DOI:10.2214/AJR.05.0695
AJR 2006; 187:W103-W106
© American Roentgen Ray Society
Opposed-Phase MRI for Fat Quantification in Fat-Water Phantoms with 1H MR Spectroscopy to Resolve Ambiguity of Fat or Water Dominance
Jerry S. Chang1,
Bachir Taouli2,
Nouha Salibi3,
Elizabeth M. Hecht2,
Deanna G. Chin4 and
Vivian S. Lee2
1 Westwood Radiology Associates, Pascack Valley Hospital, Westwood, NY
07675.
2 Department of Radiology, NYU Medical Center, 560 First Ave., TCH-HW 202, New
York, NY 10016.
3 Siemens Medical Solutions, Malvern, PA 19355.
4 Department of Radiology, Robert Wood Johnson University Hospital, East
Brunswick, NJ 08816.
Received April 23, 2005;
accepted after revision August 22, 2005.
Jerry S. Chang and Bachir Taouli contributed equally to this article.
Address correspondence to B. Taouli
(bachir.taouli{at}med.nyu.edu).
WEB
This is a Web exclusive article.
Abstract
OBJECTIVE. The purpose of the study was to quantify the fat fraction
in nine fat-water phantoms containing 0-80% fat using opposed-phase imaging
with the qualitative guidance of 1H MR spectroscopy (MRS), which
was used by observer 1 to visually assess the sizes of the water and fat peaks
to apply two alternative mathematic formulas for the calculation of the fat
fraction. In addition, the fat fraction was also quantified directly with
1H MRS as an independent method by two observers (observers 2 and
3).
CONCLUSION. The fat fraction calculated with opposed-phase imaging
(FFOPI) and that calculated with 1H MRS
(FFMRS) correlated well with the known fat fractions of the
phantoms (FFP): r = 0.99 for FFOPI; p
< 0.0001 and r = 0.96-0.98 for FFMRS; p <
0.001, for observers 2 and 3, respectively. Opposed-phase imaging should be
combined with 1H MRS to ensure accurate quantification of the fat
fraction.
Keywords: liver disease • MRI • SPECT
Introduction
The opposed-phase MRI technique is based on phase differences in images
acquired using different TEs. Fat and water signals are additive on in-phase
images and subtracted on opposed-phase images. Therefore, opposed-phase
imaging reduces the signal from tissues containing intravoxel fat.
Opposed-phase imaging is commonly used to identify tissues containing a
significant proportion of intracellular microscopic triglycerides, such as
fatty liver
[1-5].
However, quantification of the hepatic fat fraction might be limited with
opposed-phase imaging because of the ambiguity of fat or water
dominance—that is, a high fat fraction will result in a similar signal
intensity (SI) change on opposed-phase imaging as will a low fat fraction. A
high or low percentage of fat determines a signal drop on opposed-phase MRI
that looks similar at qualitative visual analysis. Consequently, to date,
opposed-phase imaging has proven inaccurate for quantification of high fat
fraction in the liver [1,
5]. To help resolve fat-water
ambiguity, qualitative assessment of a quick 1H MR spectroscopic
(MRS) acquisition can be performed. To the best of our knowledge, there is no
study that has combined the use of opposed-phase imaging with qualitative
1H MRS to quantify the fraction of fat in the liver.
The purpose of our investigation was to correlate the fraction of fat in
fat-water phantoms measured on opposed-phase imaging (FFOPI) and to
use 1H MRS to resolve the ambiguity of fat or water dominance.
Materials and Methods
Phantom Composition
Nine phantoms with various fractions of fat were produced by homogenizing a
mixture of commercially available vegetable canola oil and fresh calf liver.
Liver and oil were mixed using a blender, and a homogeneous mayonnaise-like
emulsion was subsequently obtained. The mixture was put in a plastic container
measuring 6 x 8 cm.
The phantoms contained 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, and 80% fat
by weight (using a commercially available scale). Each phantom was placed in a
water bath to augment magnetic field homogeneity. The water bath served as an
SI calibrator for opposed-phase images.
