The SIMRI project A versatile and interactive MRI simulator * H. Benoit-Cattin 1 , G. Collewet 2 , B. Belaroussi 1 , H. Saint-Jalmes 3 , C. Odet 1 1 CREATIS, UMR CNRS #5515, U 630 Inserm, INSA Lyon, UCB Lyon 2 CEMAGREF, Food process Engineering Research Unit, Rennes 3 LMRMN-MIB, UMR CNRS 5012, UCB Lyon * JMR, 173 (2005) 97-115 COST B21 Meeting, Lodz, 6-9 Oct. 2005 2/32 1. Context 2. SIMRI overview 3. Simulation results 4. Simulator implementation 5. Perspectives
Transcript
The SIMRI projectA versatile and interactive MRI
simulator *
H. Benoit-Cattin 1, G. Collewet 2, B. Belaroussi 1, H. Saint-Jalmes 3, C. Odet 1
1 CREATIS, UMR CNRS #5515, U 630 Inserm, INSA Lyon, UCB Lyon2 CEMAGREF, Food process Engineering Research Unit, Rennes3 LMRMN-MIB, UMR CNRS 5012, UCB Lyon
* JMR, 173 (2005) 97-115
COST B21 Meeting, Lodz, 6-9 Oct. 2005
2/32
1. Context
2. SIMRI overview
3. Simulation results
4. Simulator implementation
5. Perspectives
3/32
MR Imaging• Recent and complex technique• Based on protons magnetization• High static field + RF excitations• Contrast imaging adapted to soft tissue
MRI simulation• Better understanding of MRI images• Pedagogic purposes• Conception, calibration and test of MRI sequences• MRI Images with a « ground truth »
• Artifacts impact and correction• Image processing assessment (segmentation, quantification)
• Main works• 1D MRI simulation [Bittoun-81]• 2D MRI simulation [Olsson-95]• Simulation with a distributed implementation [Brenner-97]• 3D brain MRI simulation [Kwan-99]• Susceptibility and MRI simulation [Yoder-02-04]
- Sinc shaped pulse (duration, number of lobes, number of
point)
- User shaped pulse (file, constant pulse list)
• 1D signal acquisition step (number of points,
bandwidth, readout gradient, k space position)
10/32
2.5. Magnetization kernel
−−×=
10
2
2
/)(//
).(TMM
TMTM
BMdtMd
z
y
xrrr
γ
),(..).().(),( trMRRRotRotttrM RFrelaxizgzrrrr
θθ=∆+
RFacquisition
)(tMr
)( 1tMr
)( 2tMr
Magnetic event
k-space filling
Do PulseDo GradientDo Waiting
Kernel is based on the Bloch Equation
and on their discrete time solution
11/32
Lorenzian distribution of isochromats
FID weighting
Spin refocusing, Spin Echo
2.6. T2* effect, limited spin number
iBTT∆+= γ
22
1*
1tBie ∆−γ
α1 α2
t
exp(- B )∆ τi
a) b) c)τ=0
(τ τ= +t) (τ τ= +t)τ τ=-(τ τ= +t)
TE/2 TE/2
12/32
2.7. Noise and filtering
Thermal noise White Gaussian noise in the k-space
K-space filtering Hamming (or any digital filter), prevent ringing
ReconstructionFFT
13/32
3. Simulation results
3.1. Echo train
3.2. Contrast in GE and SE imaging
3.3. True Fisp imaging
3.4. Chemical shift artifact
3.5. Susceptibility artifact
14/32
3.1. Echo train and T2*
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.50
1000
2000
3000
4000
5000
6000
|M|
xy
T2
90° 180° 180° 180° 180° 180°
te
t(s)
te/2
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.090
500
1000
1500
2000
2500
3000
3500
t(s)
|M|
xy
T2
T *2
90°Gx
tete/2
Simulated signal obtained after a gradient echo pulse sequence. It is composed of a train of T2* weighted gradient echoes. The intra-voxel inhomogeneity is set to 10-6 T and the main field to 1 T.
Simulated signal obtained after a CPMG sequence. It is composed of a train of spin echoes T2 weighted. The intra-voxel inhomogeneity is set to 10-6 T and the main field to 1 T.
15/32
3.2. Contrast in SE and GE imaging
SE sequence
GE sequence
16/32
McGill Brain phantom10 tissues profiles
17/32
3.3. True Fisp imaging
18/32
3.4. Chemical shift artifact
ππ )12(2 +=⋅∆⋅ kTEf
Water and fat signals in phase opposition !
