tmr mrtbasics41 part2 sd 2sw...spatial coordinate, since frequency encoding source: liang and...
TRANSCRIPT
1
Seite 1
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 1
Basics of Magnetic Resonance Imaging (MRI)(Part 2)
Dr. Sebastian Domsch
TMR Lecture, Module 4.1
Chair in Computer Assisted Clinical MedicineFaculty of Medicine Mannheim University HeidelbergTheodor-Kutzer-Ufer 1-3D-68167 Mannheim, GermanySebastian.Domsch@MedMa.Uni-Heidelberg.dewww.ma.uni-heidelberg.de/inst/cbtm/ckm/
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 2
MR Imaging
MR Imaging
2
Seite 2
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 3
radiofrequency:ω(z) = γ (B0 + Gzz)
gradient Gz
magnetic field gradient, i.e. Gz
z
G
Gradient Field: Slice Selection
Nobelprize 2003
Paul Lauterbur (1929-2007)
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 4Gradient Coil: Design
• current through the coil pairs runs in opposite directions
• B-field of the gradient coil is added or subtracted to B0, respectively
3
Seite 3
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 5
1. slice thickness d = z2 - z1 can be varied by RF bandwidth (frequency width) or gradient strength Gz.
2. slice position can be changed by shifting the frequency spectra with constant RF bandwidth
ω ω
I(ω)
ω (z2)
ω (z1)
ω0
ω = γ (B0+Gz·z)
zz1z2
d
frequency spectraof RF pulse
Slice Selective Excitation
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 6
- RF excitation with a sin(x) / x amplitude function (sinc-pulse) creates arectangular frequency distribution
- sinc-pulse at different gradient strength leads to slices with different positions and thickness
Slice Selective Excitation: Sinc-Pulse
source: Liang and Lauterbur. “Principles of Magnetic Resonance Imaging” 2000
4
Seite 4
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 7
ω(x) = γ (B0+ Gx·x)
- iso-frequency-lines are perpendicularto G
- oscillation frequency of an activated MR-signal is linearly dependent on spatial coordinate, since
Frequency Encoding
source: Liang and Lauterbur. “Principles of Magnetic Resonance Imaging” 2000
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 8
Gz
t
Gx
t
t
signalacquisition - superposition of a gradient field
Bx= Gx·x during the acquisition phase results in:
ω(x) = γ (B0+ Gx·x)
where the Lamor frequency islinked with the spatialinformation x
- spatial information is encodedin the precision frequency of thetransversal magnetization
Frequency Encoding: Gradient Schema
RF
5
Seite 5
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 9
x
y
z Gx
ω(x)~x
I(ω)
ω(x2)ω(x1)
- all nuclei in a stripe perpendicular to thegradient direction are contributing to theNMR signal at Lamor frequency ω(x)
⇒ direct one-dimensional diagram of the spatial distribution of excited nuclei
in the slice
- Fourier-transformation of the FID-signal yields the amount of different frequency components
- I(ω) ~ number of nuclei at frequency ω
frequency spectrum
projection of spin density in direction of gradient
Principle of Frequency Encoding
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 10
φ(x) = - γ · Gx · x · Tpe
Phase Encoding
- phase encoding is performed by stamping an initial phase angle ontothe spins of an excited slice
- iso-phase-lines are perpendicularto G
- after switching off the phase encoding gradient the magnetization is continuing to precede at the same frequency ω0 but with different phase
- phase information of an activated MR-signal is linearly dependent on spatial coordinate, since
source: Liang and Lauterbur. “Principles of Magnetic Resonance Imaging” 2000
6
Seite 6
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 11
slice selectiongradient
phase encodinggradient
Gz
Gy
yTy
- phase encoding is performed before signalacquisition
- gradient field is switched on for a constant time Ty
- gradient strength is increased stepwise by ∆Gy after every sequence passage
Phase Encoding: Gradient Schema
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 12
Gz
tGy
t
tsignal acquisition
RF
Gx
t
phase encoding
frequency encoding
MR Imaging Principle
7
Seite 7
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 13
K-Space
K-Space
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 14
k-space image-space
k-Space vs Image-Space
hologramfrequency distribution
imagedensity distribution
8
Seite 8
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 15k-Raum-Darstellung
????
