DOI:10.2214/AJR.06.0748
AJR 2007; 188:W309-W316
© American Roentgen Ray Society
Gadolinium-Enhanced Multiphasic 3D MRI of the Liver with Prospective Adaptive Navigator Correction: Phantom Study and Preliminary Clinical Evaluation
Masayuki Kanematsu1,2,
Satoshi Goshima3,
Hiroshi Kondo3,
Yusuke Tsuge3,
Ryujiro Yokoyama1,
Kimihiro Kajita1,
Minoru Onozuka2,4,
Yuriko Suzuki5,
Marc Van Cauteren5 and
Noriyuki Moriyama6
1 Radiology Services, Gifu University Hospital, 1-1 Yanagido, Gifu, Gifu, Japan
501-1194.
2 Research Center of Brain and Oral Science, Kanagawa Dental College, Yokosuka,
Japan.
3 Department of Radiology, Gifu University Hospital, Gifu, Japan.
4 Department of Physiology and Neuroscience, Kanagawa Dental College, Kanagawa,
Japan.
5 Philips Electronics Japan, Ltd., Medical Systems, Tokyo, Japan.
6 Research Center for Cancer Prevention and Screening, National Cancer Center
Hospital, Tokyo, Japan.
Received June 29, 2006;
accepted after revision October 11, 2006.
Address correspondence to M. Kanematsu.
The employment status of Y. Suzuki and M. Van Cauteren did not influence
the data in this study.
WEB This is a Web exclusive article.
Supported in part by Health and Labour Sciences Research Grants for Third
Term Comprehensive Control Research for Cancer.
Abstract
OBJECTIVE. The purpose of this study was to determine whether
prospective adaptive navigator correction improves slice position
invariability while maintaining image quality and enables efficient cine
display observation on gadolinium-enhanced multiphasic thin-slice 3D MRI of
the liver.
MATERIALS AND METHODS. The study consisted of two parts: a phantom
study and a clinical study. To explore the effect of navigator correction, a
phantom was imaged in the resting state and in continuous movement. In the
clinical study, gadolinium-enhanced four-phase 3D spoiled turbo field-echo
images (3-mm thickness with no intersectional gap, 60 slices for whole liver)
were retrospectively assessed. The subjects were 83 patients with 130 focal
hepatic lesions randomized into two groups: with (n = 45) and without
(n = 38) navigator correction. Images were qualitatively assessed by
two blinded radiologists using a three-point slice position invariability
scale for liver and focal hepatic lesions. Image degradation due to motion or
artifacts was qualitatively assessed.
RESULTS. Phantom images were obtained with excellent slice position
invariability while image quality was maintained with navigator correction.
Navigator correction substantially degraded the quality of the images of two
patients (one with a large amount of ascites and the other with a large
hepatic cyst). In the other 81 patients, the degree of slice position
invariability for the liver was greater (p < 0.001) with (score,
2.84 ± 0.43 [SD]) than without (score, 2.37 ± 0.75) navigator
correction. For focal hepatic lesions, slice position invariability also was
greater (p < 0.0001) with (score, 2.95 ± 0.21) than without
(score, 2.18 ± 0.88) navigator correction. No difference in degree of
image degradation was found with or without navigator correction.
CONCLUSION. Prospective navigator correction improves slice position
invariability for cine display observation while preserving image quality for
gadolinium-enhanced mul-tiphasic thin-slice 3D MRI of the liver.
Keywords: contrast media • dynamic MRI • hemodynamics • liver • liver tumor • MRI technique • navigator
Introduction
Advances in fast MRI have led to whole-liver imaging within a single
breath-hold. Now, through the use of multiple serial data acquisitions, the
liver can be imaged immediately after IV administration of contrast material
[1-4].
Understanding and assessment of the temporarily changing hemodynamics of
hepatocellular carcinoma [5,
6], metastatic lesions
[7,
8], cavernous hemangioma
[9,
10], and other lesions have
become crucial aspects of the diagnosis of focal hepatic lesions.
MRI technology enables use of time-resolved 3D spoiled gradient-echo pulse
sequences and imaging of the entire liver in multiple serial contrast phases
using thin slices with a thickness of no more than 2-3 mm
[11]. When radiologists
interpret a large number of thin-slice images obtained in multiple contrast
phases for small focal hepatic lesions, evaluation of lesion hemodynamics with
cine displays, that is, observation of hemodynamic changes in lesions over a
series of contrast-enhanced phases at the same slice level, becomes more
necessary than ever for differentiation of small hypervascular hepatic lesions
[6,
11]. At our institution, we
have been performing cine display observations of gadolinium-enhanced
multiphasic 3D MR images of the liver in a large number of patients. We
believe that cine display observations increase radiologist confidence in the
diagnosis of small hypervascular hepatic lesions. However, slice level
misregistration between contrast phases owing to imperfect or inconstant
breath-hold makes it difficult to observe hemodynamic changes at a single
slice level, particularly in imaging of small hepatic lesions.
In 1989, Ehman and Felmlee
[12] found that adaptive
navigator correction markedly improved images degraded by voluntary motion.
They concluded that the technique showed promise for addressing the problem of
respiratory motion during thoracoabdominal imaging. Subsequent researchers
described the usefulness of navigator echo on diffusion-weighted MRI of the
brain [13], breath-hold MR
coronary angiography [14],
single-voxel proton spectroscopy of the human liver
[15], MRI-guided treatments of
the liver [16], and volumetry
of the liver [17]. In
particular, McConnell et al.
