Stanford UHF MRI Program: Engineering Challenges and Solutions

STANFORD
UHF MRI
PROGRAM
ULTRA HIGH-FIELD MRI
OPEN QUESTIONS IN ENGINEERING AND
MULTIPHYSICS
SIMONE A. WINKLER, PH.D.
MCGILL UNIVERSITY | 2014/04/16
Department of Radiology
Stanford University
ETH Zürich | 2014/12/19S. A. WINKLER

STANFORD
UHF MRI
PROGRAM
B0SHIMMINGACOUSTICSSAFETYIMAGESHADINGINTRODUCTION
OUTLINE
 Introduction to UHF MRI
 Engineering-related challenges in UHF MRI
 Image shading due to B1 inhomogeneities
 Image distortion and artefacts due to B0 inhomogeneities
 MR Safety
 MR Acoustics
ETH ZÜRICH | 2014/12/19S. A. WINKLER

STANFORD
UHF MRI
PROGRAM
STANFORD RADIOLOGICAL SCIENCES LABORATORY
JOHANNES KEPLER UNIVERSITY LINZ | 2014/05/08S. A. WINKLER
• Three 3T GE whole-body MR
systems (one being the first
GE PET-MR)
• One 7T GE whole-body MR
system
Gary
Glover
Brian
Hargreaves
Daniel
Spielman
Brian
Rutt
Rebecca
Fahrig
Michael
Moseley
Norbert
Pelc
Jennifer
McNab
Roland
Bammer
Kim Butts
Pauly

STANFORD
UHF MRI
PROGRAM
PH.D.STUDENTS
JasonSu
MihirPendse
MatthewMarzelli
PI
Prof.BrianRutt
POSTDOCS
SimoneWinkler
SCIENTIFICSTAFF
ManojSaranathan
RonWatkins
FORMER MEMBERS:
- Kimberley Brewer
- James Rioux
- Ives Levesque
- Thomas Tourdias
- Prachi Pandit
- Mehdi Khalighi
- Sui Seng Tee
- Jonathan Lu
THE
RUTT
LABORATORY
FIELDS OF
RESEARCH:
- Molecular imaging
- Pulse sequence
development
- Parallel transmit
techniques
- Hardware
development
- Multiphysics modeling
- MR safety
EXTERNAL
MEMBERS:
- Andrew Alejski
- Trevor Wade
- Janos Bartha
THE RUTT LABORATORY!
LeoTam

STANFORD
UHF MRI
PROGRAM
MR BASICS – NUCLEAR MAGNETIC RESONANCE
 Isotopes in the body with odd number of protons possess nonzero spin
 moving electrons generate small magnetic field
 Hydrogen atoms most abundant in human body (tissue water content)
TWO-STEP PROCEDURE:
1. Alignment (polarization) of spins with an external static magnetic
field B0
2. Perturbation of alignment by use of an RF (electromagnetic) field 
magnetization of protons can be measured

STANFORD
UHF MRI
PROGRAM
MR BASICS – STEP 1: SPIN ALIGNMENT
Without
external
B0-field
With
external
B0-field
Larmor frequency
depends on magnetic
properties of proton,
and is proportional to
static magnetic field
strength
Spin precession
at Larmor
frequency
Alignment Precession
No net magnetization

STANFORD
UHF MRI
PROGRAM
Transverse
component
can be
detected with
an RF coil!
MR BASICS – STEP 2: PERTURBATION/MEASUREMENT
excitationrelaxation
RF excitation:
Additive B1 pulse
in transverse
plane at Larmor
frequency
T2 T1
Precession at Larmor Frequency Added RF pulse at Larmor frequency
Tissue contrast:
RF coil
transceive
s
transverse
magnetic
field
B1
B0
Idea: measure magnetic field that spin generates!
B0
After RF pulse is turned off:

STANFORD
UHF MRI
PROGRAM
MR BASICS – SPATIAL LOCALIZATION
 Larmor frequency depends on static magnetic field strength
 Small magnetic field variation is overlaid depending on location of imaged
voxel
  frequency of received signal contains spatial information!
With linear additive
gradient:
Spatial information can be
recovered from Inverse
Fourier Transform!!
3Dlocalization
Gradient coils
DC current generates
linear B0 variation

