Post on 13-Oct-2020
M. Lustig, EECS UC Berkeley
Principles of MRIEE225E / BIO265
Lecture 15
Instructor: Miki LustigUC Berkeley, EECS
1
M. Lustig, EECS UC Berkeley
Last Time
• Practical Issues in MRI• T2 decay
– Map decay to k-space– result in artifacts
• Image weighting• blurring in the readout
• Started talking about off-resonance
2
M. Lustig, EECS UC Berkeley
Effect of T2 Decay on Imaging
• Signal decays along k-space trajectory
s(t) =
Z
~
R
Mxy
(~r, 0)e�t
T2 e�i2⇡~k(t)·~rd~r
tt=0 t=TE
Approximation:Mxy created mid-RFDecays with T2 after
3
kx
=�
2⇡G
x
(t� TE) ) t =kx
�
2⇡Gx
+ TE
e�t
T2 = e�TE
T2 e� k
x
�
2⇡G
x
T2
M. Lustig, EECS UC Berkeley
Effect of T2 Decay on Imagingk-space
kx
(t)t
t=TE
4
M. Lustig, EECS UC Berkeley
Point Spread Function
t=10ms
T2=20ms
T2=5ms
Decay PSF
T2=100msT2=20msT2=5ms
TE=10ms, Treadout=20ms
5
M. Lustig, EECS UC Berkeley
Effect of T2 Decay on Imaging
• Two effects:– Signal loss by exp(-TE/T2) (T2 weighting)– Apodization - causes blurring in readout
T2=100msT2=20msT2=5ms
TE=10ms, Treadout=20ms
6
M. Lustig, EECS UC Berkeley
Off - Resonance
• So far, assumed B0 constant. But B0 varies due to:– Main field inhomogeneity (~ 1ppm)– Object magnetic susceptibility– Chemical shift
7
M. Lustig, EECS UC Berkeley
Main Field Inhomogeneity
• Magnet is designed to be homogeneous over a (spherical) volume
• Typical numbers:– Bare magnet ~10-100 ppm– Shimmed magnet ~1 ppm
@3T 1ppm is 127Hz
• Generally, main magnet inhomogeneity is not a limitation
8
M. Lustig, EECS UC Berkeley
Object Susceptibility
• Most biological objects perturb the field
• χ is the magnetic susceptibility• Larmor frequency lower in tissue than air χwater = -9.05ppm w.r.t free-spaceχair = 0.36 ppm w.r.t free-space
�Bz,tissue ⇡ �B0,freespace
�Bz =
��B0
3
⇣ar
⌘3(3 cos
2 ✓ � 1)
for a sphere:
9
M. Lustig, EECS UC Berkeley
Object Susceptibility
• Complex behavior at boundaries.– Depends on Δχ and geometry– Typical ΔB0 ± 3 ppm (-12< χ < -6)
sphere
cylinder
cylinder
10
M. Lustig, EECS UC Berkeley
Macroscopic Effect
• Problem areas:– Brain above sinuses, auditory canals– Heart surrounded by lungs– Abdomen
Therefore, the method is likely to penalize a small re-gion (e.g., the frontal lobe area) while reducing the over-all off-resonance frequency in other regions. The result-ing field map may exhibit a large peak off-resonanceregion as shown in Fig. 1e and f. In bSSFP, this largepeak off-resonance can result in banding artifacts (theyellow arrow in Fig. 1a, and the red strip in Fig. 1f). Insome cases, one can manipulate the shim parameterssuch that the maximum off-resonance frequency is re-duced to avoid the artifacts (Fig. 1g and h). In thisexample, the center frequency and y-shim were modi-fied to decrease the peak off-resonance frequency. Theresulting field map has a larger mean square error com-pared to the least-squares shim result but the image hasno artifact (Fig. 1d). Hence, one can achieve a widerspatial coverage by reducing the peak off-resonance fre-quency.
This approach can be restated as follows: in bSSFP,better spatial coverage can be obtained by minimizing themaximum absolute off-resonance frequency of the fieldmap (min-max shim). It can be formulated as follows (2):
s0,sx,sy. . .argmin max[|F(x,y,z) ! F0s0 ! Fxsx ! Fysy ! Fzsz
! Fxysxy!Fxzsxz – !!!|], [1]
where F(x,y,z) is the field map; F0, Fx, Fy, Fz, Fxy, and Fxz
represent the center frequency, x-shim, y-shim, z-shim,xy-shim, and xz-shim induced field basis vectors (or ma-trices); and s0, sx, sy, sz, sxy, and sxz represent the variableshim values. Equation [1] can be reformulated as a linear
program (2) and solved very efficiently (23). The resultingfield map has the minimum peak off-resonance frequencywith both positive and negative peaks the same absolutevalues. Based on the shim result, one can choose a TR thatensures artifact-free images if the TR is longer than theminimum TR of the sequence.
