(Radiology. 2000;217:911-915.)
© RSNA, 2000
Origin of a Signal Intensity Loss Artifact in Fat-Saturation MR Imaging1
Leon Axel, PhD, MD,
Louis Kolman,
Riad Charafeddine, MD,
Scott N. Hwang, PhD and
Alan H. Stolpen, MD, PhD
1 From the Department of Radiology, University of Pennsylvania Medical Center, 3400 Spruce St, Philadelphia, PA 19104. From the 1999 RSNA scientific assembly. Received October 20, 1999; revision requested November 18; revision received March 24, 2000; accepted April 4. Address correspondence to L.A. (e-mail: axel@oasis.rad.upenn.edu).
 |
ABSTRACT
|
|---|
Artifactual water signal intensity loss can be observed on fat-saturation magnetic resonance (MR) images of inhomogeneous regions such as the thorax. Magnetic effects of air inclusions on fat-saturation pulses were investigated as the possible origin of this artifact. Computer simulation results agreed well with observed production of water saturation by means of nominal fat suppression in MR imaging of phantoms and a representative clinical example.
Index terms: Magnetic resonance (MR), artifact, 9*.932 • Magnetic resonance (MR), experimental studies • Magnetic resonance (MR), fat suppression, 9*.129415 • Magnetic resonance (MR), vascular studies, 9*.12942 • Test objects
|
INTRODUCTION
|
|---|
Fat-saturation radio-frequency (RF) pulses in magnetic resonance (MR) imaging take advantage of the small difference in the MR frequency of hydrogen nuclei in fat and water. By using narrow-bandwidth–exciting RF pulses centered on the average frequency of fat, these RF pulses cause selective saturation of the magnetization of the fat, so that the fat signal is suppressed in the imaging excitations that immediately follow. This can be used to suppress the otherwise relatively strong fat signal in MR imaging applications such as gadolinium-enhanced MR angiography.
Artifactual loss of signal intensity in blood vessels has been observed on some thoracic MR angiographic studies that used fat saturation (1,2). In these instances, the adjacent fat was noted to not be suppressed. It has previously been demonstrated (3–5) that magnetic field inhomogeneity due to adjacent air can cause local failure of fat saturation. We hypothesized that air in the lungs caused a local perturbation of the magnetic field sufficient in magnitude and sign to cause saturation of water instead of fat in the regions where the blood vessel artifact was observed, resulting in local loss of blood signal. We sought to test this hypothesis with computer simulations and phantom imaging studies.
|
Materials and Methods
|
|---|
A spherical air collection was considered first, for simplicity. The presence of a spherical air collection in an otherwise uniform diamagnetic medium such as water, oil, or (approximately) body tissue is equivalent to having a uniform positively magnetized sphere in a uniform unmagnetized material, if we remove the baseline uniform background magnetic field. The form of the magnetic field of a uniformly magnetized sphere is well known, and there is an analytic solution for it (6,7). The field produced by the sphere will be superimposed on the main magnetic field of the imaging system, B0, which we can consider (without loss of generality) to be directed along the z direction in an x, y, z coordinate system. The effective magnetization of the sphere will also be directed along the z direction. Because the strength of B0 is much greater than the induced magnetic field from the sphere, we need consider only the z component of the field. The net z component of the magnetic field around the sphere in the x-z plane will be given by (7):
where
is the molar or volumetric susceptibility of the medium, 
is the difference in susceptibility (relative to air), R is the radius of the sphere, 
is the density of the medium, and M is the molecular weight of the medium. By symmetry, the same field distribution would be found in any plane that includes the z axis. Equations (1) and (2) were used to calculate the effects on the magnetic field of a spherical air collection in water and two oils, stearic acid and oleic acid, by using magnetic susceptibility values from standard tables (8).
The resonance frequency is proportional to the local value of the magnetic field. Hydrogen in water resonates at a slightly higher frequency than does hydrogen in fat. At 1.5 T, the difference in resonance frequency of a hydrogen nucleus in fat and water, 220 Hz, corresponds to approximately 5.2 x 10-6 T, which is on the order of the magnetic field shifts expected from the spherical air collection. In practice, the fat-saturation pulses used have finite bandwidth. The bandwidth of the fat-saturation pulse was evaluated with the MR imaging system by measuring the effect on signal frequency from a uniform phantom of progressive offsetting of the center frequency of the saturation pulse away from the resonance frequency.