[14] described the usefulness
of prospective adaptive navigator correction in breath-hold MR coronary
angiography. We believe that this technique can be applied to breath-hold
gad-olinium-enhanced MRI of the liver to ensure slice position invariability
for slices of 3 mm or thinner. The purpose of this study was to determine
whether prospective adaptive navigator correction improves slice position
invariability while maintaining image quality for cine display observation for
breath-hold gadolinium-enhanced multiphasic thin-slice 3D MRI of the
liver.
View larger version (8K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1 —Diagram shows pulse sequence architecture of fat-suppressed 3D
spoiled turbo field-echo sequence. First fat-suppression prepulse is applied
before 2D navigator pulse to avoid fat signal contamination, which can lead to
misrecognition of lung-liver interface because 2D navigator pulse is not
chemical-shift selective. Second fat-suppression prepulse is used for 3D
imaging of liver. Signals generated by navigator pulse are Fourier transformed
in real time, and lung-liver interfaces in vivo are determined with
cross-correlation method. In craniocaudal axis, prospective correction is
achieved by updating frequency of volume-selective radiofrequency excitation
pulse. During this 200-millisecond segment, pulse excitation and data
acquisition for liver are repeated 39 times over approximately 140
milliseconds. This single segment is repeated 60 times over 12 seconds in
single phase.
|
|
Materials and Methods
Prospective Adaptive Navigator Correction
A localized vertical navigator through the top surface of an imaging object
was prepared with use of a coronal scout image. The localized navigator echo
consisted of a spiral 30-mm-diameter cylindric excitation followed by
flow-compensated readout along the long axis of the excited cylinder. The data
from this readout were Fourier transformed in real time, the signal magnitude
being used as the navigator image. The navigator correction program monitored
the air-object interface approximately every 200 milliseconds, performed
cross-correlations with the reference image, calculated the motion of the
imaged object along the axis of the cylinder, and updated the frequency of the
volume-selective excitation pulse to chase the moving object
[14]
(Fig. 1). The cross-correlation
in signal processing was a standard measure of similarity of two signals,
commonly used to find features in an unknown signal by comparing it with a
known one. In our clinical study, a reference image was obtained separately
during a 5-second breath-hold at end-inspiration immediately before the start
of gadolinium-enhanced imaging.
Because this study was a comparative one un-dertaken to assess the effects
of navigator correction, identical pulse sequences were used to obtain
navigator displays for patients with and without correction. Adaptive
correction was not executed by software setting when navigator correction was
switched off. Consequently, the acquisition times were identical with and
without correction.
Phantom Study
Two plastic syringes filled with gadolinium chelate solution (4.5 mmol/L)
were sunk into a plastic box filled with gadolinium chelate solution with a
concentration of 2.25 mmol/L. The syringes were fixed in a V shape so that the
transverse distance between the two syringes on transaxial images reflected
the level of the transaxial image plane.
A localized vertical navigator positioned through the top surface of the
phantom was prepared on the coronal scout image. A 3D spoiled turbo field-echo
sequence (TR/TE, 6.5/1.8; flip angle, 15°; matrix size, 256 x 192;
acquisition slice thickness to produce 3-mm interpolated slices with no gap,
6-mm; receiver bandwidth, 149.7 Hz/pixel; number of signals acquired, 1; 12
slices per 19 seconds) was used for the phantom study. The phantom placed on
the patient couch was scanned in three serial phases: first phase, resting
state; second phase, continuous movement at 2.5 mm/s induced by a mechanical
pusher activated by a commercially available MRI-compatible power injector;
third phase, repeat imaging of the phantom in the resting state at the final
position of the second phase. This three-phase imaging was performed with and
without navigator correction.
Clinical Study
Patients—Over the 2-month period September and October 2005,
86 consecutively enrolled patients believed to have hepatic disease on the
basis of previous sonographic, CT, or laboratory findings underwent
gadolinium-enhanced MRI of the liver in our department. All patients provided
written consent in accordance with the requirements of our institutional
review board.
The patients were randomized into two groups: with and without navigator
correction. We excluded three patients from the study because of technical
failures associated with an MR imager malfunction in one case and with
navigator display recording in two cases. These patients were excluded from
the study population because acquisition of multiphasic gadolinium-enhanced
images or measurement of right hemidiaphragmatic motion was not possible. The
other 83 patients (50 men, 33 women; age range, 21-88 years; mean age, 63.3
years) constituted the study population.
Among these 83 patients, 27 had hepatic cysts, 15 had hepatocellular
carcinoma, seven had cavernous hemangioma, five had metastasis, four had
vascular pseudolesions, and one had biliary cystadenoma. The other 24 patients
had no focal hepatic lesions. Focal hepatic lesions were diagnosed on the
basis of pathologic findings in 10 patients with hepatocellular carcinoma,
three with metastasis, and one with biliary cystadenoma. The other pathologic
conditions were diagnosed on the basis of pathognomonic radiologic findings,
the presence of lesion growth on follow-up images, or elevated levels of serum
tumor markers.
MRI protocol—MRI was performed with a 1.5-T superconducting
MRI system (Intera Achieva Nova Dual, Philips Medical Systems). The system has
a maximum gradient strength of 66 mT/m with a peak slew rate of 160 mT/m/ms.