MRI
The following sequences were performed using a 1.5-T magnet (Magnetom
Quantum Symphony, Siemens Medical Solutions) and a body coil.
Opposed-phase imaging—A dual-echo opposed-phase
gradient-recalled echo T1-weighted sequence was performed in the coronal plane
using the following parameters: TR, 167 msec; TE, 4.76 (in-phase) and 2.38
(opposed-phase) msec; flip angle, 80°; matrix, 95 x 256; field of
view, 150 x 300 mm; 1 signal averaged; 4 slices of 8-mm thickness and
2-mm gap; and acquisition time, 14 seconds.
View larger version (8K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —In-phase and opposed-phase imaging of fat-water phantoms.
Curve of signal intensity (SI) of fat-water phantoms on in-phase (•) and
opposed-phase ( ) images versus phantom fat fraction (FFP).
Arrow points to maximum SI loss on opposed-phased images for fat fraction of
50%.
|
|
View larger version (32K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —In-phase and opposed-phase imaging of fat-water phantoms.
Corresponding in-phase and opposed-phase images and 1H MR
spectroscopy (MRS) images of three fat-water phantoms placed in water bath;
phantoms are composed of 30% (B), 50% (C), or 70% (D)
fat. Water peak (short arrow) is 4.7 ppm and fat peak (long
arrow) ranges from 1 to 1.5 ppm on MRS images. Example of placement of
regions of interest (circles, C) is shown in C. Maximum
signal loss on opposed-phase images is seen for phantom with 50% fat
(C). Phantoms with 30% (A) and 70% (B) fat are
indistinguishable on opposed-phase images. Without MRS to resolve for
fat-water dominance, high fat fraction may be misinterpreted as low fat
fraction.
|
|
View larger version (30K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1C —In-phase and opposed-phase imaging of fat-water phantoms.
Corresponding in-phase and opposed-phase images and 1H MR
spectroscopy (MRS) images of three fat-water phantoms placed in water bath;
phantoms are composed of 30% (B), 50% (C), or 70% (D)
fat. Water peak (short arrow) is 4.7 ppm and fat peak (long
arrow) ranges from 1 to 1.5 ppm on MRS images. Example of placement of
regions of interest (circles, C) is shown in C. Maximum
signal loss on opposed-phase images is seen for phantom with 50% fat
(C). Phantoms with 30% (A) and 70% (B) fat are
indistinguishable on opposed-phase images. Without MRS to resolve for
fat-water dominance, high fat fraction may be misinterpreted as low fat
fraction.
|
|
View larger version (34K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1D —In-phase and opposed-phase imaging of fat-water phantoms.
Corresponding in-phase and opposed-phase images and 1H MR
spectroscopy (MRS) images of three fat-water phantoms placed in water bath;
phantoms are composed of 30% (B), 50% (C), or 70% (D)
fat. Water peak (short arrow) is 4.7 ppm and fat peak (long
arrow) ranges from 1 to 1.5 ppm on MRS images. Example of placement of
regions of interest (circles, C) is shown in C. Maximum
signal loss on opposed-phase images is seen for phantom with 50% fat
(C). Phantoms with 30% (A) and 70% (B) fat are
indistinguishable on opposed-phase images. Without MRS to resolve for
fat-water dominance, high fat fraction may be misinterpreted as low fat
fraction.
|
|
1H MRS—A single-voxel 1H MRS sequence
(point-resolved spectroscopy [PRESS]) was performed after the opposed-phase
imaging sequence using the following parameters: TR/TE, 2,000/30; voxel size,
2.5 x 2.5 x 2.5 cm; 12 signals averaged; acquisition time, 24
seconds; and spectral width, 6.2 ppm.
Image Evaluation
Qualitative evaluation—In- and opposed-phase phantom images
were visually inspected for SI changes by a single experienced observer
(observer 1) who was a clinical fellow in body MRI at time of the study.