To test the predictions of the theoretic model, spherical air phantoms were constructed for MR imaging. The phantoms consisted of two 2.5-L rectangular plastic tanks that were filled with either oil or water; an air-filled table tennis ball was attached to the bottom of each tank with string so that it floated in the center of the tank. Two additional tanks containing oil or water without table tennis balls were also imaged, for a total of four tanks (two oil filled and two water filled).
The tanks were imaged with a 1.5-T MR imaging system (Signa; GE Medical Systems, Waukesha, Wis), by using the following three-dimensional gradient-echo sequences: (a) with a saturation pulse centered on the resonance frequency of fat, (b) with a saturation pulse centered on the resonance frequency of water, and (c) without any saturation pulse. Specific imaging factors included fast multiplanar spoiled gradient-recalled acquisition in the steady state (90° flip angle; repetition time msec/echo time msec, 225/2.6; field of view, 24 cm; section thickness, 3 mm; image acquisition matrix, 256 x 192; four signals acquired). To minimize geometric distortion due to the susceptibility effects of the air in the table tennis balls (9), a wide imaging bandwidth (64 kHz) was used.
An additional phantom was constructed to more closely simulate the geometry of thoracic MR angiography. This phantom also used an oil-filled rectangular tank; two air-filled balloons in the oil were used to approximately simulate the effect of the lungs. Straws filled with gadolinium-doped water placed in the oil between and around the balloons were used to simulate the thoracic blood vessels.
For a representative clinical example, we evaluated a patient’s transverse (axial) MR angiograms (250/1.5, 256 x 125 matrix) on which the fat saturation-induced artifactual signal intensity loss was exhibited. To test the effects of lung susceptibility in a more fully realistic geometry, a surface triangulation algorithm was used to extract the outer contours of the lungs and the chest wall from the images. These geometric data were used to calculate the magnetic field distribution, by using a previously described surface-charge method (10). The voxels located in the lungs were first manually colored black by one of the authors (S.N.H.) using the image analysis software, NIH IMAGE (version 1.6.2; National Institutes of Health, Bethesda, Md; available at rsb .info.nih.gov/nih-image/). A surface model consisting of triangular elements was then generated from the segmented images. An effective magnetic surface-charge density was computed for each triangular element based on its orientation with the applied magnetic field and the susceptibility difference between tissue inside and that outside the lungs; the susceptibility difference was assumed to be that of free air and water (-9.05 x 10-6 meter-kilogram-second system), respectively. The induced magnetic field was then efficiently computed with a multipole-accelerated algorithm that estimates the sum of the field contributions from all triangular elements. The details of the algorithm are described by Hwang and Wehrli (10). If a value for lung susceptibility intermediate between those of air and tissue were to be used, the induced magnetic field would be scaled correspondingly.
|
Results
|
|---|
The vector field plot in Figure 1 illustrates the induced magnetic field due to the sphere, as calculated from Equation (1) and the equivalent expression for Bx. Again, only the component along B0 will be important. We can see that the induced magnetic field near the poles (relative to B0) is directed along B0, while the induced magnetic field near the equator is directed opposite to B0. The magnitude of the magnetic field shift at the poles is approximately twice that near the equator. For a sphere of air in water, the field shift at the poles for a B0 of 1.5 T is 8.4 x 10-6 T (5.6 ppm, or 355 Hz at 1.5 T), while for a sphere of air in stearic acid or oleic acid it is approximately 4.3 x 10-6 T (3.1 ppm, or 182 Hz at 1.5 T) at the poles. The difference between field shifts in water and fat is due to the difference in volumetric susceptibility, as seen in Equation (2).
View larger version (39K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. Vector field plot of perturbation of a magnetic field distribution around a spherical air collection in a uniform diamagnetic medium. The plot is in a plane parallel to the external magnetic field, B0, and passes through the center of the sphere (B0 is directed vertically). The largest field perturbation is at the "pole"; for a sphere of air in water the magnetic field shift is 5.6 ppm, or 355 Hz at 1.5 T.
|
|
Thus, an RF pulse designed to selectively saturate nuclei resonating near the nominal frequency of fat will fail to produce fat saturation near both the pole and the equator of the sphere; however, it may instead produce saturation of water near the equator. Similarly, a nominal water-saturation pulse may fail to saturate water near both the pole and the equator; however, it may produce saturation of fat near the pole.
