Breast Cancer Trends 1973-1999

1
Magnetic Resonance Imaging of Fast Relaxing
Spins Acquisition during Adiabatic Excitation 
November 14, 2005, CMRR Djaudat
Idiyatullin
2
Mikes crazy idea is working 
3
Interleaved excitation and sampling during a
frequency-swept pulse


BIR4

• Steady state
• Sensitive to spins with a very short T2
• It is not FID but the signal predictable by Bloch
  simulation.

4
How to extract information from this weird
sampling during swept excitation?
Least square method Monte Carlo simulation
Wavelet transform
5
How to extract information from this weird
sampling during swept excitation?Solution

1. Away from adiabatic condition

2. Linear system- Correlation method
6
Linear system
System h(t)
Output r(t)
Input x(t)

• A system is linear if
• Linearity Output C Input
• Shift invariant delaying of Input ? same
  delaying of Output

Convolution
x(t) ? x(t)h(t- t)
Fourier theorem
0 t t
7
Evolution of the isochromats during HS8 pulse
(dw10mks, ( 30 degree), R2500Hz)
8
Evolution of the isochromats during HS8 pulse
(dw10mks, ( 30 degree), R2500Hz)
Linear system
9
Correlation method for linear system
FT
Spin system h(t)
Excitation x(t)
Response r(t)

System spectrum
FT
10
Simulated data HS4 pulse 100 isochromats from
-12.5kHz step 250Hz dw10mks R1500Hz
11
SWeep Imaging with Fourier Transform (SWIFT)
HSn pulses Flip angle lt 90 degree Tr
Tp Bwsw2pN/Tp Back-projection reconstruction
12
SWIFT, characteristics

• Signal intensity depends only on T1 and spin
  density (M0)
• Maximum signal intensity
• Ernst angle
• Maximum T1 contrast
• Spin density contrast
• Sensitive to short T2

13
SWIFT, hardware problems

• Dead time after pulse
• 4.7T , 7T 3µs sw lt 130kHz
• 4T 20µs sw lt 40kHz
• FIFO underflow happens if
• Tr lt 5ms for 128 sampling
• Tr lt 10ms for 256 sampling
• sw 25-35 kHz

14
MIP of 3D image sw32kHz128x128 x 644T
Empty 16-element TEM head coil
15
3D image of thermoplastic T20.3ms
sw100kHz128x128 x 1284.7T
Sensitivity to short T2
16
MIP of 3D image plastic toy in breast coil
sw39kHz128x128 x 128D25cm4T
Sensitivity to short T2
17
Slices of 3D image of feet sw20kHz4T
First in vivo SWIFT 3D images
18
Slices of 3D image raspberryin vivo
sw100kHz128x128x128D3cm4.7T
Sensitivity to raspberry
19
Advantages Disadvantages
fast Too fast for VARIAN FIFO underflow
Sensitive to short T2 Sensitive to coil material
Reduced motion artifacts (zero echo time, back projection reconstruction) Problems with slice selection
Reduced signal dynamic range ?
quiet Too quiet
20
Another Mikes crazy idea 
21
Breast MR scanner
22
Thanks toIvan Tkac Gregor Adriany
Peter Andersen Tommy Vaughan Xiaoliang
ZhangCarl SnyderBrian Hanna John StruppJanis
Zeltins Patrick BolanLance DelaBarreUte
Goerke
all CMMR

• Fast Quiet MRI by
• Sweeping Radiofrequency
• Djaudat Idiyatullin, Curt Corum,
• Jang-Yeon Park, Michael Garwood

Macros, C programmingHardware Software Yellow
pages of CMRR Discussion