STANFORD
UHF MRI
PROGRAM
MR BASICS - SYSTEM
Main system components:
1. Magnet (generates B0
field  aligns spins  net
magnetization)
2. RF coil (excites and
receives RF magnetic
field at Larmor frequency)
3. Gradient coils (localization
by small variation of B0
field)
Typical B0 field strengths (human):
Clinical: 1.5T-3T
Research: 7T-11.7T

STANFORD
UHF MRI
PROGRAM
WHY UHF MRI?
 Increased sensitivity for proton imaging applications (SNR
increases approx. proportionally with B0)
 Allows for anatomical imaging with higher spatial resolution
 Use higher sensitivity to examine dynamic and functional
aspects (BOLD, other brain activity)
7T
• Potential for detection
of new physiology of
healthy and diseased
tissue
• New forms of contrast
Energy gap:

STANFORD
UHF MRI
PROGRAM
HIGH-FIELD MRI AT STANFORD
 GE Discovery MR950 7T
scanner
 2-ch Tx / 32-ch receive
modality
 8-ch pTx / 32-ch receive
modality
Approximately 50 7T
scanners throughout the
world only!
Not FDA approved

STANFORD
UHF MRI
PROGRAM
HARDWARE RELATED UHF MRI CHALLENGES
 Global SAR increases with B0
2
 Local SAR becomes more inhomogeneous due to
wave effects
 Local SAR is strongly position- and coil-
dependent
ETH ZÜRICH | 2014/12/19
RF heating in tissue
(SAR)
B1 inhomogeneities
Lorentz ForcesB0 inhomogeneities
128MHz 512MHz
Larmor frequency
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.2
-0.15
-0.1
-0.05
0
0.05
W/kg
0.57
1.14
1.71
2.28
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.2
-0.15
-0.1
-0.05
0
0.05
W/kg
0.57
1.14
1.71
2.28
2.25
0
W/kg
Image shading Patient safety
Acoustics &
Vibrations
Geometric distortion
& artefacts
 Lorentz forces increase with
B0
 Higher sound pressure levels
 Stronger vibrations
 Spatial encoding
depends on B0
homogeneity
 Signal loss

STANFORD
UHF MRI
PROGRAM
IMAGE SHADING – B1 INHOMOGENEITY
OPTIMIZED DIELECTRIC SHIMMING – SIMONE WINKLER, ALESSANDRO SBRIZZI, ET AL.
16-CHANNEL PTX COIL DESIGN – RICCARDO STARA, SIMONE WINKLER, ET AL.

STANFORD
UHF MRI
PROGRAM
B1 INHOMOGENEITY BASICS
 Parallel Transmission (pTx):
excitation from multiple
independent channels with different
pulse signal shapes at each
channel
 RF shimming: excitation from
multiple independent channels; only
amplitude and phase are varied
 Dielectric shimming: passive
perturbation of E- and B-fields by
insertion of dielectric material that
focuses field lines
128MHz 512MHz
Larmor frequency
Remedy:Challenge:

STANFORD
UHF MRI
PROGRAM
16-CHANNEL PTX COIL DESIGN
1. Novel coil element 2. Modular configuration
1. 8ch single row
2. 8ch checkerboard
3. 16ch with two Butler matrices
4. 8ch with Butler matrix
5. 2ch with Butler matrix
6. etc.
Inherent inter-element decoupling
B1 mapsZ-segmentation
1 2
3 4COAX4
ID=CX2
EL=90 Deg
Fo=0.1 GHz
Z=50
IND
ID=L1
L=580 nH
CAPQ
ID=C1
C=24 pF
Q=0
FQ=0 GHz
ALPH=1
CAPQ
ID=C2
C=10 pF
Q=0
FQ=0 GHz
ALPH=1
T
1
2
REST
ID=IN1
R=1-5 Ohm
T=-273.15 DegC
IND
ID=L2
L=80 nH
DIODE1
ID=Microsemi HUM2015
Nu=0
T=21.85 DegC
Io=0 mA
CAPQ
ID=C3
C=10 pF
Q=0
FQ=0 GHz
ALPH=1
CAP
ID=C4
C=0-1 pF
PORT
P=1
Z=50 Ohm
PORT
P=2
Z=50 Ohm
CT1
CT2
Coil
CM
TuningMatching
RF choke
Decoupling
• Illumination of
lower brain regions
• Better acceleration
• Better pTx
flexibility
Inter-element coupling
3. 3D segmentation Results