If the choice of TR is constrained (for example, inbSSFP fMRI studies) and the off-resonance frequency ofthe volume is larger than the coverage obtained by theminimum TR, the resulting bSSFP image will inevitablycontain banding artifacts. Moreover, the min-max shimcan potentially create more banding artifacts than least-squares shim because both positive and negative off-resonance frequency peaks may create the artifact, andtheir spatial location may appear in central portion ofthe image. For example, both orange and blue coloredareas in Fig. 1h (min-max shim) can show the bandingartifact for a certain TR, whereas only the red- to orange-colored areas will have the banding artifact in the least-squares shim (Fig. 1e).
To avoid this problem, the min-max shim methodneeds to be improved. When the banding artifact isunavoidable due to excessive off-resonance, our goal isto reduce the size of banding-artifact regions. Moreover,it is favorable to reduce the size of the banding artifactfrom the least-squares shim result without creating newbanding artifacts at arbitrary locations. To ensure re-duced banding regions compared to the least-squaresresult, we modified the min-max shim method as fol-lows. First, an ROI where shimming is performed ischosen. The least-squares shim is performed for the ROI
FIG. 1. Conceptual illustration of the advantage of the min-max shim method in bSSFP imaging. a,d: The bSSFP image. b,c: The bSSFPmagnitude profile. The flat magnitude off-resonance frequency range is referred to as the pass-band. e,h: The color-coded field map of theslice (in Hz). f,g: The off-resonance distribution of the central line in the field map (the red dotted line in e and h). When the min-max shimis used, the peak off-resonance frequency is reduced allowing an artifact-free image.
Min-Max Shimming for bSSFP 1501
Therefore, the method is likely to penalize a small re-gion (e.g., the frontal lobe area) while reducing the over-all off-resonance frequency in other regions. The result-ing field map may exhibit a large peak off-resonanceregion as shown in Fig. 1e and f. In bSSFP, this largepeak off-resonance can result in banding artifacts (theyellow arrow in Fig. 1a, and the red strip in Fig. 1f). Insome cases, one can manipulate the shim parameterssuch that the maximum off-resonance frequency is re-duced to avoid the artifacts (Fig. 1g and h). In thisexample, the center frequency and y-shim were modi-fied to decrease the peak off-resonance frequency. Theresulting field map has a larger mean square error com-pared to the least-squares shim result but the image hasno artifact (Fig. 1d). Hence, one can achieve a widerspatial coverage by reducing the peak off-resonance fre-quency.
This approach can be restated as follows: in bSSFP,better spatial coverage can be obtained by minimizing themaximum absolute off-resonance frequency of the fieldmap (min-max shim). It can be formulated as follows (2):
s0,sx,sy. . .argmin max[|F(x,y,z) ! F0s0 ! Fxsx ! Fysy ! Fzsz
! Fxysxy!Fxzsxz – !!!|], [1]
where F(x,y,z) is the field map; F0, Fx, Fy, Fz, Fxy, and Fxz
represent the center frequency, x-shim, y-shim, z-shim,xy-shim, and xz-shim induced field basis vectors (or ma-trices); and s0, sx, sy, sz, sxy, and sxz represent the variableshim values. Equation [1] can be reformulated as a linear
program (2) and solved very efficiently (23). The resultingfield map has the minimum peak off-resonance frequencywith both positive and negative peaks the same absolutevalues. Based on the shim result, one can choose a TR thatensures artifact-free images if the TR is longer than theminimum TR of the sequence.
If the choice of TR is constrained (for example, inbSSFP fMRI studies) and the off-resonance frequency ofthe volume is larger than the coverage obtained by theminimum TR, the resulting bSSFP image will inevitablycontain banding artifacts. Moreover, the min-max shimcan potentially create more banding artifacts than least-squares shim because both positive and negative off-resonance frequency peaks may create the artifact, andtheir spatial location may appear in central portion ofthe image. For example, both orange and blue coloredareas in Fig. 1h (min-max shim) can show the bandingartifact for a certain TR, whereas only the red- to orange-colored areas will have the banding artifact in the least-squares shim (Fig. 1e).
