Download Improvement of gadoxetate arterial phase capture with a high spatio

CME
JOURNAL OF MAGNETIC RESONANCE IMAGING 38:938–945 (2013)
Original Research
Improvement of Gadoxetate Arterial Phase Capture
With a High Spatio-Temporal Resolution Multiphase
Three-Dimensional SPGR-Dixon Sequence
Thomas A. Hope, MD,* Manojkumar Saranathan, PhD, Iva Petkovska, MD,
Brian A. Hargreaves, PhD, Robert J. Herfkens, MD, and Shreyas S. Vasanawala, MD, PhD
Purpose: To determine whether a multiphase method
with high spatiotemporal resolution (STR) by means of a
combination of parallel imaging, pseudorandom sampling
and temporal view sharing improves the capture and intensity of gadoxetate arterial phase images as well as
lesion enhancement.
Materials and Methods: Thirty-seven patients were
imaged with a conventional spoiled gradient echo acquisition and 48 with a high STR multiphase acquisition after
the administration of gadoxetate. Arterial phase capture,
image quality, and quality of fat suppression were qualitatively graded. Fourteen lesions in the conventional
group and 28 in the high STR multiphase group were
imaged, including 34 focal nodular hyperplasias. The ratio of lesion to parenchyma enhancement as well as relative hepatic artery enhancement were calculated. Chisquared, Mann-Whitney U and student t-tests were used
to compare differences.
Results: The high STR multiphase acquisition included
the arterial phase more frequently than conventional
acquisitions (P < 0.001), with the arterial phase missed in
17% (95% CI of 4–28%) of patients with conventional acquisition compared with 2% (95% CI of 0–6%) with the
high STR multiphase acquisition. There was no loss of
image quality or degree of fat saturation. Additionally,
there was increased relative intensity of the hepatic
arteries (P < 0.001) as well as lesion enhancement
(P ¼ 0.01).
Conclusion: The high STR multiphase acquisition
resulted in more reliable gadoxetate arterial phase
capture compared with a conventional acquisition while
preserving image quality with robust fat saturation
Key Words: gadoxetate; arterial phase; view sharing; FNH
J. Magn. Reson. Imaging 2013;38:938–945.
C 2013 Wiley Periodicals, Inc.
V
Department of Radiology, Stanford University Medical Center,
Stanford, California, USA.
*Address reprint requests to: T.A.H., Department of Radiology,
Stanford University School of Medicine, 300 Pasteur Drive, H1307,
Stanford, CA 94305-5105. E-mail: [email protected]
Received October 10, 2012; Accepted December 18, 2012.
DOI 10.1002/jmri.24048
View this article online at wileyonlinelibrary.com.
C 2013 Wiley Periodicals, Inc.
V
GADOXETATE DISODIUM (Eovist, Bayer Healthcare)
has a central role in the imaging of hepatic lesions.
Because the injected dose is a quarter of that used for
typical extracellular agents (0.025 mmol/kg versus 0.1
mmol/kg) and the injected volume is half (0.1 mL/kg
versus 0.2 mL/kg), arterial phase imaging is challenging due to the resulting brief temporal window of peak
aortic enhancement. To overcome this issue, some
groups have proposed using slower injection rates (1
mL/s), doubled the dose of the agent, and diluted the
agent to created prolonged aortic enhancement (1–5).
A different approach to addressing this problem is to
improve the temporal resolution of the arterial phase
acquisition to match the shorter duration of aortic
enhancement. One could simply decrease the spatial
resolution to improve the temporal resolution, but the
resulting images would be less diagnostic. Early
dynamic contrast-enhanced MRI (DCEMRI) approaches
to maintain spatial resolution while improving temporal resolution used keyhole-based techniques. Keyholebased techniques sample only the low spatial frequencies for the dynamic phases, which results in blurring
of small structures that have a lot of high spatial frequency content (6,7). More recently elliptical-centric
two-dimensional keyhole methods like CENTRA plus
and 4D THRIVE have been proposed, but they still suffer from loss of fine spatial resolution in fast-enhancing small structures, limiting their visualization and
quantification (8,9). Combinations of view sharing and
parallel imaging have been proposed for MR angiography, such as TRICKS, TWIST, or CAPR (10–12). A constrained reconstruction technique with a 3D variable
density spiral acquisition called TRACER has also been
proposed for high temporal resolution liver perfusion
imaging (13).
