Transcript Optical Imaging

DIDO YOVA
LABORATORY OF BIOMEDICAL OPTICS AND APPLIED
BIOPHYSICS
SCHOOL OF ELECTRICAL AND COMPUTERS
ENGINEERING
NATIONAL TECHNICAL UNIVERSITY OF ATHENS
LABORATORY OF BIOMEDICAL OPTICS AND
APPLIED BIOPHYSICS
OPTICAL
IMAGING
CONFOCAL LASER SCANNING MICROSCOPY
IMAGING AT THE CELLULAR LEVEL
TISSUE IMAGING
AFM AND SHG MICROSCOPY
IMAGING OF BIOMOLECULES
TISSUE IMAGING
3D BINOCULAR MACHINE VISION SYSTEM
FLUORESCENCE MOLECULAR IMAGING
Imaging at the Cellular Level
Various imaging technologies are developing to understand
and optimize PDT process.
New developments in microscopy are providing crucial information and
essential approaches for understanding the structure and function of
cells and molecules.
Combined with:
Recent developments in computing
and
Molecular probes
Offer great promise for delivery of vital new information.
IMAGING AT THE CELLULAR LEVEL
IN PDT
The mechanism of tumor destruction by PDT is very complex and is
still under investigation. Photoactivation initiates photochemical
reactions generating highly cytotoxic reactive oxygen species (ROS)
The initial insult is a form of oxidative stress which triggers a variety
of events contributing to the inactivation of cancerous cells.
A very interesting problem is to image the cascade of
events of induced oxidative stress at the cellular level.
Imaging at the Cellular Level
Monitoring early events of cellular response to oxidative stress
We investigated the cascade of early intracellular phenomena
evoked by oxidative stress in real time at the single cell level.
Oxidative stress was induced by photosensitization of ZnPc in
Human Fibroblasts using the 647 nm laser line, using a dose
that did not lead to apoptosis or necrosis.
By :
• Confocal Laser Scanning Microscopy
• Vital Fluorescent Probes
• Photosensitive Molecules
• Advanced Image Analysis and Processing
Fibroblasts coincubated with ZnPc
and MitoTracker Green.Fluorescence
image of ZnPc λexc:647nm, λem:680
nm
Fibroblast incubated with
MitoTracker Green λexc:488nm,
λem:522 nm
Merged image of the red and green
fluorescence. By advanced
colocalization algorithm, ZnPc is
above 85% localized in the
mitochondria.
Detection of intracellular ROS (Reactive Oxygen Species)
generated by ZnPc photosensitization using H2DCFDA.
Fibroblasts incubated with ZnPc
+ H2DCFDA (after oxidation by
ROS produces DCF)
Pseudocolored image
Mitochondrial membrane potential (ΔΨm) decrease after
ZnPc photosensitization + JC-1.
0 min
30 s
3 min
8 min
1 min
15 min
Resting ΔΨm =140 mV
 ΔΨm =90  5 mV after oxidative stress
Intracellular pH changes after ZnPc photosensitization
using the membrane permeable (BCECF-AM) probe .
0 min
30 s
2 min
3 min
5 min
10 min
Resting pHi = 7.45  0.03
ΔpHi=0.40  0.08 after oxidative stress
Spatiotemporal global Ca2+ oscillations evoked by ZnPc
photosensitization monitored by Fluo-3 (pseudocolored images).
0 min
30 s
4 min
30 s
1 min
1 min
2 min
2 min
Time course experiment of intracellular Ca2+concentration.
Resting [Ca2+ ] 60nM
[Ca2+ ] 0.25μM after oxidative stress
TISSUE OPTICAL IMAGING
Development of animal models.
Research related to small animals optical imaging
NMSC ANIMAL MODEL
Non-melanoma carcinomas in SKH-1 mice
Typical series of confocal images obtained horizontally, at 0, 20, 40 and 60 μm
from skin surface, of a healthy hairless mouse 1 hour after topical application of
AlClPc. Images were acquired with excitation at 647 nm and emission at 680 nm.
Penetration
depth 1490 μm
Confocal image obtained from a cross-section of a non-melanoma skin
carcinoma topically applied with AlClPc for 1 hour. Images were acquired with
excitation at 647 nm and emission at 680nm. The yellow line indicates the
penetration depth. Scale bar: 100 μm.
PDT in DERMATOLOGY
OPTICAL IMAGING MONITORING
Answers to be given:
 Accurately imaging tumors smaller than 1 cm.
