Embed Size (px)
Citation preview
1
MULTI-SCALE PHOTOACOUSTIC TOMOGRAPHY AND ITS APPLICATIONS TO CANCER RESEARCH
By
LEI XI
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2012
2
© 2012 Lei Xi
3
To my wife, Hui Jin; My mom and my dad;
And all who have been supportive to me in my life
4
ACKNOWLEDGMENTS
Firstly, I would like to thank Dr. Huabei Jiang for his support , guidance and
encouragement for me during all my years as a graduate student. Without his full
financial support, constructive advice and discussions, I would not have the opportunity
to make progress in my projects.
Secondly, I wish to thank my committee members, Dr. Huikai Xie, Dr. Sihong Song
and Dr. Rosalind Sadleir. Their comments and suggestions were very helpful for my
research. I also appreciate their valuable time spent on reading my proposal and
dissertation.
Thirdly, thanks for help from Dr. Stephen Grobmyer, an associate professor on
Department of Surgery, on guidance and suggestions to our intraoperative
photoacoustic tomography system. As our research partner, Dr. Huikai Xie provided
indispensable MEMS devices for our imaging systems. Additionally, I would like to thank
Dr. Lily Yang providing nanoparticles for our photoacoustic molecular study and Dr. Jun
Cai's valuable collaboration on anti-angiogenesis medicine study using photoacoustic
microscopy technique.
Last but not the least, I would like to take the opportunity to express my thanks to
Qizhi Zhang, Yao Sun, Zhen Yuan, Lei Yao, Ruixin Jiang, Hao Yang, Xiaoqi Li, Lijun Ji,
Bin He, Bo Wang, Jingjing Sun, Can Duan, Wenjun Liao and all members in our lab and
our collaboration lab for their assistance, suggestions and the great time we spent
together.
5
TABLE OF CONTENTS page
ACKNOWLEDGMENTS .................................................................................................. 4
LIST OF TABLES ............................................................................................................ 8
LIST OF FIGURES .......................................................................................................... 9
LIST OF ABBREVIATIONS ........................................................................................... 12
ABSTRACT ................................................................................................................... 16
CHAPTER
1 INTRODUCTION .................................................................................................... 18
1.1 Background and Motivations ............................................................................. 18
1.2 Overview of Photoacoustic Tomography for Cancer Research ........................ 19
2 OPTICAL-RESOLUTION PHOTOACOUSTIC MICROSCOPY AND ITS APPLICATION ........................................................................................................ 23
2.1 Optical-resolution Photoacoustic Microscopy ................................................... 23
2.2 Materials and Methods ...................................................................................... 23
2.3 In Vivo Animal Experiments Demonstration ...................................................... 24
2.4 Miniature Hybrid Probe Combining ORPAM and OCT for Endooscopic Vascular Visualization.......................................................................................... 24
2.4.1 Motivations .............................................................................................. 24
2.4.2 Materials and Methods ............................................................................ 25
2.4.3 Results and Discussion ........................................................................... 27
3 ACOUSTIC-RESOLUTION PHOTOACOUSTIC MICROSCOPY ........................... 33
3.1 Motivations ........................................................................................................ 33
3.2 Materials and Methods ...................................................................................... 33
3.3 Quantitative ARPAM ......................................................................................... 34
3.4 Phantom Validation ........................................................................................... 37
3.5 In Vivo Animal Experiments Validation ............................................................. 38
4 APPLICATIONS OF ACOUSTIC-RESOLUTION PHOTOACOUSTIC MICROSCOPY TO PRECLINICAL CANCER RESEARCH .................................... 44
4.1 Photoacoustic Imaging of Tumor Vasculature Development for Breast Cancer Research ................................................................................................. 44
4.1.1 Motivations .............................................................................................. 44
4.1.2 Materials and Methods ............................................................................ 47
6