To avoid this problem, the min-max shim methodneeds to be improved. When the banding artifact isunavoidable due to excessive off-resonance, our goal isto reduce the size of banding-artifact regions. Moreover,it is favorable to reduce the size of the banding artifactfrom the least-squares shim result without creating newbanding artifacts at arbitrary locations. To ensure re-duced banding regions compared to the least-squaresresult, we modified the min-max shim method as fol-lows. First, an ROI where shimming is performed ischosen. The least-squares shim is performed for the ROI
FIG. 1. Conceptual illustration of the advantage of the min-max shim method in bSSFP imaging. a,d: The bSSFP image. b,c: The bSSFPmagnitude profile. The flat magnitude off-resonance frequency range is referred to as the pass-band. e,h: The color-coded field map of theslice (in Hz). f,g: The off-resonance distribution of the central line in the field map (the red dotted line in e and h). When the min-max shimis used, the peak off-resonance frequency is reduced allowing an artifact-free image.
Min-Max Shimming for bSSFP 1501
Hz
Hz
Field inhomogeneity brain@1.5T
11
M. Lustig, EECS UC Berkeley
Macroscopic Effects - Metal Artifacts
3.Illustration of severe through-plane distortions caused by distorted excitation profiles. (a)depicts a simulated dipole pattern of metal-induced field inhomogeneities from a cylindricalobject, where the sample field inhomogeneities along the x-axis (solid line) and along the z-axis (dashed line) are plotted in (b). When the field inhomogeneities superimpose upon thefrequency induced by the slice-select gradient, the resulting excitation profiles correspond tothe frequency bands shown in (c). The bottom three sample distorted excitation profiles in(d, e, f)) show that the spins are excited when their precession frequencies fall into thefrequency bands (highlighted in gray in (c)) corresponding to the excited slices. As theexcited slices contain spins from different slice locations, the distorted excitation profileslead to through-plane distortions.
Lu et al. Page 15
Magn Reson Med. Author manuscript; available in PMC 2010 July 1.
NIH
-PA
Author M
anuscriptN
IH-P
A A
uthor Manuscript
NIH
-PA
Author M
anuscript
SEMAC: Slice Encoding for Metal Artifact Correction in MRI
Wenmiao Lu1,2,*, Kim Butts Pauly1, Garry E. Gold1, John M. Pauly2, and Brian A.Hargreaves1
1 Department of Radiology, Stanford University, Stanford, California2 Department of Electrical Engineering, Stanford University, Stanford, California
AbstractMagnetic resonance imaging (MRI) near metallic implants remains an unmet need due to severeartifacts, which mainly stem from large metal-induced field inhomogeneities. This work addressesMRI near metallic implants with an innovative imaging technique called “Slice Encoding forMetal Artifact Correction” (SEMAC). The SEMAC technique corrects metal artifacts via robustencoding of each excited slice against metal-induced field inhomogeneities. The robust sliceencoding is achieved by extending a view-angle-tilting (VAT) spin-echo (SE) sequence withadditional z-phase encoding. While the VAT compensation gradient suppresses most in-planedistortions, the z-phase encoding fully resolves distorted excitation profiles that cause through-plane distortions. By positioning all spins in a region-of-interest to their actual spatial locations,the through-plane distortions can be corrected by summing up the resolved spins in each voxel.The SEMAC technique does not require additional hardware and can be deployed to the largeinstalled base of whole-body MRI systems. The efficacy of the SEMAC technique in eliminatingmetal-induced distortions with feasible scan times is validated in phantom and in vivo spine andknee studies.
Keywordsdistortion correction; metallic implants; metal artifacts; metal-induced distortions; susceptibilityartifact; field inhomogeneities; distorted excitation profiles; view-angle-tilting
IntroductionMetallic implants, such as pedicle screws, are commonly used in orthopedic surgery tofixate fractures, replace arthritic joints, and to align and immobilize vertebra. In the UnitedStates alone, there were 325,000 spinal fusions performed in 2003 and 450,000 primary orrevision total knee arthroplasties performed in 2002 [1]. The current standard imaging testfor complications associated with metallic implants is a plain radiography. For accuratediagnosis, a radiography requires the x-ray beam to be oriented exactly parallel to the bone-implant interface: any obliquity of the x-ray beam can obscure the radiolucent area [2]. Abetter substitute for a two-dimensional radiograph is cross-sectional imaging that images theentire bone-implant interface in three dimensions. However, the physical characteristics ofmetallic implants cause difficulties with cross-sectional imaging techniques. Computedtomography (CT) and invasive CT myelography suffer from metal-induced streak/beam-
Correspondence to: Wenmiao Lu, Department of Radiology, Lucas Center MC5488, Stanford, CA 94305-5488, Phone: (650)723-5945, Fax: (650) 723-8473, wenmiao.lu@gmail.com.A preliminary account of this work was presented at the 16th Meeting of the International Society for Magnetic Resonance inMedicine, May, 2008, page 838.