Fat saturation is a significant issue with postcontrast DCEMRI. The most common approach to suppress signal from fat is chemical shift selective fat saturation (14). Most fast imaging techniques, including
TRACER and THRIVE, use chemical shift based fat saturation, which is often suboptimal due to B1 and B0
heterogeneity, especially at 3 Tesla (T). Field inhomogeneities result in both reduction of water signal as well
as regions of incomplete fat suppression that can be
938
Improvement of Gadoxetate Arterial Phase Capture With a High Spatio-Temporal
Resolution Multiphase Three-Dimensional SPGR-Dixon Sequence
939
Figure 1. Graphical representation of the high STR multiphase acquisition (DISCO) k-space sampling strategy during the arterial breathhold. The periphery of k-space is subdivided into three equally distributed pseudo-random selections (labeled B1,
B2, and B3). These are sequentially ordered with interspersed acquisitions of the center of k-space (labeled A). Each phase is
reconstructed from a single center of a k-space (A) combined with the nearest peripheral k-space neighbors. This results in a
temporal resolution of A þ Bn.
artifactual. A two-point Dixon based fat–water separation technique has been proposed, which is somewhat
insensitive to field inhomogeneities, and provides very
robust fat saturation in short breathhold times (15,16).
A DCEMRI technique that combined TRICKS with a
Dixon-based fat–water separation scheme was proposed for hepatic imaging (17). This method relied on
consistent breathholding due to the larger temporal
footprint of the TRICKS acquisition, which forced view
sharing across breathholds. Recently, a high spatiotemporal resolution (STR) DCEMRI technique which
combined nearest neighbor view sharing of pseudorandomly under-sampled peripheral k-space with parallel imaging (DIfferential Sub-sampling with Cartesian
Ordering or DISCO) was proposed (18). This technique
restricted view sharing to within a breathhold while
minimizing motion artifacts due to its pseudo-random
traversal of peripheral k-space.
No clinical study to our knowledge has evaluated
the effect of improving the temporal resolution of the
MR acquisition to match the more compact bolus of
gadoxetate. In this study, we sought to determine
whether a high STR imaging method that uses a combination of parallel imaging and view sharing (DISCO)
improves capture of the gadoxetate arterial phase.
MATERIALS AND METHODS
High STR Multiphase Pulse Sequence
The DIfferential Sub-sampling with Cartesian Ordering
(DISCO) sequence has been previously described (high
STR multiphase acquisition) (18). In short, the
sequence is based on a three dimensional (3D) fast
SPGR-Dixon sequence with an elliptically ordered ky-kz
that is segmented into four regions: the first representing the center of k-space (labeled A) and the remaining
three representing equally distributed pseudo-random
sub-samples of the periphery of k-space (labeled B1,
B2, and B3) (Fig. 1). During each breathhold, the three
peripheral k-space segments are alternated with the
central portion of k-space (e.g., acquisition order of
AB1AB2AB3AB1A). View sharing is restricted to nearest
temporal neighbors and to k-space segments acquired
within the same breathhold to minimize temporal blurring and motion misregistration. Additionally, two
dimensional (in ky and kz) self-calibrated hybrid space
parallel imaging (following view-sharing) was incorporated to further accelerate the acquisition. A typical acquisition takes 30 s, with a temporal phase reconstructed every 5.4 s (SD ¼ 1.4 s) depending on the
volume acquired. The temporal footprint, or the time
taken to acquire all three peripheral k-space segments,
was 16.2 s. A 2-point Dixon reconstruction algorithm
with a region growing fat–water separation method is
used to create separate fat and water images (16).
Patient Groups
Informed consent was waived for this retrospective
study by our local institutional review board. For both
acquisitions, all imaging was performed on a 3T
MR750 system (GE Healthcare, Waukesha, WI) using
the upper 20 elements of a 32-channel torso phased
array coil (NeoCoil, Pewaukee, WI). In each patient,
0.025 mmol/kg of gadoxetate (Eovist, Bayer Healthcare) was administered at a rate of 2.5 mL per second,
followed by 20 mL saline at the same rate. A standard
940
15-s delay was used for all patients without using
bolus tracking. The patients were instructed to hold
their breath just before the end of the 15-s delay.
Patients with an estimated GFR < 30 mL/min/1.73
m2 were excluded.