The effective bandwidth of the saturating pulse was found to be on the order of half the frequency difference between the fat and water resonance frequencies. The results of imaging the table tennis ball phantoms in oil and water, without and with fat and water saturation, are demonstrated in Figure 2. As predicted, a fat-saturation pulse failed to suppress fat near the poles and the equator of the table tennis ball in the oil-filled tank, but the nominal fat-saturation pulse produced water suppression near the equator in the water-filled tank. The effect is demonstrated schematically in Figure 3. Similarly, water saturation failed near both the poles and the equator of the table tennis ball in the water-filled tank but produced fat suppression near the poles in the oil-filled tank.
View larger version (149K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. Gradient-echo MR images (225/2.6, 90° flip angle) of table tennis balls in tank phantoms; B0 is directed vertically on these nominal sagittal plane images. (a) Oil-filled tank, no saturation. Note the uniform oil signal intensity. (b) Oil-filled tank, fat-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows). (c) Oil-filled tank, water-saturation pulse. Note the signal intensity loss near the "poles" (arrows). (d) Water-filled tank, no saturation. Note the uniform signal intensity. (e) Water-filled tank, fat-saturation pulse. Note the signal intensity loss near the "equator" (arrows). (f) Water-filled tank, water-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows).
|
|
View larger version (157K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. Gradient-echo MR images (225/2.6, 90° flip angle) of table tennis balls in tank phantoms; B0 is directed vertically on these nominal sagittal plane images. (a) Oil-filled tank, no saturation. Note the uniform oil signal intensity. (b) Oil-filled tank, fat-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows). (c) Oil-filled tank, water-saturation pulse. Note the signal intensity loss near the "poles" (arrows). (d) Water-filled tank, no saturation. Note the uniform signal intensity. (e) Water-filled tank, fat-saturation pulse. Note the signal intensity loss near the "equator" (arrows). (f) Water-filled tank, water-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows).
|
|
View larger version (143K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2c. Gradient-echo MR images (225/2.6, 90° flip angle) of table tennis balls in tank phantoms; B0 is directed vertically on these nominal sagittal plane images. (a) Oil-filled tank, no saturation. Note the uniform oil signal intensity. (b) Oil-filled tank, fat-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows). (c) Oil-filled tank, water-saturation pulse. Note the signal intensity loss near the "poles" (arrows). (d) Water-filled tank, no saturation. Note the uniform signal intensity. (e) Water-filled tank, fat-saturation pulse. Note the signal intensity loss near the "equator" (arrows). (f) Water-filled tank, water-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows).
|
|
View larger version (134K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2d. Gradient-echo MR images (225/2.6, 90° flip angle) of table tennis balls in tank phantoms; B0 is directed vertically on these nominal sagittal plane images. (a) Oil-filled tank, no saturation. Note the uniform oil signal intensity. (b) Oil-filled tank, fat-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows). (c) Oil-filled tank, water-saturation pulse. Note the signal intensity loss near the "poles" (arrows). (d) Water-filled tank, no saturation. Note the uniform signal intensity. (e) Water-filled tank, fat-saturation pulse. Note the signal intensity loss near the "equator" (arrows). (f) Water-filled tank, water-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows).
|
|
View larger version (136K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2e. Gradient-echo MR images (225/2.6, 90° flip angle) of table tennis balls in tank phantoms; B0 is directed vertically on these nominal sagittal plane images. (a) Oil-filled tank, no saturation. Note the uniform oil signal intensity. (b) Oil-filled tank, fat-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows). (c) Oil-filled tank, water-saturation pulse. Note the signal intensity loss near the "poles" (arrows). (d) Water-filled tank, no saturation. Note the uniform signal intensity. (e) Water-filled tank, fat-saturation pulse. Note the signal intensity loss near the "equator" (arrows). (f) Water-filled tank, water-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows).