STANFORD
UHF MRI
PROGRAM
OPTIMIZED DIELECTRIC SHIMMING
 Electromagnetic scattering is ANALYTICALLY described for dielectric-
absorptive spheres! [1]
  model dielectric pad as a sphere!
 Model head as a dielectric sphere
 Model dielectric shimming pads as additional spheres
 Optimization to determine optimized position of the sphere(s)
Requirement:
 Electromagnetic fields of unloaded RF coil expanded into
vector spherical harmonics
S. A. WINKLER
[1] Gustav Mie, ”Beiträge zur Optik trüber Medien, speziell kolloidaler Metallösungen,“ Annalen der Physik, Vierte Folge, Band 25, 1908, No. 3, p 377-445.
[2] M. I. Mishchenko, L. D. Travis, and D. W. Mackowski, T-matrix computations of light scattering by nonspherical particles: A review, J. Quant. Spectrosc. ETH ZÜRICH | 2014/12/19
Improvement:
41.25%
Dielectric
shimming
also
increases
average B1+
field!
Applications:
• 3T breast MR
• 7T neuroimaging

STANFORD
UHF MRI
PROGRAM
B0 INHOMOGENEITIES
RF SHIM COIL AT 7T – SIMONE WINKLER, JASON STOCKMANN, ET AL.

STANFORD
UHF MRI
PROGRAM
B0 INHOMOGENEITY BASICS
 B0 field has to be extremely homogeneous (order of
ppm, for 7T in the 0.25μT range)
 Spatial encoding depends on B0 homogeneity
 Geometric distortions and/or signal loss
 Sources of B0 inhomogeneity:
1. Magnet construction
2. Magnetic susceptibility differences
in tissues; in particular:
air cavities (ears, sinuses, etc.)
Remedy:Challenge:
7.00001
6.99999
7.00000
B0 shimming:
Air cavities
water
Passive
Active
Spherical
harmonic
s
Matrix
shim
Iron
inserts

STANFORD
UHF MRI
PROGRAM
HISTORY
 High-order B0 shimming is essential for modern MR neuroimaging
 Conventional approach: spherical harmonic shimming
ISMRM 2015 TORONTO, CANADA | 2015/06/04S. A. WINKLER
Multi-Coil Shimming Multi-Coil + RF shimming
Alternative to spherical harmonic
shimming
+ Efficient
+ Switch shim dynamically without eddy
currents
- Restricted space for RF arrays
- Interactions with RF coils
Uses RF chokes to bridge DC shim
current into RF loop
+ Larger efficiency due to single-turn
loop
+ Dynamic switching
- Construction complexity
Juchem C, JMR 212:280–288 (2011) Truong TK, Neuroscience 103:235-240 (2014)
7T 3TDrastic B0 homogeneity improvements

STANFORD
UHF MRI
PROGRAM
RF-SHIM HELMET - CONCEPT
RF-Shim Helmet - Concept:
32chRF-shimcoil
32chRF-onlycoil
Conventional RF coil Shim-RF coil
Coilelement
Combining Rx (RF) and B0 (DC)
shimming functions on the same
conductor
Array of single-turn loops
Helmet shape for closer proximity
(increased Rx SNR and B0 shim efficiency)
+
Stockmann, MRM early view (2015)
Chokes to bridge
capacitor breaks
RF-path only Added DC path
3T

STANFORD
UHF MRI
PROGRAM
RF-SHIM HELMET - RESULTS
z
Rx-Shim Helmet – 3T Results:
B0 – shimming results:
32ch commercial c32ch RF-shim coil
SNRCoilcorrelation
Precursor 32ch coil
a.u.
32ch commercial coil32ch RF-shim coil
0
100
200
300
400
SNRCoilcorrelation
Precursor 32ch coil
0
1
0.5
RF - SNR results:
RF-shim coil exhibits modest
SNR loss as compared to
precursor RF-only coil
Stockmann, MRM early view (2015)
+ Reduced distortion and closer alignment for blip-up and blip-
down
+ Predicted and measured field maps agree
+ Low current requiremnts (2.5 A max per coil)

