MR Contrast Mechanism

1
MR Contrast Mechanism

• BioE 297 pulse sequence seminar
• 7/7/2005 330
• Donghyun Kim

2
Welcome,

• Objective of todays talk
• Know the different contrast generating mechanisms
• Inter-relate different pulse sequences (that we
  will be looking into) in terms of contrast
  perspective
• Think about contrast

3
Contrast

• Measure of image quality
• SNR, CNR (contrast to noise ratio)
• Maximize CNR
• Signal difference ?, noise ?
• Imaging time, coil design, etc
• Manipulate imaging parameters for ?signal
• Often, CNR is sacrificed for SNR
• Ex) filtering

4
What generates the contrast? (difference in
signal intensity)

• Basic physical parameters
• Proton density, T1, T2
• Can we change these basic physical parameters?
• B0 field
• Contrast agents
• Ex) Gd-DPTA dose of unpaired electrons
• T1 shortening, T2 magnetic susceptibility effect
• Magnetic field inhomogeneity
• T2
• BOLD (fMRI)
• Flow of blood in active regions changes local
  magnetic field
• dynamic susceptibility (ex perfusion imaging)
• Contrast agent changes T2 with time
• Motion
• Phase contrast imaging
• Phase sensitizing gradients are used to generate
  flow dependant phase
• Diffusion
• Brownian motion of water molecules depend on
  microstructure
• Rich set of contrast mechanism is one of the most
  powerful aspects of MRI.

5
Basic physical parameters
Derivation from basic quantum physics with simple
assumptions
6
T1 weighted imageex) gray and white matter

• Achieve contrast based on the difference of T1
• Gray matter 920 ms, white matter 790 ms.
• Signal amplitude is dependant on T1
• Signal amplitude ? 1-exp(-t/T1)

TRmax ? (T1gray T1white)/2

• Note WM (which has shorter T1) will have greater
  signal than GM (which has longer T1)
• In T1 weighted image, longer T1 components are
  darker! (reversed intensity)

7
T1 weighted image
8
T2 weighted imageex) gray matter and white matter

• Achieve contrast based on the difference of T2
• Gray matter 101 ms, white matter 92 ms.
• Signal amplitude is dependant on T2
• Signal amplitude ? exp(-t/T2)

TEmax ? (T2gray T2white)/2

• Sample at TE 95 ms (near the T2) to get maximum
  difference between gray and white matter
• Engineering general rule of thumb
• For exponential modeling, sample at the time
  constant

9

• So, TE T2 then everything is fine
• But, what is T2?
• T2 varies with location
• T2 varies with age (myelination)

10

• CSF has too much signal
• Suppress it! (FLAIR Fluid attenuated inversion
  recovery)
• More dynamic range
• Lecture 2

T2
FLAIR
11
T1, T2 summary
Gradient echo sequences, spin echo sequences

• No contrast?
• Is there any case where we want NO contrast?

Spin echo

• Rapid GRE sequences (SSFP)
• Complicated contrast composed of T1 and T2
  contributions
• Ex) b-SSFP contrast dependant on T1/T2
• Lecture 3

12
k-space perspective on contrast

• Contrast mostly determined by samples near the
  k-space origin
• Why?

• k-space origin sum of intensity in image space
• Eg) object intensity 100, noise intensity 0
• k(0) 100, CNR 100
• This sample point needed for accurate intensity
  measure
• if k(0) 80 (due to sampling error), CNR 80
• Therefore, samples near the origin critical for
  contrast.

13

• If center k-space region so important for
  contrast, could oversampling of center k-space be
  used to manipulate contrast?
• yes
• Ex) KWIC algorithm
• FSE radial sampling
• oversampling
• Use only data from certain echo train for center
  k-space region to manipulate contrast
• Use different filter functions for more elegant
  contrast

T2
KWIC
KWIC algorithm, Song et al. MRM
14
Fat

• Fat is always bright on T1 and T2 weighted image
• Fat tells you nothing about weighting
• On FSE (fast spin echo) image, fat is even
  brighter
• Why?

15
Magnetic field inhomogeneity generated contrast
(T2 weighted)

• Local susceptibility difference causes
  inhomogeneous field

• T2 effect
• local inhomogeneity
• Ex)
• Injection of magnetic particles (contrast agent)
• Activation resulting in deoxyhemoglobin, which is
  paramagnetic

16
T2 weighting

• Gradient echo sensitive to (local) inhomogeneity
• Used for T2 weigthing
• Signal decays exponentially as T2
• Engineering rule of thumb TE T2
• How do you measure T2?
• Can you induce T2 weighting on spin echo?

17
weighting vs. mapping

• Weighted images provide contrast from the
  difference in the measured property
• Looking at relative values, not absolute values
• T1 weighted image has contributions from T2
  weighting.
• Mapping get absolute values
• T1, T2, T2
• Other physiological mapping T1-rho, pH,
  temperature
• Lecture 7

18
Summary

• T1 weighted image is inverted
• Fat is always bright unless suppressed
• Contrast dominated by center k-space region
• Resolution dominates outer k-space region
• Engineering general rule of thumb
• Exponential modeling case
• Sample at time constant for maximum sensitivity
• TR T1 for T1 weighted image 1-exp(-t/T1)
• TE T2 for T2 weighted image exp(-t/T2)
• TE T2 for T2 weighted image exp(-t/T2)
• b value corresponding to D for dwi exp(-bD)

19
(No Transcript)