Feasibility of FAIR imaging for evaluating tumor perfusion.
ABSTRACT To evaluate the feasibility of flow-sensitive alternating inversion recovery (FAIR) for measuring blood flow in tumor models.
In eight mice tumor models, FAIR and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) was performed. The reliability for measuring blood flow on FAIR was evaluated using the coefficient of variation of blood flow on psoas muscle. Three regions of interest (ROIs) were drawn in the peripheral, intermediate, and central portions within each tumor. The location of ROI was the same on FAIR and DCE-MR images. The correlation between the blood flow on FAIR and perfusion-related parameters on DCE-MRI was evaluated using the Pearson correlation coefficient.
The coefficient of variation for measuring blood flow was 9.8%. Blood flow on FAIR showed a strong correlation with Kep (r = 0.77), percent relative enhancement (r = 0.73), and percent enhancement ratio (r = 0.81). The mean values of blood flow (mL/100 g/min) (358 vs. 207), Kep (sec(-) (1)) (7.46 vs. 1.31), percent relative enhancement (179% vs. 134%), and percent enhancement ratio (42% vs. 26%) were greater in the peripheral portion than in the central portion (P < 0.01).
As blood flow measurement on FAIR is reliable and closely related with that on DCE-MR, FAIR is feasible for measuring tumor blood flow.
[show abstract] [hide abstract]
ABSTRACT: The T1 perfusion model has worked well in brain functional studies where flow changes are measured. Using selective and nonselective inversion pulses, a new method has been developed to study steady-state brain blood flow. The authors obtained flow-sensitive images using selective inversion and flow-insensitive images using nonselective inversion. Subtraction of flow-insensitive images from flow-sensitive images gave us flow-weighted images with good gray-white flow contrast in cortical gray matter as well as in the thalamus and basal ganglia. Fitting T1s of flow-insensitive and flow-sensitive images allowed us to obtain preliminary results of brain blood flow maps. Two specific problems can seriously affect the accuracy of the brain blood flow values and the gray-white flow contrast of brain blood flow maps. These are the problems of the partial volume effect of CSF and gray matter, and the difference between blood T1 and white matter T1. The authors discuss in detail the character of these problems and present a number of approaches to manage such problems.Magnetic Resonance in Medicine 01/1996; 34(6):878-87. · 2.96 Impact Factor
Article: Arterial spin labeling blood flow magnetic resonance imaging for the characterization of metastatic renal cell carcinoma(1).[show abstract] [hide abstract]
ABSTRACT: This study sought to assess the feasibility of arterial spin labeling (ASL) blood flow (BF) magnetic resonance imaging (MRI) for the study of metastatic renal cell carcinoma (RCC) in the body, where the respiratory, cardiac, and peristaltic motions present challenges when applying ASL. ASL was performed using a background-suppressed single-section flow-alternating inversion recovery (FAIR) preparation and a single-shot fast spin-echo imaging sequence on a 3.0-T whole body imager. Tumor BF was evaluated for 26 patients with RCC metastatic to the liver, bone, lung, or lymph nodes before VEGF receptor inhibitor therapy. Two cases with tumor size change after treatment were also scanned 1 month after therapy. For validation, kidney cortex BF in five normal volunteers was measured with the same technique and compared with literature values. ASL was successfully performed in all normal volunteers and in 20 of 26 patients. The six failures resulted from a systematic error, which can be avoided in future