View
8
Download
0
Category
Preview:
Citation preview
3/24/10
1
Lecture 2010-03-23
• Multislice acquisition (MS, 3D) • GRE: Simple Steady-State model • RF pulse train (multiple echoes)
– Eight-Ball Echo – Stimulated Echo – Multiples echoes
• GRE FID and ECHO components • Mathematical Model of GRE • GRE contrast properties and techniques
FYS‐KJM 4740 1
• Multislice acquisition (MS, 3D) • GRE: Simple Steady-State model • RF pulse train (multiple echoes)
– Eight-Ball Echo – Stimulated Echo – Multiples echoes
• GRE FID and ECHO components • Mathematical Model of GRE • GRE contrast properties and techniques
FYS‐KJM 4740 2
3/24/10
2
Spin Echo – k-space sampling
slice selec5on
Read‐out direc5on (frequency encoding)
phase encoding
Spa5al informa5on in “y” direc5on
Gz
Gx
Gy
acquisi5on of all profiles
FYS‐KJM 4740 3
Gradient Echo – k-space sampling
slice selec5on
Read‐out direc5on (frequency encoding)
phase encoding
Spa5al informa5on in “y” direc5on
Gz
Gx
Gy
acquisi5on of all profiles
FYS‐KJM 4740 4
3/24/10
3
TE and TR
TR
TE
FYS‐KJM 4740 5
Multislice acquisition
6 FYS‐KJM 4740
Sl 1
Sl 2
Sl 3
TR
Slice order: “Interleaved”
3/24/10
4
3D acquisition
7 FYS‐KJM 4740
Same principle of phase encoding is applied to the slice direc5on (refocusing lobe of the slice selec5on)
2D versus 3D acquisition
8 FYS‐KJM 4740
3/24/10
5
• Multislice acquisition (MS, 3D) • GRE: Simple Steady-State model • RF pulse train (multiple echoes)
– Eight-Ball Echo – Stimulated Echo – Multiples echoes
• GRE FID and ECHO components • Mathematical Model of GRE • GRE contrast properties and techniques
FYS‐KJM 4740 9
GRE properties
• Requires less time than SE to form an Echo (i.e. TE shorter for 1 k-space profile acquisition)
• RF pulse angle α ≤ 90° • SI = f(T1, T2, T2*)
10 FYS‐KJM 4740
3/24/10
6
GRE: Simple steady-state model
TR short, T2* << TR, pulse angle α
Repeated RF pulses dynamic equilibrium called “Steady State”
11 FYS‐KJM 4740
Steady‐state reached at the 2nd pulse if α = 90deg.
GRE steady-state
12 FYS‐KJM 4740
Note that the highest signal is obtained for α = 20°
3/24/10
7
GRE: Simple steady-state model
13 FYS‐KJM 4740
TR
Simple mathematical model
14 FYS‐KJM 4740
We know from basic NMR intro that
And the solu5on along z‐axis to this equa5on is (with the nota5on from last figure):
Looking at the rota5on by the α RF pulse along x’ axis, we get:
We assume that which is true if T2*<<TR
3/24/10
8
Simple mathematical model (cont.)
15 FYS‐KJM 4740
We have now the following equa5ons
and
At equilibrium (steady‐state):
and wri5ng
Which then gives
Applying an RF pulse of angle α Eq. 4‐8 (Compendium)
Moreover, the net magne5za5on of GRE readout is modulated by exp(‐TE/T2*)
Simple mathematical model (cont.)