Fat quantification analysis on opposed-phase imaging—With
care to avoid bubbles, the same observer (observer 1) used five manually drawn
regions of interest (ROIs) randomly selected on each phantom to measure the
mean SI on in-phase images (SIIP) and the mean SI on opposed-phase
images (SIOP) of the ROIs, with an exact matching location, and
measurements were averaged. Random measure was possible after ensuring
homogeneous preparation of the phantom using a blender. ROIs included at least
200 pixels. SIIP and SIOP were normalized to the SI of
the water bath on in- and opposed-phase images. Opposed-phase images show SI
loss because of cancellation of fat and water signals due to differences in
precession frequencies so that, theoretically, signal loss on opposed-phase
images should be highest for a fat fraction of 50% (Figs.
1A,
1B,
1C, and
1D).
View larger version (4K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2 —Correlation between fat fraction calculated using
opposed-phase imaging (FFOPI) with 1H MR spectroscopy
for qualitative analysis to resolve ambiguity of fat or water dominance in
phantom with known fat fraction (FFP). y = 1.09x - 0.02,
R2 = 0.99, p < 0.001.
|
|
The fat fraction on opposed-phase imaging was calculated as a percentage
using the Dixon technique [1]
with 1H MRS to resolve the ambiguity of fat or water dominance
based on a visual analysis of the size of the fat and water peaks, located at
1.0-1.5 and 4.7 ppm, respectively. If water > fat, FFOPI =
(SIIP / SIIP water bath - SIOP / SIOP
water bath) / (2 x SIIP / SIIP water bath)
x 100. If fat 
water, then FFOPI = 1 - (SIIP /
SIIP water bath - SIOP / SIOP water bath) /
(2 x SIIP / SIIP water bath) x 100.
Fat quantification analysis on MRS—As an independent method,
the areas under the water and fat peaks on 1H MRS were analyzed
quantitatively by two observers (observers 2 and 3), different from the
observer mentioned earlier, who were blinded to the known phantom composition
and opposed-phase imaging results. They calculated the fat fraction with
1H MRS using commercially available software (Syngo, Siemens
Medical Solutions): FFMRS (%) = [fat peak / (fat peak +
water peak)] x 100. The fat peak was measured at 1.0-1.5 ppm, and the
water peak was measured at 4.7 ppm. The 1H MRS postprocessing time
was less than 1 minute for both qualitative and quantitative use.
Statistical Analysis
Statistical analysis was performed using Minitab software (version 14,
Minitab Inc.). The error range and mean absolute error of the fat fractions
calculated using opposed-phase imaging and 1H MRS were measured.
The FFOPI (for observer 1) and FFMRS (for observers 2
and 3) were correlated with the known phantom fat fraction (FFP)
using Spearman's rank correlation coefficient test. A correlation was defined
as strong when r > 0.8, moderate for r ranging between
0.6 and 0.8, and weak when r < 0.6. In addition, FFMRS
for observers 2 and 3 were correlated using the Spearman's rank correlation
coefficient test and by calculating the mean absolute percentage difference
between the two observers as FFMRS for observer 2 -
FFMRS for observer 3.
Results
Opposed-Phase Imaging
By visual inspection, the phantoms with 30-70% fat (Figs.
1A,
1B,
1C, and
1D) and 40-60% fat were
indistinguishable on in- and opposed-phase images. The relationship between SI
on in-phase and opposed-phase images and the known phantom fat fraction
(FFP) is illustrated on the curve in Figures
1A,
1B,
1C, and
1D, which shows maximum signal
loss on opposed-phase images when FFP is 50% and comparable degrees
of SI loss in phantoms with small or large amounts of fat.
With the visual analysis of 1H MRS fat-water peak size to guide
for fat and water dominance on in-phase and opposed-phase images, there was a
strong correlation (r = 0.99, p < 0.0001) between the fat
fraction calculated with opposed-phase imaging (FFOPI) and the
known fat fraction (FFP) (Fig.
2). The mean absolute error was 3% (range, -3% to 6%). When
opposed-phase imaging was used to quantify the fat fraction without
1H MRS, there was a weak correlation between the fat fraction
calculated with opposed-phase imaging and the known fat fraction (r =
0.47, p = 0.2).