studies. For normal volunteers, measured kidney cortex BF was 275 +/- 14 mL/min/100 g, a value consistent with the literature. ASL determined tumor BF averaged across tumor volume and subjects was 194 mL/min/100 g (intersubject SD = 100), resulting in high perfusion signal and conspicuity of lesions. Bright signal was also seen in large vessels and occasionally in bowel. In the two cases studied 1 month after therapy, ASL perfusion changes were consistent with tumor size changes. With background suppression, ASL MRI is a feasible method for quantifying BF in patients with renal cell carcinoma. This technique may be useful for evaluating tumor response to antiangiogenic agents.Academic Radiology 04/2005; 12(3):347-57. · 1.69 Impact Factor
[show abstract] [hide abstract]
ABSTRACT: Three major models (from Tofts, Larsson, and Brix) for collecting and analyzing dynamic MRI gadolinium-diethylene-triamine penta-acetic acid (Gd-DTPA) data are examined. All models use compartments representing the blood plasma and the abnormal extravascular extracellular space (EES), and they are intercompatible. All measure combinations of three parameters; (1) kPSp is the influx volume transfer constant (min-1), or permeability surface area product per unit volume of tissue, between plasma and EES; (2) ve is the volume of EES space per unit volume of tissue (0 < ve < 1); and (3) K(ep), the efflux rate constant (min-1), is the ratio of the first two parameters (k(ep) = kPSp/ve). The ratio K(ep) is the simplest to measure, requiring only signal linearity with Gd tracer concentration or, alternatively, a measurement of T1 before injection of Gd (T10). To measure the physiologic parameters kPSp and ve separately requires knowledge of T10 and of the tissue relaxivity R1 (approximately in vitro value).Journal of Magnetic Resonance Imaging 7(1):91-101. · 2.70 Impact Factor
Feasibility of FAIR Imaging for Evaluating
Jee-Hyun Cho, MS,1,6Gyunggoo Cho, PhD,1,6Youngkyu Song, MS,1Chulhyun Lee, PhD,1
Bum-Woo Park, BS,2Chang Kyung Lee, MS,2Namkug Kim, PhD,2Sung Bin Park, MD,3
Jong Soon Kang, PhD,4Moo Rim Kang, PhD,4Hwan Mook Kim, PhD,4
Young Ro Kim, PhD,5Kyoung-Sik Cho, MD,2and Jeong Kon Kim, MD2*
Purpose: To evaluate the feasibility of flow-sensitive alter-
nating inversion recovery (FAIR) for measuring blood flow
in tumor models.
Materials and Methods: In eight mice tumor models,
FAIR and dynamic contrast-enhanced magnetic resonance
imaging (DCE-MRI) was performed. The reliability for
measuring blood flow on FAIR was evaluated using the
coefficient of variation of blood flow on psoas muscle.
Three regions of interest (ROIs) were drawn in the periph-
eral, intermediate, and central portions within each tu-
mor. The location of ROI was the same on FAIR and DCE-
MR images. The correlation between the blood flow on
FAIR and perfusion-related parameters on DCE-MRI was
evaluated using the Pearson correlation coefficient.
Results: The coefficient of variation for measuring blood
flow was 9.8%. Blood flow on FAIR showed a strong corre-
lation with Kep (r ¼ 0.77), percent relative enhancement (r
¼ 0.73), and percent enhancement ratio (r ¼ 0.81). The
mean values of blood flow (mL/100 g/min) (358 vs. 207),
Kep (sec?1) (7.46 vs. 1.31), percent relative enhancement
(179% vs. 134%), and percent enhancement ratio (42%
vs. 26%) were greater in the peripheral portion than in
the central portion (P < 0.01).
Conclusion: As blood flow measurement on FAIR is reli-
able and closely related with that on DCE-MR, FAIR is
feasible for measuring tumor blood flow.
Key Words: flow-sensitive alternating inversion recovery;
arterial spin labeling; blood flow; angiogenesis; magnetic
J. Magn. Reson. Imaging 2010;32:738–744.
C 2010 Wiley-Liss, Inc.
