2010 - University of Western Ontario

Report
Six-week project
Lauren Villemaire
MBP 3970Z
Department of Medical Biophysics
University of Western Ontario
Outline
 Introduction
MRI
Imaging at high magnetic fields
Phantoms
 Objective
 Methods
 Results
 Discussion
 Conclusion
 Acknowledgements
 References
2
Introduction to MRI
 Uses nuclear magnetic resonance of protons to produce
proton density images
 Magnetism
- proton spin
 Larmor Frequency
- rate of precession
The 7T MRI at Robarts Research Institute.
3
 The primary magnet
- main magnetic field
- the tesla
 The gradient magnet
- alters magnetic field
- focuses magnetic field
 The RF coil
- alters direction of proton spin
- detects precession energy
 T1 and T2 relaxation times
- characteristic of tissues
- image contrast
4
Imaging at ultra high magnetic fields
Advantages
Disadvantages
SNR
RF heterogeneities
Spatial Resolution
Magnetic susceptibility
artifacts
Tissue Contrast
Specific Absorption
Rate (SAR)
5
Human hippocampus
Images at 7T have
much higher spatial
resolution and SNR
than at 1.5T
1.5 T
7T
Brain images at 7T have shading and
bright spots that compromise image
homogeneity.
6
What are phantoms?
An artificial object of known size and composition that
is imaged to test, adjust or monitor an MRI system’s
- homogeneity
- imaging performance
- orientation aspects.
7
Objective
 To develop a brain-mimicking phantom for use in the 7T
MRI with the following characteristics in common with
the brain:
- Grey matter/white matter T1 and T2 relaxation times
- Electrical and wave properties
- Anatomical structure and size (not symmetrical)
8
Method
The relaxivity of varying concentrations of GdCl3 and agarose were
measured
T1 and T2 values that match average human
grey matter and white matter values were
determined via measurements done by MRI
CSF
CSF was mimicked by a 50-mM NaCl
solution
A concentric phantom was fabricated
White matter
Grey matter
Coil loading was measured and B1+ effects were empirically determined
9
Axial images of brain slices
were obtained from Brain
Web – Simulated Brain
Database.
These images represent the
standard size and structure
of the human brain.
Number of slices  36
Modality  T1
Slice thickness  5mm
Noise  0%
Intensity non-uniformity
(RF)  0%
10
14 brain slices, each
1cm apart, were
selected
Images were, then,
modified using
Image J to sharpen
and enhance
contrast between
grey matter, white
matter, and CSF.
11
Each image was manually outlined to distinguish between
the different compartments.
12
Tracings were scanned and made binary using ImageJ and
then converted to SAT using Solid Works.
jpg
pdf
dxf  sat
13
Images were then formatted to open in the MasterCam
Mill 9 program where they were modified.
14
Results
An agarose gel and saline solution phantom was developed to
mimic properties of the human brain for imaging at 7T.
Tissue
Target T1
(ms)
Target T2
(ms)
% agarose
[GdCl3]
(uM)
Grey matter
2000
55
2.1%
8
White matter
1300
45
2.2%
22
15
TI = 500ms
T1 = 1400ms to null GM
T1 = 900ms to null WM
T1W MP RAGE images of the same slice of the phantom with
different inversion times.
16
Comparison of RF
interference patterns.
Single element
transmitting
(located at back of head)
All elements transmitting
with random phases to
produce interferences.
17
Now that each brain slice is compartmentalized into
MasterCam, they can be milled out of plastic and
eventually filled with the appropriate brain mimicking
substances.
18
Discussion
I’ve successfully designed a head-mimicking phantom for use in the
7T MRI.
 The phantom exhibits very similar dielectric properties
(conductivity and permittivity) to the human brain
 The phantom is the same size and shape of the average
brain
 The phantom has similar anatomical structure to the
average brain
 The phantom has grey matter/ white matter contrast with
the same T1 and T2 relaxation times as human brain tissue
imaged at 7T
19
Conclusion
Such a phantom is unique.
It would...
(1) allow the ability to instrument the phantom and
measure RF power deposition (SAR)
and
(2) optimize RF shimming techniques using multiple
transmitters.
Both of these are major challenges currently.
20
Acknowledgements
Supervisor: Dr. Ravi S. Menon
Post-doctoral student: Kyle Gilbert
Graduate student: Andrew Curtis
21
References
 Brain Web: Simulated Brain Database
http://mouldy.bic.mni.mcgill.ca/brainweb/
 Rooney WD, et al. Magn Reson Med 2007; 57:308-318
 Wright PJ et al. MAGMA 2008; 21:121-130
 Yoshida A et al. Int J Hyperthermia 2004; 20:803-814
22

similar documents