The first group of patients, which included 37 consecutive patients imaged between September 2009 and
June 2011, were imaged using a routine clinical 3D
SPGR-Dixon (conventional, LAVA-FLEX) sequence with
the following parameters: 340 256 matrix, 6167 kHz
bandwidth, 12 flip angle, TR/TE1/TE2 4.1/1.2/2.4,
3.6 mm section thickness reconstructed to 1.8 mm
sections, and 2 2 ARC (Autocalibrating Reconstruction for Cartesian imaging) acceleration (19). This
conventional sequence takes 25 s to acquire, with
acquisition time varying slightly based on the number
of slices required to cover the liver. For each of these
subjects, a separate breathhold was used for the arterial, portal venous, and delayed phases. Identical
parameters were used for the precontrast sequence
was acquired before the administration of contrast.
The second group included 48 consecutive patients
imaged between July 2011 and July 2012, and were
imaged using the high STR multiphase sequence
described above. The following parameters were used:
320 224 matrix, 6167 kHz bandwidth, 15 flip
angle, 3.0- to 4.4-mm section thickness reconstructed
to 1.5–2.2 mm sections, TR/TE1/TE2 4.1/1.2/2.4,
typical number of slices acquired per volume 60, 2 2 outer acceleration, and five time points reconstructed during each arterial breath hold. For the precontrast sequence, a single time point, which included
all four segments of k-space, was acquired with identical parameters before the administration of contrast.
Qualitative Image Quality Analysis
The percent of studies with adequate arterial phase
capture was calculated for each patient group. For
each exam, all phases imaged were classified as one of
the following: angiographic referring to hepatic artery
enhancement without any portal venous enhancement,
early arterial referring to hepatic artery enhancement
with early filling of the portal vein, late arterial referring to complete opacification of the portal vein without
enhancement of the hepatic veins, and portal venous
referring to opacification of both the portal and hepatic
veins. The percentage of patients where imaging
included an early arterial, late arterial or both early
and late arterial phase were calculated for each group.
Image quality was graded by a board-certified radiologist on a 5-point scale for overall quality (0: Nondiagnostic; 1: Only large and intensely enhancing lesions
would be detectable; 2: Only large or intensely enhancing lesions would be detectable; 3: Image quality for
detection of small subtly enhancing lesions; 4: No artifacts and minimal image noise) and the quality of fat
saturation (0: Complete failure of fat suppression; 1:
Large regional fat–water swaps due to failure of the
Dixon algorithm; 2: Regional fat–water failures but still
interpretable; 3: Minimal failures in image periphery; 4:
Perfect fat–water separation). We additionally tabulated
the number of phases for the high STR multiphase and
Hope et al.
conventional acquisitions that demonstrated any evidence fat–water swapping. For high STR multiphase
sequences, the highest quality grading for either an early
or late arterial phase was also calculated. Additionally,
for each phase imaged in the high STR multiphase and
conventional acquisitions respiratory artifact was graded
(0: no respiratory artifact; 1: minimal respiratory artifact
allowing for detection of small enhancing lesions; 2: respiratory artifact was significant enough to allow detection of only large lesions; 3: severe respiratory artifact
rendering study nondiagnostic). The grading was
grouped as minimal respiratory artifact (grades 0 and 1)
and significant respiratory artifact (grades 2 and 3).
Quantitative Arterial Enhancement Analysis
All quantitative analysis was performed using OsiriX
software (20). Regions of interests (ROIs) were placed
over the hepatic artery and the closest adjacent hepatic parenchyma that did not include any major vessels. For the high STR multiphase acquisitions, ROIs
were propagated between phases so that the size of
each ROI remained identical, although each ROI had
to be repositioned depending on patient motion
between phases. The ratio of hepatic artery intensity
to precontrast intensity within the hepatic artery as
well as to adjacent hepatic parenchyma was determined using both average and maximum intensities
within the ROIs. For the high STR multiphase acquisitions, the results from the phase with the peak
enhancement ratio were selected for analysis.
Lesion Analysis
Only lesions greater than 1 cm were included for analysis, and a maximum of three lesions were included
per patient. Within the patients imaged with the conventional acquisition, there were 14 hypervascular
lesions in 12 patients, which included 9 focal nodular
hyperplasias (FNH). In the patients imaged with the
high STR multiphase acquisition, there were 28
hypervascular lesions in 19 patients, which included
25 FNHs. The diagnosis of FNH was based on typical
imaging findings, including relative hyperintensity
with respect to adjacent liver on the hepatobiliary
phase. At each time point imaged, an ROI was placed
within the hypervascular lesion without inclusion of
scar and an ROI was placed in directly adjacent hepatic parenchyma, taking care to not include any vascular structures. Lesion enhancement relative to precontrast images (lesionpre) and relative to adjacent
parenchyma (lesionadj) were calculated. For the high
STR multiphase acquisitions, the maximum lesionpre
and lesionadj were selected for analysis. Additionally,
for the high STR multiphase acquisitions, the temporal phase at which lesionadj occurred was recorded.