K-Space Quiz
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 16K-Space: Mona Lisa
9
Seite 9
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 17
k-space
image
Fouriertransformation
kx
ky
y
x
hologram
image
Imaging: k-space
Jean Baptiste Fourier (1768–1830)
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 18
transmitter
receiver
gradient
shim
imag
e pr
oces
sor350 MHz
control panel
computer
350 MHz
radio- gradients Gxyz static field B0frequency RF shim coils
MRI Components: Physical Parameters
static field B0 � M0
radiofreq. RF � signal
gradients Gxyz � image
technicalcomponent
physicalparameter�
10
Seite 10
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 19
T1 Measurement
T1 Measurement
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 20Experiment: T1-Measurement #1
Inversion Recovery Technique
Time t
90°
180°
TE
180°
TI
Signal
x
y
z
-M0
t = 0
x
y
z
t = TI
SI~Mz
TI
Time t
1,00
0
-1,00
Mz
(1 - 2e-t/T1)
TI: Inversion TimeTE: Spin-Echo Time
11
Seite 11
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 21Experiment: T1-Measurement #2
Measurement #2
x
y
z
-M0
t = 0
SI~Mz
TI2
Time t
1,00
0
-1,00
Mz
(1 - e-t/T1)
180°
TI2
Time t
90°
180°
TE
Signal
t = TI2
x
y
z#1
#2
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 22T1 Measurement: Inversion Recovery
inversion recovery (Mz(0) = -M0):
Mz(t) = M0 (1 – 2 exp(-TI/T1))
with Mz = 0 at TI = TI0:0.5 = exp(-TI0/T1)
→ T1 = -TI0 / ln(0.5) = TI0 / 0.7
T1WM = 400 ms / - 0.7 = 570 ms
TI0
White Matter (WM)
12
Seite 12
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 23
T2 Measurement
T2 Measurement
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 24
Signal:S = S0·e-t/T2
WM: T2 ≅ 90 msGM: T2 ≅ 100 msCF: T2 > 500 ms
Experiment: T2-Measurement
Spin-Echo Technique
TR: Repetition TimeTE: Spin-Echo Time
Time t
90°
180°
TE2
180°
Signal Signal
180°
Signal
TE3TE1
90°
TRSignal [a.u.]
Time t
Sig
nal I
nten
sity
T2 ~ e-t/T2T2*
13
Seite 13
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 25T2 Measurement: Spin Echo
spin-echo (Mxy(0) = M0):
Mxy(t) = M0 exp(-t/T2)
→ slope of straight-line in semi-logarithmic scale
T2WM = 90 ms
T2 Measurement: Spin-Echo
White Matter (WM)
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 26
Measurement of Brain Relaxation Times T1 and T2
1. Evaluate from an Inversion Recovery measurement the T1-relaxation of White Matter (WM), Grey Matter (GM), and Cerebrospinal Fluid (CF).
Experiment #1: 6 images were measured (IR_01.ima ... IR_06.ima, data at:http://www.ma.uni-heidelberg.de/inst/cbtm/ckm/lehre/ “Medical Physics: Lab Rotation MR-Radiology“) with TI = 50, 400, 550, 750, 1200, 2000 ms. Plot signal intensity (= pixel mean value of ROI) as a function of TI and calculate T1 of WM, GM and CF.
2. Evaluate from a spin-echo measurement the T2-relaxation of WM, GM and CF.
Experiment #2: 11 images were measured (SE_01.ima ... SE_11.ima) with TE = 25, 50 ... 275.0 ms. Plot signal intensity (= pixel mean value of ROI) semi-logarithm as a function of TE and calculate T2 of WM, GM and CF.
Exercise: Measurement of T1 and T2
- group 1: WM
- group 2: GM
- group 3: CF
14
Seite 14
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 27Exercise: Measurement of T1 and T2
Data to find @
http://www.umm.uni-heidelberg.de/inst/cbtm/ckm/lehre/index.html
Use e.g. Matlab or Excel for data evaluation
Use e.g. ImageJ (freeware) for data postprocessing (ROI analysis)
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 28
Functional MRI
Functional MRI
15
Seite 15
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 29Motivation
FMRI shows activated brain areas based on blood oxygenation (BOLD-effect)
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 30
T2* Relaxation
T2* Relaxation
16
Seite 16
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 31T2* Relaxation
Signal [a.u.]
Time tSig
nal I
nten
sity
~ e-t/T2~ e-t/T2*
Microscopic field inhomogeneities
Mesoscopic field inhomogeneities
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 32BOLD-Effekt*
� Deoxyhämoglobin: paramagnetic (0<χ<1) � generates mesoscopic field inhomogeneities
� Oxyhämoglobin: diamagnetic (χ<0)
*Ogawa 1990
T2*~1/∆Bloc
17
Seite 17
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 33BOLD Response
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 34Single-Shot Gradien-Echo EPI Sequence*
*Mansfield 1977
+ T2*-sensitive � high BOLD contrast
+ fast: 30-40 slices in 2-3s!