4.1.2.1 Construction of scAAV2 vector for delivering siRNAs against the UPR proteins .......................................................................................... 47
4.1.2.2 scAAV2 infection of cells in vitro .................................................... 48
4.1.2.3 Microscopy and fluorescence activated cell sorting (FACS) analysis ................................................................................................... 48
4.1.2.4 Cell culture ..................................................................................... 48
4.1.2.5 In vitro angiogenesis assay ............................................................ 50
4.1.2.6 Apoptosis assay ............................................................................. 50
4.1.2.7 Proliferation assay ......................................................................... 51
4.1.2.8 Mice breast cancer xenograft models ............................................ 51
4.1.2.9 Photoascoustic (PA) imaging systems and noninvasive monitoring in vivo tumor angiogenesis .................................................... 52
4.1.2.10 Statistics analysis ......................................................................... 52
4.1.3 Results .................................................................................................... 53
4.1.3.1 scAAV2 sept mut exhibits higher transduction efficiency in MMECs ................................................................................................... 53
4.1.3.2 Septuplet tyrosine-mutations improve the transduction efficiency of scAAV2-mediated siRNAs infection in MMECs .................................. 54
4.1.3.3 siRNA knockdown of the UPR proteins decreased NeuT EMTCL2-induced in vitro angiogenic activity of endothelial cells ............ 54
4.1.3.4 siRNA knockdown of the UPR proteins down-regulated the survival of endothelial cells ..................................................................... 55
4.1.3.5 Different malignant mice breast cancer cells induced differential agngiogenic responses in breast cancer xenograft models were monitored by serial photoacoustic (PA) imaging ..................................... 56
4.1.3.6 Knockdown of the UPR proteins significantly inhibited Neut EMTCL2-induced in vivo angiogenesis ................................................... 57
4.1.4 Discussion ............................................................................................... 57
4.2 uPAR Targeted Magnetic Iron Oxide as a Contrast Agent for In Vivo Molecular Photoacoustic Tomography of Breast Cancer ..................................... 62
4.2.1 Motivations .............................................................................................. 62
4.2.2 Material and Methods .............................................................................. 64
4.2.2.1 Cell line .......................................................................................... 64
4.2.2.2 Preparation of NIR830-ATF-IONP ................................................. 64
4.2.2.3 Animal preparation ......................................................................... 65
4.2.2.4 Photoacoustic microscopy imaging system .................................... 65
4.2.2.5 Near-infrared planar fluorescence imaging system ........................ 66
4.2.2.6 Image processing ........................................................................... 67
4.2.2.7 Histologic analysis.......................................................................... 67
4.2.2.8 Statistical analysis .......................................................................... 67
4.2.3 Results .................................................................................................... 67
4.2.4 Discussion and Summary ........................................................................ 69
4.3 Photoacoustic and Fluorescence Tomography of HER-2/Neu Positive Ovarian Cancers Using Receptor-Targeted Nanoprobes In An Orthotopic Human Ovarian Cancer Xenograft Model ............................................................ 71
4.3.1 Motivations .............................................................................................. 71
7
4.3.2 Materials and Methods ............................................................................ 73
4.3.2.1 Cell lines ........................................................................................ 73
4.3.2.2 HER-2/neu specific affibody conjugation to IONP .......................... 73
4.3.2.3 Orthotopic human ovarian cancer xenograft model ....................... 73
4.3.2.4 Fluorescence molecular tomography imaging System. .................. 74
4.3.2.5 In vivo and ex vivo planar near infrared fluorescence imaging ....... 74
4.3.2.6 Histological analysis ....................................................................... 75