Statistical Analysis
A chi-squared test was used to compare the percentage
of conventional versus high STR multiphase studies
that contained either an early or late arterial phase,
and 95% confidence intervals (CI) were calculated. A
Improvement of Gadoxetate Arterial Phase Capture With a High Spatio-Temporal
Resolution Multiphase Three-Dimensional SPGR-Dixon Sequence
941
Figure 2. Images from representative high STR multiphase and conventional acquisitions. Conventional acquisitions provide
a single arterial phase, while the high STR multiphase acquisition provides multiple arterial phases within one breathhold,
increasing the chances that an adequate arterial phase is acquired. Notice the robust fat saturation across the entire image
with preserved image quality in the high STR multiphase acquisition.
chi-squared test was also used to compare the number
of phases in each group that had significant respiratory artifact and fat–water swapping. A Mann-Whitney
U test was used to compare image quality and quality
of fat saturation between the two groups. A Student ttest was used for comparison of means of lesion
intensity and relative hepatic artery intensity between
conventional and high STR multiphase acquisitions. A
Wilcoxon signed rank test was used determine if there
was a difference in the timing of lesionpre and lesionadj
phases within the high STR multiphase group. Excel
12.1.0 (Microsoft, Redmond, WA) and SPSS Statistics
17.0 (SPSS, Chicago, IL) were used for statistical analysis. A P-value of < 0.05 was used to determine significance for all tests.
RESULTS
during the early arterial phase, 21 during the late arterial phase, and the bolus was missed in two cases.
Overall 31/37 or 84% (95% CI: 72–96%) of patients
imaged with the conventional acquisition had an early
or late arterial phase image. Of the 48 patients imaged
with the high STR multiphase acquisition, 25 had an
angiographic phase, 43 an early arterial phase, and 42
a late arterial phase. Overall 47/48 or 98% (95% CI:
94–100%) of patients had an early or late arterial
phase using high STR multiphase acquisitions (Fig. 2
and 3: P < 0.001 for comparing conventional versus
high STR multiphase acquisitions for percent of
patients with either early or late arterial phase images).
On average, high STR multiphase acquisitions provided 1.88 6 0.81 early arterial phases and 2.08 6
1.20 late arterial phases per study. 83% of patients
imaged with high STR multiphase acquisitions had
both early and late arterial phase images.
Reliability of Arterial Phase Capture
The use of the high STR multiphase acquisition
resulted in a higher rate of arterial phase capture. Of
the 37 patients imaged with the conventional acquisition, 4 were imaged during the angiographic phase, 10
Image Quality
Image quality was not significantly different between
the conventional and high STR multiphase groups.
When including all phases from the high STR
942
Hope et al.
with fat–water swapping did not differ between the
high STR multiphase and the conventional acquisitions (percent of imaged phases in the conventional
group with fat–water swapping was 14% with a 95%
CI of 3–25%, and for the high STR multiphase group
the percentage of phases imaged with fat–water swapping was 13% with a 95% confidence interval of 9–
18%; P ¼ 0.95). The rate of significant respiratory artifact did not differ between the two groups (percent of
imaged phases in the conventional group with significant respiratory artifact was 16% with a 95% CI of 4–
28%, and for the high STR multiphase group the percentage of phases imaged with significant respiratory
artifact was 13% with a 95% confidence interval of 8–
17%; P ¼ 0.55).
