Allen W. Song, PhD Brain Imaging and Analysis Center Duke University MRI: Contrast Mechanisms and Pulse Sequences
Jan 14, 2016
Allen W. Song, PhDBrain Imaging and Analysis CenterDuke University
MRI: Contrast Mechanismsand Pulse Sequences
Image Contrasts
The Concept of ContrastContrast = difference in signals emitted by water protons
between different tissuesFor example, gray-white contrast is possible because T1
is different between these two types of tissue
Two Types of Contrast
Static Contrast: Image contrast is generated from the static properties of biological systems (e.g. density).
Motion Contrast: Image contrast is generated from movement (e.g. blood flow, water diffusion).
T2 Decaytransverse
MRSignal
T1 Recoverylongitudinal
MRSignal
50 ms50 ms 1 s1 s
Static Contrast Imaging Methods
timetime timetime
1. Weighted by the Proton Density
2. Weighted by the Transverse Relaxation Times (T2 and T2*)
3. Weighted by the Longitudinal Relaxation Time (T1)
Most Common Static Contrasts
Proton Density Contrast
Contrast solely dependent on proton density,without influence from relaxation times.
The Effect of TR and TE onProton Density Contrast
0 10 20 30 40 50 60 70 80 90 1000
0.5
1
1.5
2
2.5
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
0.5
1
1.5
2
2.5
T2 Decay
MR
Sig
nal
t (ms)t (s)
MR
Sig
nal
TR TE
T1 Recovery
Optimal Proton Density Contrast
Technique: use very long time between RF shots (large TR) and very short delay between excitation and readout window (short TE)
Useful for anatomical reference scans Several minutes to acquire 256256128 volume ~1 mm resolution
Proton Density Weighted Image
T2 and T2* Contrasts
Contrast dominated by the difference inT2 and T2* (transverse relaxation times).
T2Cars on the same track
T2*Cars on different tracks
Transverse Relaxation Times
To get pure T2 contrast, we need perfectlyhomogeneous magnetic field. Thisis difficult to achieve, as sometimeeven if the actual field is uniform,the presence of biological tissue will still change the homogeneity.
So how do we then remove the influenceof the magnetic field inhomogeneity?
180o turnt = TE/2
180o turnt = TE/2
TE/2
TE/2
t=0
t=TE
t=0
t=TE
Fast Spin Fast Spin
Fast Spin Fast Spin
Slow Spin Slow Spin
Slow SpinSlow Spin
Time Reversal Using 180o RF Pulse
TE/2
TE/2
The Effect of TR and TE onT2* and T2 Contrast
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 10 20 30 40 50 60 70 80 90 1000
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
T2 Decay
MR
Sig
nal
MR
Sig
nal T1 Recovery
TR TE
T1 ContrastT1 Contrast T2 ContrastT2 Contrast
Optimal T2* and T2 Contrast
Technique: use large TR and intermediate TE
Useful for functional (T2* contrast) and anatomical (T2 contrast to enhance fluid contrast) studies
Several minutes for 256 256 128 volumes, or second to acquire 64 64 20 volume
1mm resolution for anatomical scans or 4 mm resolution [better is possible with better gradient system, and a little longer time per volume]
T2 Weighted Image
T2* Weighted Image
PD ImagesPD Images
T2* ImagesT2* Images
T1 Contrast
Contrast dominated by the T1 (longitudinal relaxation time) differences.
The Effect of TR and TE on T1 Contrast
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 10 20 30 40 50 60 70 80 90 1000
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
T1 contrast T2 contrast
T2 Decay
MR
Sig
nal
MR
Sig
nal
T1 Recovery
TR TE
Optimal T1 Contrast
Technique: use intermediate timing between RF shots (intermediate TR) and very short TE, also use large flip angles
Useful for creating gray/white matter contrast for anatomical reference
Several minutes to acquire 256256128 volume
~1 mm resolution
T1 Weighted Image
-S-So
SSo
S = SS = Soo * (1 – 2 e * (1 – 2 e –t/T1–t/T1))
S = SS = Soo * (1 – 2 e * (1 – 2 e –t/T1’–t/T1’))
Inversion Recovery to Boost T1 Contrast
IR-Prepped T1 Contrast
In summary, TR controls T1 weighting and TE controls T2 weighting. Short T2 tissues are dark on T2 images, but short T1 tissues are bright on T1 images.