All MR images were obtained with a four-channel phased-array multicoil and a
field of view of 38 cm. The MRI protocol consisted of the following sequences:
dual-echo T1-weighted fast field-echo (220/4.6 in phase; 220/2.3 opposed
phase; matrix size, 320 x 224 [frequency x phase encoding];
receiver band-width, 523.2 Hz/pixel; parallel imaging reduction factor, 2;
number of signals acquired, 1; breath-hold acquisition time for 30 slices each
for in-phase and opposed-phase images, 23 seconds); fat-suppressed
respiration-triggered T2-weighted turbo spin-echo
(TReff/TEeff, 1,200-3,600/80; echo-train length, 21;
matrix size, 512 x 256; receiver band-width, 210 Hz/pixel; reduction
factor, 2; number of signals acquired, 2; acquisition time, 3-5-minutes); and
breath-hold gadolinium-enhanced double hepatic artery phase imaging with
fat-suppressed 3D spoiled turbo field-echo (3.3/1.1, flip angle, 15°;
matrix size, 352 x 230; acquisition slice thickness to produce 3-mm
interpolated slices with no gap, 6 mm; receiver bandwidth, 434.3 Hz/pixel;
reduction factor, 2; number of signals acquired, 1; 60 slices/12 seconds).
View larger version (9K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2 —Diagram shows time course of gadolinium-enhanced multiphasic spoiled
turbo field-echo imaging. Scans were started at test-bolus peak enhancement,
and central k-space data were acquired 6, 18, 55, and 180 seconds after
arrival of contrast medium in abdominal aorta, estimated with test-bolus
imaging. Echoes sampled during acquisitions and filled in central k-space
lines markedly affected entire image contrast. HAP = hepatic artery-dominant
phase, PVP = portal venous phase, EP = equilibrium phase.
|
|
Test bolus imaging—Test bolus imaging was performed to
determine the aortic transit time of a test bolus in the abdominal aorta at
the level of the first lumbar vertebral body. Coronal 20-mm-thick
single-section fast field-echo images (14/1.0; acquisition time, 1 second)
were obtained every second after initiation of an IV bolus injection of 1 mL
of gadopentetate dimeglumine (Magnevist, Schering) followed by a flush of
sterile saline solution. The volume of the flush was calculated from patient
body weight according to the following equation: V = 15 + 0.2 x
Wbody - 1, where V is the volume of the saline
solution in milliliters and Wbody is patient body weight
in kilograms. A power injector was used to inject the contrast medium and
saline chaser at a rate of 3 mL/s. The aortic transit time in seconds was
determined as the time from initiation of the contrast injection to peak
enhancement in the abdominal aorta.
Gadolinium-enhanced multiphasic MRI of the liver—Breath-hold
gadolinium-enhanced MRI of the liver was performed with a field of view of 38
cm, and inhomogeneity corrections of calculated images were made with
reference data measured before MRI. This option, termed clear, was integrated
into the parallel imaging reconstruction. A localized vertical navigator of a
30-mm-diameter cylindric excitation followed by flow-compensated readout was
prepared through the right hemidiaphragm on a coronal scout image. MR images
were obtained before and after an IV bolus injection of gadopentetate
dimeglumine at 0.1 mmol/kg body weight followed by a 15-mL flush of sterile
saline solution. The amount of sterile saline solution flushed was fixed in
all patients. The power injector used for test bolus imaging also was used for
injecting contrast material and saline solution at a rate of 3 mL/s. Patients
were instructed to perform a breath-hold at end-inspiration because early and
late hepatic artery phase imaging was serially and seamlessly performed over
24 seconds. Breath-hold acquisition times were identical with and without
correction.
Figure 2 illustrates a
multiphasic MRI timing scheme. Gadolinium-enhanced MRI was initiated at peak
enhancement of a test bolus. Early and late hepatic artery phase images were
serially and seamlessly acquired over 24 seconds; portal venous phase imaging
was initiated after a 21-second breathing interval; and equilibrium phase
imaging was initiated 174 seconds after test bolus peak enhancement.
Consequently, the central k-space lines of each phase were filled 6, 18, 55,
and 180 seconds after arrival of contrast medium in the abdominal aorta.
Image review by cine display observation—The findings for
two patients—one with massive ascites and one with a large hepatic cyst
in the right upper lobe of the liver—for whom image quality was highly
degraded owing to an adverse effect of navigator correction were assessed
separately and were excluded from the blind image review. On the 81 sets of
images deemed acceptable for evaluation, we assessed slice position
invariability and image quality. We evaluated a maximum of five lesions per
liver. We eventually evaluated a total of 130 focal hepatic lesions: 74
hepatic cysts (size range, 3-80 mm; mean size, 12.4 ± 11.1 mm [SD]),
nine cavernous hemangiomas (5-20 mm; 12.7 ± 5.5 mm), 32 hepatocellular
carcinomas (5-100 mm; 21.6 ± 23.0 mm), eight metastatic lesions (5-60
mm; 22.8 ± 16.7 mm), five vascular pseudolesions (4-6 mm; 4.8 ±
0.8 mm), and two biliary cystadenomas (both 80 mm).
Two gastrointestinal radiologists with 6 and 9 years of clinical experience
and blinded to clinical information at the time of image review independently
evaluated the gadolinium-enhanced MR images on a DICOM viewer. Using cine
display mode, the radiologists reviewed enhanced multiphasic images at a given
slice level. Each reviewer used a three-point scale to evaluate the slice
position invariability of livers. A score of 3 was assigned for almost perfect
invariability regarding liver contours and hepatic vessels throughout the
phases, a score of 2 for invariability that was moderately impaired but
acceptable for evaluation of liver hemodynamics and blood vessels, and a score
of 1 for severely impaired invariability and unacceptability for evaluation of
liver hemodynamics and blood vessels. Each reviewer then used a three-point
scale to evaluate the degree of image degradation due to artifacts caused by
motion or susceptibility. A score of 3 was assigned for almost no degradation,
2 for moderate degradation but acceptable quality for image interpretation,
and 1 for severe degradation and unacceptable quality for image
interpretation. When disagreement occurred, consensus was reached by
discussion.
