theoretically, the plane of data acquisition should not affect the perfusion parameters derived or measured, and former studies 5,33,34 revealed identical trends in using different planes of vertebral perfusion, and no significant effects were found among the sagittal or axial plane perfusion imaging. Thus, we believe that the differences generated by different scanning planes between ASL and DCE-MRI imaging should not have influenced the results. Second, the specificity and sensitivity of SBM FAIR perfusion imaging on benign and malignant lesions of the spine need further studies and need to be confirmed by pathophysiology or other perfusion technologies. In conclusion, when TI 5 1200 msec the SBM ASL perfusion imaging has the maximum value of DM and SNR, and the measurement of BFASL of SBM has good reproducibility. Moreover, the correlation of ASL and DCEMRI suggested that the ASL MRI technology is a promising method in spinal bone marrow perfusion evaluation.
with chronic exertional compartment syndrome: preliminary experience. AJR Am J Roentgenol 2009;193:W327–W333. 11.
Fenchel M, Konaktchieva M, Weisel K, et al. Early response assessment in patients with multiple myeloma during anti-angiogenic therapy using arterial spin labeling: first clinical results. Eur Radiol 2010; 20:2899–2906.
12.
Wu WC, Mohler ER, Ratcliffe SJ, Wehrli FW, Detre JA, Floyd TF. Skeletal muscle microvascular flow in progressive peripheral artery disease: assessment with continuous arterial spin-labeling perfusion magnetic resonance imaging. J Am Coll Cardiol 2009;53:2372–2377.
13.
West A, Meyer C, Epstein F, et al. Arterial spin labeling MRI to measure peak-exercise calf muscle perfusion reproducibly discriminates peripheral arterial disease from normal. J Cardiov Magn Reson 2011; 13(Suppl 1):P347.
14.
Buxton RB, Frank LR, Wong EC, Siewert B, Warach S, Edelman RR. A general kinetic model for quantitative perfusion imaging with arterial spin labeling. Magn Reson Med 1998;40:383–396.
15.
Fan L, Liu SY, Sun F, Xiao XS. Assessment of pulmonary parenchyma perfusion with FAIR in comparison with DCE-MRI—initial results. Eur J Radiol 2009;70:41–48.
16.
Paldino MJ, Barboriak DP. Fundamentals of quantitative dynamic contrast-enhanced MR imaging. Magn Reson Imaging Clin N Am 2009;17:277–289.
17.
Tofts PS, Brix G, Buckley DL, et al. Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging 1999;10: 223–232.
18.
Floyd TF, Ratcliffe SJ, Wang J, Resch B, Detre JA. Precision of the CASL-perfusion MRI technique for the measurement of cerebral blood flow in whole brain and vascular territories. J Magn Reson Imaging 2003;18:649–655.
19.
Padhani AR, Hayes C, Landau S, Leach MO. Reproducibility of quantitative dynamic MRI of normal human tissues. NMR Biomed 2002;15: 143–153.
20.
Hendrikse J, van Osch MJ, Rutgers DR, et al. Internal carotid artery occlusion assessed at pulsed arterial spin-labeling perfusion MR imaging at multiple delay times. Radiology 2004;233:899–904.
21.
Petersen ET, Zimine I, Ho YC, Golay X. Non-invasive measurement of perfusion: a critical review of arterial spin labelling techniques. Br J Radiol 2006;79:688–701.
22.
Yang Y, Engelien W, Xu S, Gu H, Silbersweig DA, Stern E. Transit time, trailing time, and cerebral blood flow during brain activation: measurement using multislice, pulsed spin-labeling perfusion imaging. Magn Reson Med 2000;44:680–685.
23.
Gonzalez-At JB, Alsop DC, Detre JA. Cerebral perfusion and arterial transit time changes during task activation determined with continuous arterial spin labeling. Magn Reson Med 2000;43:739–746.
24.
Thomas DL, Lythgoe MF, van der Weerd L, Ordidge RJ, Gadian DG. 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 2006;26: 274–282.
25.
Qiu M, Paul MR, Arora J, 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.
26.
Goh V, Schaeffter T, Leach M. Reproducibility of dynamic contrastenhanced MR imaging: why we should care. Radiology 2013;266: 698–700.
27.
Yen YF, Field AS, Martin EM, et al. Test-retest reproducibility of quantitative CBF measurements using FAIR perfusion MRI and acetazolamide challenge. Magn Reson Med 2002;47:921–928.
28.