16 FYS‐KJM 4740
Finding the maximum of this func5on:
αe is the Ernst angle
contrast weigh5ng in a spoiled GRE sequence • T1‐weighted • T2*‐weighted (weigh5ng depends on the TR/T1 at given α)
Short TR (TR/T1 << 1) Large α T1 weigh5ng
Long TR (TR/T1 ≈1) Low α T2* weigh5ng
3/24/10
9
Steady-State
17 FYS‐KJM 4740
Steady-State
18 FYS‐KJM 4740
3/24/10
10
• Multislice acquisition (MS, 3D) • GRE: Simple Steady-State model • RF pulse train (multiple echoes)
– Eight-Ball Echo – Stimulated Echo – Multiples echoes
• GRE FID and ECHO components • Mathematical Model of GRE • GRE contrast properties and techniques
FYS‐KJM 4740 19
GRE: General behavior
20 FYS‐KJM 4740
TR
Since TR is short
We have then:
• Longitudinal component
• Transverse component
with mul5ple RF pulses repe55on
3/24/10
11
Effect of repeated RF excitations: eight-ball echo
21 FYS‐KJM 4740
Eight-Ball Echo
22 FYS‐KJM 4740
TR TR
dephasing
3/24/10
12
Eight-Ball Echo
23 FYS‐KJM 4740
TR TR
dephasing
Eight-Ball Echo
24 FYS‐KJM 4740
TR TR
2nd 90° pulse
3/24/10
13
Eight-Ball Echo
25 FYS‐KJM 4740
TR TR
rephasing
Eight-Ball Echo
26 FYS‐KJM 4740
TR TR
Eight‐ball echo Locus
3/24/10
14
Eight-Ball Echo
27 FYS‐KJM 4740
TR TR
dephasing
Stimulated Echo • two 90° pulses eight-ball echo
• The 2nd 90°pulse flipped x’ component of the transverse magnetization along z-axis (this z-magnetization carries magnetization history from before 2nd RF pulse, i.e. gradient dephasing as function of position)*
• 3rd 90° pulse then flips back this longitudinal component into the transverse plane. The 3rd Gx gradient will refocus these components previously stored along z-axis
28 FYS‐KJM 4740
* The longitudinal component is not affected by the second gradient ager RF pulse
3/24/10
15
Stimulated Echo
29 FYS‐KJM 4740
“stored” z‐component with dephasing history from the 1st gradient (not affected by the second)
Will create eight_ball echo
Stimulated Echo
30 FYS‐KJM 4740
3/24/10
16
Stimulated Echo (simulation)
31 FYS‐KJM 4740
Multiple Echoes
• From Steady-State simplified model / Ernst angle α < 90°
• Both transverse and longitudinal components are present at any time
• RF pulse train generates multiple echoes
32 FYS‐KJM 4740
3/24/10
17
3 RF-pulses gives a total of 5 echoes
α1 α2 α3
e1 e2 e3 e4 e5
FYS‐KJM 4740 33
α1 α2 α3
e1 e2 e3 e4 e5
e1: eight‐ball echo of α1 and α2
FYS‐KJM 4740 34
3/24/10
18
α1 α2 α3
e1 e2 e3 e4 e5
e2: S5mulated echo of α1,α2 and α3
FYS‐KJM 4740 35
α1 α2 α3
e1 e2 e3 e4 e5
e3: refocused e1 by α3
FYS‐KJM 4740 36
3/24/10
19
α1 α2 α3
e1 e2 e3 e4 e5
e4: eight‐ball echo of α2 and α3
FYS‐KJM 4740 37
α1 α2 α3
e1 e2 e3 e4 e5
e5: eight‐ball echo of α1 and α3
FYS‐KJM 4740 38
3/24/10
20
• Multislice acquisition (MS, 3D) • GRE: Simple Steady-State model • RF pulse train (multiple echoes)
– Eight-Ball Echo – Stimulated Echo – Multiples echoes
• GRE FID and ECHO components • Mathematical Model of GRE • GRE contrast properties and techniques
FYS‐KJM 4740 39
FID and ECHO components
• α < 90° • Transverse and longitudinal components • Echoes can be formed (eight-ball/ stimulated) contain magnetic history from former RF pulses/gradients
40 FYS‐KJM 4740
3/24/10
21
FID
FID
41 FYS‐KJM 4740
αn αn+1
Longitudinal magne5za5on available • “fresh” component fid (small lejers) • Contains also: transverse magne5za5on components flipped to z‐axis by αn‐i (see also echoes flipped to z‐axis)
ECHO
Available transverse component preceding an RF pulse • Refocused echoes (posi'on dependent history / gradients) • These echoes are par5ally rotated out of the transverse plane (α < 90°) and a part remains in the transverse plane ager the RF Pulse αn+1.
