Magnetic resonance imaging in biomedical research

Report
Magnetic resonance imaging in
biomedical research
Igor Serša
Ljubljana, 2011
History of Nuclear Magnetic
Resonance (NMR)
Purcell, Torrey, Pound (1946)
Bloch, Hansen, Packard (1946)
1D NMR spectroscopy (CW)
Pulsed NMR
Emergence of computers
P.C. Lauterbur (1973)
P. Mansfield (1973)
R.R. Ernst (1975)
Multidimensional
NMR spectroscopy
NQR ,
Solid state NMR,
NMR in Eart‘s field
Biomedcial use of
NMR, magnetic
resonance imaging
(MRI)
Nobel Laureates in MRI
R.R. Ernst
1991 chemistry
For a discovery of multidimensional NMR and
setting foundations of Fourier transform MRI
methods
P. Mansfield
2003 medicine
For the development of fast MRI (Echo planar
imaging)
P.C. Lauterbur 2003 medicine
First who succeded to get a MR image
MRI in early days
Lauterbur, P.C. (1973). Nature 242, 190.
… and MRI now
MRI statistics
• MRI Equipment Market of 5.5 Billion Dollars in 2010
• 91.2 MRI exams are performed per 1,000 population per year in USA
• 41.3 MRI exams are performed per 1,000 population per year in OECD countries
• 22.2 MRI exams are performed per 1,000 population per year in Slovenia
• 7,950 MRI scanners in USA (25.9 MRI scanners per million population)
• 18 MRI scanners in Slovenia (9 MRI scanners per million population)
Opening ceremony of the last
MRI scanner in Slovenia
(Murska Sobota)
Investment of 1,200,000 €
MRI systems
Clinical MRI system
Use in radiology
B0 = 1,5 T, opening 60 cm
High-reolution NMR/MRI system
Use in chemistry, MR microscopy
B0 = 7 T, opening 3 cm
Nuclear magnetization
M
p
mi
i
V
Nuclear precession
Mz
B0
M0
M0/2
RF pulse
B1 field
T1 ln(2)
   B1 t p
0  B0
100 MHz proton precession frequency in 2.35 T
t
MR signal
FID signal
Ui
U0
t
M
FT

Ui

spectrum
Magnetic field gradients
B0
x
Gx x
+
x
=
B
Sedle coil
Maxwell pair
x
MR imaging in one dimension
B
x
0
 ( x )  0