Parkes LM, Rashid W, Chard DT, Tofts PS. Normal cerebral perfusion measurements using arterial spin labeling: reproducibility, stability, and age and gender effects. Magn Reson Med 2004;51: 736–743.
REFERENCES 1.
Zha Y, Li M, Yang J. Dynamic contrast enhanced magnetic resonance imaging of diffuse spinal bone marrow infiltration in patients with hematological malignancies. Korean J Radiol 2010;11:187–194.
2.
Khadem NR, Karimi S, Peck KK, et al. Characterizing hypervascular and hypovascular metastases and normal bone marrow of the spine using dynamic contrast-enhanced MR imaging. AJNR Am J Neuroradiol 2012;33:2178–2185.
3.
Biffar A, Dietrich O, Sourbron S, Duerr HR, Reiser MF, Baur-Melnyk A. Diffusion and perfusion imaging of bone marrow. Eur J Radiol 2010; 76:323–328.
4.
Chen BB, Hsu CY, Yu CW, et al. Dynamic contrast-enhanced MR imaging measurement of vertebral bone marrow perfusion may be indicator of outcome of acute myeloid leukemia patients in remission. Radiology 2011;258:821–831.
5.
Ma HT, Griffith JF, Yeung DK, Leung PC. Modified brix model analysis of bone perfusion in subjects of varying bone mineral density. J Magn Reson Imaging 2010;31:1169–1175.
6.
Martirosian P, Boss A, Schraml C, et al. Magnetic resonance perfusion imaging without contrast media. Eur J Nucl Med Mol Imaging 2010; 37(Suppl 1):S52–S64.
7.
Wu WC, Su MY, Chang CC, Tseng WY, Liu KL. Renal perfusion 3-T MR imaging: a comparative study of arterial spin labeling and dynamic contrast-enhanced techniques. Radiology 2011;261:845–853.
8.
Fenchel M, Konaktchieva M, Weisel K, Kraus S, Claussen CD, Horger M. Response assessment in patients with multiple myeloma during antiangiogenic therapy using arterial spin labeling and diffusionweighted imaging: a feasibility study. Acad Radiol 2010;17:1326– 1333.
9.
Spira D, Kotter I, Henes J, et al. MRI findings in psoriatic arthritis of the hands. AJR Am J Roentgenol 2010;195:1187–1193.
10.
Andreisek G, White LM, Sussman MS, et al. T2*-weighted and arterial spin labeling MRI of calf muscles in healthy volunteers and patients
November 2015
1319
Journal of Magnetic Resonance Imaging 29.
Cho JH, Cho G, Song Y, et al. Feasibility of FAIR imaging for evaluating tumor perfusion. J Magn Reson Imaging 2010;32: 738–744.
32.
Cai W, Li F, Wang J, et al. A comparison of arterial spin labeling perfusion MRI and DCE-MRI in human prostate cancer. NMR Biomed 2014;27:817–825.
30.
Buxton RB. Quantifying CBF with arterial spin labeling. J Magn Reson Imaging 2005;22:723–726.
33.
31.
Ludemann L, Jedrzejewski G, Heidenreich J, Han ET, Bruhn H. Perfusion imaging of cerebral arteriovenous malformations: a study comparing quantitative continuous arterial spin labeling and dynamic contrast-enhanced magnetic resonance imaging at 3 T. Magn Reson Imaging 2011;29:1157–1164.
Griffith JF, Yeung DK, Antonio GE, et al. Vertebral bone mineral density, marrow perfusion, and fat content in healthy men and men with osteoporosis: dynamic contrast-enhanced MR imaging and MR spectroscopy. Radiology 2005;236:945–951.
34.
Griffith JF, Yeung DK, Antonio GE, et al. Vertebral marrow fat content and diffusion and perfusion indexes in women with varying bone density: MR evaluation. Radiology 2006;241:831–838.