Figure 3. Percent of patients with each vascular phase for
conventional and high STR multiphase acquisitions. The far
right column (‘‘overall’’) is the percent of patients with either
an early or late arterial phase image in both groups. Overall
84% of patients imaged with the conventional and 98% of
patients imaged with the high STR multiphase acquisition
included both an early and late arterial phase (*refers to
P < 0.001).
multiphase arterial breathhold compared with the
conventional arterial phase, image quality and fat saturation did not differ (average image quality score of
3.72 6 0.59 for high STR multiphase versus 3.59 6
0.72 for conventional, Z ¼ 1.00 and P ¼ 0.32; average score of quality of fat saturation of 3.84 6 0.41
for high STR multiphase versus 3.76 6 0.43 for conventional, Z ¼ 1.55 and P ¼ 0.12). Additionally,
when comparing the best quality arterial phases from
the high STR multiphase acquisitions there was no
significant difference between the groups (average
image quality of 3.71 6 0.87 for high STR multiphase
versus 3.59 6 0.72 for conventional, Z ¼ 1.57 and P
¼ 0.12), and the overall quality of fat saturation was
not significantly different in either group (average
quality of fat saturation of 3.73 for high STR multiphase 6 0.84 versus 3.76 6 0.43 for conventional, Z
¼ 1.012 and P ¼ 0.31). The percentage of phases
Quantitative Arterial Enhancement
The high STR multiphase acquisition resulted in
increased relative intensity of the hepatic arteries. The
ratio of hepatic artery intensity to precontrast hepatic
artery intensity was greater with high STR multiphase
compared with conventional acquisitions (Fig. 4: average intensity ratio: 4.22 6 1.54 for high STR multiphase versus 2.94 6 1.04 for conventional, P <
0.001; maximum intensity: 6.18 6 2.44 for high STR
multiphase versus 4.03 6 1.64 for conventional, P <
0.001). When comparing the ratio of hepatic artery intensity to adjacent parenchyma, there was only a significant increase for average intensities likely reflecting focal vessel enhancement within the selected
parenchymal ROIs that affected only maximum intensity values (Fig. 4: average intensity: 1.77 6 0.62 for
high STR multiphase versus 1.32 6 0.54 for conventional, P < 0.001; maximum intensity: 1.99 6 0.67 for
high STR multiphase versus 1.80 6 0.78 for conventional, P ¼ 0.22).
Lesion Analysis
The maximum relative enhancement of lesions with
respect to adjacent parenchyma was greater in lesions
Figure 4. Quantitative analysis of arterial enhancement. In
comparison to conventional
acquisitions, high STR multiphase acquisitions resulted in
higher hepatic artery intensity
compared with adjacent hepatic parenchyma (HA/adj)
and compared with precontrast hepatic arteries (HA/pre),
both when measured as an average across the ROI and
when comparing maximum
values within the ROI. (*refers
to P < 0.01).
Improvement of Gadoxetate Arterial Phase Capture With a High Spatio-Temporal
Resolution Multiphase Three-Dimensional SPGR-Dixon Sequence
943
Figure 5. Thick slab maximum intensity projections from a high STR multiphase exam with focal nodular hyperplasia. Note
the early arterial enhancement of the lesion (black circle) during the angiographic phase with strong hepatic artery enhancement (white arrow) before enhancement of the portal vein (black arrow). The maximum relative enhancement between the
lesion and adjacent parenchyma is during phase 3, the early arterial phase.
imaged with high STR multiphase compared with conventional acquisitions (lesionadj of 2.16 6 0.67 for
high STR multiphase versus 1.62 6 0.46 for conventional, P ¼ 0.01). There was no significant difference
in the precontrast intensity of lesion relative to adjacent parenchyma (lesionadj of 0.96 6 0.33 for high
STR multiphase versus 0.90 6 0.17 for conventional,
P ¼ 0.55). There was a trend toward a greater maximum relative enhancement in lesions with respect to
precontrast images with the high STR multiphase acquisition, but this did not reach statistical significance (lesionpre of 2.85 6 0.66 for high STR multiphase versus 2.50 6 0.42 for conventional, P ¼ 0.07).
Of note, the time point at which lesionadj occurred
was earlier than lesionpre in high STR multiphase
lesions (mean phase of lesionadj 2.0 6 1.1 and mean
phase of lesionpre 3.4 6 1.8, Z ¼ 4.24 and P <
0.001). Additionally within the high STR multiphase
group, lesionadj occurred during the early arterial
phase in 20 lesions, the late arterial phase in 6
lesions and the angiographic phase in two lesions
(Fig. 5). Four of the 6 lesions where the lesionadj
occurred in the late arterial phase did not have early
arterial phase images due to early bolus timing.
Although not qualitatively analyzed, spatial resolution
was preserved allowing delineation of small subcentimeter arterial hyperenhancing lesions (Fig. 6).