Each reviewer also used a three-point scale to evaluate the slice position
invariability of focal hepatic lesions. A score of 3 was assigned for almost
perfect invariability with regard to lesion contours and internal structures
throughout the multiple phases, 2 for moderately impaired invariability but
acceptable quality for evaluating the hemodynamics of lesions and internal
structures, and 1 for severely impaired invariability and unacceptable quality
for evaluating the hemodynamics of lesions and internal structures.
View larger version (125K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —Transaxial phantom images. First (A), second (B), and
third (C) phase MR images obtained without prospective navigator
correction. Without correction, distance between two syringes increases from
24 to 30 to 60 mm in the first, second, and third phases, and transaxial scan
levels in first (A) and third (C) phases are clearly different.
Image during movement (B) is substantially degraded.
|
|
View larger version (127K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —Transaxial phantom images. First (A), second (B), and
third (C) phase MR images obtained without prospective navigator
correction. Without correction, distance between two syringes increases from
24 to 30 to 60 mm in the first, second, and third phases, and transaxial scan
levels in first (A) and third (C) phases are clearly different.
Image during movement (B) is substantially degraded.
|
|
View larger version (126K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C —Transaxial phantom images. First (A), second (B), and
third (C) phase MR images obtained without prospective navigator
correction. Without correction, distance between two syringes increases from
24 to 30 to 60 mm in the first, second, and third phases, and transaxial scan
levels in first (A) and third (C) phases are clearly different.
Image during movement (B) is substantially degraded.
|
|
View larger version (113K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3D —Transaxial phantom images. First (D), second (E), and
third (F) phase MR images obtained with prospective navigator
correction. With correction, distance between two syringes is constant 24 mm
and transaxial scan levels in first (D) and third (F) phases are
matched perfectly. Image is not degraded during movement (E).
|
|
View larger version (113K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3E —Transaxial phantom images. First (D), second (E), and
third (F) phase MR images obtained with prospective navigator
correction. With correction, distance between two syringes is constant 24 mm
and transaxial scan levels in first (D) and third (F) phases are
matched perfectly. Image is not degraded during movement (E).
|
|
View larger version (112K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3F —Transaxial phantom images. First (D), second (E), and
third (F) phase MR images obtained with prospective navigator
correction. With correction, distance between two syringes is constant 24 mm
and transaxial scan levels in first (D) and third (F) phases are
matched perfectly. Image is not degraded during movement (E).
|
|
At the completion of the qualitative review of each patient, using
navigator display exhibiting movement of the right hemidiaphragm during image
acquisition in each phase, the two radiologists in concert measured in
millimeters the distance of right hemidiaphragmatic movement from the starting
point of the early hepatic artery phase to the starting points of the late
hepatic arterial, portal venous, and equilibrium phases for evaluation of
inconstant breath-holds. They also recorded the continuous distance of right
hemidiaphragmatic movement (in millimeters) from the start to the end of each
phase for the evaluation of imperfect breath-holds. All measurements were
obtained with an electronic caliper on the DICOM viewer.
Statistical Analysis
The Mann-Whitney U test, a nonparametric test for comparing two
independent groups of sampled data, was used to compare the following
determinants in the groups with and without navigator correlation: patient
age, body weight, summation of distance of right hemidiaphragmatic movement,
size of focal hepatic lesions, degree of slice position invariability for
liver, degree of slice position invariability for focal hepatic lesions, and
degree of image degradation. Similar analyses were performed in the following
subgroups: 36 patients in whom the summation of distances of right
hemidiaphragmatic movement was greater than the median value (8 mm)
(inconstant breath-hold group), seven patients in whom the continuous distance
of right hemidiaphragmatic movement from the start to the end of any phase was
greater than 5 mm (imperfect breath-hold group), and 39 patients with focal
hepatic legions that were 10 mm in diameter or smaller (66 lesions).
For assessment of interobserver variability in terms of interpreting
images, kappa statistics were used to measure degree of agreement. A kappa
value up to 0.20 represented little agreement; 0.21-0.40, fair agreement;
0.41-0.60, moderate agreement; 0.61-0.80, substantial agreement; and 0.81 or
greater, almost perfect agreement.
Results
In the phantom study, without navigator correction, image quality was
substantially degraded during the second phase owing to continuous phantom
motion, and substantial slice level misregistration occurred between the first
and third phases. In contrast, slice position invariability was excellent in
all three phases, and no image quality degradation was found in any phase when
prospective adaptive navigator correction was used (Fig.
3A,
3B,
3C,
3D,
3E,
3F).
In the clinical study, in one patient with massive ascites extending over
the right hepa-todiaphragmatic interface and in another patient with a large
hepatic cyst measuring 70 mm in diameter in the right upper lobe of the liver
immediately beneath the right hemidiaphragm, a substantial degree of image
degradation occurred when navigator correction was used. In these patients,
recognition of the lung-liver interface by the navigator correction program
was hampered by intervening signals from the ascites fluid or hepatic cyst.
These findings clearly contraindicate navigator correction in such
patients.