ARTERIAL SPIN LABELING (ASL) is a technique to
measure blood flow using an intrinsic contrast, ie,
tagged spin. Its most striking advantage is that it
does not require intravenous contrast injection, which
may cause a potential problem during repeated longi-
tudinal studies, such as difficult vascular access or
the influence of residual contrast material in the tis-
sue on the subsequent images. Although there have
been studies using ASL for measuring blood flow,
most of them could only estimate the blood flow deliv-
ered in a single direction (1). The flow-sensitive alter-
nating inversion recovery (FAIR) sequence, a modified
ASL, allows estimation of multidirectional blood flow
as it measures the difference of T1 recovery between
slice selective and nonselective inversion recovery
preparations (2). Although investigators have proven
that this technique can accurately estimate the blood
flow in various noncerebral organs (3), there has been
no trial to measure the blood flow of tumors in non-
Dynamic contrast enhanced magnetic resonance
imaging (DCE-MRI) is a widely used method for evalu-
ating tumor angiogenesis (4). Gd-DTPA is a low-mo-
lecular intravenous (IV) contrast material and its theo-
retical distribution is blood flow-limited, as this agent
has freely diffusible tracers in the vascular wall and
therefore is thought to perform similarly with regard
to its first-pass kinetics and clearance (5). Therefore,
if the estimation of blood flow on FAIR is reliable and
shows a close relationship with the perfusion-related
1Division of Magnetic Resonance, Korea Basic Science Institute, 804-1
Ochang, Cheongwon, Chungbuk, 363-883, Korea.
2Department of Radiology, Research Institute of Radiology, Medical
Imaging Laboratory, Asan Medical Center, University of Ulsan College
of Medicine, Songpa-gu, Seoul, Korea.
3Department of Radiology, Ulsan University Hospital, University of
Ulsan, Ulsan, Korea.
4Bioevaluation Center, Korean Research Institute of Bioscience and
Biotechnology, Ochang, Cheongwon, Chungbuk, Korea.
5Athinoula A. Martinos Center for Biomedical Imaging, Department
of Radiology, Massachusetts General Hospital, Charlestown,
6Authors equally contributed to this study.
Contract grant sponsor: Korea Healthcare Technology R&D Project,
Ministry for Health, Welfare & Family Affairs; Contract grant number:
A090754; Contract grant sponsor: Basic Science Research Program
through the National Research Foundation of Korea (NRF) funded by
the Ministry of Education, Science and Technology; Contract grant
number: 2009-0066963; Contract grant sponsor: Korean Basic
Science Institute; Contract grant number: T30403.
*Address reprint requests to: J.K. Kim , Department of Radiology,
Research Institute of Radiology, Medical Imaging Laboratory, Asan
Medical Center, University of Ulsan College of Medicine, 388-1 Poong-
nap dong, Songpa-gu, Seoul 138-736, Korea.
Received June 30, 2009; Accepted June 7, 2010.
View this article online at wileyonlinelibrary.com.
JOURNAL OF MAGNETIC RESONANCE IMAGING 32:738–744 (2010)
C 2010 Wiley-Liss, Inc.
parameters of DCE-MRI, FAIR can be used for evalu-
ating tumor angiogenesis.
The purpose of this study was to evaluate the feasi-
bility of FAIR for measuring tumor blood flow by ana-
lyzing its reliability and correlation to perfusion pa-
rameters on DCE-MRI in mouse tumor models.
MATERIALS AND METHODS
This study was approved by the Institutional Ethics
Committee for Animal Care and Use.
In eight 5–6-week-old female BALB/c-nu mice (SLC,
Shizuoka, Japan), U-118 MG cells (ATCC HTB-15)
were subcutaneously implanted into the flank region
in all mice. When the tumor diameter reached 7–10
mm, MR examinations were performed.
The mice were examined using a 4.7-T Biospec MRI
system (Bruker, Karlsruhe, Germany), equipped with
a maximum gradient strength of 400 mT/m (diameter,
12 cm). Radiofrequency transmission and reception
were performed with a quadrature volume coil (dia-
meter, 35 mm).
In each animal, FAIR imaging was performed ahead
of DCE-MR. FAIR images were obtained using a true
fast imaging with steady precession data-acquisition
sequence. The imaging parameters were as follows:
TR, 1000 msec; TE, 1.65 msec; flip angle, 20?; dura-
tion of inversion pulse, 10 msec; slice thickness, 2
mm; thickness of selective inversion, 4 mm; field of
view (FOV), 4.0 ? 4.0 cm2; matrix size, 128 ? 128;
and number of average 48. For fitting the T1 curve,
15 signals were measured according to the inversion
time: 265 msec þn? 422.4 msec (0 ? n ? 15, n is in-
teger). Scanning time per tumor was ?16 minutes.