 As PDT is a repeatable technique to monitor tumour
shrinkage, after each PDT treatment, will facilitate
the optimization of therapy.
3-D Binocular Machine Vision System for Gauging
Small Tumors
3D Binocular Machine Vision System for Gauging
Small Tumors
Animal model for NMSC
Normal
area
Tumour
area
3-D Binocular Machine Vision System for Gauging Small
Tumors
3-D Binocular Machine Vision System for Gauging Small
Tumors
 Successful reconstruction and gauging of tumours
smaller than 1 cm maximum diameter via a fully
automated software package.
 Surface rendering and gauging tool for skin tumours
imaging and following of their shrinkage after PDT
treatment.
 Prospects of other medical applications like in burn
depth estimation, by introducing an articulated arm.
 Useful in a variety of other 3-D gauging applications like
in archeology.
FLUORESCENCE MOLECULAR IMAGING
in PDT
Non-invasive monitoring of molecular targets is
particularly relevant to photodynamic therapy
(PDT), including the delivery of photosensitizer in
the treatment site and monitoring of molecular and
physiological changes following treatment.
WHAT ABOUT DEEP SEATED TUMORS?
PROSTATE CANCER ANIMAL MODEL
Palpable tumors appear 2 weeks after
inoculation.
Once they are formed, they grow
rapidly.
Tumors reach the appropriate size
(thickness 4 – 6 mm) approximately
3 – 5 weeks after the inoculation.
Animals survive up to 100 days after
injection.
Tumors 9 weeks post inoculation
FLUORESCENCE MOLECULAR IMAGING
One of the most challenging problems in medical imaging is to
see a tumour embedded in tissue which is a diffusive medium.
 Light in the range of ~650 nm – ~950 nm can penetrate up to
several centimeters into tissue because of the low photon
absorption in this region of the spectrum, enabling imaging at
greater depths.
 Tissue autofluorescence is very low in this spectral region as
well.
However, these photons are highly scattered into tissue and
become diffuse.
FLUORESCENCE MOLECULAR IMAGING
Progress has been enabled by:
 The development of new probes that emit at the near
IR region and they have increased photostability and
selectivity.
 Development of new imaging modalities.
Fluorescence Molecular Imaging
PROSTATE CANCER
In our Laboratory we use:
 Fluorescence probes for labeling prostate tumours at:
λexc = 680nm
λem = 700nm
 Free-space, non-contact geometry for excitation (red diode
laser) and detection of light
 Direction of excitation and detection from the same side of
the tissue
Inverse Problem
Forward problem: image x
Inverse problem: data y
data y.
image x.
The inverse problem is ill-posed because the solution
is non-unique and does not depend continuously on
the data.
FORWARD SOLVER
‣ Discretization scheme
๏ Use of the Delaunay Triangulation
Method.
๏ Construction of Triangulation Matrix.
‣ Fluorophore distribution mapping
๏ Use of the Super-Ellipsoid Models.
๏ Mapping of the absorption coefficient
based on interior/exterior position
determination relative to the SuperEllipsoid surface.
‣ Finite Elements
๏ Application of the Galerking Finite
Element Method.
๏ Definition of the Spatial and Angular
distribution basis functions.
INVERSE SOLVER
‣ Data fitting process
๏Intensity adjustment.
๏Simulated and acquired
image coordinates
correlation.
๏Feature extraction.
๏Image registration.
‣ Image fine-tuning process
๏Least squares method.
๏Levenberg-Marquardt
optimization.
๏Database update.
DA
RTE
Coupled RTE-DA
•3072 elements
•8 sec
•3072 elements
•8 directions
•1.5 h
•3072 elements
•8 directions
•24 min
This configuration was chosen to match the corresponding properties of Indocyanine Green (ICG) dye. The
absorption and isotropic scattering properties of 1% Liposyn solution were chosen to mimic the background of
the phantom. The excitation source had been simulated as a point source (Dirac function).
Fluorescence Molecular Imaging
The three figures represent the photon density
magnitude of the excitation light (top row, marked as
a) and the emission light (bottom row, marked as b)
at the y = 0 plane. The outcomes are from 3D
experiments. The least squares relative residual was
in the order of 10-14 for both DA and RTE and in the
order of 10-13 for the coupled model.
Inverse Problem Solution
The data fitting procedure provides the
initial fluorophore distribution.
Input
Intensity adjustment
Denoising
Segmentation
THE SYSTEM