1320
Volume 42, No. 5
Feasibility of ASL Spinal Bone Marrow Perfusion Imaging With Optimized Inversion Time Dong Xing, MD,1 Yunfei Zha, MD, PhD,1* Liyong Yan, MD,1 Kejun Wang, MD,1 Wei Gong, MD,1 and Hui Lin, MS2 Purpose: To assess the correlation between flow-sensitive alternating inversion recovery (FAIR) and dynamic contrastenhanced magnetic resonance imaging (DCE-MRI) in the measurement of spinal bone marrow (SBM) perfusion; in addition, to assess for an optimized inversion time (TI) as well as the reproducibility of SBM FAIR perfusion. Materials and Methods: The optimized TI of a FAIR SBM perfusion experiment was carried out on 14 volunteers; two adjacent vertebral bodies were selected from each volunteer to measure the change of signal intensity (DM) and the signal-to-noise ratio (SNR) of FAIR perfusion MRI with five different TIs. Then, reproducibility of FAIR data from 10 volunteers was assessed by the reposition SBM FAIR experiments. Finally, FAIR and DCE-MRI were performed on 27 subjects. The correlation between the blood flow on FAIR (BFASL) and perfusion-related parameters on DCE-MRI was evaluated. Results: The maximum value of DM and SNR were 36.39 6 12.53 and 2.38 6 0.97, respectively; both were obtained when TI was near 1200 msec. There were no significant difference between the two successive measurements of SBM BFASL perfusion (P 5 0.879), and the within-subject coefficients of variation (wCV) of the measurements was 3.28%. The BFASL showed a close correlation with Ktrans (P < 0.001) and Kep (P 5 0.004), and no correlation with Ve (P 5 0.082) was found. Conclusion: 1200 msec was the optimal TI for the SBM ASL perfusion image, which led to the maximum DM and a good quality perfusion image. The SBM FAIR perfusion scan protocol has good reproducibility, and as blood flow measurement on FAIR is reliable and closely related with the parameters on DCE-MRI, FAIR is feasible for measuring SBM blood flow. J. MAGN. RESON. IMAGING 2015;42:1314–1320.
V
ertebrae have a high frequency of predilection sites for degeneration, metastases, bone marrow infiltrates, and other diseases 1,2; those diseases can lead to serious harm to the health of patients. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), which can quantitatively evaluate microvascular structure and function utilizing the information about blood flow, permeability, and perfusion of the microcirculation, has been widely applied to investigate spinal bone marrow perfusion with degeneration and to discriminate between benign and malignant lesions of the musculoskeletal system, and has proven to be a powerful tool in monitoring tumor response to chemotherapy and antiangiogenic therapy.1–5 Arterial spin labeling (ASL) uses magnetically labeled blood as an endogenous tracer, and is another technique
for quantification of tissue perfusion.6,7 This simple and reproducible technique requires no contrast administration, has a preliminary application in the musculoskeletal system, and its research scope covers assessing the perfusion of skeletal muscle, bone, and joint, and monitoring the response to antiangiogenic therapy in patients with multiple myeloma.8–13 Thus, we hypothesized that ASL can be used to assess the perfusion of spinal bone marrow (SBM) as well.
MATERIALS AND METHODS Subjects This study was approved by the Institutional Review Board. Subjects with BMI (body mass index) >27 kg/m2, with hematologic, diabetes mellitus, or other systemic disorders, underlying malignant
View this article online at wileyonlinelibrary.com. DOI: 10.1002/jmri.24891 Received Dec 24, 2014, Accepted for publication Mar 8, 2015. *Address reprint requests to: Y.Z., 99, Zhangzhidong Rd, Wuhan, 430060, P.R. China. E-mail: [email protected] From the 1Department of Radiology, Renmin Hospital of Wuhan University, Wuhan, Hubei, China; and 2MR Research, GE Healthcare China, Shanghai, China
1314
C 2015 Wiley Periodicals, Inc. V
Xing et al.: ASL Perfusion of Spinal Bone Marrow
FIGURE 1: ASL perfusion of spinal bone marrow. The signal intensity–acquisitions curves (A–E) and the perfusion images (F–J) of five inversion times (TIs).
disease, radiographic findings of fractures or degeneration of the lumbar spine, or recent trauma were excluded. Fifty-six volunteers were recruited in this work. Among them, 14 healthy volunteers (mean age, 26.28 years 6 3.63 [SD]; age range, 25–34 years) including 12 men and 2 women, were enrolled for the inversion time (TI) optimization study. Ten healthy volunteers (men; mean age, 41.67 years 6 11.88 [SD]; age range, 26–52 years) were recruited for ASL a test–retest reproducibility study. For the comparison study, 32 volunteers were recruited at first, among which five volunteers were excluded due to the motion artifacts. Therefore, 27 healthy volunteers (mean age, 52.22 years 6 17.28 [SD]; age range, 25–75 years) including 12 men and 15 women underwent back-to-back ASL and DCE-MRI successfully. All subjects signed written informed consent.