4.3.2.7 Statistical analysis .......................................................................... 75
4.3.3 Results .................................................................................................... 75
4.3.4 Discussion ............................................................................................... 77
5 INTRAOPERATIVE PHOTOACOUSTIC TOMOGRAPHY ...................................... 96
5.1 Motivations ........................................................................................................ 96
5.2 Materials and Methods ...................................................................................... 97
5.2.1 Animal Protocol ....................................................................................... 97
5.2.2 Intraoperative Photoacoustic Tomography System ................................. 97
5.2.3 Imaging Procedure .................................................................................. 99
5.2.4 Histology .................................................................................................. 99
5.3 Spatial Resolution of the System .................................................................... 100
5.4 Phantom Experiments ..................................................................................... 100
5.5 Human Blood Vessels Experiments ................................................................ 101
5.6 Preclinical Evaluation of Intraoperative Photoacoustic Tomography .............. 101
5.7 Discussion ...................................................................................................... 103
6 CIRCULAR ARRAY-BASED PHOTOACOUSTIC TOMOGRAPHY AND ITS APPLICATION TO BREAST CANCER DETECTION ........................................... 117
6.1 Motivations ...................................................................................................... 117
6.2 Materials and Methods .................................................................................... 118
6.2.1 System Description ................................................................................ 118
6.2.2 Quantitative Reconstruction Methods of PAT and DOT ........................ 120
6.3 Performance Evaluation and Phantom Experiments....................................... 122
6.3.1 Performance of PAT .............................................................................. 123
6.3.2 Phantom Experiments ........................................................................... 124
6.4 Ex Vivo Experiment......................................................................................... 126
6.5 Multilayer Ultrasound Transducer with Improved Sensitivity and Bandwidth .. 126
6.6 Discussion and Future Directions. .................................................................. 130
7 CONCLUSION AND FUTURE WORK .................................................................. 146
LIST OF REFERENCES ............................................................................................. 147
BIOGRAPHICAL SKETCH .......................................................................................... 163
8
LIST OF TABLES
Table page 3-1 Comparison of exact and reconstructed values (absorption coefficient and
size) in the target area for all 6 experimental cases ........................................... 43
6-1 Exact size (diameter), location (off center), depth, and absorption and reduced scattering coefficients of the five targets ............................................. 144
6-2 Exact and reconstructed target size (diameter), location (off center) and absorption and reduced scattering coefficients for the phantom experiments .. 145
9
LIST OF FIGURES Figure page 2-1 Schematic of ORPAM system ............................................................................ 30
2-2 MAP photoacoustic images of mice ear ............................................................. 30
2-3 Schematic and photograph of the hybrid probe .................................................. 31
2-4 Schematic of the integrated ORPAM and OCT system ...................................... 31
2-5 Resolution test for ORPAM and OCT ................................................................. 32
2-6 In vivo imaging of mouse ear by the integrated probe ........................................ 32
3-1 Our experimental PAM system ........................................................................... 40
3-2 Reconstructed absorption coefficient images (mm-1) from all 6 experimental cases .................................................................................................................. 40
3-3 Reconstructed absorption coefficient profiles ..................................................... 41
3-4 In vivo imaging of the blood vessels in a rat ear ................................................. 42
4-1 Comparative analysis of scAAV2-mediated transduction of MMECs .................. 80