19/32
3.5. Susceptibility artifact
Geometric distortions+ Intensity loss
20/32
4. Simulator implementation
3.1. Code organization
3.2. Sequence programming
3.3. Acquisition programming
3.4. 1D interactive simulation
3.5. Distributed implementation
21/32
3.1. Code organization
• ANSI C language• Running on Linux and windows OS• Independent module organization• Linked in a dll wrapped for being used with python for the 1D
interface• Parallelized using MPI to run on grid, cluster, multiprocessors.
22/32
3.2. Sequence programming
23/32
3.3. Acquisition programming
24/32
3.4. 1D interactive simulation
25/32
The Spin Player
26/32
3.5. Distributed implementationTime is the enemy !
• Simulation time ])..).[(...( exs NPNMZYXcTacqTexcT +=+≈
• 2D Image (NxN) - Nx2 time x16- 10242 = 2.3 days- High resolution : small cluster
• 3D image(NxNxN) - Nx2 time x 64- 1283 = 9.3 days / 5123 = 104 years !- High resolution large scale data grid
27/32
• A “Divide & Conquer “ parallelisation scheme • Based on the free library MPI, transparent at a user level• Allowing run on single PC, PC cluster, grid architecture, massively parallel machine
• Work done in the context of European grid projects- DATAGRID Project (2001-03)- EGEE Project (2003-05)
Simulationweb portal
28/32
5. Conlusion
An operational MRI simulator• 1D-2D-3D • Sequences (SE, GE, Turbo, TFisp, STIR)• Artifacts : Chemical shift, susceptibility, static field• B0 , T2*, Off-On resonance• Versatile, Parallelized, Interactive and GPL !
29/32
6. Perspectives
Main perspectives• Anatomic object design Sequence tuning
Image processing evaluation• Molecular imaging susceptibility and Cell. Con. agent
• New sequences / Interface with ODIN project• RF / Antennas• Artifact correction• Flow , Diffusion & Perfusion ?
30/32
31/32
Field map estimation
Object( ki )
Boundary Integral
SIMRI
−= ∫∑
= k Face
NbFace
1kΩ q)dsq(P,C oB n∆Xm
4π1(P)δ Xm(P)oB(P)B
Boundary element
[Yoder02]
[Balac04]
)(rBsr
∆
sB∆
Acquisition [Kanayama_96]- 2 EG phase images with TE1,
TE2- ∆Φ + unwrap [Jenkinson_03] > ∆B0
32/32
−=
1000cossin0sincos
)( θθθθ
θzRot∫∆+
=tt
tg dGr ττγθ )(..
rr
trBi ∆∆= ).(. rγθ
−
=∆
−
∆−
∆−
)(
)(
)(
1
2
2
100
00
00
rTt
rTt
rTt
relax
e
e
e
Rr
r
r
)().().( φαφ −= zxzRF RotRotRotR
( )2
2.'
+∆−=ταωτα
gradient effect
Relaxation effect
Inhomogeneity effect
Pulse effect
Off-resonance effect
33/32
• K-space acquisition - One point is obtained by summation all over the object
- Next point is obtained after a time step ∆t, the sampling rate
∑∑ +=rr
ytrMjxtrMtsrr
rrrrrr).,().,(][
34/32
3.5. Distributed implementationTime is the enemy !
• Simulation time ])..).[(...( exs NPNMZYXcTacqTexcT +=+≈
• 2D Image (NxN) - Nx2 time x16- 10242 = 2.3 days- High resolution : small cluster
• 3D image(NxNxN) - Nx2 time x 64- 1283 = 9.3 days / 5123 = 104 years !- High resolution large scale data grid
35/32
• A “Divide & Conquer “ parallelisation scheme • Based on the free library MPI• Transparent at a user level• Allowing run on single PC, PC cluster, grid architecture, massively parallel machine
Virtualobject
Master node
Computing node N1
Computing node Ni
Computing node NN
Seq.Seq.
Seq.Seq.
Seq.Seq.
11
ii
N-1
ΣRec.FFT
Master node
RF signalcontributions
RF signalcontributions
s1
si
sN-1
Virtual objectportions
Virtual objectportions
MRISeq.
MRIimage
k space
36/32
• Some test results :- PC cluster CREATIS : 8 (PIII-1Ghz) + 10 (PIV-2,6 Ghz)- CINES parallel machine : 64 to 128 nodes, RI 4000-500 Mhz- IN2P3 through EGEE grid interface : 9 AMD Opteron 2.2 Ghz
- 1283 : 8h30 on a 128 proc. CINES machine
• Work done in the context of European grid projects- DATAGRID Project (2001-03)- EGEE Project (2003-05)