ECHO
42 FYS‐KJM 4740
αn αn+1
3/24/10
22
FID is mainly T1 weighted (based on available longitudinal magne5za5on before the RF pulse)
ECHO is mainly T2 weighted (persistence of the transverse components that form the echoes (e.g. eight‐ball etc.)
ECHO
FID & ECHO
43 FYS‐KJM 4740
αn αn+1
FID
Now considering the readout gradients in this figure:
Both FID and ECHO can give a gradient echo (signal that will be encoded into k‐space to form the GRE image)
ECHO
FID & ECHO
44 FYS‐KJM 4740
αn αn+1
FID
3/24/10
23
FID & ECHO
45 FYS‐KJM 4740
αn αn+1
FID
FID & ECHO
46 FYS‐KJM 4740
αn αn+1
FID
TE
3/24/10
24
FID & ECHO
47 FYS‐KJM 4740
αn αn+1
TE
ECHO
Imaging Echo (from ECHO) appears
when
between TE and TR.
Echo appears at the next a pulse but
is represented normally in apparent
“reversed” 5me
Both FID and ECHO signals are used in Imaging
FYS‐KJM 4740 48
3/24/10
25
FID and ECHO proper5es:
FID carries par5ally the spa5al info contains in the ECHO component which is flipped to the x‐y plane by the αn pulse
It can be shown that ECHO can add parallel to the FID of an5‐parallel depending on the previously acquired phase angle (i.e. posi5on) Periodic “band” structure in the image with different T1/T2 weigh5ng)
ECHO
FID & ECHO
49 FYS‐KJM 4740
αn αn+1
FID
• Multislice acquisition (MS, 3D) • GRE: Simple Steady-State model • RF pulse train (multiple echoes)
– Eight-Ball Echo – Stimulated Echo – Multiples echoes
• GRE FID and ECHO components • Mathematical Model of GRE • GRE contrast properties and techniques
FYS‐KJM 4740 50
3/24/10
26
Precession: MT is position dependent (in-plane gradient)
FYS‐KJM 4740 51
Steady-state signal behavior
Excitation Precession
Relaxation
M = [Mx,My,Mz]T , M0=[0,0,M0]T, E2 = exp(-t/T2) and E1=exp(-t/T1).
From Chap 2:
FYS‐KJM 4740 52
3/24/10
27
Steady-state signal behavior Relaxation and Precession occurs simultaneously and can be combined:
FYS‐KJM 4740 53
Steady-state
U = unit matrix
FYS‐KJM 4740 54
3/24/10
28
Steady-state
MT = Mx’ + jMy’
FYS‐KJM 4740 55
Steady-state
The equation reduces to the ‘spoiled’ GRE expression in when E2=exp(-t/T2)=0 (TR >> T2 or active spoiling). In this case, the factor D in the denominator depends on T1-weighting (through E1) only.
E2=0
Transverse magn. immediately after nth RF pulse:
FYS‐KJM 4740 56
3/24/10
29
Steady-state
The term 1-E2exp(-jθ) in the numerator implies a varying degree of T2-weighting, dependent on θ, and hence position. This means that, if E2 >0 the image will contain ‘bands’ of varying degree of T1- and T2-weighting depending on the value of exp(-jθ(r)) at a given position .
FYS‐KJM 4740 57
Steady-state: magnetization evolution between RF-excitations Transverse magn. during nth TR-interval:
relaxation Gradient effects
FYS‐KJM 4740 58
3/24/10
30
The FID signal
FYS‐KJM 4740 59
FYS‐KJM 4740 61
Recommended