NIH Public AccessAuthor ManuscriptMagn Reson Med. Author manuscript; available in PMC 2010 July 1.
Published in final edited form as:Magn Reson Med. 2009 July ; 62(1): 66–76. doi:10.1002/mrm.21967.
NIH
-PA Author Manuscript
NIH
-PA Author Manuscript
NIH
-PA Author Manuscript
12
M. Lustig, EECS UC Berkeley
Macroscopic Effects - Metal Artifacts
8.Comparison of in vivo results of the subject with stainless steel screws in his lower leg. Ascompared to the SE (the first row) and the VAT-SE (the second row) sequences, theSEMAC technique (the bottom row) greatly suppresses the metal-induced distortions. Ascan be seen from the in-plane sample results and the reformatted coronal views, the SEMACtechnique enables the clear visualization of the screws, which cannot be seen with the SEand VAT-SE sequences. Due to the sheer field inhomogeneities induced by the stainlesssteel, some spins fall outside the volume covered by the prescribed slices. This leads topartial recovery of the rounded shapes of the drilled holes. In addition, the remaininguncorrected in-plane distortions manifest in the forms of ripple patterns (dotted arrows),which need to be addressed with stronger slice-select and/or readout gradients.
Lu et al. Page 20
Magn Reson Med. Author manuscript; available in PMC 2010 July 1.N
IH-P
A A
uthor Manuscript
NIH
-PA
Author M
anuscriptN
IH-P
A A
uthor Manuscript
8.Comparison of in vivo results of the subject with stainless steel screws in his lower leg. Ascompared to the SE (the first row) and the VAT-SE (the second row) sequences, theSEMAC technique (the bottom row) greatly suppresses the metal-induced distortions. Ascan be seen from the in-plane sample results and the reformatted coronal views, the SEMACtechnique enables the clear visualization of the screws, which cannot be seen with the SEand VAT-SE sequences. Due to the sheer field inhomogeneities induced by the stainlesssteel, some spins fall outside the volume covered by the prescribed slices. This leads topartial recovery of the rounded shapes of the drilled holes. In addition, the remaininguncorrected in-plane distortions manifest in the forms of ripple patterns (dotted arrows),which need to be addressed with stronger slice-select and/or readout gradients.
Lu et al. Page 20
Magn Reson Med. Author manuscript; available in PMC 2010 July 1.
NIH
-PA
Author M
anuscriptN
IH-P
A A
uthor Manuscript
NIH
-PA
Author M
anuscript
metal artifact corrected (SEMAC)
could be a project...
SEMAC: Slice Encoding for Metal Artifact Correction in MRI
Wenmiao Lu1,2,*, Kim Butts Pauly1, Garry E. Gold1, John M. Pauly2, and Brian A.Hargreaves1
1 Department of Radiology, Stanford University, Stanford, California2 Department of Electrical Engineering, Stanford University, Stanford, California
AbstractMagnetic resonance imaging (MRI) near metallic implants remains an unmet need due to severeartifacts, which mainly stem from large metal-induced field inhomogeneities. This work addressesMRI near metallic implants with an innovative imaging technique called “Slice Encoding forMetal Artifact Correction” (SEMAC). The SEMAC technique corrects metal artifacts via robustencoding of each excited slice against metal-induced field inhomogeneities. The robust sliceencoding is achieved by extending a view-angle-tilting (VAT) spin-echo (SE) sequence withadditional z-phase encoding. While the VAT compensation gradient suppresses most in-planedistortions, the z-phase encoding fully resolves distorted excitation profiles that cause through-plane distortions. By positioning all spins in a region-of-interest to their actual spatial locations,the through-plane distortions can be corrected by summing up the resolved spins in each voxel.The SEMAC technique does not require additional hardware and can be deployed to the largeinstalled base of whole-body MRI systems. The efficacy of the SEMAC technique in eliminatingmetal-induced distortions with feasible scan times is validated in phantom and in vivo spine andknee studies.
