Transcript Diapositive 1

A general study of optimal geometric parameters for time resolved
diffuse optical tomography.
Jérôme Boutet, Ludovic Lecordier, Simon Rehn, Mathieu Debourdeau, Lionel Hervé, Jean-Marc Dinten
LETI - CEA Minatec - 17 rue des Martyrs - 38054 Grenoble Cedex 9, France.
e-mail: [email protected]
Abstract
Fluorescence Diffuse Optical Tomography is a promising technique for cancer diagnostic.
It provides a way to characterize and localize tumors with a good accuracy and without
ionization of tissues[1]. This technique can be performed by continuous or time resolved
measurements. Time-resolved imaging is based on the use of subnanosecond laser
pulses for excitation combined to fast response detection, providing photon time of flight .
This poster presents an optimization of the acquisition geometry for breast cancer
diagnosis.
The influence of the sources step and the relative position between the source-detectors
and the inclusions are demonstrated by simulations.
After setting the optimal parameters, experimental acquisition was performed on a breast
phantom made of gelatin and other components reproducing both its optical and
geometrical properties. A fluorescent inclusion was placed inside the phantom to
simulate a marked tumor.
The fluorescence yield is reconstructed by processing the first two moments of the
temporal fluorescence signals obtained for each source detector combination. A novel
reconstruction method was used to localize the fluorescent inclusion with a millimetric
resolution.
Contribution of the time of flight information in reflection geometry
z localisation
1
0.9
0.8
Simulated reconstruction of a fluorescent
inclusion in a cubic medium with the use of
Intensity only
from intensity measurements
from temporal measurements
0.7
agreement
•Intensities contribution
Reconstruction from intensity only fails due
to poorer signal to noise and the fact that
shallow fluorophores provide a similar
intensity contribution as more concentrated
deep fluorophores.
Time of flight contribution
In reflection geometry, the time of
flight is highly dependent on the
distance between the inclusion and
the surface.
Thus, time of flight
provide a good resolution along Z
axis.
0.6
0.5
0.4
0.3
0.2
0.1
0
Simulated reconstruction of a fluorescent
inclusion in a cubic medium with the use
of the time of flight and intensities
0
0.5
1
1.5
depth (cm)
2
2.5
3
Probability of localization of the
inclusion as a function of Z axis.
Influence of geometrical parameters
Influence of the probe position: The best resolution is obtained for the shortest
source-inclusion distance.
•We simulated a fluorescent inclusion inside the
breast and compared different geometries to
study the influence of the number, step and
position of the sources and detectors
Influence of the step between sources :These series of simulations show that
larger source steps provide better localization accuracy. However, we have to
find a compromise between the quality of reconstruction and the size of the
probe
Simulated phantom and fluorescence obtained for different source and detector positions.
1st line corresponds to an inclusion positioned at the centre of the breast. 2nd line to an
inclusion at the external boundary.
Simulated phantom and fluorescence obtained for increasing source step values
Reconstruction technique
Acquisition system and experimental results
camera
Signals: This algorithm uses the first
two moments of the signals : mean
intensity M0 and mean time of flight
M1/M0
200
7

M 0 ( x, y)   I ( x, y, t ).dt
0
Algorithm: A novel reconstruction
method was developed to achieve
localization of the inclusion. This
algorithm finds the best position and
concentration of one fluorophore in
the medium according to a Chisquare criterion.

M1 ( x, y)   I ( x, y, t ).t.dt
0
•Discretized volume: We work on a
tetrahedral breast-shaped mesh
obtained by a Delaunay tessellation
Discretized Breast volume
For each voxel of the discretized
medium, we assume the voxel
contains the fluorescent inclusion
that we are looking for. fs laser
Then, we calculate the fluorescence
yield that best matches the signals
and determine the value of the
criterion for the considered voxel.
At last, the set of voxels for which
the criterion is optimal is displayed
Photo of the breast
phantom mold and its
fluorescent inclusion
This phantom was made
by a combination of gelatin
[3] , intrlipid and bovine
hemoglobine.
Concentration
of
the
components were tuned to
fit the absorption and
diffusion coefficient of
breasts
(µa=0.03cm1|µs’=7cm-1)
The cancerous region was
simulated by a capillary
tube containing 0.1 cm3
solution of ICG dye.
lens
HRI
lens
filters
Experimental set-up based on a titanium sapphire laser producing 50 fs pulses at
775 nm. The detector, a time gated camera, is connected to the probe through an
optical fiber network
• Reconstruction: The algorithm had been able to reconstruct the fluorescence yield where
expected with experimental measurements.
Detailed description: This algorithm
will be presented during Scientific
Session 16: Advances in Optical
Imaging ~ September 25, 2009 from
1:00 PM to 2:30 PM.
3D view of the reconstructed fluorescence map based on experimental acquisitions.
Conclusion
•An analysis of the geometrical parameters showed that larger source steps provide better
localization. The quality of the reconstruction is highly dependent on the relative position of the
source and the inclusion.
•An optical time resolved system allowed us to achieve localization of a fluorescent inclusion
inside a breast-mimicking phantom which is representative both on its optical absorption (µa)
and diffusion (µs’) properties. The associated reconstruction algorithm uses the mean intensity
and the time of flight of the photons to get a 3D estimation of the tumor localization.
References
[1]
A.P. Gibson, J.C. Hebden, et S.R. Arridge, “Recent advances in diffuse optical imaging,” Phys. Med. Biol,
vol. 50, 2005, pp. 1–43.
[2]
J. Hebden, H. Veenstra, H. Dehghani, E. Hillman, M. Schweiger, S. Arridge, et D. DELPY, “Threedimensional time-resolved optical tomography of a conical breast phantom,” APPLIED OPTICS, vol. 40, Jul.
2001, pp. 3278-3287.
[3]
A. De Grand, S. Lomnes, D. Lee, M. Pietrzykowski, S. Ohnishi, T. Morgan, A. Gogbashian, R. Laurence,
et J. Frangioni, “Tissue-like phantoms for near-infrared fluorescence imaging system assessment and the training
of surgeons,” JOURNAL OF BIOMEDICAL OPTICS, vol. 11, Fév. 2006.
www.leti.fr
© CEA 2008. Tous droits réservés.
Toute reproduction totale ou partielle sur quelque support que ce soit ou utilisation du
contenu de ce document est interdite sans l’autorisation écrite préalable du CEA
All rights reserved. Any reproduction in whole or in part on any medium or use of the
information contained herein is prohibited without the prior written consent of CEA