DCE-MR images were obtained using a dynamic se-
ries of gradient echo images with the following param-
eters: TR, 20 msec; TE, 2.14 msec; flip angle, 30?; and
number of repetitions, 400. The other acquisition pa-
rameters were the same as those for FAIR. The con-
trast agent (Dotarem, Guerbet, France) was injected as
a bolus into the tail vein at a dose of 0.1 mmol/kg 30
seconds following the start of image acquisition. Total
scanning time per tumor was ?17 minutes as the tem-
poral resolution for each scanning was 2.55 seconds.
DR1between selective and nonselective inversions was
calculated using a nonlinear least square fit. Accord-
ing to extended Bloch equations (6), the use of FAIR
techniques allows measurement of the blood flow as
the difference in the T1 relaxation according to the fol-
lowing equation (7):
R1sel¼ R1non?selectiveþ f=l
where R1selis the T1 relaxation rate of selective inver-
sion, R1nonselectivelis the T1 relaxation rate of nonse-
lective inversion, f is the blood flow in milliliters per
100 g per minute and is the blood–tissue water parti-
tion coefficient, which is constant at 80 mL/100 g (8).
A time-intensity curve was fitted according to a two-
compartment model, as described previously (9,10):
Sð0Þ¼ 1 þ AHkep
where S(0) is signal intensity before contrast injection,
S(t) is signal intensity at a certain time of t, Kepis the
efflux rate constant from extravascular extracellular
space to blood plasma, thereby indicating the perme-
ability, Kelis the first-order rate constant for elimina-
tion of Gd-DTPA from the blood plasma, and AHis
constant depending on the properties of the tissue
and represents the size of the extravascular extracel-
lular space (9).
From this fit, six additional perfusion-related pa-
rameters were estimated as follows (4): 1) the time of
arrival which was defined as the onset time of con-
trast enhancement; 2) the time to peak indicating the
duration from the time of arrival to the peak signal in-
tensity; 3) the percent relative enhancement indicat-
ing peak enhancement/peak signal intensity, where
the peak enhancement was the difference between
peak signal intensity and baseline signal intensity; 4)
the percent enhancement ratio which indicated peak
signal intensity/baseline signal intensity; 5) the wash-
in rate which was defined as the maximum slope
between the time of arrival and the time of peak
enhancement during the entire MR acquisition; and 6)
the wash-out rate which was defined as the slope
between the peak enhancement and the last signal in-
tensity in an examination.
Three regions of interest (ROIs) per tumor were placed
by a radiologist according to the following rules: 1)
ROIs were selected on DCE-MR so as to include three
enhancing areas covering peripheral, intermediate,
and central portions within a tumor, while attempting
to avoid a necrotic area; 2) each ROI contained 12 or
more pixels; and 3) the location and the area of ROIs
on the X, Y, and Z axes was the same in FAIR and
In order to evaluate the accuracy of curve fitting on
FAIR and DCE-MR, the root mean square error
(RMSE) was calculated, and then the ratio of RMSE
over the peak signal intensity was calculated. The cor-
relation of the blood flow on FAIR to the perfusion-
related parameters on DCE-MR was evaluated using
the Pearson correlation coefficient.
FAIR for Tumor Perfusion 739
The blood flow as well as perfusion-related parame-
ters was compared between the ROI locations, ie, the
peripheral, intermediate, and central portions within
a tumor, using the Generalized Estimating Equations,
which allow the comparison of measurement repeated
within each subject (11).
Reliability of Curve Fitting and Blood Flow
In fitting the T1 recovery curve on FAIR, the ratios of
RMSE over peak signal intensity were 6.0% 6 2.8
(range, 1.0%–9.6%) for selective inversion and 5.4% 6
3.1 (range, 0.7%–9.7%) for nonselective inversion,
respectively. In fitting the time-intensity curve on
DCE-MR, the ratio of RMSE over the peak signal in-
tensity was 3.6% 6 1.3 (range, 1.8%–6.7%).