Motion Contrast Imaging Methods
Prepare magnetization to make signal sensitive to different motion properties
Flow weighting (bulk movement of blood) Diffusion weighting (water molecule random motion) Perfusion weighting (blood flow into capillaries)
Flow Weighting: MR Angiogram
• Time-of-Flight Contrast
• Phase Contrast
Time-of-Flight Contrast
No Flow
Medium Flow
High Flow
No Signal
Medium Signal
High Signal
Vessel
AcquisitionSaturation Excitation
Vessel Vessel
90o
Excitation
ImageAcquisition
RF
Gx
Gy
Gz
90o
Saturation
Time to allow fresh flow enter the slice
Pulse Sequence: Time-of-Flight Contrast
Phase Contrast (Velocity Encoding)
Externally AppliedSpatial Gradient G
Externally AppliedSpatial Gradient -G
Blood Flow v
2
0
2)()(
GvT
dtvtxGdtvtxGT T
T
Time
T2T0
90o
Excitation
Phase
Image
Acquisition
RF
Gx
Gy
Gz
G
-G
Pulse Sequence: Phase Contrast
MR Angiogram
Random Motion: Water Diffusion
Diffusion Weighting
Dtl 2
Externally AppliedExternally AppliedSpatial Gradient Spatial Gradient GG
Externally AppliedExternally AppliedSpatial Gradient -Spatial Gradient -GG
TimeTime
TT2T2T00
322
3
2TGD
oeSS
Pulse Sequence: Gradient-Echo Diffusion Weighting
90o
Excitation
Image
Acquisition
RF
Gx
Gy
Gz
G
-G
Large Lobes
90o
Excitation
Image
Acquisition
RF
Gx
Gy
Gz
G
180o
G
Pulse Sequence: Spin-Echo Diffusion Weighting
Diffusion Anisotropy
Determination of fMRI Using the Directionality of Diffusion Tensor
Advantages of DWI
1. The absolute magnitude of the diffusion coefficient (ADC) can help determine proton pools with different mobility
2. The diffusion direction can indicate fiber tracks
ADCADC AnisotropyAnisotropy
Fiber Tractography
DTI and fMRI
AB
C
D
Perfusion
The injection of fluid into a blood vessel in order to reachan organ or tissue, usually to supply nutrients and oxygen.
In practice, we often mean capillary perfusion as most delivery/exchanges happen in the capillary beds.
Perfusion Weighting: Arterial Spin Labeling
TransmissionTransmission
Imaging PlaneImaging Plane
Labeling CoilLabeling Coil
AlternatingAlternatingInversionInversion
Pulsed LabelingPulsed Labeling
AlternatingAlternatingInversionInversion
Imaging PlaneImaging Plane
FAIRFAIRFlow-sensitive Alternating IRFlow-sensitive Alternating IR
EPISTAREPISTAREPI Signal Targeting with Alternating RadiofrequencyEPI Signal Targeting with Alternating Radiofrequency
Arterial Spin Labeling Can Also Be Achieved Without Additional Coils
RF
Gx
Gy
Gz
Image
90o 180o
Alternating oppositeDistal Inversion
OddScan
EvenScan
180o
RF
Gx
Gy
Gz
Image
90o180o180o
AlternatingProximal Inversion Odd Scan
Even Scan
Pulse Sequence: Perfusion Imaging
FA
IRF
AIR
EP
IST
AR
EP
IST
AR
Advantages of ASL Perfusion Imaging
1. It is non-invasive2. Combined with proper diffusion weighting
to eliminate flow signal first, it can be used to assess capillary perfusion
Perfusion Contrast
PerfusionPerfusionDiffusionDiffusion
Perfusion Map
Some fundamental acquisition methods commonly used to generate
static and motion contrasts,and their k-space views
k-Space Recap
Kx = /20tGx(t) dt
Ky = /20tGx(t) dt
Equations that govern k-space trajectory:
These equations mean that the k-space coordinatesare determined by the area under the gradient waveform
dxdyeyxIkkS ykxkiyx
yx )(2),(),(
Gradient Echo Imaging
Signal is generated by magnetic field refocusing mechanism only (the use of negative and positive gradient)