|
|
View larger version (128K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2f. Gradient-echo MR images (225/2.6, 90° flip angle) of table tennis balls in tank phantoms; B0 is directed vertically on these nominal sagittal plane images. (a) Oil-filled tank, no saturation. Note the uniform oil signal intensity. (b) Oil-filled tank, fat-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows). (c) Oil-filled tank, water-saturation pulse. Note the signal intensity loss near the "poles" (arrows). (d) Water-filled tank, no saturation. Note the uniform signal intensity. (e) Water-filled tank, fat-saturation pulse. Note the signal intensity loss near the "equator" (arrows). (f) Water-filled tank, water-saturation pulse. Note the loss of signal intensity except near the "poles" (long arrows) and the "equator" (short arrows).
|
|
View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3. Schematic of the effect of susceptibility-induced magnetic field shift on the results of a nominal fat-saturation pulse applied to fat and water. Dashed line = center frequency of saturation pulse, f = frequency. A, Frequency distribution of a nominal fat-saturation RF pulse. B, Resonance frequencies of fat and water in a homogeneous portion of the magnetic field remote from the susceptibility-induced disturbance. Fat signal will be suppressed by the fat-saturation pulse, but water will be unaffected. C, Resonance frequencies of fat and water near the "pole" of an air collection, where the magnetic field is locally increased. Neither fat nor water will be affected by the fat-saturation pulse, due to the resonance frequency shifts. D, Resonance frequencies of fat and water near the "equator" of an air collection where the magnetic field is locally decreased. Fat will be unaffected but water signal may be suppressed, due to the resonance frequency shifts.
|
|
In the double-balloon phantom, the effects of local magnetic field lowering near the equator were additive in the region between the two balloons (roughly simulating the mediastinum), resulting in a relatively larger region of failure of fat saturation. There was production of water saturation in the water-filled straws by the nominal fat-saturation pulse in a smaller portion of this region, corresponding to the blood vessel signal intensity loss artifact observed at fat-suppressed thoracic MR angiography (Fig 4).
View larger version (109K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. Gradient-echo MR images (225/2.6, 90° flip angle) of a phantom with two air-filled balloons in an oil-filled tank simulating lungs and mediastinum and with water-filled straws simulating vessels between the balloons. (a) Image without fat-saturation pulse shows uniform signal intensities in the oil and in the water-filled tubes. (b) Image with fat-saturation pulse shows failure of fat signal suppression in regions near the "apices" (long arrows) and in the "mediastinum" (short arrow), with water signal intensity suppression in part of the "mediastinum."
|
|
View larger version (116K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. Gradient-echo MR images (225/2.6, 90° flip angle) of a phantom with two air-filled balloons in an oil-filled tank simulating lungs and mediastinum and with water-filled straws simulating vessels between the balloons. (a) Image without fat-saturation pulse shows uniform signal intensities in the oil and in the water-filled tubes. (b) Image with fat-saturation pulse shows failure of fat signal suppression in regions near the "apices" (long arrows) and in the "mediastinum" (short arrow), with water signal intensity suppression in part of the "mediastinum."
|
|
A representative image of the chest from the set of MR images acquired with fat-saturation during enhancement with a gadolinium-based contrast material that demonstrated the artifact is shown in Figure 5a. The magnetic field, B0, is perpendicular to the image plane. Note the failure of fat saturation in the left side of the mediastinum, with saturation of the blood in the subclavian artery in the middle of this region. As seen in Figure 5b, obtained at a different level of the chest in the same study, fat saturation did not produce the artifact elsewhere, where the susceptibility-induced magnetic field shifts were smaller. A gray-scale image display of the corresponding calculated magnetic effect of the air in the lungs at the same level as Figure 5a is shown in Figure 5c. The average shift of the magnetic field in the region corresponding to the left side of the mediastinum, outlined in Figure 5d, is -4.8 x 10-6 T (-3.2 ppm, or -203 Hz at 1.5 T), enough to shift the local water resonance frequency into the frequency range of the nominal fat-saturation pulse. For a lower susceptibility difference between lung and mediastinum, this will scale correspondingly. For example, if the lung is assumed to have a density of approximately 0.2 g/mL, the average local resonance shift would be -2.6 ppm, or -165 Hz at 1.5 T, still within the range of the fat-saturation pulse.