FAIR vs. DCE-MR
The blood flow on FAIR showed a positive correlation
with the Kep, percent relative enhancement, and per-
cent enhancement ratio as the correlation coefficient r
was 0.77 for Kep, 0.73 for the percent relative
enhancement, and 0.81 for the percent enhancement
ratio (Table 1, Figs. 1–4). The mean 6 SD values of
blood flow, Kep, percent relative enhancement, and
percent enhancement ratio according to the location
are summarized in Tables 2 and 3.
Correlation Coefficient Between Blood Flow on FAIR and Each of
the Perfusion-Related Parameters on DCE-MR
Time of arrival
Time to peak (sec)
% relative enhancement*
% enhancement ratio*
Wash-in rate (sec?1)
Wash-out rate (sec?1)
Parameters labeled with * show positive correlation between FAIR
and DCE-MR images.
Figure 1. Scatterplot and regression line show a positive
correlation between blood flow and Kep.
Figure 2. Scatterplot and regression line show a positive
Figure 3. Scatterplot and regression line show a positive
correlation between blood flow and percent enhancement
740Cho et al.
All tumors showed heterogeneity in both blood flow
on FAIR images and contrast enhancement on DCE-
MR images as the blood flow, Kep, percent relative
enhancement, and percent enhancement ratio showed
a wide variation according to the location of ROIs
(Table 2). The mean value of blood flow, Kep, percent
relative enhancement, and percent enhancement ratio
were greater in the peripheral portion than in the cen-
tral portion (P < 0.001 for blood flow, Kep, and per-
cent enhancement ratio and P ¼ 0.005 for percent rel-
ative enhancement). The peripheral portion showed
significantly greater blood flow than the intermediate
portion (P ¼ 0.006 for blood flow, P ¼ 0.005 for Kep,
and P ¼ 0.026 for percent enhancement ratio),
whereas the percent relative enhancement tended to
be greater in the central portion than in the interme-
diate portion (P ¼ 0.072).
In all mice the blood flow and Kep were greater in
the peripheral portion than in the central portion. In
seven (88%) mice, the percent relative enhancement
and the percent enhancement ratio were greater in
the peripheral portion than in the central portion. In
one mouse the percent relative enhancement and the
percent enhancement were similar between the pe-
ripheral (152% and 34%) and central portion (155%
Investigators have attempted to validate the blood
flow on ASL in the brain (12) and have shown that the
blood flow on the ASL is linearly correlated with the
blood flow obtained on other confirming methods. Our
between the blood flow on FAIR and various perfu-
sion-related parameters on DCE-MR, including the
enhancement ratio. Moreover, those parameters from
FAIR and DCE-MR simultaneously depicted more
active angiogenesis in the peripheral portion than in
the rest of the tumor. These results imply that FAIR
can potentially be used for evaluating tumor angio-
genesis in noncerebral organs.
As the measurement of blood flow on ASL was ini-
tially introduced in the brain using the continuous
inversion of incoming blood from the carotid arter-
ies (2), more qualitative steady-state flow-weighted
Blood Flow, Kep, % Relative Enhancement, and % Enhancement Ratio
MouseLocationFlowKep % Relative enhancement% Enhancement ratio
Blood Flow, Kep, % Relative Enhancement, and % Enhancement
Parameter Mean 6 SD (range)
Blood flow (mL/100 g/min)
% relative enhancement
% enhancement ratio
358 6 72 (234–450)
226 6 122 (30–406)
207 6 111 (20–377)
7.46 6 3.89 (2.64–11.78)
3.64 6 3.71 (0.23–9.85)
1.31 6 2.04 (0.01–5.56)
179 6 43 (132–257)
154 6 28 (117–202)
134 6 19 (107–158)
42 6 13 (24–61)
33 6 12 (15–51)
26 6 9 (15–37)
Peripheral portion showed significantly greater blood flow, Kep, %
relative enhancement, and % enhancement ratio (P < 0.05)
according to Generalized Estimating Equations.