Keywordsdistortion correction; metallic implants; metal artifacts; metal-induced distortions; susceptibilityartifact; field inhomogeneities; distorted excitation profiles; view-angle-tilting
IntroductionMetallic implants, such as pedicle screws, are commonly used in orthopedic surgery tofixate fractures, replace arthritic joints, and to align and immobilize vertebra. In the UnitedStates alone, there were 325,000 spinal fusions performed in 2003 and 450,000 primary orrevision total knee arthroplasties performed in 2002 [1]. The current standard imaging testfor complications associated with metallic implants is a plain radiography. For accuratediagnosis, a radiography requires the x-ray beam to be oriented exactly parallel to the bone-implant interface: any obliquity of the x-ray beam can obscure the radiolucent area [2]. Abetter substitute for a two-dimensional radiograph is cross-sectional imaging that images theentire bone-implant interface in three dimensions. However, the physical characteristics ofmetallic implants cause difficulties with cross-sectional imaging techniques. Computedtomography (CT) and invasive CT myelography suffer from metal-induced streak/beam-
Correspondence to: Wenmiao Lu, Department of Radiology, Lucas Center MC5488, Stanford, CA 94305-5488, Phone: (650)723-5945, Fax: (650) 723-8473, wenmiao.lu@gmail.com.A preliminary account of this work was presented at the 16th Meeting of the International Society for Magnetic Resonance inMedicine, May, 2008, page 838.
NIH Public AccessAuthor ManuscriptMagn Reson Med. Author manuscript; available in PMC 2010 July 1.
Published in final edited form as:Magn Reson Med. 2009 July ; 62(1): 66–76. doi:10.1002/mrm.21967.
NIH
-PA Author Manuscript
NIH
-PA Author Manuscript
NIH
-PA Author Manuscript
• χtitanium =182• χstainless steal (nonmagnetic) = 3520-6700
13
M. Lustig, EECS UC Berkeley
Microscopic Effects
• Lungs:–Approx: 1/6 tissue, 5/6 air
• Result:– Distribution of field/frequencies
–Tells a lot about microstructure of tissue
ppm -9.05
airlungs H2O (37℃)
0.36 ≃ -6
14
M. Lustig, EECS UC Berkeley
Blood
• +Water χ=-9.05• +Hemoglobin molecule (deoxy) χ=0.15• +Red Blood cells (deoxy) χ=-6.52• *Deoxy blood χ=-8.77• *Oxy blood χ=-9.05
*Magnetic Resonance in Medicine 68:863–867 (2012)+
15
M. Lustig, EECS UC Berkeley
Chemical Shift
• Protons in complex molecules are “shimmed” by adjacent spins and electrons
• σ is a shielding constant - depends on molecular structure
• Example: Lipids
Bcs = B0(1� �)
16
M. Lustig, EECS UC Berkeley
Chemical Shift
• Example: Lipids
C C C C C=_ _ _ __ _ _
_ _ _
_ _
H
H H H
H
H
H
H
HC_
_H
H
... C_
_
H
H
ppm 0.9
CHTMS
01.34.75.4
CH3H2O
CH2
Lipids from CH2 is 3.4ppm below water@3T ≈ 440 Hz
J Magn Reson Imaging. 2009 June; 29(6): 1332–1339.
17
M. Lustig, EECS UC Berkeley
Heterogeneous Tissue
• Tissue is a combination of– Chemical shift– Susceptibility– Geometry
• Results are complex– Interesting cases:
• Blood (fMRI)• Lungs • Trabecular bone• Iron in brain / liver
18
M. Lustig, EECS UC Berkeley
Effect on Imaging
• Magnitude• Phase• Geometric• Blurring
• All depend on spatial scale we look at and acquisition strategy
19
e�t/T2
M. Lustig, EECS UC Berkeley
Off Resonance Effect on FID
• Simple spectroscopy experiment– On resonance
Signal decays with T2
A/D
RF
20
M. Lustig, EECS UC Berkeley
Off Resonance Effect on FID
• Simple spectroscopy experiment– With Susceptibility variation, FID is:
A/D
RF
Why?