The two groups with and without navigator correction comprised 43 and 38
patients, respectively. These groups were similar in terms of background
factors, that is, patient age (62.7 ± 15.3 and 63.9 ± 11.8
years; p =0.96), body weight (55.0 ± 10.6 and 56.6 ±
11.8 kg; p = 0.65), size of focal hepatic lesions (16.8 ± 20.1
mm [64 lesions] and 15.4 ± 14.3 mm [66 lesions]; p = 0.56),
and summation of diaphragmatic motion distances (9.5 ± 7.7 and 7.8
± 7.0 mm; p = 0.28). The degree of slice position
invariability for the liver was greater with navigator correction (mean score,
2.84 ± 0.43) than without (mean score, 2.37 ± 0.75) (p
< 0.001). Moreover, the degree of slice position invariability for focal
hepatic lesions over-all was greater with (mean score, 2.95 ± 0.21)
than without (mean score, 2.18 ± 0.88) navigator correction (p
< 0.0001). No difference was found between the two groups in terms of
degree of image degradation (mean scores, 2.88 ± 0.32 and 2.92 ±
0.27, respectively) (p = 0.58)
(Table 1).
View larger version (159K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A —72-year-old man with multiple small hepatocellular carcinomas (HCCs)
in moderate cirrhosis. Transverse gadolinium-enhanced spoiled turbo field-echo
(TR/TE, 3.3/1.1) MR images obtained with prospective adaptive navigator
correction during early hepatic artery-dominant (A), late hepatic
artery-dominant (B), portal venous (C), and equilibrium
(D) phases. Small hypervascular HCC (arrow), measuring 8 mm in
diameter, is constantly imaged in its maximum cross-section, and cine display
shows coronal enhancement (arrow, B and C) and washout
(arrow, D) of this small hypervascular HCC.
|
|
View larger version (165K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B —72-year-old man with multiple small hepatocellular carcinomas (HCCs)
in moderate cirrhosis. Transverse gadolinium-enhanced spoiled turbo field-echo
(TR/TE, 3.3/1.1) MR images obtained with prospective adaptive navigator
correction during early hepatic artery-dominant (A), late hepatic
artery-dominant (B), portal venous (C), and equilibrium
(D) phases. Small hypervascular HCC (arrow), measuring 8 mm in
diameter, is constantly imaged in its maximum cross-section, and cine display
shows coronal enhancement (arrow, B and C) and washout
(arrow, D) of this small hypervascular HCC.
|
|
View larger version (170K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4C —72-year-old man with multiple small hepatocellular carcinomas (HCCs)
in moderate cirrhosis. Transverse gadolinium-enhanced spoiled turbo field-echo
(TR/TE, 3.3/1.1) MR images obtained with prospective adaptive navigator
correction during early hepatic artery-dominant (A), late hepatic
artery-dominant (B), portal venous (C), and equilibrium
(D) phases. Small hypervascular HCC (arrow), measuring 8 mm in
diameter, is constantly imaged in its maximum cross-section, and cine display
shows coronal enhancement (arrow, B and C) and washout
(arrow, D) of this small hypervascular HCC.
|
|
View larger version (167K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4D —72-year-old man with multiple small hepatocellular carcinomas (HCCs)
in moderate cirrhosis. Transverse gadolinium-enhanced spoiled turbo field-echo
(TR/TE, 3.3/1.1) MR images obtained with prospective adaptive navigator
correction during early hepatic artery-dominant (A), late hepatic
artery-dominant (B), portal venous (C), and equilibrium
(D) phases. Small hypervascular HCC (arrow), measuring 8 mm in
diameter, is constantly imaged in its maximum cross-section, and cine display
shows coronal enhancement (arrow, B and C) and washout
(arrow, D) of this small hypervascular HCC.
|
|
Right hemidiaphragmatic motion summations ranged from 0 to 33 mm (mean, 8.7
± 7.4 mm) with a median value of 8 mm. In 36 patients in whom right
hemidiaphragmatic motion summation was greater than 8 mm (inconstant
breath-hold group), the two subgroups with and without navigator correction
comprised 21 and 15 patients, respectively. Moreover, these subgroups were not
significantly different in terms of background factors: patient age (65.8
± 14.3 and 65.1 ± 8.1 years, respectively; p = 0.56),
body weight (60.0 ± 11.7 and 54.2 ± 9.8 kg; p = 0.11),
size of focal hepatic lesion (19.4 ± 21.0 mm [33 lesions] and 15.1
± 12.2 mm [26 lesions]; p = 0.61), or diaphragmatic motion
summation (15.6 ± 6.4 and 14.5 ± 6.1 mm; p = 0.58). In
these 36 patients, the degree of slice position invariability for the liver
was greater with (mean score, 2.76 ± 0.54) than without (mean score,
2.07 ± 0.80) navigator correction (p = 0.004). The degree of
slice position invariability for focal hepatic lesions also was greater with
(mean score, 2.91 ± 0.29) than without (mean score, 1.62 ± 0.70)
navigator correction (p < 0.0001). No difference was found between
the two groups in terms of degree of image degradation (mean score, 2.86
± 0.36 and 2.87 ± 0.35, respectively) (p = 0.94)
(Table 2).
In 39 patients with small hepatic lesions (
10 mm in diameter), the
degree of slice position invariability for focal hepatic lesions was greater
with (mean score, 2.97 ± 0.17) than without (mean score, 2.16 ±
0.93) navigator correction (p < 0.0001) (Fig.
4A,
4B,
4C,
4D). In 18 of these 39
patients, the right hemidiaphragmatic movement distance summations were larger
than 8 mm. In these 18 patients, the degree of slice position invariability
for focal hepatic lesions 10 mm or smaller was greater with (mean score, 2.94
± 0.25) than without (mean score, 1.64 ± 0.81) navigator
correction (p < 0.0001) (Table
3).