FAIR for Tumor Perfusion741
images have been made using either the continuous
inversion method or with a single inversion pulse.
However, the measurement of steady state flow leads
to two possible artifacts as follows: 1) tagged spins
may relax before they reach the imaging slice of inter-
est, depending on the distance that the tagged blood
should pass; and 2) tagged blood outside the imaging
slice of interest may generate magnetization transfer
artifacts that can alter the signal intensity many times
larger than does the perfusion signal. The FAIR tech-
nique can largely reduce these problems and allows
the simple increment of inversion times, thereby pro-
viding more accurate measurement of blood flow.
Moreover, as almost all noncerebral tumors have mul-
tidirectional blood flow in contrast to the brain, the
FAIR technique, which is independent of flow direc-
tion, seems to be suitable for measuring blood flow in
approach for quantifying angiogenesis, our study
used DCE-MR findings as a reference standard for
evaluating the feasibility of FAIR rather than micro-
vascular density, according to the following rationale.
First, actually it is difficult to get the same cross-sec-
whereas the same slice can be readily imaged between
DCE-MR and FAIR. Therefore, we suggest that the
comparison would be more accurate between FAIR
and DCE-MR than between FAIR and microvascular
transverse FAIR and DCE-MR
DCE-MR images, three round
ROIs were drawn in the pe-
ripheral (P), intermediate (I),
and central (C) portions within
tumor (white line). The T1sel
msec and 1504 msec in the
peripheral portion (c), 1402
msec and 1510 msec in the
intermediate portion (d), and
1373 msec and 1435 msec in
the central portion (e), respec-
then 277 mL/100g/min in the
peripheral portion, 245 mL/
100g/min in the intermediate
portion, and 151 mL/100g/
min in the central portion,
respectively. The Kep, percent
was 10.95/sec?1, 175% and
45% in the peripheral portion,
3.93/sec?1, 146% and 31% in
the intermediate portion, and
0.89/sec?1, 125% and 20% in
the central portion.
coded map of blood flow from
FAIR (f) shows the distribution
of the blood flow, which corre-
distribution on DCE MR.
742 Cho et al.
DCE-MR, such as Kep and the amount of contrast
enhancement, have been shown to strongly correlate
with angiogenesis activity(13,14). Third, microvascu-
lar density may be variable according to observer and
identifying blood vessels in histological sections do
not discern whether the vessels provide routes of
blood flow (15,16).
Although our study showed a close relationship of
blood flow on FAIR to Kep, percent relative enhance-
ment, and percent enhancement ratio on DCE-MR,
While the blood flow on FAIR means the volume of
blood flow per 100 g of tissue per minute, Kepindi-
cates the permeability from extravascular space to
enhancement and percent enhancement ratio are
determined by the flow. Therefore, the simple pre-
sumption that the blood flow correlates with the per-
meability in all circumstances may be incorrect. We
suggest that our strong correlation of the blood flow
to the Kep, percent relative enhancement, and per-
cent enhancement ratio results from the distribution
property of Gd-DTPA. Actually, the transfer of this
plasma, ie, permeability, is heavily dependent on the
molecular size of contrast material. As Gd-DTPA is a
low-molecular contrast material with an approximate
molecular weight of 500 Da, Gd-DTPA has a large
first-pass extraction because it can transfer through
both normal and large endothelial gaps in the blood
vessels. Therefore, the permeability of this small mo-
lecular contrast material is almost determined by
bloodflow (4) andall
enhancement, and percent enhancement ratio are
strongly correlated with the blood flow on FAIR.