MRI All MRI examinations were performed at 3T (Signa HDxt, GE Healthcare, Milwaukee, WI) with a spine-array coil. Conventional MRI including sagittal T1-FLAIR (TR 5 2000 msec, TE 5 8.4/Ef, TI 5 1080 msec, ETL 5 8) and FRFSE T2WI (TR 5 2140 msec, TE 5 120 msec, ETL 5 25) sequence with thickness of 4.8 mm, gap of 2.4 mm, and field of view (FOV) of 320 3 320 mm were scanned. The FAIR-EPI (echo-planar imaging) sequence (TR 5 800 msec, TE mini, FOV 5 300 3 300 mm, matrix 5 128 3 96, NEX 5 1, average 5 32 [16 pairs of images], slice thickness 5 5 mm, gap 5 1.5 mm) was chosen for ASL SBM perfusion, and FAIR images were acquired in the axial plane with one slice placed in the middle portion of the each lumbar vertebral body. To ascertain the optimal TI value, FAIR with five different TI values (800, 1000, 1200, 1400, 1600 msec) were performed on two adjacent lumbar bodies orderly selected from L1–L5; 28 vertebral bodies were eventually included for TI comparison, and the results will be used in subsequent experiments. Reproducibility of FAIR data was assessed by a test–retest experiment. After the first scan, subjects left the scanner, then returned to it and were rescanned in the same position and location as far as possible, and one vertebral body (L4 or L5) was chosen from each subject. November 2015
In the correlation study, two adjacent lumbar vertebral bodies were selected for FAIR imaging as mentioned above, and DCE-MRI experiments were performed. Sagittal DCE-MR images were obtained with a LAVA-XV (3D-SGRE-T1WI) sequence (TR 5 2.7 msec, TE 5 1.2 msec, flip angle 5 10 , FOV 5 320 3 320 mm, matrix 5 256 3 160, slab thickness 5 5 mm, gap 5 2.4 mm) then a bolus injection of 0.2 mmol/kg of Gd-BOPTA (Multihance, BRACCOSINE) by means of an automatic injector at a rate of 3 mL/s followed by a 15-mL saline flush. Each slab acquisition took 7 seconds, and the whole DCE-MRI lasted for 105 seconds.
Data Processing The axial FAIR raw images of SBM were analyzed with Functool with an AW 4.4 workstation. Each vertebral body was covered by one region of interest (ROI) (sizes ranged form 30 to 95 pixels, an average of 63 pixels), and care was taken to make sure the ROI did not exceed the border of the lumbar body. The average signal intensity curve and perfusion map of SBM were drawn with Functool automatically (Fig. 1). For each vertebral body ROI, the signal intensities of both control (Mcon) and tagging images (Mtag) were taken from the average signal intensity curve, and the BFASL of SBM was taken from the perfusion map. The signal intensity change (DM) was calculated according to Eq. [1],14 and the SNRs of SBM FAIR images were calculated as in Eq. [2].15 SDMcon-noise is the standard deviation of the background noise, and the ROI of noise (sizes ranged from 190 to 197 pixels, an average of 192 pixels) was set in the lower right corner of the unorganized background area of the control image.15 The data were divided into five groups according to five different TIs. FAIR DATA PROCESSING.
DM 5 Mcon 2Mtag SNR 5 Mcon 2Mtag =SDMcon2noise
Eq. [1] Eq. [2]
Quantification of the BFASL was performed using Standard Kinetic Mode,14 as follows: 1315
Journal of Magnetic Resonance Imaging
FIGURE 2: DCE-MRI: The Ktrans, Kep, and Ve map of L2.
DM ðtÞ50 52aM0a f ðt2dtÞe
2t=T1a
0 < t < dt qðtÞ
dt < t < dt1s
52aM0a f se 2t=T1a qðtÞ
Eq. [3]
t > dt1s
where M0a is the steady state equilibrium value of arterial blood; f is blood flow; T1a 5 longitudinal relaxation time of arterial blood; a is the tagging efficiency, which means the fraction of the maximum change in the longitudinal magnetization that was achieved; transit time (dt) and trail time (s) are the times required for the head and tail arrival the imaging level, respectively. The above processes were repeated once, and the average value should be taken as the final data. We measured the DCE perfusion by drawing an ROI encompassing the entire vertebral bodies manually with Cine tool software at the same vertebral level with FAIR imaging (Fig. 2). Three quantitative parameters (Kep, Ktrans, Ve) were calculated with the bicompartmental Toft model,16,17 as follows: Kep 5 Ktrans / Ve. Ktrans is the volume transfer constant, kep is the flux rate constant between extravascular extracellular space (EES) and plasma, and Ve represents the volume of EES per unit volume of tissue. Those parameters were measured twice and the average values were taken. DCE DATA PROCESSING.