View this table:
[in this window]
[in a new window]
|
TABLE 3: Degree of Slice Position Invariability for Focal Hepatic Lesions in 39
Patients with Small Focal Hepatic Lesions 10 mm in Diameter or Smaller
|
|
On navigator displays, the right hemidiaphragm moved continuously more than
5 mm during one or more phases in seven patients (four men, three women; age
range, 56-86 years; mean age, 67.0 ± 10.5 years) (imperfect breath-hold
group). Four of the seven patients had hepatocellular carcinoma, one had
hepatic cysts, one had biliary cystadenomas, and one had no focal hepatic
lesion. No slice level misregistration occurred in the four patients with
navigator correction (Fig. 5A,
5B,
5C,
5D,
5E), and mild image
degradation occurred in one patient (Fig.
6A,
6B,
6C,
6D,
6E). Slice level
misregistration occurred in all three patients without navigator correction,
and image degradation occurred in one of these patients.
View larger version (46K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A —70-year-old man with hypervascular hepatocellular carcinoma (HCC) in
moderate cirrhosis. Navigator display shows that lung-liver interface
(arrow) continuously moved 14 mm in cranial direction during serial
early and late hepatic artery-dominant phases. HAP = hepatic artery-dominant
phase, PVP = portal venous phase, and EP = equilibrium phase.
|
|
View larger version (148K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5B —70-year-old man with hypervascular hepatocellular carcinoma (HCC) in
moderate cirrhosis. Gadolinium-enhanced multiphasic 3D spoiled turbo
field-echo MR images (TR/TE, 3.3/1.1) obtained with navigator correction.
Hemodynamics in hypervascular 30-mm HCC (arrow, B) imaged in
maximum cross-section during early hepatic artery-dominant phase (B)
were well observed on late hepatic artery-dominant (C), portal venous
(D), and equilibrium (E) images. Portal venous branches in
umbilical portion of liver (arrowhead, B) are constantly
observed in same configuration throughout phases.
|
|
View larger version (162K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5C —70-year-old man with hypervascular hepatocellular carcinoma (HCC) in
moderate cirrhosis. Gadolinium-enhanced multiphasic 3D spoiled turbo
field-echo MR images (TR/TE, 3.3/1.1) obtained with navigator correction.
Hemodynamics in hypervascular 30-mm HCC (arrow, B) imaged in
maximum cross-section during early hepatic artery-dominant phase (B)
were well observed on late hepatic artery-dominant (C), portal venous
(D), and equilibrium (E) images. Portal venous branches in
umbilical portion of liver (arrowhead, B) are constantly
observed in same configuration throughout phases.
|
|
View larger version (159K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5D —70-year-old man with hypervascular hepatocellular carcinoma (HCC) in
moderate cirrhosis. Gadolinium-enhanced multiphasic 3D spoiled turbo
field-echo MR images (TR/TE, 3.3/1.1) obtained with navigator correction.
Hemodynamics in hypervascular 30-mm HCC (arrow, B) imaged in
maximum cross-section during early hepatic artery-dominant phase (B)
were well observed on late hepatic artery-dominant (C), portal venous
(D), and equilibrium (E) images. Portal venous branches in
umbilical portion of liver (arrowhead, B) are constantly
observed in same configuration throughout phases.
|
|
View larger version (155K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5E —70-year-old man with hypervascular hepatocellular carcinoma (HCC) in
moderate cirrhosis. Gadolinium-enhanced multiphasic 3D spoiled turbo
field-echo MR images (TR/TE, 3.3/1.1) obtained with navigator correction.
Hemodynamics in hypervascular 30-mm HCC (arrow, B) imaged in
maximum cross-section during early hepatic artery-dominant phase (B)
were well observed on late hepatic artery-dominant (C), portal venous
(D), and equilibrium (E) images. Portal venous branches in
umbilical portion of liver (arrowhead, B) are constantly
observed in same configuration throughout phases.
|
|
View larger version (44K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6A —86-year-old man with multiple hepatocellular carcinomas (HCCs) in
severe cirrhosis. Navigator display shows that patient did not hold his breath
at beginning (arrow) of portal venous phase and that liver moved
substantially at beginning of portal venous phase. HAP = hepatic
artery-dominant phase, PVP = portal venous phase, and EP = equilibrium
phase.
|
|
View larger version (163K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6B —86-year-old man with multiple hepatocellular carcinomas (HCCs) in
severe cirrhosis. Gadolinium-enhanced multiphasic 3D spoiled turbo field-echo
MR images (TR/TE, 3.3/1.1) obtained with navigator correction. Hypervascular
20-mm HCC (arrow, B) imaged in maximum cross-section during
early hepatic artery-dominant phase is depicted in same cross-section in late
hepatic artery-dominant (C), portal venous (D), and equilibrium
(E) phases. Image quality in portal venous phase (D) is
moderately degraded despite navigator correction, but slice position is
maintained, allowing observation of hemodynamics of tumor.
|
|
View larger version (166K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6C —86-year-old man with multiple hepatocellular carcinomas (HCCs) in
severe cirrhosis. Gadolinium-enhanced multiphasic 3D spoiled turbo field-echo
MR images (TR/TE, 3.3/1.1) obtained with navigator correction. Hypervascular
20-mm HCC (arrow, B) imaged in maximum cross-section during
early hepatic artery-dominant phase is depicted in same cross-section in late
hepatic artery-dominant (C), portal venous (D), and equilibrium
(E) phases. Image quality in portal venous phase (D) is
moderately degraded despite navigator correction, but slice position is
maintained, allowing observation of hemodynamics of tumor.