Therefore, if macromolecular contrast material is
used for DCE-MR, more careful interpretation should
be made regarding the correlation between FAIR and
In FAIR there are two methods for calculating blood
flow. The first method, which was used in this study,
is to calculate the blood flow based on the difference
of T1 relaxivity between slice selective and nonselec-
tive inversions. The second method is to calculate the
blood flow by subtraction of an image with nonselec-
tive inversionfrom that
according to the following equations (17):
where DM indicates the difference of longitudinal
magnetization between selective inversion and nonse-
lective inversions. M0is the tissue equilibrium mag-
longitudinal relaxation time of tissue, TI is the inver-
sion time, l is the blood–tissue water partition coeffi-
cient, which is constant at 80 mL/100 g, and f is the
blood flow in milliliters per 100 g per minute. (8).
The first method has an advantage over the second
method, as T1 measurement is more independent of
any systematic bias caused by the use of two different
sequences, ie, selective and nonselective inversions,
than simple subtraction of two images (2). As noted in
a previous study (1) in which the blood flows on FAIR
images with various TIs were correlated with that on
I-123-iodoamphetamin, the blood flow calculated by
subtracting a flow-insensitive image from a flow-sen-
sitive image may vary according to TI.
On the other hand, our method to calculate the
blood flow based on the difference of T1 relaxivity
between selective and nonselective inversions requires
longer scan times than the second method does.
Therefore, because of the time limitations, our study
only measured the blood flow in a single slice at the
center level of tumor. We suggest that the scanning
time can be reduced by applying an alternative
method for T1 measurement, such as the Look-Locker
method, which uses a series of low flip-angle, echo
planar imaging readout modules to monitor the recov-
ery of longitudinal magnetization following a single
inversion pulse (18).
Our study showed heterogeneous angiogenic activ-
ities according to location within the tumor, as the
blood flow, Kep, percent relative enhancement, and
percent enhancement ratio were greater in the periph-
eral portion of tumor than in the central portion. Our
(19,20). Tumor angiogenesis increases transvascular
leakiness, induces vasodilatation, and expands the
increases the blood flow and permeability; however,
impaired or insufficient remodeling of vasculature
within a tumor causes heterogeneous flow and holes
(diameter, 400–600 nm) in the capillary wall (19,20).
Moreover, the pressure of the interstitial fluid pres-
sure varies within a tumor as interstitial fluid pres-
sure increases toward the deeper core layers of the tu-
mor by a large gradient. As increased tissue pressure
leads to reduction of perfusion and permeability, the
peripheral portion of tumor shows greater blood flow
and permeability on FAIR and DCE-MR than the other
The recent development of antiangiogenesis agents
has allowed inhibition of specific components of the
angiogenic cascade (2). Microcirculatory alterations
induced by antiangiogenic treatment will depend on
the angiogenesis activity. Therefore, based on our
results, we suggest that FAIR can be used for meas-
uring blood flow repeatedly in a longitudinal study to
monitor the change of tumor blood flow by antiangio-
Our study has limitations. First, our study meas-
ured blood flow and perfusion-related parameters
only in three ROIs in each tumor. Theoretically, the
correlation of results between two imaging studies is
most accurately evaluated by pixel-wise correlation af-
method needs correction of image distortion for com-
plete coregistration of two images, which requires
dedicated software. Unfortunately, because of the lack
of such dedicated technique, our study could not per-
form pixel-wise correlation. Instead, our study drew
three ROIs with reasonable areas in each tumor while
applying the same location and area of ROIs between
FAIR and DCE-MR.
FAIR for Tumor Perfusion743
Second, despite showing heterogeneous perfusion-
related parameters within tumors on both FAIR and
DCE-MR, as our study included animals with the
same tumor cell line, our results merely show the
intratumoral angiogenic activity but do not necessar-
ily indicate that FAIR and DCE-MR can reflect the
tumors. To prove the correlation of perfusion-related
parameters on FAIR and DCE-MR to angiogenic activ-
ity, it will be necessary to compare those parameters
in tumors with different angiogenic activities.
In conclusion, as tumor blood flow measurement on
FAIR is reliable and closely related with that on DCE-
MR, FAIR is feasible for measuring tumor blood flow.