DISCUSSION
The incorporation of temporal view sharing and parallel imaging to allow the acquisition of multiple arterial
phases within one breathhold resulted in more reli-
able gadoxetate arterial phase capture compared with
a conventional acquisition while preserving image
quality with robust fat saturation. Additionally, relative hepatic artery enhancement and lesion enhancement improved significantly. In our study, 17% of
patients imaged with the conventional acquisition
missed both the early and late arterial phases while
only 2% of patients imaged with the high STR multiphase acquisition missed both phases.
Proper arterial phase timing is difficult with gadoxetate due to its brief bolus duration, and capturing the
arterial phase is critical for many hypervascular hepatic lesions such as hepatocellular carcinoma, adenoma, FNH, and hypervascular neuroendocrine metastasis.
Typical
extracellular
agents,
such
as
gadopentetate dimeglumine (Magnevist, Bayer Healthcare), have a larger bolus size and broader peak arterial enhancement; therefore the chance that the central
k-space acquisition coincides with the arterial bolus is
more likely. Further work is being performed to analyze the effect on arterial timing with typical extracellular contrast agents. One additional reason that arterial
phase capture may be limited with gadoxetate administration is secondary to a recently reported association
with acute transient dyspnea (21).
The reason that arterial phase capture and relative
enhancement is improved with the high STR multiphase acquisition is likely due to the oversampling of
the center of k-space. Other techniques, such as keyhole, result in significant blurring artifact which limits
the diagnostic accuracy of the image (6). One unique
benefit of the DISCO implementation of the multiphase acquisition is that due to the pseudo-random
selection of the periphery of k-space, motion artifacts
944
Hope et al.
Figure 6. Multiple subcentimeter arterial enhancing lesions in a patient with numerous FNHs imaged using the high STR
multiphase acquisition. First row demonstrates a 6 mm arterial enhancing FNH an axial images near the liver tip (white
arrow). The second row demonstrates a 5 mm FNH in segment 8 near the dome of the liver (white circle). The third row demonstrates a third lesion on coronal reformations measuring 3 mm along the inferior surface of the liver (white arrow head),
which is presumed to represent a small FNH. Note the robust fat saturation throughout the imaged volume.
are dispersed resulting in preserved image quality.
With the high STR multiphase acquisition, the center
of k-space is sampled roughly every 5 s, increasing
the probability that the center of k-space is acquired
during peak arterial enhancement.
Bolus tracking software is often used to optimize the
timing of the arterial phase (22) and fluoroscopic triggering has previously been shown to aid in arterial
phase capture with gadoxetate (23). By incorporating
triggering or using slower injections rates, the arterial
capture rates likely would have been improved with
the conventional technique (2). What is remarkable is
that using the high STR multiphase acquisition
resulted in arterial phase capture in nearly all patients
without the use of bolus tracking or changing the dose
or injection rate. This could allow for improved workflow without the need for bolus tracking. Of note, bolus
tracking with gadoxetate can be problematic as the
bolus is short and getting patients to suspend their
respiration between aortic arrival and hepatic arrival
can be difficult. Additionally, the use of a separate test
bolus can confound arterial phase images due to hepatobiliary retention of gadoxetate. Nonetheless bolus
tracking techniques are another method of acquiring
the appropriate arterial phase.
The optimal hepatic arterial phase depends on the
pathology being imaged. For example, with hepatocellular carcinoma, the late arterial phase has the highest sensitivity for detecting lesions (24,25). Compari-
son of lesion conspicuity on different arterial phases
for FNH has not previously been performed, but it has
been noted that lesion to adjacent parenchyma ratios
are greatest during the arterial phase as compared
with the portal venous phase (26). In our series, 71%
of lesions had a maximum enhancement ratio (lesionadj) during the early arterial phase, suggesting that
early arterial phases might provide the greatest lesion
conspicuity for FNH. Additionally, there likely is variability within each disease entity as to which phase
has the highest sensitivity; therefore, the usage of a
technique which acquires numerous phases during
the arterial phase breathhold may be preferred. For
the evaluation of FNH in the setting of gadoxetate, the
arterial phase is less critical than with conventional
extracellular agents as the hepatobiliary phase provides additional diagnostic information (27,28). As
gadoxetate becomes used more frequently in the evaluation of hepatocellular carcinoma and hypervascular
metastasis, accurate arterial phase timing will become
critical (29,30).
One main limitation of this study is the lack of
intra-patient and intra-lesion comparisons. To perform intra-patient and intra-lesion comparisons,
patients would have to be injected and imaged twice,
which was not possible in this retrospective study.