It reflects the uniformity of the magnetic field
Signal intensity is governed by
S = So e-TE/T2*
where TE is the echo time (time from excitation to
the center of k-space)
Can be used to measure T2* value of the tissue
MRI Pulse Sequence for Gradient Echo Imaging
digitizer ondigitizer on
ExcitationExcitation
SliceSliceSelectioSelectionnFrequencyFrequency EncodingEncoding
PhasePhase EncodingEncoding
ReadoutReadout
K-space view of the gradient echo imaging
Kx
Ky
123.......n
Multi-slice acquisition
Total acquisition time =Total acquisition time = Number of views * Number of excitations * TRNumber of views * Number of excitations * TR
Is this the best we can do?Is this the best we can do?
Interleaved excitation methodInterleaved excitation method
readoutreadout
ExcitationExcitation
SliceSliceSelectioSelectionnFrequencyFrequency EncodingEncoding
PhasePhase EncodingEncoding
ReadoutReadout
readoutreadout readoutreadout
…………
…………
…………
TRTR
…………
Spin Echo Imaging
Signal is generated by radiofrequency pulse refocusing mechanism (the use of 180o pulse )
It doesn’t reflect the uniformity of the magnetic fieldSignal intensity is governed by S = So e-TE/T2
where TE is the echo time (time from excitation to
the center of k-space)Can be used to measure T2 value of the tissue
MRI Pulse Sequence for Spin Echo Imaging
digitizer ondigitizer on
ExcitationExcitation
SliceSliceSelectioSelectionnFrequencyFrequency EncodingEncoding
PhasePhase EncodingEncoding
ReadoutReadout
9090 180180
K-space view of the spin echo imaging
Kx
Ky
123.......n
Fast Imaging Sequences
How fast is “fast imaging”?
In principle, any technique that can generate an entire image with sub-second temporal resolution can be called fast imaging.
For fMRI, we need to have temporal resolution on the order of a few tens of ms to be considered “fast”. Echo-planar imaging, spiral imaging can be both achieve such speed.
Echo Planar Imaging (EPI)
Methods shown earlier take multiple RF shots to readout enough data to reconstruct a single image Each RF shot gets data with one value of phase
encoding If gradient system (power supplies and gradient coil) are
good enough, can read out all data required for one image after one RF shot Total time signal is available is about 2T2* [80 ms]
Must make gradients sweep back and forth, doing all frequency and phase encoding steps in quick succession
Can acquire 10-20 low resolution 2D images per second
......
...
Pulse Sequence K-space View
Echo Planar Imaging (EPI)
Why EPI?
Allows highest speed for dynamic contrast
Highly sensitive to the susceptibility-induced field changes --- important for fMRI
Efficient and regular k-space coverage and good signal-to-noise ratio
Applicable to most gradient hardware
Gradient-Recalled EPI Images Under Homogeneous Field
Distorted EPI Images with Imperfect Field
x imperfectionx imperfection
y imperfectiony imperfection
z imperfectionz imperfection
Spiral Imaging
t = TERFRF
GxGx
GyGy
GzGz
t = 0
K-Space Representation of Spiral Image Acquisition
Why Spiral?
• More efficient k-space trajectory to improve throughput.• Better immunity to flow artifacts (no gradient at the center of k-space)• Allows more room for magnetization preparation, such as diffusion weighting.
Gradient Recalled Spiral Images Under Homogeneous Field
Distorted Spiral Images with Imperfect Field
x imperfectionx imperfection
y imperfectiony imperfection
z imperfectionz imperfection