|
|
View larger version (164K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6D —86-year-old man with multiple hepatocellular carcinomas (HCCs) in
severe cirrhosis. Gadolinium-enhanced multiphasic 3D spoiled turbo field-echo
MR images (TR/TE, 3.3/1.1) obtained with navigator correction. Hypervascular
20-mm HCC (arrow, B) imaged in maximum cross-section during
early hepatic artery-dominant phase is depicted in same cross-section in late
hepatic artery-dominant (C), portal venous (D), and equilibrium
(E) phases. Image quality in portal venous phase (D) is
moderately degraded despite navigator correction, but slice position is
maintained, allowing observation of hemodynamics of tumor.
|
|
View larger version (158K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6E —86-year-old man with multiple hepatocellular carcinomas (HCCs) in
severe cirrhosis. Gadolinium-enhanced multiphasic 3D spoiled turbo field-echo
MR images (TR/TE, 3.3/1.1) obtained with navigator correction. Hypervascular
20-mm HCC (arrow, B) imaged in maximum cross-section during
early hepatic artery-dominant phase is depicted in same cross-section in late
hepatic artery-dominant (C), portal venous (D), and equilibrium
(E) phases. Image quality in portal venous phase (D) is
moderately degraded despite navigator correction, but slice position is
maintained, allowing observation of hemodynamics of tumor.
|
|
The kappa values for the two reviewers ranged from 0.76 to 0.90 (mean,
0.82) for rating images independently, indicating substantial to almost
perfect agreement.
Discussion
The use of a 3D field-echo sequence increased the possible number of
imaging slices over the whole liver and reduced slice thickness to no more
than 2-3 mm, enabling observation of the hemodynamics of fairly small focal
hepatic lesions [6,
11]. Prospective adaptive
navigator correction adjusts craniocaudal axis deviation of the liver by
referring to the lung-liver interface as determined by navigator echoes
[14,
18]. This technique has
allowed evaluation of the vascularity and hemodynamics of fairly small lesions
over serial multiphasic images. Use of this technique will enable us to
efficiently differentiate tiny hypervascular benign and malignant lesions such
as cavernous hemangioma, focal nodular hyperplasia, hepatocellular carcinoma,
metastatic lesions, and vascular pseudolesions under less stressful
conditions. Furthermore, the slice position invariability assured by
correction may enable efficient image subtraction for detection of subtle
enhancement.
Slice position invariability on cine display observations of the liver and
of focal hepatic lesions was significantly better with the use of navigator
correction in the overall group of 81 patients with 130 focal hepatic lesions
and in 39 patients with 66 small ( 10 mm) lesions. Moreover, in the
subgroup of patients in whom the level of the right hemidiaphragm was
inconstant through multiple phases, slice position invariability with
correction was better in 36 patients with 59 lesions and in 18 patients with
27 small ( 10 mm) lesions. Our results indicate that navigator correction
significantly improved slice position invariability, even in patients with
imperfect or inconstant breath-holds and in patients with small hepatic
lesions. These findings suggest that the use of this program in clinical
practice would improve evaluation of the hemodynamics of focal hepatic lesions
in cine display observations over multiphasic images.
In our study 36 (44%) of the patients had an inconstant breath-hold. In
most of these patients, the level of the right hemidiaphragm deviated caudally
in the portal venous phase because these patients inhaled deeper for portal
venous phase imaging after the somewhat longer respiratory suspension required
for the hepatic arterial phase imaging. In the differential diagnosis of small
enhancing lesions in the hepatic artery phase of imaging, evaluation of portal
venous and equilibrium phase images is crucial. The excellent slice position
invariability for cine display observations enabled by navigator correction
helps radiologists observe subtle hemodynamic changes, which often are
different for hepatic tumors with different pathologic characteristics and for
a variety of vascular abnormalities
[5-7,
11].
Because we monitored the lung-liver interface every 200 milliseconds and
adapted the excitation pulse frequency during liver data acquisition
[14], the volume excited by
radiofrequency followed the liver, which moved unexpectedly owing to imperfect
breath-hold. In our phantom study, use of navigator correction resulted in
excellent slice position invariability and image quality, even for a
continuously moving phantom. In some of the patients, however, we encountered
image degradation due to imperfect breath-hold despite the use of navigator
correction. Two causes are inferred. First, recognition of the air-phantom
interface was more accurate than that of the lung-liver interface in patients,
presumably because of differences in internal magnetization homogeneity, the
intensities of navigator echoes, and the homogeneity of navigator echo data.
Second, the phantom moved at a constant speed, whereas in patients, motion was
not uniform.
In our study, no significant improvement in image quality was achieved with
navigator correction. This finding indicates that the clinical advantage of
this program in terms of improving the quality of images obtained during a
breath-hold has yet to be determined. Our subanalysis on the seven patients
with imperfect breath-holds indicated that there was no significant difference
in image quality between patients with and those without correction. Even in
such patients, however, navigator correction significantly improved slice
position invariability, preserving image quality. We inferred that at least
the fact that image quality was not significantly degraded with patient motion
indicated an advantage of the use of navigator correction.
In a patient with massive ascites beneath the right hemidiaphragm as the
result of peritoneal tumor dissemination and in another patient with a large
hepatic cyst immediately beneath the right hemidiaphragm, prospective adaptive
navigator correlation substantially degraded image quality. These adverse
effects occurred because signal intensity differences between lung and liver
were obscured by the pathologic conditions. When a patient has a lesion
beneath the right hemidiaphragm that can obscure signal intensity differences
between lung and liver, the navigator should be positioned clear of the lesion
or not be used.