B
x

 ( x )  0   G x x
MR imaging in two dimensions
back projection reconstruction method
Pulse sequences
RF
p
p/2
AQ
Gx
Gy
Gz
TE
MRI in biomedicine
Research on clinical MR scanners
Hardware development
• RF coils
• Gradient coils
• Amplifiers
• Spectrometers
Imaging sequences
• Standard MRI
• Contrast
• Speed
• Resolution
• Spectroscopic
Data processing
• New reconstruction algorithms
• Image filtering
• Mathematical modelling
Rsearch on other MRI systems
MR microscopy
• MRI of wood
• Pharmaceutical
studies
• Porous materials
• Biologoical Tissue
properties
• MRI of food
Small anaimal MRI
• Development of new MRI
contrast agents
• Study of new drugs
Hardware development
Multi channel RF coils
(32 channel head coil)
Gradient amplifiers
• Gradients up to 45 mT/m
• Gradient rise time of 200 T/m/ms
• 600 A @ 2000 V = 1.2 MW !
RF amplifiers
• 35 kW
MRI magnets
• 1.5 T, 3 T, 7 T
• Low weight
• Compact dimensions
• Low helium consumption
Imaging sequences
Type of sequence
Spin echo (SE)
Multiecho SE
Fast SE
Ultrafast SE
IR
STIR
FLAIR
Gradient echo (GE)
GE with spoiled residual transverse
magnetization
Ultrafast GE
Ultrafast GE with magnetization
preparation
Steady state GE
Contrast enhanced steady state GE
Balanced
steady state GE
Echoplanar
Hybrid echo
Principles
Advantages
Disadvantages
simple, SE
T1, T2, DP contrast
SE several TE, several images
Contrast
Slow (especially in T2)
DP + T2 images
Slow, even if acquisition of the 2nd image does
not lengthen acquisition
SE, echo train
effctive TE
SE, long echo train, half-Fourier
Faster than simple SE simple
ES contrast
Even faster
Fat shown as a hypersignal
RF 180°, TI + ES/ESR/EG
T1 weighting
Tissue suppression signal if TI is adapted to T1
Longer TR / acquisition time
IR, short TI 150 ms
Fat signal suppression
Longer TR / acquisition time
IR, long TI 2200 ms
CSF signal suppression
Longer TR / acquisition time
< 90° α and short TR
No rephasing pulse
TR < T2
Gradients / RF dephasers
+ speed
T2* not T2
small α and very short TR
Gradients / RF dephasers
k-space optimization
++ speed
cardiac perfusion
+ preparation pulse:
- IR (T1weighted)
- T2 sensibilization
TR < T2
Rephasing gradients
FID
Rephasing gradients
Hahn echo ( trueT2)
Balanced gradients in all 3 directions
T2/T1contrast
++ speed
AngioMRI Gado
Cardiac perfusion / viability
+ signal
++ speed
Single GE or multi shot
Preparation by SE (T2), GE (T2*), IR (T1), DW
Exacting for gradients
++++ speed
Perfusion
MRIf BOLD
Diffusion
Fast SE
+ intermediary GE
++ speed
SAR reduction
Low signal to noise ratio
T1, DP weighting
Poor T1 weighting
Complex contrast
Not much signal
T2 weighted
++ signal, ++ speed
Flow correction
Limited resolution
Artifacts
Clinical MR images
fMRI
Fiber tracking
MRI of spine
MR angiography
DWI - stroke
MRI – brain tumors
New reconstruction methods
0
Sa’
0
=
Sc’
Sd’
0
0
Sb’
R=4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S
Small animal MRI
Experimental mice
Anaesthesia
Placement in the probe
Resolution
Resolution × SNR
 const($$$)
Time
Time
Signal
Noise
Multiple sclerosis model
• Mice having Theiler’s Murine Encephalitis Virus infection (TMEV) may develop
symptoms similar to that of multiple sclerosis
• Intracerebral injection causes demyelinating disease
• CD8 cell mediated disease
7 days post
infection
Cr
Cho
NAA
Before
infection
Normal cord
MS cord
6
T2-weighted images
MS lesions (demyelinated choppy structures)
appear bright
5
4
3
2
1
ppm 0