STANFORD
UHF MRI
PROGRAM
RF-SHIM CONCEPT AT 7T
Using the RF-shim concept at 7T appears compelling:
1. Increased in-vivo B0 distortion at 7T  greater impact for
• EPI (fMRI, Diffusion)
• QSM
• Inversion pulses
2. SNR impact at 7T might be lower
• A well designed receive array is body noise dominated
• We hypothesize that at 7T added copper noise will have less impact
However:
 Conversion from 3T RF-only to RF-shim array caused 10-
15% SNR loss
 We seek to minimize SNR loss at 7T
Optimized loop design
becomes essential
The resulting 7T RF-
shim array would offer:
- Helmet design for
SNR and shim
efficiency
- Minimal interference
of RF and shim
functions
Optimized compact tool for
modern high-field MR
neuroimaging
Here: Exploratory work to choose optimized loop design

STANFORD
UHF MRI
PROGRAM
MOTIVATION
Larger number of capacitors
required for 7T  larger
number of toroidal bridging
chokes
Challenges:
Coaxial loop
element
Novel combined conductor concepts:
Concentric loop element
- Inner conductor:
DC shim current
- Outer conductor
(RF shield):
RF current
- Inner loop:
DC shim current
- Outer loop:
RF current
 Reduced number of capacitors
 No chokes required
• Additional loss
• Heating
• RF field perturbation
• Construction complexity in
confined space
Magnetic coupling between conductors lowers self-inductance of RF loopChokes
DC
path
DC
path

STANFORD
UHF MRI
PROGRAM
THEORY
Magnetic coupling between inner and outer conductor
 Two conjugate poles
 Two conjugate zeroes
Resonance + anti-resonance
𝑓𝑟,shim
𝑓𝑟,standalone
=
1
1 − 𝑘2 k = inter-loop coupling factor
𝑓𝑎𝑟,shim
𝑓𝑎𝑟,standalone
=
1
1 + α2
Resonance:
Anti-resonance:
α =
𝐿 𝑠
𝐿 𝑅𝐹
−
2𝐿 𝑚
𝐿 𝑅𝐹
LS = shim loop inductance
LRF = RF loop inductance
Lm = mutual inductance
Trade-off between
Q-factor reduction
and number of
capacitors to be
found
Coaxial loop
element
Concentric
loop element
Effects of added loop:
1. Resonance frequency increases
because of lowered self inductance
2. Added loop acts as a resistive load
in time-quadrature to the original
loop
+ Fewer capacitors are required
(and therefore bridging chokes)
- Reduction in Q-factor (and thus
SNR) by insertion of additional
loop

STANFORD
UHF MRI
PROGRAM
MR SAFETY
THERMOACOUSTIC SAR MAPPING – SIMONE WINKLER, PAUL PICOT, MICHAEL THORNTON,
BRIAN RUTT
SAR-CONSTRAINED PTX PULSE DESIGN – MIHIR RAJENDRA PENDSE, SIMONE WINKLER, BRIAN
RUTT

STANFORD
UHF MRI
PROGRAM
MR SAFETY BASICS
 Main concern: RF power deposition in tissue (metric: specific absorption rate (SAR))
Low field MRI:
Wave effects are not prominent and local SAR is therefore easier to predict
High-field MRI:
1. SAR increases with B02
2. variation in tissue properties AND fields strong variations in local SAR –
SAR is strongly dependent on:
o Patient
o Coil
o Position
3. pTx techniques may cause worst-case constructive interference of individual
local SAR patterns  great concerns!
To date there is no
experimental procedure
to measure local SAR!
RF power deposition over
entire anatomy of interest
Global SAR
local variation of SAR that
leads to prominent
hotspots
Local SAR
IEC regulation:
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.2
-0.15
-0.1
-0.05
0
0.05
W/kg
0.57
1.14
1.71
2.28
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.2
-0.15
-0.1
-0.05
0
0.05
W/kg
0.57
1.14
1.71
2.28
2.25
0
W/kg
3.2 W/kg 10 W/kg
Alternative metric:
temperature in tissue (max.
1°C heating)
• SAR-minimized pTx
pulse design
• Find methods to
experimentally
determine SAR
• Find alternative
metric
(temperature)
Remedy:Challenge:

STANFORD
UHF MRI
PROGRAM
SAR-MINIMIZED PTX PULSE DESIGN
ETH ZÜRICH | 2014/12/19S. A. WINKLER Pendse, Winkler and Rutt, ISMRM 2014
B1 mapping
MR scan
Localizer Query database for body
model matching patient
Anatomical
parameters
minSAR pTx scan
minSAR
RF Pulse
Optimize pulse
sequence
parameters
minSAR
MLS
Local SAR
dependent
Local SAR
independent
RF Pulse Design
Non pTx
scanning
E-field maps VOPs
…
Body models
SAR database (computed offline)
EM simulationofflinereal-time
Compressed
dataset

STANFORD
UHF MRI
PROGRAM
THERMOACOUSTIC SAR CONCEPT
IDEA:
 absorbed RF energy ≡ SAR!
 MR RF coil as RF transmitter  generates same local
SAR pattern as in an MR scan
 RF-induced thermoacoustic signal is proportional
to local SAR  use MR coil with short bursts of RF
energy
𝛻2 −
1
𝑣𝑠
2
𝜕2
𝜕𝑡2
𝒑 = −
𝛽𝜌
𝐶 𝑝
𝐒𝐀𝐑 𝐫
𝜕𝐼
𝜕𝑡
Pulsed RF absorptionAcoustic pressure wave
US patent: Simone Winkler, Brian Rutt, Paul Picot, Michael Thornton. “In-Vivo Specific Absorption Rate
Mapping using the Thermoacoustic Effect", May 7, 2014

STANFORD
UHF MRI
PROGRAM
MODELING
Delay-and-
Sum
reconstruction
IEC/CENELEC SAM head model
with registered SAR pattern from Ella.
10
0
5
2 01 10 05 4 02
2D SAR pattern in W/kg
originalreconstructed
SEMCAD
Matlab
COMSOL
𝛻2 −
1
𝑣𝑠
2
𝜕2
𝜕𝑡2
𝒑 = −
𝛽𝜌
𝐶 𝑝
𝐒𝐀𝐑 𝐫
𝜕𝐼
𝜕𝑡
Pulsed RF absorptionAcoustic pressure wave
Winkler, Picot, Thornton and Rutt, ISMRM 2014, ISMRM 2015

STANFORD
UHF MRI
PROGRAM
TEST PLATFORM
 Proof of concept with thermoacoustic
imaging platform at Stanford
(collaboration with Endra, Inc.)
S. A. WINKLER
By varying
1. Conductivity
2. Electric Field
in a phantom, we can emulate the
local SAR variation in tissue in UHF
MRI
1) OriginalTan
k
A
B
0°
2) One excitation
channel off
Tan
k
A
X
Agar (σ = 0)
Saline
1.25%
2.25%
3.75%
5%
1 – conductivity variation 2 – E-field variation
5%
3%
2.25%
1.25%
Point source experiments
E-Fieldvariation
RF horns
Ultrasound
transducer

STANFORD
UHF MRI
PROGRAM
EXPERIMENTS – FIRST ATTEMPT
Agar (σ = 0)
Saline
1.25%
2.25%
3.75%
5%
Varying conductivity:
Concen-
tration
Image
value
1.25% 536±91
2% 876±125
2.25% 960±65
3.75% 1768±353
5% 2423±254
ThermoacousticImage
0
500
1000
1500
2000
2500
3000
0.00% 2.00% 4.00% 6.00%
Image value over
saline concentration

STANFORD
UHF MRI
PROGRAM
POINT SOURCE EXPERIMENTS – OVERALL SAR VARIATION
S. A. WINKLER
R² = 0.5224
R² = 0.8458
0
2
4
6
8
10
12
0 2 4 6 8 10
Horizontal SAR
Vertical SAR
Linear (Horizontal SAR)
Linear (Vertical SAR)
Tan
k
A
X
Tan
k
A
B
0°
5%
3%
2.25%
1.25%
2.25%
2.25%
2.25%
2.25%
Conductivity variation
Constant
conductivity
Varied
conductivity
E-Field variation
+ rotated
ultrasound
transducer
to average
over
perceived E-
field
Without
rotation to
maximize E-
field
variation
E-field was
measured with a
coaxial probe
Reference test
SAR variation
E-Fieldvariation
Conduc-
tivity
variation
Conductivity
variation
Conductivity
variation
E-Fieldvariation
UStransducer
RF
horn
Image values fitted to
SAR equation:

STANFORD
UHF MRI
PROGRAM
MR SYSTEM INTEGRATION
 Fast rise-time pulse shape
required  MR system might have
to be upgraded with a larger
bandwidth auxiliary signal
generator
 It is estimated that the power
amplifier of the MR system will be
delivering sufficient peak power for
the experiment
 An ultrasound bandwidth of 200
kHz will be used for penetrating
skull bone
 Ultrasound transducers will be
mounted at two lateral acoustic
windows in the skull
S. A. WINKLER
Power
amplifier
MR
signal
generato
rAuxiliary
signal
generato
r
Control
room
workstatio
n
Data
acquisition
system
Patient head
Ultrasoun
d
transduce
r
RF
coil

STANFORD
UHF MRI
PROGRAM
MR ACOUSTICS
UHF-MRI ACOUSTICS IN HEAD AND BODY GRADIENTS – SIMONE
WINKLER, TREVOR WADE, ANDREW ALEJSKI, CHARLES MCKENZIE,
BRIAN RUTT

STANFORD
UHF MRI
PROGRAM
HEAD GRADIENT INSERT
ETH ZÜRICH | 2014/12/19
 Smaller field of linear region (approx. 20
cm)
 Higher slew rate
 Higher gradient strength (100 – 300 mT/m)
 Inserted into bore
190
170
220
450
490 338
Body gradient coil Head gradient insert coil
Faster or higher resolution imaging
 Large linear region (60 cm)
 Low slew rate
 Low gradient strength (50 – 80 mT/m)
 Built into MR system
• Attractive at
high field
• Louder
acoustics

STANFORD
UHF MRI
PROGRAM
FUNDAMENTALS OF VIBROACOUSTICS
Time-
varying
pressure
inside bore
Structural
vibration of
gradient coil
B0
.
F(t)
i(t)
Fundamentals:
Time-varying
current in gradient
conductor, in
presence of strong
B0 field, causes
Lorentz forces
Structur
e
Flui
d
(air)
vibrations
Pressure change = sound wave
Structural vibrations:
n=0 n=1
n=2 n=3
Breathing Bending
Ovaling
Acoustic coupling:
structure modes in the coil
structure
acoustic modes in the coil
bore
Interaction
MUTUAL
COUPLING!
Avg displacement (3T, 50A): 1μm
Total displacement is weighted
sum of different modes

STANFORD
UHF MRI
PROGRAM
MODELING
No displacement in
vertical direction
Hard acoustic
boundary
1
2
2
1. Structural vibration of coil
2. Acoustic wave propagation in
air
Infinite half space of air on both
bore ends
Perfectly
Matched
Layer
2
Insert gradient modeling: Surrounding air modeling:
Acousticwavepropagation:LorentzForces:
150
0
N/m
Fx
Fy
YX Z
820 Hz
Highly accurate prediction of SPL in gradient coils
 acoustic-based design strategies for SPL reduction become possible
New features in gradient vibroacoustic modeling:
as compared to existing published modalities
-----------------------------------------------------------------------
1. Accurate wire patterns
2. Realistic bore shape with patient bridge
3. Acoustic propagation outside bore
4. Full coupling of vibrations and acoustics
5. Lorentz damping

STANFORD
UHF MRI
PROGRAM
RESULTS REALISTIC HEAD GRADIENT
50A,3T
X-gradient Y-gradient Z-gradient
isocenter
20.6 g
26.8 g
8.7 g
3.7 g
15.6 g
12.4 g
Innerborewall,+16cm
98.7 dB
99.2 dB
95.4 dB
104.3 dB
97.6 dB
96.9 dB
50A,3TAccelerationsSoundpressurelevel

STANFORD
UHF MRI
PROGRAM
 Modes are shifted to
higher frequencies
and out of the band
of interest for RF
pulses
 Extremely short axial
length of our head
gradient design
pushes
circumferential
modes to much
higher frequencies
 These are not
excited by the z-
gradient  short is
good!
S. A. WINKLER
(0,0,1)
(0,0,1)
(0,1,1)
(0,0,2)
(0,1,2)
(0,0,3)
(0,1,3) (1,0,2)
(coupled)
(0,5,4)
(0,0,2)
(0,1,5)
(1,0,3)
(2,0,3)
(1,0,3)
(2,0,3)
(2,2,1)
(3,0,3)
(4,0,0)
axial modes
radial modes
COMPARISON HEAD-BODY GRADIENTS
ETH ZÜRICH | 2014/12/19