View larger version (142K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a. (a) Transverse MR image of the upper chest acquired as part of a three-dimensional fat-saturation gadolinium-enhanced gradient-echo (250/1.5, 20° flip angle) MR angiographic study. Artifacts of poor fat suppression and loss of blood signal intensity in the subclavian artery (arrow) are seen on the left side of the mediastinum. (b) Transverse MR image of the chest at a higher level (near the left pulmonary apex) from the same study as in a shows the expected uniformly low signal intensity from fat and high signal intensity from blood in vessels. (c) Altered magnetic field distribution due to air in lungs, calculated with the surface-charge method (9) for the lung geometry of the chest in a. Lighter regions denote locally reduced magnetic field, and darker regions denote increased magnetic field; lungs and airways are shown as black. (d) Same image as in a but with highlighted outline of the region over which the mean left mediastinal magnetic field shift was calculated (-3.2 ppm, or -203 Hz at 1.5 T).
|
|
View larger version (148K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b. (a) Transverse MR image of the upper chest acquired as part of a three-dimensional fat-saturation gadolinium-enhanced gradient-echo (250/1.5, 20° flip angle) MR angiographic study. Artifacts of poor fat suppression and loss of blood signal intensity in the subclavian artery (arrow) are seen on the left side of the mediastinum. (b) Transverse MR image of the chest at a higher level (near the left pulmonary apex) from the same study as in a shows the expected uniformly low signal intensity from fat and high signal intensity from blood in vessels. (c) Altered magnetic field distribution due to air in lungs, calculated with the surface-charge method (9) for the lung geometry of the chest in a. Lighter regions denote locally reduced magnetic field, and darker regions denote increased magnetic field; lungs and airways are shown as black. (d) Same image as in a but with highlighted outline of the region over which the mean left mediastinal magnetic field shift was calculated (-3.2 ppm, or -203 Hz at 1.5 T).
|
|
View larger version (85K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5c. (a) Transverse MR image of the upper chest acquired as part of a three-dimensional fat-saturation gadolinium-enhanced gradient-echo (250/1.5, 20° flip angle) MR angiographic study. Artifacts of poor fat suppression and loss of blood signal intensity in the subclavian artery (arrow) are seen on the left side of the mediastinum. (b) Transverse MR image of the chest at a higher level (near the left pulmonary apex) from the same study as in a shows the expected uniformly low signal intensity from fat and high signal intensity from blood in vessels. (c) Altered magnetic field distribution due to air in lungs, calculated with the surface-charge method (9) for the lung geometry of the chest in a. Lighter regions denote locally reduced magnetic field, and darker regions denote increased magnetic field; lungs and airways are shown as black. (d) Same image as in a but with highlighted outline of the region over which the mean left mediastinal magnetic field shift was calculated (-3.2 ppm, or -203 Hz at 1.5 T).
|
|
View larger version (136K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5d. (a) Transverse MR image of the upper chest acquired as part of a three-dimensional fat-saturation gadolinium-enhanced gradient-echo (250/1.5, 20° flip angle) MR angiographic study. Artifacts of poor fat suppression and loss of blood signal intensity in the subclavian artery (arrow) are seen on the left side of the mediastinum. (b) Transverse MR image of the chest at a higher level (near the left pulmonary apex) from the same study as in a shows the expected uniformly low signal intensity from fat and high signal intensity from blood in vessels. (c) Altered magnetic field distribution due to air in lungs, calculated with the surface-charge method (9) for the lung geometry of the chest in a. Lighter regions denote locally reduced magnetic field, and darker regions denote increased magnetic field; lungs and airways are shown as black. (d) Same image as in a but with highlighted outline of the region over which the mean left mediastinal magnetic field shift was calculated (-3.2 ppm, or -203 Hz at 1.5 T).
|
|
|
Discussion
|
|---|
The results of these computational and phantom studies have confirmed our hypothesis that the artifact of blood vessel signal intensity loss observed on some thoracic MR angiographic studies with fat saturation is due to local magnetic field shifts lowering the regional water resonance frequency into the bandwidth of the saturation pulse. These magnetic field shifts are due to susceptibility effects of air in the lungs. Not only was the effect duplicated with the experimental phantom studies, but the theory also correctly predicted the magnitude and location of the effect in a representative clinical example.