1. Arbab AS, Aoki S, Toyama K, et al. Optimal inversion time for
acquiring flow-sensitive alternating inversion recovery images to
quantify regional cerebral blood flow. Eur Radiol 2002;12:
2. Kwong KK, Chesler DA, Weisskoff RM, et al. MR perfusion studies
with T1-weighted echo planar imaging. Magn Reson Med 1995;
3. de Bazelaire C, Rofsky NM, Duhamel G, Michaelson MD, George
D, Alsop DC. Arterial spin labeling blood flow magnetic resonance
imaging for the characterization of metastatic renal cell carci-
noma1. Acad Radiol 2005;12:347–357.
4. Tofts PS. Modeling tracer kinetics in dynamic Gd-DTPA MR imag-
ing. J Magn Reson Imaging 1997;7:91–101.
5. Tofts PS, Brix G, Buckley DL, et al. Estimating kinetic parame-
ters from dynamic contrast-enhanced T1-weighted MRI of a dif-
fusable tracer: standardized quantities and symbols. J Magn
Reson Imaging 1999;10:223–232.
6. Roberts DA, Detre JA, Bolinger L, et al. Renal perfusion in
humans: MR imaging with spin tagging of arterial water. Ra-
7. Leithner C, Gertz K, Schrock H, et al. A flow sensitive alternating
inversion recovery (FAIR)-MRI protocol to measure hemispheric
cerebral blood flow in a mouse stroke model. Exp Neurol 2008;
8. Karger N, Biederer J, Lusse S, et al. Quantitation of renal perfu-
sion using arterial spin labeling with FAIR-UFLARE. Magn Reson
9. Hoffmann U, Brix G, Knopp MV, Hess T, Lorenz WJ. Pharmacoki-
netic mapping of the breast: a new method for dynamic MR mam-
mography. Magn Reson Med 1995;33:506–514.
10. Buckley DL, Kerslake RW, Blackband SJ, Horsman A. Quantita-
tive analysis of multi-slice Gd-DTPA enhanced dynamic MR
images using an automated simplex minimization procedure.
Magn Reson Med 1994;32:646–651.
11. Kenward MG, Jones B. Alternative approaches to the analysis of
binary and categorical repeated measurements. J Biopharm Stat
12. Buxton RB. Quantifying CBF with arterial spin labeling. J Magn
Reson Imaging 2005;22:723–726.
13. Barrett T, Brechbiel M, Bernardo M, Choyke PL. MRI of tumor
angiogenesis. J Magn Reson Imaging 2007;26:235–249.
14. Padhani AR, Husband JE. Dynamic contrast-enhanced MRI stud-
ies in oncology with an emphasis on quantification, validation
and human studies. Clin Radiol 2001;56:607–620.
15. Mcdonald DM, Choyke PL. Imaging of angiogenesis: from micro-
scope to clinic. Nat Med 2003;9:713–725.
16. Hlatky L, Hahnfeldt P, Folkman J. Clinical application of antian-
giogenic therapy: microvessel density, what it does and doesn’t
tell us. J Natl Cancer Inst 2002;94:883–893.
17. Kim SG. Quantification of relative cerebral blood flow change by
flow-sensitive alternating inversion recovery (FAIR) technique:
application to functional mapping. Magn Reson Med 1995;34:
18. Francis ST, Bowtell R, Gowland PA. Modeling and optimization of
Look-Locker spin labeling for measuring perfusion and transit
time changes in activation studies taking into account arterial
blood volume. Magn Reson Med 2008;59:316–325.
19. de Lussanet QG, Langereis S, Beets-Tan RGH, et al. Dynamic
contrast-enhanced MR imaging kinetic parameters and molecular
weight of dendritic contrast agents in tumor angiogenesis in
mice. Radiology 2005;235:65–72.
20. Patan S, Munn LL, Jain RK. Intussusceptive microvascular
growth in a human colon adenocarcinoma xenograft: a novel
mechanism of tumor angiogenesis. Microvasc Res 1996;51:
744Cho et al.