Statistical Analysis Statistical analysis was performed with SPSS Statistics 17.0 software (Chicago, IL) with a significance level of 0.05.
First, the mean and SD of DM and SNR of each TI group were calculated. Second, in order to evaluate the test-retest reproducibility, paired t-test was used to describe the difference in BFASL between the first and the second experiments. Third, the withinsubject coefficient of variation was also calculated for the BFASL as follows 18,19: To begin with, three statistical assumptions should be satisfied, 1) the difference of BFASL between the repeated examinations are normally distributed (tested by boxplots and the Shapiro– Wilk W-test); 2) the difference value of BFASL between examinations are independent of the mean value of it (tested by Kendall’s rank correlation coefficient); 3) the distribution of BFASL between examinations are comparable (tested by Wilcoxon’s signed ranks test). Then several parameters were calculated: 1) the squared root of the mean squared difference (dSD), is the standard deviation of the difference between test–retest measurements; 2) the withinsubject SD (wSD) 5 dSD/冑2; 3) repeatability 5 2.77 3 wSD; 4) within-subject coefficient of variation (wCV) 19 5 wSD/subject mean BFASL. Finally, the correlation of the BFASL to the DCE perfusion parameters were evaluated using the two-tailed Pearson correlation coefficient.
RESULTS TI Optimization The values of DM and SNR of each TI group are summarized in Table 1 and the curves of the mean value of DM and SNR response to TIs are illustrated in Fig. 3. We learned from the curves that when TI (msec) was 800, 1000, 1200, 1400, 1600 successively, the mean value of DM and SNR increases first and then decreases; the peak value was reached with a TI near 1200 msec. Reproducibility Analysis The repeat measurements of BFASL (ml/100 g/min, mean 6 SD) of SBM were 108.936 6 4.607 and 109.186 6 4.617 respectively, and no significant difference between them was found (t 5 –0.157, P 5 0.879). For the reposition experiment, the BFASL of SBM demonstrated a wCV of 3.28% (Table 2). FAIR vs. DCE-MRI (Correlation Analysis) In this part, data collected in five subjects were excluded because of aortic motion artifacts, body motion artifacts, or
TABLE 1. Signal Intensity Change (DM) and SNR of Spinal Bone Marrow With Different Inversion Time (TI) (Mean 6 SD)
TI (msec) 800 28
1000 28
1200 28
1400 28
1600 28
DM
20.76 6 9.04
28.97 6 10.91
36.39 6 12.53
26.17 6 10.14
23.82 6 11.53
SNR
2.02 6 1.02
2.31 6 0.82
2.38 6 0.97
2.26 6 0.77
1.95 6 0.71
Sample size
1316
Volume 42, No. 5
Xing et al.: ASL Perfusion of Spinal Bone Marrow
FIGURE 3: The curves of the mean value of DM and SNR response to TIs: With the increment of TI from 800 msec to 1600 msec, DM and SNR both increased first and then decreased; the maximum value of both happened at TI approximately equal to 1200 msec.
both. The range and mean 6 SD values of BFASL, Ktrans, Kep, and Ve calculated from 54 vertebral bodies are represented in Table 3. Correlation analysis showed that there was a strong positive correlation between the BFASL and Ktrans (r 5 0.646, P < 0.001) and a weak positive correlation with Kep (r 5 0.387, P 5 0.004), but no correlation was observed with Ve (r 5 0.239, P 5 0.082) (Figs. 4, 5).