4-2 scAAV2 encoding siRNAs against UPR proteins ................................................ 81
4-3 Pro-angiogenic and survival role of the UPR proteins on endothelial cells ......... 82
4-4 Serial photoacoustic imaging of the developing tumor vasculature and quantitative analysis for mice breast cancer xenograft models .......................... 83
4-5 Knockdown of the siRNAs against the UPR protein resulted in decreased tumor growth and tumor vasculature in mice breast cancer xenografts .............. 84
4-6 Illustration of NIR-830 dye and ATF-conjugated IONP probe ............................. 85
4-7 Schematic of NIR fluorescence imaging system ................................................. 85
4-8 In vivo and in vitro test of targeted nanoprobe using different wavelengths ....... 86
4-9 In vivo photoacoustic MAP and fluorescence images before and after injection .............................................................................................................. 87
4-10 Quantitative plot and comparison of photoacoustic and fluorescent signals ....... 88
4-11 In vivo photoacoustic images with adding chicken breast between detector and tumor ........................................................................................................... 89
10
4-12 Schematic and spectrum of targeted nanoprobe and FMT system description .. 90
4-13 Multi-mode in vivo imaging and histological validation ....................................... 91
4-14 Quantitative analysis based on different wavelengths ........................................ 92
4-15 Image depth ability evaluation ............................................................................ 93
4-16 3D results for both PAMT and FMT .................................................................... 94
4-17 Quantitative comparison of spatial resolution between photoacoustic and fluorescence molecular tomography images with increased imaging depth ....... 95
5-1 System description of the experimental platform .............................................. 106
5-2 Schematic representation of MEMS-based photoacoustic imaging system ..... 107
5-3 Schematic and performance of a ring-shaped ultrasound transducer .............. 108
5-4 Schematic of the miniaturized probe ................................................................ 109
5-5 Performance evaluation of the system ............................................................. 110
5-6 Phantom experiment with single target ............................................................. 111
5-7 Diagram of the phantom and imaging result with multiple targets .................... 111
5-8 Ex vivo experiment with single target ............................................................... 112
5-9 In vivo experimental evaluation using human hand .......................................... 112
5-10 In vivo three-dimensional (3D) tumor mapping in a mouse model .................... 113
5-11 Quantitative analysis of the photoacoustic slices and H&E stained sections .... 114
5-12 Quantitative analysis of photoacoustic images in correlation with H&E sections with increasing image depth ............................................................... 115
5-13 Design of future imaging probe......................................................................... 116
6-1 System description ........................................................................................... 132
6-2 Performance evaluation of the hybrid system ................................................... 133
6-3 System resolution evaluation ............................................................................ 134
6-4 System performance evaluation by phantom experiments ............................... 135
6-5 Quantitative PAT and DOT images of targets................................................... 136
11
6-6 Reconstructed optical properties of ex vivo tumor ........................................... 140
6-7 Schematic of the multi-layered transducer and performance evaluation systems ............................................................................................................ 141
6-8 Calibration of multi-layered transducer ............................................................. 141
6-9 Frequency response of multilayered and single-layered PVDF transducer ...... 142
6-10 In vivo experimental evaluation of multi-layered transducer ............................. 143
6-11 In vivo photoacoutic imaging of mouse head using multilayered and single-layered transducer ............................................................................................ 144
12
LIST OF ABBREVIATIONS
1D One-dimensional
2D Two-dimensional
3D Three-dimensional
AAV Adeno-associated Virus
ANOVA Analysis of Variance
ARPAM Acoustic-resolution Photoacoustic Microscopy
ATF Amino-terminal Fragments