Decrease in NAA/Cr ratio in
early stage of MS.
Superparamagnetic labells
•
•
•
•
•
Superparamagnetic
antibodies under scanning
electron microscope
attached to CD8 cells.
USPIO - Ultrasmall Super Paramagnetic Iron
Oxide particle: 50 nm in diameter
Highly specific superparamagnetically labeled
antibodies: targeted USPIO-s
Venous administration
Signal persists for days, excellent specificity
A single labeled cell can theoretically provide
adequate signal to be visualized
MS lesions detected by CD8 labeling
B6 strain mice (acute demyelinating disease, full recovery in 4-6 weeks)
Day 0
Day 3
Day 7
Day 21
Day 45
What is MR Microscopy?
MR microscopy is essentially identical to conventional MRI (most of MR sequences of
clinical MRI can be used) except that resolution is at least an order of magnitude
higher.
Signal -> Signal / 100
Conventional MRI
1 mm / pixel
2D
10 fold resolution increase
3D
10-100 µm / pixel
MR microscopy
Signal -> Signal / 1000
How to compensate the signal loss?
•
•
•
•
By using stronger magnets
By lowering the sample temperature (not an
option)
By signal averaging
By reducing RF coil size
RF coils in sizes from 2 mm – 25 mm
Signal  B 0
7 – 14 T
2
How to achieve high resolution?
By the use of stronger gradients
GR
Δt
45 mT/m @ 750 A
Conventional MRI
FOV 
2p
 GR t
GR
1500 mT/m @ 60 A
MR microscopy
Δt
MRI laboratory at JSI
100 MHz (proton frequency)
2.35 T
Horizontal bore superconducting magnet
Accessories for MR microscopy
Top gradients of 250 mT/m, RF probes 2-25 mm
Our research using MR microscopy
Electric current density imaging
NMR of porous materials
MRI of wood
NMR in studies of thrombolysis
Volume selective excitation MRI in pharmaceutical research
http://titan.ijs.si/MRI/index.html
MRI in dental research
NMR in studies of thrombolysis
blood clot
magnet
0,7 mm
3 mm
30 mm
• ηk = 1.8·ηH20 = 0.0018 Pas
• ρk = 1035 kg/m3
0,5 l
plazma +
rt-PA
3 mm
pump
p = 15 kPa (113 mmHg), arterial system
p = 3 kPa (22 mmHg), venous system
Flow regime
v [m/s]
Re
Fast
flow
begining
4,26
1660
end
0,86
1430
Slow
flow
begining
0,19
75
end
0,01
18
NMR in studies of thrombolysis
TE = 12 ms
TR = 400 ms
SLTH = 2 mm
FOV = 20 mm
Matrix: 256 x 256
Dynamical 2D MR microscopy using
spin-echo MRI sequence
Fast flow
0 min
4 min
8 min
12 min
16 min
4 min
8 min
12 min
16 min
Slow flow
0 min
NMR in studies of thrombolysis
1
x
0.8
Hiter
tok
Fast flow
0.6
Slow flowtok
Počasen
x
1
x  1  S / S
0.4
0.2
T
S0
0
0
500
1000
1500
2000
2500
t [s]
S0
t
S
S∞
SERŠA, Igor, TRATAR, Gregor, MIKAC, Urška, BLINC, Aleš. A mathematical model for the dissolution of non-occlusive blood clots in fast tangential blood
flow. Biorheology (Oxf.), 2007, vol. 44, p. 1-16.
NMR in studies of thrombolysis
•
3D RARE MRI (fast flow, ∆p = 15 kPa)
0 min
36 min
1,2
1,0
t = 0 min
0,8
0,6
0,4
0,2
0,0
0
2
4
6
8
10
12
entrance length z [mm]
channel radius R [mm]
channel radius R [mm]
1,2
t = 36 min
1,0
0,8
0,6
0,4
0,2
0,0
0
2
4
6
8
10
12
entrance length z [mm]
NMR in studies of thrombolysis
•
Blood clot dissolution progresses radially with regard to the perfusion channel
along the clot.
2R∞
•
•
2R
v 0   V / (p R )
2
Volume blood flow through the clot is constant.
Mechanical forces to the surface of the clot have viscous origin and are
therefore proportional to the shear velocity of blood flow along the clot.
Confocal microscopy of thrombolysis
F
λ
5 μm
J. W. Weisel, Structure of fibrin: impact on clot stability, J
Thromb Haemost 2007
 v
v   v 