DISCUSSION The accuracy of the BFASL calculation is dependent on the appropriate selection of a TI value, which leads to not only
a better perfusion image quality, but also a perfect match of the transit time.20,21 Transit time, dt, is thought to be the major source of error in the quantitative estimate of perfusion due to an improper TI.21 Thus, considering that dt differs according to the histology and hemodynamics of the tested tissue,22–25 it is highly necessary to ascertain the optimal TI which matches the dt of SBM the best. Our results showed that the magnitude of DM and SNR of ASL SBM perfusion imaging changes along with the different value of TI, and that when a TI of 1200 msec was used the mean magnitude of both reached a peak. In our research, the five sequences with different TIs were performed consecutively in a short period of time; thus, the five sequences with different TIs roughly have the same noise intensity and the value of SNR differs mainly according to the change of ASL intensity (ie, signal intensity change DM). DM represents the signal difference between the labeling and control images in specified ROIs; its magnitude depends on the inflowing of labeled protons, which are restricted by three parameters: inversion time (TI), transit time (dt), and trail time (s). TI represents the delay between labeling and image acquisition, which allowed the labeled blood spins to reach the capillaries to exchange with tissue water and thereby bear the perfusion signal change.6 According to the general kinetic model proposed by Buxton et al, at the short TIs (TI < dt) no DM can be detected, as there is no arrival of labeled blood; when TI is between dt and dt1s, DM increases along with the tracer inflow; and
TABLE 2. Reproducibility of BF Evaluated In Lumbar Bone Marrow
Parameter
Lumbar bone marrow
BF Means 6 SD (ml100g21min21) The former experiment (N 5 10)
108.94 6 4.61
The later experiment (N 5 10)
109.19 6 4.62
Global measurements (N 5 20)
109.06 6 4.49
Assessment of model assumptions Shapiro-Wilk Kendall’s s
a
P 5 0.87
b
P 5 0.97
Wilcoxon’s signed rank test
c
P 5 0.65
Reproducibility analysis dSD (ml100g21min21) wSD (ml100g
21
min
5.05
21
)
3.57 21
Repeatability(a 5 0.05) (ml100g wsCV(%)
min
21
)
9.89 3.28
a
The original data demonstrated an abnormal distribution. A significant increase in difference between measurements as the magnitude of the CBF measurement increased. c A significant difference in distributions between the initial and 1-week measurements. b
November 2015
1317
Journal of Magnetic Resonance Imaging
TABLE 3. Perfusion Parameter of ASL and DCE-MRI Obtained in Lumbar Bone Marrow
Perfusion parameter BFASL (ml/100g/min) trans
K
(min
Kep (min Ve (%)
21
)
21
)
Sample size
Mean 6 SD
Range
54
124.457 6 8.772
102.420145.825
54
0.387 6 0.155
0.0770.782
54
1.662 6 0.494
0.2762.915
54
0.244 6 0.131
0.1000.621
the DM would decrease when TI exceeds dt1s as a result of washout of the labeled spins and T1 relaxation.14,20,22,23 Good reproducibility is an essential element to ensure measurement accuracy and is the prerequisite to judge the response of surgery, drugs, and other interventions.26 The reproducibility of ASL perfusion imaging is influenced by three aspects 18,19,27: physiologic fluctuation, scanning and postprocessing technique, and environmental variation. For test–retest experiments, the longer the time span between experiments, the greater the variation of them.18,28 Thus, Floyd et al,18 in a reproducibility analysis of CBF measurements using the CASL sequence, found that for the whole brain CBF case, the result of a 1-hour frame experiment indicated a CBF wCV of 5.8%, while the CBF wCV of a 1week frame increased to 13%. The paired t-test showed that there was no statistically significant difference between the former and latter measurement of BFASL of SBM and illustrate good agreement between them. Also, the BFASL of SBM demonstrated a wsCV of 3.28%, slightly less than the wCV of whole brain in a 1-hour frame (5.8%) measured by Floyd et al.18 This is mainly because the time interval of our experiments was shorter, which effectively reduced physiologic and environmental variation, and this also demonstrated that the scanning and postprocessing technique itself has good reproducibility.
Several studies had demonstrated the feasibility of ASL for measuring the perfusion of brain, lung, kidney, prostate, and skeletal muscle, etc., and proved the correlation between the BFASL and perfusion-related parameters of other perfusion technologies.7,15,29–32 Our results found correlations between ASL and DCE measurements in the SBM, and indicated the feasibility of ASL for measuring the SBM blood flow. The correlations of BFASL to Ktrans and Kep have different interpretations. The BFASL means the volume of blood flow per 100 g of tissue per minute, while the explanation of Ktrans depends on the physiologic situation of tissue.17 In high permeability situations, Ktrans means the blood flow per unit volume of tissue, which is determined by the flow; and in low limited situations, Ktrans means the surface area product between blood plasma and the EES, which is limited by capillary permeability. The anatomical basis of bone marrow blood flow is sinusoid, which only has a single layer of endothelial cells and high permeability. In addition to the high permeability of sinusoid, Cho et al 29 thought that when using low-molecular contrast material as the contrast agent, the contrast agent can freely perfuse through the vascular endothelial gaps, and the vascular permeability is mostly determined by the blood flow. Thus, Ktrans of SBM equal to the blood flow per unit volume of SBM, and for the same reason, Kep, which represents the
FIGURE 4: Graph shows a positive correlation between blood flow (BFASL) and Ktrans (Pearson r 5 0.646, P < 0.001).
