Title: Multimodality small animal imaging: registration of functional EPR images with MRI anatomy
1Multimodality small animal imaging registration
of functional EPR images with MRI anatomy
Chad R. Haney, Adrian Parasca, Charles A.
Pelizzari, Greg S. Karczmar, Howard J.
Halpern Department of Radiation and Cellular
Oncology and Department of Radiology The
University of Chicago
Supported by grants DAMD17-02-1-0034 (DoD) and
P41EB002034(NIBIB)
2In Vivo EPR Imaging Topic of NIBIB Research
Resource(PI Howard Halpern, MD, PhD)
- Long term goal - develop EPR imaging techniques
which provide functional information that can be
of use in designing, delivering, and assessing
cancer therapy.
3Biological imaging to enhance targeting of
radiation therapy oxygen imaging
- Intensity modulated radiation therapy allows
sophisticated control over spatial distribution
of radiation dose
- Areas of hypoxia could be given extra dose if we
could identify them
4Why EPR Imaging?
- Spectroscopic Imaging Specific quantitative
sensitivity to Oxygen, Temperature, Viscosity,
pH, Thiol - No water background obscures spectrum of interest
(vs MRI) - 600 times stronger coupling to magnetic field,
environment (vs MRI) - Deep sensitivity at lower frequency (vs optical)
- Noninvasive (vs probes)
5EPR in vivo oximetry techniques
- Localized spectroscopy with implanted particulate
probes (Dartmouth) - Spectroscopic imaging with stepped fixed
gradients, water soluble probes - CW (Chicago, OSU, Aberdeen, LAquila)
- pulsed (NCI, Chicago)
- OMRI (NCI, Aberdeen)
- dynamic nuclear polarization using EPR spin
probes
6EPR is analogous to NMR
Zeeman splitting of electron spin energy states
in magnetic field
7EPRI is not identical to MRI
- Relaxation times 10-6 as long
- pulsed gradient techniques not applicable
- FID correspondingly short ? demanding of pulsed
measurement techniques - p/2 pulse 50 ns long, FID lasts few µs
- have to introduce spin probe no endogenous
signal - frequency 660 times higher for given field
- (or, field 660 times lower for given frequency)
8RF penetration favors lower frequency
proton Larmor frequency 4258 Hz/gauss 42.6 MHz
at 1 Tesla electron Larmor frequency 2.80
MHz/gauss 28 GHz at 1 Tesla
250 MHz 6 T MRI, 90 G EPR S/N w0.8
ratio meas/calc
Nw 1.2 IN LOSSY, CONDUCTIVE TISSUE
9Continuous wave spectral spatial imaging each
voxel yields a spectrum whose linewidth increases
linearly with local oxygen concentration fixed
stepped field gradients, swept magnetic field
EPR line broadening for current narrow line spin
probes approximately 0.5 mG/torr O2
10Line width pO2 calibration
Oxygen dependence of lorentzian line width
obtained in a series of homogenous solutions of
OX31spin probe
11Spectral-spatial projection
(c) A field sweep now corresponds to a projection
along a direction rotated in the spectral-spatial
plane. Larger gradients correspond to larger
rotation angles. Pure spatial projection would
require infinite gradient.
12250 MHz Spectrometer Magnetsvarying diameter
homogeneous field regions (90 G)
Intermediate 15 cm diam.
Large 30 cm diam.
Small 8 cm diam.
13Mouse Image using OX063 spin probe
PC3 human prostate cancer xenograft on nude mouse
hind limb
14Registration of EPR with MRI for anatomically
aided analysis
Registration based on - Fiducials - Surfaces -
Intensity distribution
Note high intensity due to poor clearance of spin
probe from tumor, and low oxygen tension in same
region
15Early fiducial markers filled with dilute spin
probe solution. Problem need to remove during 4D
image to avoid artifacts
16Immobilization cast, fiducial markers for serial
and intermodality registration
17Alignment of MRI and EPRI (red) fiducial surfaces
18Manual refinement of initial registration
estimate based on fiducials
19Application radiation inducible antivascular
gene therapy
20PC3 tumor treated with Ad.CMV.null virus (control)
Pre treatment mean pO2 in tumor 44.6 torr, std
35.1, SEM 1.62. tumor volume from MRI 0.160 mL
4 days post treatment (right) mean pO2 in tumor
28.7 torr, std 29.1, SEM 1.065. tumor volume
from MRI 0.422 mL
21PC3 tumor treated with Ad.EGR-TNFa virus 10 Gy
Pre treatment mean pO2 in tumor 27.3, std 36.1,
SEM 1.122 tumor volume from MRI 0.524 mL
4 days post treatment mean pO2 in tumor 31.7
torr, std 17.1, SEM 0.472. tumor volume from
MRI 0.417 mL
22Conclusions
- 4D EPR Images can be obtained with 1 mm spatial
resolution and 1.5mG (3 torr pO2) spectral
resolution - Preliminary images of increased and decreased
regional oxygenation levels following radiation
adeno-EGR-TNF? anti-vascular therapy have been
seen. - These images may have potential for
biologically-based planning and assessment of
radiation therapy - Registration of these functional images with
anatomic images such as MRI is key to accurate
interpretation and to eventual clinical
applications
23Chicago EPRI Lab Howard Halpern Martyna
Elas Colin Mailer Chad Haney Charles
Pelizzari Kazuhiro Ichikawa Gene Barth Ben
Williams Kang-Hyun Ahn Adrian Parasca VS
Subramanian
Chicago MRI Lab Greg Karczmar Jonathan
River Xiaobing Fan Marta Zamora
Denver EPR Lab Gareth Eaton Sandra
Eaton Richard Quine George Rinard
EGRF-TNF? radiation therapy Ralph
Weichselbaum Helena Mauceri Michael Beckett