21
M. Lustig, EECS UC Berkeley
Off Resonance Effect on FID
• The field is:
• Error is spatial and spectral• In the rotating frame w0
B(~r) = B0 + E(~r)uniform error-field
~B = E(~r)k̂
!E(~r) = �E(~r)
fE(~r) =�
2⇡E(~r)
22
M. Lustig, EECS UC Berkeley
Off Resonance Effect on FID
• Received signal is:
– with time, phase dispersion causes signal cancellation and loss.
mx,y
(~r, t) = mxy
(~r, 0)e�i!E(~r)te�t
T2(~r)
s(t) =
Z
~
R
mxy
(~r, 0)e�i!E(~r)te�t
T2(~r) d~r
magnetization dephasing relaxation
23
M. Lustig, EECS UC Berkeley
Off - Resonance Effects on FID
24
M. Lustig, EECS UC Berkeley
Off Resonance Effects on FID
25
M. Lustig, EECS UC Berkeley
Off Resonance Effects on FID
26
M. Lustig, EECS UC Berkeley
Off Resonance Effects on FID
• T2*: Not a good model for Large-Scale variations (dephasing near sinuses)
• T2*: Is a good model for small scale distributed variations (dephasing near capilaries)
27
M. Lustig, EECS UC Berkeley
Example fMRI
28
mxy
(~r, t) = mxy
(~r, 0)e�i!
E
(~r)te�t
T2(~r) e�i2⇡kx
(t)xd~r
M. Lustig, EECS UC Berkeley
Off -Resonance Effects on Imaging
kx
(t)t
t=TE
kx
=�
2⇡G
x
(t� TE) ) t =kx
�
2⇡Gx
+ TE
The x-verse magnetization is:
Text
29
mxy
(~r, t) = mxy
(~r, 0)e�i!
E
(~r)
✓k
x
(t)�
2⇡G
x
+TE
◆
e�i2⇡kx
(t)xd~r
mxy
(~r, t) = mxy
(~r, 0)e�i!
E
(~r)TEe�i2⇡k
x
(t)⇣x+
!
E
(~r)�G
x
⌘
d~r
M. Lustig, EECS UC Berkeley
Off -Resonance Effects on Imaging
neglecting T2 and substituting for t:
Or,
phase/dephasing displacement
30
M. Lustig, EECS UC Berkeley
Off -Resonance Effects on Imaging
x
0 = x+!
E
(~r)
�G
x
phase/dephasing displacement
An on-resonance spin at position:
Produces the same signal as a spin at x, with off-resonance!
mxy
(~r, t) = mxy
(~r, 0)e�i!
E
(~r)TEe�i2⇡k
x
(t)⇣x+
!
E
(~r)�G
x
⌘
d~r
31
M. Lustig, EECS UC Berkeley
Examples:
shape
signal loss
readout direction?
32
M. Lustig, EECS UC Berkeley
Examples: Phase vs Echo Timewater and fat are “in phase” “ out of phase”
Fat on Gradient Echo Images
water and lipids in phase water and lipids out-of-phase
33
M. Lustig, EECS UC Berkeley
Example: Chemical Shift
x
0 = x+!
E
(~r)
�G
x
Increasing Gx reduces chemical shift artifact
34
�
2⇡G
x
= 1
E(~r) = 3
!E(~r) = 2⇡ · 200
�x = x
0 � x =!
E
(~r)
�G
x
=��2⇡ · 200 Hz
��2⇡ · 1 KHz/cm= 0.2cm
�x
= 1 mm
M. Lustig, EECS UC Berkeley
Example:
Let,
KHz/cm
ppm
Hz
At: this is a shift of two pixels!
35
2 cyc
220 Hz⇡ 9.1 ms
M. Lustig, EECS UC Berkeley
Effects in Spin-Warp
• Off-Resonance–Modest spatial distortions ( few pixels)–Relatively benign artifacts–Reduce artifacts with large Gx
• Chemical-Shift– Fat shift of -220Hz @ 1.5T– Fat image is displaced from Water– In practice F/W shift limited to ~2pixels2 pixels shift are two cycles of linear phase across k-space
36
M. Lustig, EECS UC Berkeley
Off-Resonance in EPI
point-spread function
Example: Lipids
phase accumulation
1ms ⇒ 0.44cyc
0.44cyc ⇐ 1ms
1ms ⇒ 0.44cyc
0.44cyc ⇐ 1ms 96ms ⇒
42.24cyc
37
M. Lustig, EECS UC Berkeley
Off-Resonance in EPI
point-spread function simulation Leg
38
M. Lustig, EECS UC Berkeley
Off-Resonance in Spiralpoint-spread functionphase accumulation
2 ms ⇒ 0.88cyc
realimag
39
M. Lustig, EECS UC Berkeley
Off-Resonance in SpiralReadout time @3T:
2ms 5ms 13ms
Spiral scanwith linear off-resonance
40