FIGURE 5: Graph shows a positive correlation between blood flow (BFASL) and Kep (Pearson r 5 0.387, P 5 0.004).
1318
Volume 42, No. 5
Xing et al.: ASL Perfusion of Spinal Bone Marrow
permeability from EES to intravascular space, is restricted by the blood flow. Our study has two limitations. First, perfusion data were acquired in a different plane, the sagittal plane for DCE-MRI and the axial plane for ASL. For SBM DCEMRI perfusion imaging, both the axial and sagittal image acquisition methods are commonly used. Theoretically, the plane of data acquisition should not affect the perfusion parameters derived or measured, and former studies 5,33,34 revealed identical trends in using different planes of vertebral perfusion, and no significant effects were found among the sagittal or axial plane perfusion imaging. Thus, we believe that the differences generated by different scanning planes between ASL and DCE-MRI imaging should not have influenced the results. Second, the specificity and sensitivity of SBM FAIR perfusion imaging on benign and malignant lesions of the spine need further studies and need to be confirmed by pathophysiology or other perfusion technologies. In conclusion, when TI 5 1200 msec the SBM ASL perfusion imaging has the maximum value of DM and SNR, and the measurement of BFASL of SBM has good reproducibility. Moreover, the correlation of ASL and DCEMRI suggested that the ASL MRI technology is a promising method in spinal bone marrow perfusion evaluation.
with chronic exertional compartment syndrome: preliminary experience. AJR Am J Roentgenol 2009;193:W327–W333. 11.
Fenchel M, Konaktchieva M, Weisel K, et al. Early response assessment in patients with multiple myeloma during anti-angiogenic therapy using arterial spin labeling: first clinical results. Eur Radiol 2010; 20:2899–2906.
12.
Wu WC, Mohler ER, Ratcliffe SJ, Wehrli FW, Detre JA, Floyd TF. Skeletal muscle microvascular flow in progressive peripheral artery disease: assessment with continuous arterial spin-labeling perfusion magnetic resonance imaging. J Am Coll Cardiol 2009;53:2372–2377.
13.
West A, Meyer C, Epstein F, et al. Arterial spin labeling MRI to measure peak-exercise calf muscle perfusion reproducibly discriminates peripheral arterial disease from normal. J Cardiov Magn Reson 2011; 13(Suppl 1):P347.
14.
Buxton RB, Frank LR, Wong EC, Siewert B, Warach S, Edelman RR. A general kinetic model for quantitative perfusion imaging with arterial spin labeling. Magn Reson Med 1998;40:383–396.
15.
Fan L, Liu SY, Sun F, Xiao XS. Assessment of pulmonary parenchyma perfusion with FAIR in comparison with DCE-MRI—initial results. Eur J Radiol 2009;70:41–48.
16.
Paldino MJ, Barboriak DP. Fundamentals of quantitative dynamic contrast-enhanced MR imaging. Magn Reson Imaging Clin N Am 2009;17:277–289.
17.
Tofts PS, Brix G, Buckley DL, et al. Estimating kinetic parameters from dynamic contrast-enhanced T(1)-weighted MRI of a diffusable tracer: standardized quantities and symbols. J Magn Reson Imaging 1999;10: 223–232.
18.
Floyd TF, Ratcliffe SJ, Wang J, Resch B, Detre JA. Precision of the CASL-perfusion MRI technique for the measurement of cerebral blood flow in whole brain and vascular territories. J Magn Reson Imaging 2003;18:649–655.
19.
Padhani AR, Hayes C, Landau S, Leach MO. Reproducibility of quantitative dynamic MRI of normal human tissues. NMR Biomed 2002;15: 143–153.
20.
Hendrikse J, van Osch MJ, Rutgers DR, et al. Internal carotid artery occlusion assessed at pulsed arterial spin-labeling perfusion MR imaging at multiple delay times. Radiology 2004;233:899–904.
21.
Petersen ET, Zimine I, Ho YC, Golay X. Non-invasive measurement of perfusion: a critical review of arterial spin labelling techniques. Br J Radiol 2006;79:688–701.
22.
Yang Y, Engelien W, Xu S, Gu H, Silbersweig DA, Stern E. Transit time, trailing time, and cerebral blood flow during brain activation: measurement using multislice, pulsed spin-labeling perfusion imaging. Magn Reson Med 2000;44:680–685.
23.
Gonzalez-At JB, Alsop DC, Detre JA. Cerebral perfusion and arterial transit time changes during task activation determined with continuous arterial spin labeling. Magn Reson Med 2000;43:739–746.