The interfering effect of air collections on fat saturation, due to their magnetic susceptibility difference from surrounding tissue, has been previously described (3–5). However, the effect we described of actually producing water saturation with a nominal fat-saturation pulse in a suitable imaged object geometry has not, to our knowledge, been previously demonstrated and mathematically modeled.
The failure of fat saturation in the presence of magnetic field inhomogeneity may be due to the local field being either lower or higher than the average field in an amount sufficient to move the local fat resonance frequency out of the range of the bandwidth of the nominal fat-saturation pulse. However, for local magnetic field inhomogeneity to lead to water suppression by a nominal fat-saturation pulse, the local field must specifically be lower than the average field in an amount sufficient to move the local water resonance frequency into the range of the bandwidth of the nominal fat-saturation pulse.
The clinical importance of this effect is that it must be kept in mind to avoid misinterpreting the resulting artifactual signal intensity loss as a lesion, such as vascular stenosis or occlusion. This artifact is fundamentally different from the signal intensity loss that can be seen in the presence of strong magnetic field inhomogeneity (eg, near metal or even locally concentrated gadolinium-based contrast agent in a vein), which is independent of fat saturation, whereas the artifact studied here depends intrinsically on the use of a fat-saturation pulse.
Although invoked here to explain a thoracic MR angiography artifact that can be seen when imaging with fat saturation, this effect can potentially be seen when using fat-saturation MR imaging of any magnetically inhomogeneous portion of the body. In the specific case of thoracic MR angiography, it may be best not to routinely use fat saturation when studying mediastinal vessels. If the effect is suspected on fat-saturation MR images, the MR image acquisition can be repeated without fat saturation to see if it goes away, as it will when the fat-saturation pulse is the cause of such artifactual signal loss.
|
FOOTNOTES
|
|---|
*. Vascular system, location unspecified 
Abbreviation: RF = radio frequency
Author contributions: Guarantor of integrity of entire study, L.A.; study concepts, L.A.; study design, L.A., L.K., A.H.S.; definition of intellectual content, L.A.; literature research, L.A., L.K.; experimental studies, L.A., L.K., R.C.; data acquisition, L.A., L.K., R.C.; data analysis, L.A., S.N.H.; manuscript preparation, L.A.; manuscript editing, L.A.; manuscript review and final version approval, all authors.
|
REFERENCES
|
|---|
-
Siegelman ES, Charafeddine R, Stolpen AH, Axel L. Suppression of intravascular signal on fat-saturated contrast-enhanced thoracic MR arteriograms. Radiology 2000; 217:115-118.[Abstract/Free Full Text]
-
Stolpen AH, Charafeddine R, Siegelman ES, Axel L. Fat suppression causes artifactual signal loss in the aortic arch and great vessels (abstr). Proceedings of the Seventh Meeting of the International Society for Magnetic Resonance in Medicine Berkeley, Calif: International Society for Magnetic Resonance in Medicine, 1999; 1987.
-
Anzai Y, Lufkin RB, Jabour BA, Hanafee WN. Fat suppression failure artifacts simulating pathology on frequency-selective fat-suppression MR images of the head and neck. AJNR Am J Neuroradiol 1992; 13:879-884.[Abstract]
-
Yoshimitsu K, Varma DG, Jackson EF. Unsuppressed fat in the right anterior diaphragmatic region on fat-suppressed T2-weighted fast spin-echo MR images. J Magn Reson Imaging 1995; 5:145-149.[Medline]
-
Borges AR, Lufkin RB, Huang AY, Farahani K, Arnold AC. Frequency-selective fat suppression MR imaging: localized asymmetric failure of fat suppression mimicking orbital disease. J Neuroophthalmol 1997; 17:12-17.[Medline]
-
Stratton J. Electromagnetic theory New York, NY: McGraw-Hill, 1941.
-
Jackson J. Classical electrodynamics New York, NY: Wiley, 1962.
-
Lide D. Handbook of chemistry and physics 76th ed. New York, NY: CRC, 1995.
-
Ludeke KM, Roschmann P, Tischler R. Susceptibility artifacts of NMR imaging. Magn Reson Imaging 1985; 3:329-343.[Medline]
-
Hwang S, Wehrli F. The calculation of the susceptibility-induced magnetic field from 3D NMR images with applications to trabecular bone. J Magn Reson B 1995; 109:126-145.
TOP
ABSTRACT
INTRODUCTION