1H MRS
There was a strong correlation between the fat fraction calculated using
1H MRS (FFMRS) and the known fat fraction
(FFP) for both observers (r = 0.96, p < 0.001
for observer 2; and r = 0.98, p < 0.001 for observer 3).
FFMRS from observers 2 and 3 were strongly correlated (r =
0.96, p < 0.001). The mean absolute percentage difference for
FFMRS measurement between observers was 13% (range, 6-20%), which
shows good agreement.
Discussion
Our in vitro study showed that fat quantification is inaccurate when
opposed-phase imaging is used alone—that is, without 1H MRS
to resolve the ambiguity of fat or water dominance. Opposed-phase imaging is
known to be an accurate method to detect microscopic fat
[6]; however, data about the
use of opposedphase imaging for liver fat quantification are limited
[1-3,
5]. One study combined in vitro
and in vivo data [3] and showed
a strong correlation between FFOPI and FFP
(R2 = 0.98); however, phantoms had a maximum fat fraction
of 40%. Two previous studies
[1,
5] using opposed-phase imaging
in patients showed a good correlation between FFOPI and histologic
fat score (r = 0.84-0.86), but opposed-phase imaging underestimated
fat fraction in subjects with high histologic fat scores, likely related to
the lack of resolution for fat-water ambiguity with opposed-phase imaging, as
discussed by Levenson et al.
[1]. 1H MRS allows
the examination of the resonance frequencies of all protons within a volume of
interest and can provide an accurate assessment of liver fat using
commercially available software
[7-11].
The number of protons contributing to a given spectral peak is directly
proportional to the area under the peak. It allows in vivo quantitative
evaluation of metabolites such as water and fat within a selected volume of
interest. 1H MRS provides also a quick method with which to
determine fat or water dominance for simple fat calculation using
opposed-phase imaging.
However, 1H MRS is slightly more time consuming and operator
dependent than opposed-phase imaging, with some degree of interobserver
variability, and it evaluates only a portion of liver parenchyma. In addition,
quantitative analysis requires the correction for signal decay from T2
relaxation. Signal saturation from incomplete T1 relaxation is minimized by
using a long TR. Allowing for these corrections, Longo et al.
[7,
9] studied a population of
patients with liver steatosis and found a good correlation between fat
fraction measured with 1H MRS and CT and the histologic results. In
another study in patients with alcoholic steatosis, Thomsen et al.
[8] reported a strong
correlation between MRS and histology (r = 0.9).
Our study has some limitations: First, it included phantom data only,
without correction for T2 decay; a clinical validation of the combined use of
opposed-phase imaging with 1H MRS is needed. Second, qualitative
analysis of opposed-phase images was performed by a single observer, who was a
clinical fellow at that time and was interpreting opposed-phase images
routinely.
The underlying mechanism of liver inflammation and fibrosis in nonalcoholic
steatohepatitis is still unclear, and it has not been clearly proven that
simple steatosis without associated inflammation or fibrosis increases the
risk of liver damage. Despite these facts, opposed-phase imaging and
1H MRS could have important potential clinical applications by
facilitating diagnosis and monitoring disease progression in a large
population of patients including liver donors, obese patients, and diabetic
patients and could obviate liver biopsy
[5,
11-14].
In conclusion, an accurate estimation of fat fraction can be obtained using
opposed-phase imaging with qualitative 1H MRS to resolve the
ambiguity of fat or water dominance.
References
- Levenson H, Greensite F, Hoefs J, et al. Fatty infiltration of the
liver: quantification with phase-contrast MR imaging at 1.5 T vs biopsy.
AJR 1991; 156:307
-312[Abstract/Free Full Text]
- Mitchell DG, Kim I, Chang TS, et al. Fatty liver: chemical shift
phase-difference and suppression magnetic resonance imaging techniques in
animals, phantoms, and humans. Invest Radiol1991; 26:1041
-1052[Medline]
- Fishbein MH, Gardner KG, Potter CJ, Schmalbrock P, Smith MA.