24.
Thomas DL, Lythgoe MF, van der Weerd L, Ordidge RJ, Gadian DG. 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 2006;26: 274–282.
25.
Qiu M, Paul MR, Arora J, 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.
26.
Goh V, Schaeffter T, Leach M. Reproducibility of dynamic contrastenhanced MR imaging: why we should care. Radiology 2013;266: 698–700.
27.
Yen YF, Field AS, Martin EM, et al. Test-retest reproducibility of quantitative CBF measurements using FAIR perfusion MRI and acetazolamide challenge. Magn Reson Med 2002;47:921–928.
28.
Parkes LM, Rashid W, Chard DT, Tofts PS. Normal cerebral perfusion measurements using arterial spin labeling: reproducibility, stability, and age and gender effects. Magn Reson Med 2004;51: 736–743.
REFERENCES 1.
Zha Y, Li M, Yang J. Dynamic contrast enhanced magnetic resonance imaging of diffuse spinal bone marrow infiltration in patients with hematological malignancies. Korean J Radiol 2010;11:187–194.
2.
Khadem NR, Karimi S, Peck KK, et al. Characterizing hypervascular and hypovascular metastases and normal bone marrow of the spine using dynamic contrast-enhanced MR imaging. AJNR Am J Neuroradiol 2012;33:2178–2185.
3.
Biffar A, Dietrich O, Sourbron S, Duerr HR, Reiser MF, Baur-Melnyk A. Diffusion and perfusion imaging of bone marrow. Eur J Radiol 2010; 76:323–328.
4.
Chen BB, Hsu CY, Yu CW, et al. Dynamic contrast-enhanced MR imaging measurement of vertebral bone marrow perfusion may be indicator of outcome of acute myeloid leukemia patients in remission. Radiology 2011;258:821–831.
5.
Ma HT, Griffith JF, Yeung DK, Leung PC. Modified brix model analysis of bone perfusion in subjects of varying bone mineral density. J Magn Reson Imaging 2010;31:1169–1175.
6.
Martirosian P, Boss A, Schraml C, et al. Magnetic resonance perfusion imaging without contrast media. Eur J Nucl Med Mol Imaging 2010; 37(Suppl 1):S52–S64.
7.
Wu WC, Su MY, Chang CC, Tseng WY, Liu KL. Renal perfusion 3-T MR imaging: a comparative study of arterial spin labeling and dynamic contrast-enhanced techniques. Radiology 2011;261:845–853.
8.
Fenchel M, Konaktchieva M, Weisel K, Kraus S, Claussen CD, Horger M. Response assessment in patients with multiple myeloma during antiangiogenic therapy using arterial spin labeling and diffusionweighted imaging: a feasibility study. Acad Radiol 2010;17:1326– 1333.
9.
Spira D, Kotter I, Henes J, et al. MRI findings in psoriatic arthritis of the hands. AJR Am J Roentgenol 2010;195:1187–1193.
10.
Andreisek G, White LM, Sussman MS, et al. T2*-weighted and arterial spin labeling MRI of calf muscles in healthy volunteers and patients
November 2015
1319
Journal of Magnetic Resonance Imaging 29.
Cho JH, Cho G, Song Y, et al. Feasibility of FAIR imaging for evaluating tumor perfusion. J Magn Reson Imaging 2010;32: 738–744.
32.
Cai W, Li F, Wang J, et al. A comparison of arterial spin labeling perfusion MRI and DCE-MRI in human prostate cancer. NMR Biomed 2014;27:817–825.
30.
Buxton RB. Quantifying CBF with arterial spin labeling. J Magn Reson Imaging 2005;22:723–726.
33.
31.
Ludemann L, Jedrzejewski G, Heidenreich J, Han ET, Bruhn H. Perfusion imaging of cerebral arteriovenous malformations: a study comparing quantitative continuous arterial spin labeling and dynamic contrast-enhanced magnetic resonance imaging at 3 T. Magn Reson Imaging 2011;29:1157–1164.
Griffith JF, Yeung DK, Antonio GE, et al. Vertebral bone mineral density, marrow perfusion, and fat content in healthy men and men with osteoporosis: dynamic contrast-enhanced MR imaging and MR spectroscopy. Radiology 2005;236:945–951.
34.
Griffith JF, Yeung DK, Antonio GE, et al. Vertebral marrow fat content and diffusion and perfusion indexes in women with varying bone density: MR evaluation. Radiology 2006;241:831–838.
1320
Volume 42, No. 5