Additionally, the lesion analysis included a small
number of lesions within the conventional group, and
it is possible that the differences in lesion-to-
Improvement of Gadoxetate Arterial Phase Capture With a High Spatio-Temporal
Resolution Multiphase Three-Dimensional SPGR-Dixon Sequence
parenchyma enhancement may reflect differences in
lesion pathology between the two groups rather than
differences in technique. Also the high injection rate
used (2.5 mL/s) and the lack of bolus tracking limited
the conventional technique. The quantitative analysis
is also limited as image noise is not taken into
account in the analysis. It has been well demonstrated that noise calculations in sequences accelerated with parallel imaging are difficult to measure
(31), although numerous previous articles have taken
approaches similar to ours in analyzing relative
enhancement data (5).
In conclusion, the high spatiotemporal resolution
multiphase DISCO acquisition provides clinically acceptable arterial phase capture without the use of
bolus timing techniques or compromising spatial resolution or coverage.
ACKNOWLEDGMENT
This research was performed in collaboration with
and with support from GE Healthcare.
REFERENCES
1. Chung S-H, Kim M-J, Choi J-Y, Hong H-S. Comparison of two
different injection rates of gadoxetic acid for arterial phase MRI of
the liver. J Magn Reson Imaging 2010;31:365–372.
2. Zech CJ, Vos B, Nordell A, et al. Vascular enhancement in early
dynamic liver MR imaging in an animal model: comparison of two
injection regimen and two different doses Gd-EOB-DTPA (gadoxetic acid) with standard Gd-DTPA. Invest Radiol 2009;44:
305–310.
3. Tamada T, Ito K, Yoshida K, et al. Comparison of three different
injection methods for arterial phase of Gd-EOB-DTPA enhanced
MR imaging of the liver. Eur J Radiol 2011;80:e284–e288.
4. Motosugi U, Ichikawa T, Sou H, et al. Dilution method of gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (Gd-EOBDTPA)-enhanced magnetic resonance imaging (MRI). J Magn
Reson Imaging 2009;30:849–854.
5. Motosugi U, Ichikawa T, Sano K, et al. Double-dose gadoxetic
Acid-enhanced magnetic resonance imaging in patients with
chronic liver disease. Invest Radiol 2011;46:141–145.
6. Bishop JE, Santyr GE, Kelcz F, Plewes DB. Limitations of the
keyhole technique for quantitative dynamic contrast-enhanced
breast MRI. J Magn Reson Imaging 1997;7:716–723.
7. Kagawa Y, Okada M, Kumano S, et al. Optimal scanning protocol
of arterial dominant phase for hypervascular hepatocellular carcinoma with gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid-enhanced MR. J Magn Reson Imaging 2011;33:
864–872.
8. Beck GM, De Becker J, Jones AC, von Falkenhausen M, Willinek
WA, Gieseke J. Contrast-enhanced timing robust acquisition
order with a preparation of the longitudinal signal component
(CENTRA plus) for 3D contrast-enhanced abdominal imaging.
J Magn Reson Imaging 2008;27:1461–1467.
9. Coenegrachts K, Ghekiere J, Denolin V, et al. Perfusion maps of
the whole liver based on high temporal and spatial resolution
contrast-enhanced MRI (4D THRIVE): feasibility and initial
results in focal liver lesions. Eur J Radiol 2010;74:529–535.
10. Lim RP, Shapiro M, Wang EY, et al. 3D time-resolved MR angiography (MRA) of the carotid arteries with time-resolved imaging
with stochastic trajectories: comparison with 3D contrastenhanced Bolus-Chase MRA and 3D time-of-flight MRA. Am J
Neuroradiol 2008;29:1847–1854.
11. Haider CR, Hu HH, Campeau NG, Huston J, Riederer SJ. 3D
high temporal and spatial resolution contrast-enhanced MR angiography of the whole brain. Magn Reson Med 2008;60:749–760.
12. Korosec FR, Frayne R, Grist TM, Mistretta CA. Time-resolved contrast-enhanced 3D MR angiography. Magn Reson Med 1996;36:
345–351.
945
13. Xu B, Spincemaille P, Chen G, et al. Fast 3D contrast enhanced
MRI of the liver using temporal resolution acceleration with constrained evolution reconstruction. Magn Reson Med 2012 [Epub
ahead of print].
14. Schwartz LH, Seltzer SE, Tempany CM, et al. Prospective comparison of T2-weighted fast spin-echo, with and without fat suppression, and conventional spin-echo pulse sequences in the
upper abdomen. Radiology 1993;189:411–416.