ATF6 Activating Transcription Factor 6
ATF-IONP Amino-terminal fragments of uPA conjugated to iron oxide nanoparticles
BCT Barker Coded Transducer
BSA Bovine Serum Albumin
bZIP Basic Leucine Zipper
CBA CMV-chicken β-actin
CO2 Carbon Dioxide
CT Computed Tomography
CW Continuous Wave
DAQ Data Acquisition
DOT Diffuse Optical Tomography
DRS Diffuse Reflectance Spectroscopy
EDAC Ethyl-3-dimethyl Amino Propyl Carbodiimide
EMS Electromagnetic Shield
ER Endoplasmic Reticulum
ESS Elastic Scattering Spectroscopy
FACS Fluorescence Activated Cell Sorting
13
FCS Fetal Calf Serum
FD Frequency Domain
FE Finite Element
FEM Finite Element Method
FMT Fluorescent Molecular Tomography
FT Folded Transducer
FWHM Full Width Half Maximum
G1 Generation one
G2 Generation two
GFP Green Fluorescent Protein
GRIN Graded-Index
H&E Hematoxylin and Eosin
Hb Hemoglobin
His6-ZHER2:342-Cys Histidine-tagged HER-2-specific Affibody
IACUC Institutional Animal Care and Use Committee
ICG Indocyanin Green
IONP Magnetic Iron Oxide Nanoparticles
iPAI Intraoperative Photoacoustic Imaging
iPAT Intraoperative Photoacoustic Tomography
IRE1α Inositol-requiring Protein 1 α
IUPUI Indiana University Purdue University at Indianapolis
LOIS Laser-based Optoacoustic Imaging System
MAP Maximum Amplitude Projection
MEMS Microelectromechanical System
MMECs Mice Microvascular Endothelial Cells
14
MOI Multiplicity of Infection
MRI Magnetic Resonance Imaging
NIR Near-infrared
NIR-830 Near-infrared 830
NIR830-ATF-IONP Near-infrared Dye Amino-terminal Fragments Iron Oxide Nanoparticles
NIR830-ZHER2:342-IONP Near-infrared Dye HER-2/neu-specific Affibody IONP
NTA Nitrilotri-acetic Acid
OCT Optical Coherence Tomography
ORPAM Optical-resolution Photoacoustic Microscopy
PAM Photoacoustic Microscopy
PAT Photoacoustic Tomography
PCR Polymerase chain reaction
PERK PKR-like ER Kinase
PET Positron Emission Tomography
PET Positron Emission Tomography
PMN-PT Pb(Mg1/3Nb2/3)O3–PbTiO3
PO Prophylactic Oophorectomy
PVDF Polyvinylidene fluoride
PZT Lead zirconate titanate
RBC Red Blood Cell
RSOD Rapid Scanning Optical Delay
RTE Radiative Transfer Equation
SBCT Switchable Barker Coded Transducer
scAAV Self-complementary Adeno-associated Virus
scAAV2 Self-complementary Adeno-associated Virus Serotype 2
15
SEM Standard Error of the Mean
SNR Signal to Noise Ratio
SPECT Single Photon Emission Computed Tomography
sulfoNHS Sulfo-N-Hydroxysuccinimide
TAT Thermoacoustic Tomography
TD Time Domain
uPAR Urokinase plasminogen receptor
UPR Unfolded Protein Response
UV Ultraviolet
VEGF Vascular Endothelial Growth Factor
XBP-1 X-box-binding Protein 1
Y-F Tyrosine-to-phenylalanine
16
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
MULTI-SCALE PHOTOACOUSTIC TOMOGRAPHY AND ITS APPLICATIONS TO
CANCER RESEARCH
By
Lei Xi
December 2012
Chair: Huabei Jiang Major: Biomedical Engineering
Photoacoustic tomography (PAT) is an emerging non-ionizing, non-invasive
biomedical imaging modality combining high optical contrast with high ultrasound
resolution. It offers a unique ability of multi-scale imaging from microscopic, mesoscopic,
to macroscopic for biological tissues.
This dissertation intends to develop a spectrum of strategies that aim to explore
and demonstrate such multi-scale imaging ability of PAT and its applications to cancer
research. The first part of my dissertation discusses the development of optical-
resolution photoacoustic microscopy (ORPAM) owning high-resolution and high-
sensitivity in vivo imaging. Here, we first describe the conventional ORPAM system
including the principles, system design, experimental procedures and a simple
application of visualizing microvascular structure in a mouse ear. Subsequently, a
miniature hybrid probe combing ORPAM with optical coherence tomography (OCT) is
proposed and demonstrated with in vivo animal experiments. This hybrid ORPAM/OCT
technique shows the potential for endoscopic cancer research. The second part of the
dissertation focuses on developing acoustic-resolution photoacoustic microscopy
(ARPAM) that breaks the limitation of imaging depth associated with ORPAM. We
17
demonstrate the applications of ARPAM to anti-angiogenesis medicine study of breast
cancer and to molecular imaging of breast and ovarian cancers with targeted
nanoprobes. The third part of my dissertation is concerned with intraoperative
photoacoustic tomography (iPAT) miniature probe for image-guided surgery of breast
cancer. The probe is evaluated with phantom and in vivo experiments in an animal
model. The fourth part of my dissertation focuses on integrating circular array-based
photoacoustic tomography system with diffuse optical tomography system. Several sets
of phantom experiments with embedded tumor mimicking targets and ex vivo tumor are
used to evaluate the performance of this hybrid system. In the last part of the
dissertation, conclusions are made and future directions are discussed.