- susceptible to macroscopic field inhomogeneities � distortions, signal drop outs
- blurring due to long signal read out
Domsch PhD Thesis 2013, Heidelberg University
18
Seite 18
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 35Functional MRI: Principal
source: Reiser and Semmler. “Magnetresonanztomographie” 2002
morphological imagingslice selection
acquisition offMRI series
stimulation: off on off on off on
parameter image
overlay withmorphological images
quantificationsignal-time-curve
time [s]MR
sig
nal i
nten
sity
[a.u
.]
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 36
fMRI Example 1: EPI Sequence
FOV 220mm: dx = 2.3mm�1.1mm
19
Seite 19
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 37
motoric stimulation visual stimulation
parameters FLASH:- 9 segments TR = 93 ms, TE = 40 ms, α = 40°, BW = 260 Hz/pixel- MA = 192 x 256 pixel, TH = 2-3 mm, TA = 2.4 s, pixel size 0.8 x 0.8 mm2
3.0 x 0.8 x 0.8 mm3 2.0 x 0.8 x 0.8 mm3
Heiler Diploma Thesis 2007, Universität Heidelberg
fMRI Example 2: Segmented FLASH Sequence
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 38
Imaging Contrast
Imaging Contrast
20
Seite 20
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 39Time-Of-Flight Angiography (TOF MRA)
flow direction
excitingvolume
maximum intensityprojection (MIP)
problems:- slow flowing spins- saturation in exciting volume- resolution 0.5 x 0.5 x 1.0 mm3
d`Avila. MRI 2005
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 40Time-Of-Flight Angiography (TOF MRA)
original FLASH image
high signal ofinflowing spins
patients: arterio venous malformation (AVM)
maximum intensity projection (MIP)
21
Seite 21
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 41
kx
k-space: after 1. TR k-space: after 2. TR k-space: after 256. TR
ky
Spin-Echo Sequence
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 42
PwTR = 2775 ms
TE = 17 ms
T2wTR = 2775 msTE = 102 ms
T1wTR = 575 msTE = 14 ms
Spin-Echo Images
- SE gold standard technique for T1 and T2 morphology
22
Seite 22
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 43
TR
TE
90
-
60ms
40
-
10ms
T1-weighted
T2-weighted
proton-weighted
300 - 800 ms 1500 - 3000 ms
no contrastSNR low
Spin-Echo Contrast: TR, TE
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 44Spin-Echo: T1 Dependency
source: Reiser and Semmler. “Magnetresonanztomographie” 2002
T1-factor:[1-exp(-TR/T1)]
GM: T1 = 970 msWM: T1 = 600 ms
contrast
T1 - factor
WM
T1 - contrast
GM
23
Seite 23
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 45Spin-Echo: T2 Dependency
T2-factor:exp(-TE/T2)
GM: T2 = 110 msWM: T2 = 90 msT2 - factor
WM GM
T2 - contrast
source: Reiser and Semmler. “Magnetresonanztomographie” 2002
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 46Spin-Echo: TR, TE
source: Schlegel and Mahr. “3D Conformal Radiation Therapy: A Multimedia Introduction to Methods and Techniques" 2007
Which Weighting??
24
Seite 24
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 47Inversion Recovery Sequence
Mz
|Mz|
source: Reiser and Semmler. “Magnetresonanztomographie” 2002
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 48Inversion Recovery Images
25
Seite 25
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 49
αExample: α = 20°
Mz - reduction by 6%Mxy- value 34% of Mz!!
M
x´,y´
z
Mz
Mxy
Gradient-Echo Sequence
spoiler
dephasing oftransversal
magnetization
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 50Gradient-Echo: Steady-State
number of RF
example: WMT1 ~ 600 msTR = 25 ms
Mz steady-state:
Mxy steady-state:
source: Reiser and Semmler. “Magnetresonanztomographie” 2002
26
Seite 26
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 51Gradient-Echo: FLASH
Haase et al. JMR 1986
FLASH signal: with E1 = exp(-TR/T1) FLASH: “fast low angle shot”
Ernst-angle:
Signal Strength:
RUPRECHT-KARLS-UNIVERSITY HEIDELBERG
Computer Assisted Clinical MedicineProf. Dr. Lothar Schad
10/19/2016 | Page 52
SE GE
α = 90° / 180°TE = 20 ms
TR = 600 ms
Taq: minutes
α = 25°TE = 7 ms
TR = 20 ms
Taq: seconds !!
Gradient-Echo: Measuring Time