15. Dixon WT. Simple proton spectroscopic imaging. Radiology 1984;
153:189–194.
16. Ma J. Breath-hold water and fat imaging using a dual-echo twopoint Dixon technique with an efficient and robust phase-correction algorithm. Magn Reson Med 2004;52:415–419.
17. Saranathan M, Rettmann D, Bayram E, Lee C, Glockner J. Multiecho time-resolved acquisition (META): a high spatiotemporal resolution Dixon imaging sequence for dynamic contrast-enhanced
MRI. J Magn Reson Imaging 2009;29:1406–1413.
18. Saranathan M, Rettmann DW, Hargreaves BA, Clarke SE, Vasanawala SS. DIfferential Subsampling with Cartesian Ordering
(DISCO): a high spatio-temporal resolution Dixon imaging
sequence for multiphasic contrast enhanced abdominal imaging.
J Magn Reson Imaging 2012;35:1484–1492.
19. Brau ACS, Beatty PJ, Skare S, Bammer R. Comparison of reconstruction accuracy and efficiency among autocalibrating datadriven parallel imaging methods. Magn Reson Med 2008;59:
382–395.
20. Rosset A, Spadola L, Ratib O. OsiriX: an open-source software for
navigating in multidimensional DICOM images. J Digit Imaging
2004;17:205–216.
21. Davenport MS, Viglianti BL, Al-Hawary MM, et al. Comparison of
acute transient dyspnea after intravenous administration of
gadoxetate disodium and gadobenate dimeglumine: effect on arterial phase image quality. Radiology 2012 [Epub ahead of print].
22. Kanematsu M, Semelka RC, Matsuo M, et al. Gadoliniumenhanced MR imaging of the liver: optimizing imaging delay for
hepatic arterial and portal venous phases–a prospective randomized study in patients with chronic liver damage. Radiology 2002;
225:407–415.
23. Haradome H, Grazioli L, Tsunoo M, et al. Can MR fluoroscopic
triggering technique and slow rate injection provide appropriate
arterial phase images with reducing artifacts on gadoxetic acidDTPA (Gd-EOB-DTPA)-enhanced hepatic MR imaging? J Magn
Reson Imaging 2010;32:334–340.
24. Mori K, Yoshioka H, Takahashi N, et al. Triple arterial phase
dynamic MRI with sensitivity encoding for hypervascular hepatocellular carcinoma: comparison of the diagnostic accuracy among
the early, middle, late, and whole triple arterial phase imaging.
AJR Am J Roentgenol 2005;184:63–69.
25. Murakami T, Kim T, Takamura M, et al. Hypervascular hepatocellular carcinoma: detection with double arterial phase multi-detector row helical CT. Radiology 2001;218:763–767.
26. Terkivatan T, van den Bos IC, Hussain SM, Wielopolski PA, de
Man RA, IJzermans JNM. Focal nodular hyperplasia: lesion characteristics on state-of-the-art MRI including dynamic gadoliniumenhanced and superparamagnetic iron-oxide-uptake sequences
in a prospective study. J Magn Reson Imaging 2006;24:864–872.
27. Suh Y, Kim M, Choi J, Park Y, Park M, Kim K. Differentiation of
hepatic hyperintense lesions seen on gadoxetic acid-enhanced
hepatobiliary phase MRI. AJR Am J Roentgenol 2011;197:
W44–W52.
28. Grazioli L, Bondioni M, Haradome H, et al. Hepatocellular adenoma and focal nodular hyperplasia: value of gadoxetic acidenhanced MR imaging in differential diagnosis. Radiology 2012;
262:520–529.
29. Park MJ, Kim YK, Lee MW, et al. Small hepatocellular carcinomas: improved sensitivity by combining gadoxetic acid-enhanced
and diffusion-weighted MR imaging patterns. Radiology 2012;
264:761–770.
30. Onishi H, Kim T, Imai Y, et al. Hypervascular hepatocellular carcinomas: detection with gadoxetate disodium-enhanced MR imaging and multiphasic multidetector CT. Eur Radiol 2012;22:
845–854.
31. Reeder SB, Wintersperger BJ, Dietrich O, et al. Practical
approaches to the evaluation of signal-to-noise ratio performance
with parallel imaging: application with cardiac imaging and a 32channel cardiac coil. Magn Reson Med 2005;54:748–754.