18
CHAPTER 1 INTRODUCTION
1.1 Background and Motivations
Photoacoustic tomography (PAT) referred to as thermoacoustic tomography (TAT)
is a promising biomedical imaging technique based on light-generated acoustic wave
over last decade. It combines the high contrast of pure optics imaging modalities with
the high ratio of imaging depth to spatial resolution of ultrasound. The research history
of photoacoustic phenomenon study is relatively long. In 1880, Alexander Graham Bell
observed the generated acoustic wave in solid due to absorption of rapid interrupted
sunlight.1-2 The same effect was observed in gases and liquids by other researchers in
their following experiments. The invention of lasers in the 1970s revived the applications
of photoacoustic effect.3 Until the mid-1990s, researchers began to employ
photoacoustic effect for biomedical imaging.4-8 From the beginning of 2000s, this field
had a rapid growth in terms of the development of ultrasound transducer9 and imaging
reconstruction algorithms10-13, realization of functional imaging with multi-wavelength
excitation source 14-15 and molecular imaging with molecular probes16-17 as well as the in
vivo clinical tries for human breast cancer detection.18-22
During photoacoustic experiments, short pulsed laser is commonly used to
generate ultrasound waves. For photoacoustic, optical wavelengths from visible to near-
infrared (NIR) are used, where NIR spectrum range (700-900nm) lies in optical
transparent window providing the greatest penetration depth.23-24 Owning to the use of
visible or NIR light, there is no radiation issues compared with conventional X-ray
imaging techniques.25-26 From the fundamental of photoacoustic, absorption of light by
specific tissue absorbers such as hemoglobin, melanin, water et al. results in a rapid
19
temperature rise that leads to an initial pressure which subsequently releases
generating the emission of acoustic signals. Hence, PA imaging is very sensitive to
absorption from intrinsic contrast for optics which provide the potential for excavating
the functional parameters such as concentration of hemoglobin, oxygen saturation,
blood flow et al..27-35 In addition, exogenous absorbers such as gold nanoparitcles, iron
oxide, indocyanin green (ICG) et. al. can provide external contrast enhancement for
photoacoustic imaging providing the opportunities to recover molecular information.36-41
The motivation drives us to investigate photoacoustic imaging due to some
challenges for pure optics imaging modalities. There are two main challenges,
diffraction and diffusion. For diffraction, it limits the spatial resolution of ballistic imaging
techniques such as confocal microscopy, two-photon microscopy and optical coherence
tomography (OCT).42-43 This has been overcome by super-resolution techniques.44
Diffusion limits the penetration of ballistic imaging techniques to be 1mm in human
tissue due to its high-scattering properties. Photoacoustic imaging has the potential to
overcome this limitation. In photoacoustic imaging, the photons diffuse inside the
scattering tissue too, however, all the photons arriving at the targets are useful photons
which will be absorbed by the targets. Then most part of the absorbed energy is
converted to acoustic signals. Additionally, the scattering of acoustic signal inside tissue
is 100 times weaker than that of light.
1.2 Overview of Photoacoustic Tomography for Cancer Research
There are two main types of photoacoustic imaging modalities including
photoacoustic microscopy and reconstruction based photoacoustic tomography.
For photoacoustic microscopy, the photoacoustic imaging is obtained through
mechanically scanning of a focused transducer or a focused laser beam. The acquired
20
A-lines are used to form an image without any reconstruction algorithms. If a focused
laser beam is used, it is called optical-resolution photoacoustic microscopy (ORPAM).45
If using a focused transducer without light beam focusing, it is termed acoustic-
resolution microscopy (ARPAM).46 Typically, the resolutions for ORPAM is from sub-
micro to several micrometers while the imaging depth is limited to be less than 1mm.