Introduction of fast MR imaging in the assessment of hepatic steatosis.
Magn Reson Imaging 1997;15
: 287-293[CrossRef][Medline]
- Earls JP, Krinsky GA. Abdominal and pelvic applications of
opposed-phase MR imaging. AJR 1997;169
: 1071-1077[Free Full Text]
- Rinella ME, McCarthy R, Thakrar K, et al. Dualecho, chemical shift
gradient-echo magnetic resonance imaging to quantify hepatic steatosis:
implications for living liver donation. Liver Transpl2003; 9:851
-856[CrossRef][Medline]
- Outwater EK, Blasbalg R, Siegelman ES, Vala M. Detection of lipid
in abdominal tissues with opposed-phase gradient-echo images at 1.5 T:
techniques and diagnostic importance. RadioGraphics1998; 18:1465
-1480[Abstract]
- Longo R, Ricci C, Masutti F, et al. Fatty infiltration of the
liver: quantification by 1H localized magnetic resonance
spectroscopy and comparison with computed tomography. Invest
Radiol 1993; 28:297
-302[Medline]
- Thomsen C, Becker U, Winkler K, Christoffersen P, Jensen M,
Henriksen O. Quantification of liver fat using magnetic resonance
spectroscopy. Magn Reson Imaging 1994;12
: 487-495[CrossRef][Medline]
- Longo R, Pollesello P, Ricci C, et al. Proton MR spectroscopy in
quantitative in vivo determination of fat content in human liver steatosis.
J Magn Reson Imaging 1995;5
: 281-285[Medline]
- Szczepaniak LS, Nurenberg P, Leonard D, et al. Magnetic resonance
spectroscopy to measure hepatic triglyceride content: prevalence of hepatic
steatosis in the general population. Am J Physiol Endocrinol
Metab 2005; 288:E462
-E468[Abstract/Free Full Text]
- Thomas EL, Hamilton G, Patel N, et al. Hepatic triglyceride content
and its relation to body adiposity: a magnetic resonance imaging and proton
magnetic resonance spectroscopy study. Gut2005; 54:122
-127[Abstract/Free Full Text]
- Anderwald C, Bernroider E, Krssak M, et al. Effects of insulin
treatment in type 2 diabetic patients on intracellular lipid content in liver
and skeletal muscle. Diabetes 2002;51
: 3025-3032[Abstract/Free Full Text]
- Seppala-Lindroos A, Vehkavaara S, Hakkinen AM, et al. Fat
accumulation in the liver is associated with defects in insulin suppression of
glucose production and serum free fatty acids independent of obesity in normal
men. J Clin Endocrinol Metab 2002;87
: 3023-3028[Abstract/Free Full Text]
- Tiikkainen M, Bergholm R, Vehkavaara S, et al. Effects of identical
weight loss on body composition and features of insulin resistance in obese
women with high and low liver fat content. Diabetes2003; 52:701
-707[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
N. Pineda, P. Sharma, Q. Xu, X. Hu, M. Vos, and D. R. Martin
Measurement of Hepatic Lipid: High-Speed T2-Corrected Multiecho Acquisition at 1H MR Spectroscopy--A Rapid and Accurate Technique
Radiology,
August 1, 2009;
252(2):
568 - 576.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
F. H. Cassidy, T. Yokoo, L. Aganovic, R. F. Hanna, M. Bydder, M. S. Middleton, G. Hamilton, A. D. Chavez, J. B. Schwimmer, and C. B. Sirlin
Fatty Liver Disease: MR Imaging Techniques for the Detection and Quantification of Liver Steatosis1
RadioGraphics,
January 1, 2009;
29(1):
231 - 260.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. P. O'Regan, M. F. Callaghan, M. Wylezinska-Arridge, J. Fitzpatrick, R. P. Naoumova, J. V. Hajnal, and S. A. Schmitz
Liver Fat Content and T2*: Simultaneous Measurement by Using Breath-hold Multiecho MR Imaging at 3.0 T--Feasibility
Radiology,
May 1, 2008;
247(2):
550 - 557.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