Our study had several limitations. The liver not only moves up and down
during respiration but also tilts, twists, and transforms during respiratory
movement because it is secured by central tendons and the falciform and
triangular ligaments and is compressed by the right kidney, the spleen, the
colon, and rib deformation. The navigator correction program used in this
study adjusts the imaging volume to follow the movement of the liver in the
craniocaudal axis, and the correction of excitation volume only in the
craniocaudal axis may be insufficiently precise to correct such complicated
liver movement during respiration. Moreover, because of the preliminary nature
of this study, we did not determine with appropriate statistical analysis the
clinical effect of gadolinium-enhanced MRI of the liver with navigator
correction or the effect of this technique on radiologist performance, and
such study is warranted.
In conclusion, prospective adaptive navigator correction was found to
improve liver and focal hepatic lesion slice position invariability on
time-resolved multiphasic images obtained during multiple breath-holds.
Moreover, the technique enables observation of hemodynamic changes in small
hepatic lesions by cine display with preservation of image quality. We
recommend that prospective adaptive navigator correction be used for
thin-slice gadolinium-enhanced multiphasic hepatic imaging with 3D MR
sequences.
References
- Semelka RC, Shoenut JP, Kroeker MA, et al. Focal liver disease:
comparison of dynamic contrast-enhanced CT and T2-weighted fat-suppressed,
FLASH, and dynamic gadolinium-enhanced MR imaging at 1.5 T.
Radiology 1992;183
: 687-694
- Low RN, Francis IR, Sigeti JS, Foo TK. Abdominal MR imaging
comparison of T2-weighted fast and conventional spin-echo, and
contrast-enhanced fast multiplanar spoiled gradient-recalled imaging.
Radiology 1993;186
: 803-811[Abstract/Free Full Text]
- Whitney WS, Herfkens JR, Jeffrey RB, et al. Dynamic breath-hold
multiplanar spoiled gradient-recalled MR imaging with gadolinium for
differentiating hepatic hemangiomas from malignancies at 1.5 T.
Radiology 1993;189
: 863-870[Abstract/Free Full Text]
- Mitchell DG, Saini S, Weinreb JC, et al. Hepatic metastases and
cavernous hemangiomas: distinction with standard- and triple-dose
gadoteridol-enhanced MR imaging. Radiology1994; 193:49
-57[Abstract/Free Full Text]
- Ueda K, Matsui O, Kawamori Y, et al. Hypervascular hepatocellular
carcinoma: evaluation of hemodynamics with dynamic CT during hepatic
arteriography. Radiology 1998;206
: 161-166[Abstract/Free Full Text]
- Shimizu A, Ito K, Koike S, Fujita T, Shimizu K, Matsunaga N.
Cirrhosis or chronic hepatitis: evaluation of small (< or = 2-cm)
early-enhancing hepatic lesions with serial contrast-enhanced dynamic MR
imaging. Radiology 2003;226
: 550-555[Abstract/Free Full Text]
- Semelka RC, Hussain SM, Marcos HB, Woosley JT. Perilesional
enhancement of hepatic metastases: correlation between MR imaging and
histopathologic findings—initial observations.
Radiology 2000;215
: 89-94[Abstract/Free Full Text]
- Danet IM, Semelka RC, Leonardou P, et al. Spectrum of MRI
appearances of untreated metastases of the liver. AJR2003; 181:809
-817[Abstract/Free Full Text]
- Yamashita Y, Ogata I, Urata J, Takahashi M. Cavernous hemangioma of
the liver: pathologic correlation with dynamic CT findings.
Radiology 1997;203
: 121-125[Abstract/Free Full Text]
- Jeong MG, Yu JS, Kim KW. Hepatic cavernous hemangioma: temporal
peritumoral enhancement during multiphase dynamic MR imaging.
Radiology 2000;216
: 692-697[Abstract/Free Full Text]
- Holland AE, Hecht EM, Hahn WY, et al. Importance of small (< or
= 20-mm) enhancing lesions seen only during the hepatic arterial phase at MR
imaging of the cirrhotic liver: evaluation and comparison with whole explanted
liver. Radiology 2005;237
: 938-944[Abstract/Free Full Text]
- Ehman RL, Felmlee JP. Adaptive technique for high-definition MR
imaging of moving structures. Radiology1989; 173:255
-263[Abstract/Free Full Text]
- Ordidge RJ, Helpern JA, Qing ZX, Knight RA, Nagesh V. Correction of
motional artifacts in diffusion-weighted MR images using navigator echoes.
Magn Reson Imaging 1994;12
: 455-460[CrossRef][Medline]
- McConnell MV, Khasgiwala VC, Savord BJ, et al. Prospective adaptive
navigator correction for breath-hold MR coronary angiography. Magn
Reson Med 1997; 37:148
-152[Medline]
- Tyszka JM, Silverman JM. Navigated single-voxel proton spectroscopy
of the human liver. Magn Reson Med 1998;39
: 1-5[Medline]
- Tokuda J, Morikawa S, Dohi T, Hata N. Motion tracking in MR-guided
liver therapy by using navigator echoes and projection profile matching.
Acad Radiol 2004;11
: 111-120[CrossRef][Medline]
- Kim ID, Azuma T, Ido A, et al. Navigator-echobased MR provides
high-resolution images and precise volumetry of swine livers without breath
holding or injection of contrast media. Liver Transpl2006; 12:72
-77[CrossRef][Medline]
- Stuber M, Botnar RM, Danias PG, Kissinger KV, Manning WJ.
Submillimeter three-dimensional coronary MR angiography with real-time
navigator correction: comparison of navigator locations.
Radiology 1999;212
: 579-587[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