dA  F ds   S 

dt


S


 r    r 

 r
2

( z )  dt
rR

NMR in studies of thrombolysis
•
Mechanical work needed for the removal of the clot segment is
proportional to its volume.
λ
Layer of the clot that is well
perfused with the
thrombolytic agent
dA  c dV  c S dR
2R
Layer of the clot that is
removed in time dt
•
dR
Start of thrombolytic biochemical reactions is delayed (τ) and gradual (Δ)
1/c
1/c∞
1
c (t )

1
1
c  1  exp((  t ) /  )
τ
Δ
t
NMR in studies of thrombolysis
Perfussion channel profile


 R

R ( z, t  tD )  




1
 R 
 1  exp(( t   ) /  ) 

 0  
ln 

T7
  R  
 1  exp(  /  ) 
7
1  1 
z z0

4

2
7


;
z  z0
;
z  z0
1
7
 R 
 1  exp(( t   ) /  )  

0
R 
ln 
 

T7
  R  
 1  exp(  /  )  
7
z  z0
;
.
Thrombolytic time
7


T 
 R0  
7
   ln  exp 
1  
  1


 

 R   




tD ( z )  

T 

 R

   ln exp  7  1   0

  


 R





1 



7
z z0

4

2



   

 1  exp 
 1



  




   


  1  exp 
 1


  


;
z  z0
.
SERŠA, Igor, VIDMAR, Jernej, GROBELNIK, Barbara, MIKAC, Urška, TRATAR, Gregor, BLINC, Aleš. Modelling the effect of laminar axially
directed blood flow on the dissolution of non-occlusive blood clots. Phys. Med. Biol., 2007, vol. 52, p. 2969-2985.
NMR in studies of thrombolysis
Current density imaging
Externally applied electric field is used to induce cell permeability by transient or
permanent structural changes in membrane
The aim of this study was to monitor current density during high-voltage
electroporation (important for electrode design and positioning)
Current density imaging
Electroporation phantom
Current density imaging
Effect of electric pulses
Current density imaging
Electric pulses
CDI calculation
• Two 20 ms pulses @ 15 V
• Eight 100 μs pulses @ 1000 V
1. Phase is proportional to Bz

CDI
x, y   
Bz x, y  t
2. Ampere law

1  B z B y
j CDI 

,

 0   y
z
B x
z

B z
x
,
B y
x
Thin-sample approximation

1  B z
j CDI 
,

 0   y
Current encoding
part
Imaging
part

B z
x
,

0 


B x 

y 
Current density imaging
Vector field (jx,jy)
30
16
14
25
12
20
y
10
8
15
6
10
4
2
5
2
5
10
Vector field (jx,jy)
15
20
25
4
6
8
10
12
14
30
x
30
16
14
25
12
20
10
y
8
15
6
4
10
2
5
Electrode
setup
16
30
25
experiment
20
15
10
5
14
12
10
6
4
simulation
x
Phase image
8
2D current density field
2
16
MRI of wood
On a 3m high beech tree,
transplanted in a portable
pot, a branch of 5mm
diameter was topped. The
topped branch was then
inserted in the RF coil and
then in the magnet.
MRI of wood
Pith, xylem rays, early wood vessels and cambial zone
6 mm
21 mm
MRI of wood
• Trees do not have a mechanism to heal wounds like higher organisms
(animals, humans), i.e., wounds are not gradually replaced by the
original tissue.
• In trees wounds are simply overgrown by the new tissue, while the
wounded tissue slowly degrades.
Wound
Dehydration
and dieback
new grown tissues
Formation of
the reaction
zone
MRI of wood
Day 1
Day 3
Day 8
MRI of wood
Day 14
Day 28
Day 168
MRI in dental research
enamel
Premolars
1-2 root channels
periodontal
communications
dentin
pulp
bifurcation
Molars
3-4 root channels
(in the literature was
reported even up to 7
root channels)
root
channel
MRI in dental research
Root channels are
not clearly visible.
Root channels after
endodontic treatment.
•
Standard X-ray image corresponds to 2D projection of hard dental
tissues (enamel and dentin) into a plane of image.
•
It is impossible to accurately determine the exact number of root
channels since they may overlap in the projection.
•
Fine details (periodontal communications and anastomosis) cannot be
seen due to limited resolution.
•
X-ray scanning is harmful due to X-ray radiation.
MRI in dental research
X-ray image
Hard dental tissues are
bright on the images, soft
tissues cannot be seen.
MR image obtained after co-addition of all slices
Soft dental tissues are bright on the images, hard
tissues cannot be seen. Frontal (bucco-lingual) as
well as side (mesio-distal) view is possible.
MRI in dental research
MRI in dental research
Conclusion
• MRI is very versatile.
• Its applications range from clinical routine in
radiology to research in medicine, biology as well as
in material science.
• Close collaboration between scientists and industrial
engineers enabled an enormous development of MRI
from an unreliable imaging modality to the new
radiological standard.

similar documents