STANFORD
UHF MRI
PROGRAM
LORENTZ DAMPING
B0
FL v
FL = I x B0
JCL = σ (v x B0)
ICL
I
ICL = JCL * A
FCL = ICL x B0
Total force
F = FL + FCL
Total force explicit for B0
F = FL - σAB0
2v
I
B0
.
F(t)
i(t)

STANFORD
UHF MRI
PROGRAM
EDDY CURRENT DAMPING
Eddy current density induced in moving conductor
JCL = σ*(v x B0)
Lorentz force induced due to Eddy current
FCL = JCL*A x B0
Total force
F = FL + FCL
Total force explicit for B0 in z-direction:
F = FL - σAB0
2*v
Equation of motion
This damping term is
dependent on B0
2 
nonlinear increase in SPLs
with field strength

STANFORD
UHF MRI
PROGRAM
NOISE REDUCTION METHODS
 Moderate noise
reduction achieved
 Current efforts for
sound reduction are
ongoing
 Bore acts as acoustic waveguide
 Terminations to outside air are highly reflective  bore
acts as resonator
 1. impedance matching of waveguide to free space
impedance lets acoustic energy travel outward (horn)
 2. absorbing endcap reduces resonances
Traveling wave concept using a horn.
Proposed noise reduction method
Winkler, Alejski, Wade, McKenzie, Rutt, ISMRM 2014

STANFORD
UHF MRI
PROGRAM
GRADIENT COIL ACOUSTICS – FIELD DEPENDENCE
Lorentz damping includedNo Lorentz damping
7.5 dB from 3T to 7T
10.8 dB from 3T to 10.5 T
91.2 dB
97.5 dB
100.8 dB
3T:
7T:
10.5T:
92.1 dB
89.8 dB
90.5 dB
3T:
7T:
10.5T:
Linear scaling with main
field strength
Nonlinear scaling with
main field strength

STANFORD
UHF MRI
PROGRAM
GRADIENT COIL ACOUSTICS – FIELD DEPENDENCE
Lorentz damping includedNo Lorentz damping
91.2 dB
97.5 dB
100.8 dB
3T:
7T:
10.5T:
92.1 dB
89.8 dB
90.5 dB
3T:
7T:
10.5T:
Coupled
resonanc
e
Linear scaling with main
field strength
Nonlinear scaling with
main field strength

STANFORD
UHF MRI
PROGRAM
CONCLUSION
 UHF MRI increases SNR and resolution
 Many technical challenges are faced before clinical adoption
 This talk showed innovative approaches on:
 B1 inhomogeneities / image shading
 B0 inhomogeneities
 MR Safety
 Acoustics

STANFORD
UHF MRI
PROGRAM
OTHER RESPONSIBILITIES AND THINGS I DO
 Management of RF
and electronics lab
 3D printer and
mechanical
manufacturing facility
management
 Peripheral nerve
stimulation
 Transcranial magnetic
stimulation
 Molecular imaging –
magnetogenetics
 Genetic activation for
cancer treatment by
alternating magnetic
field induced heating
…and many more!

THANK YOU!
ACKNOWLEDGEMENTS:
o ALL ANONYMOUS VOLUNTEERS
o MOHAMMAD MEHDI KHALIGHI
o KEVIN & KARLA EPPERSON
o IVES LEVESQUE
o THOMAS TOURDIAS
o ANNE SAWYER
o BRIAN HARGREAVES
o BRUCE DANIEL
o THOMAS WRIEDT
o JUERG FROEHLICH
o ZMT/SPEAG GMBH
COLLABORATORS:
o LARRY WALD
o JASON STOCKMANN
o BORIS KEIL
o MIHIR RAJENDRA PENDSE
o RICCARDO STARA
o ALESSANDRO SBRIZZI
o PAUL PICOT
o MICHAEL THORNTON
o TREVOR WADE
o ANDREW ALEJSKI
o CHARLES MCKENZIE
o JANOS BARTHA

Stanford UHF MRI Program: Engineering Challenges and Solutions

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (13)

Similar to Stanford UHF MRI Program: Engineering Challenges and Solutions

Similar to Stanford UHF MRI Program: Engineering Challenges and Solutions (20)

Recently uploaded

Recently uploaded (20)

Stanford UHF MRI Program: Engineering Challenges and Solutions

Editor's Notes