For ARPAM, the resolutions is various determined by the transducer's aperture,
bandwidth and focal length, usually from 30 micrometers to 1mm, meanwhile the depth
is from several millimeters to centimeters.47
Currently, ORPAM is the rapidest developed photoacoustic technique among all
photoacoustic modalities.48-55 However, the application of ORPAM to cancer research is
limited due to the shallow penetration depth. The most common application is to
visualize the neovascularization inside mice or rat ears to investigate some anti-
angiogenetic factors.54-55 This is very useful for study of some anti-angiogenesis
medicine during cancer treatment research. Another application of cancer-related
research is employing ORPAM to analyze histology sections which is still under
investigation.
The applications of ARPAM are much wider than those of ORPAM.56-62 If a high
frequency focused transducer is used, ARPAM can be used to visualize the
neovascularization not only inside mouse/rat ear but also around the tumor in animal
model because the imaging depth is deeper (up to 3mm if 50MHz transducer is used)
compared with ORPAM. If a low frequency focused transducer is used, this technique
can be used to imaging large targets such as lymph node, brain tumor, breast tumor
located deep in tissues. Some groups also reported their studies to employ ARPAM to
21
image various targets with contrast agent such as ICG, blue dye, gold nanorod, iron
oxide et al..
Reconstruction based photoacoustic tomography is regarded as the traditional
mode of photoacoustic imaging. It is also the most commonly used and least restrictive
photoacoustic imaging technique. In PAT, a large diameter pulsed laser beam
illuminates the full imaging field and NIR wavelengths enable the deep penetration.
Various methods for reconstructing the PAT image from the detected signals have been
developed such as back-projection, Fourier transform, P-transform, k-wave method and
finite element method (FEM) based reconstruction algorithm.63-65 Among all these
reconstruction methods, FEM can recover accurate quantitative acoustic and optical
properties. The first and most important clinical application of PAT is breast cancer
detection. Also the detection and analysis of lymph node by reflection-mode PAT has
the great potential to be used in clinical. Meanwhile, some researchers also used three-
dimensional (3D) PAT to monitor angiogenesis inside human breast. For endoscopic
area, several endoscopic probes have been reported for ovarian cancer and colon
cancer detection.66-68
In this study, we developed several photoacoustic imaging systems covering from
macro-scale to micro-scale area for various cancer applications. In Chapter 2, we will
introduce ORPAM and its application for endoscopic cancer research integrated with
OCT. In Chapters 3 and 4, the ARPAM system and its applications on anti-angiogenesis
medicine studies on breast tumor, molecular imaging of breast cancer and ovarian
cancer using targeted contrast agent. Quantitative ARPAM will be described in detail
too. In Chapter 5, we will move to pre-clinical evaluation of intraoperative photoacoustic
22
tomography (iPAT) during breast cancer surgery on an animal model. In Chapter 6, we
will introduce newest results of combining circular array-based PAT integrated with
diffuse optical tomography (DOT) for breast cancer detection. Finally, in Chapter 7, the
summary of my dissertation and future directions are presented.
23
CHAPTER 2 OPTICAL-RESOLUTION PHOTOACOUSTIC MICROSCOPY AND ITS APPLICATION
2.1 Optical-resolution Photoacoustic Microscopy
ORPAM uses optical confinement for localization purpose and is similar to many
conventional optical microscopy techniques where the lateral resolution is determined
by the dimensions of a focused light spot. Due to high optical scattering in most tissues,
the imaging depth is limited to be less than 1mm. However, it has the highest lateral
resolution of 5µm or even better compared with that of ARPAM and circular scanning
photoacoustic tomography. As a powerful optical absorption based microscopy
technique and a valuable complement to the existing microscopy techniques, ORPAM
has been demonstrated to be used in wide biomedical areas.
In this chapter, we will introduce the ORPAM system and show the potential
application for cancer research using ORPAM.
2.2 Materials and Methods
As shown in Figure 2-1, the ORPAM system employs optical focused laser beam
from a Nd:YAG pulsed laser to achieve micro level lateral resolution. Laser pulses with
duration of 6ns and repetition rate of 10Hz are spatially filtered by a 50µm diameter
pinhole. Then the laser beam is collimated by a convex lens and focused by an
objective lens to obtain a focal spot with a diameter of 5µm in lateral. The laser energy
on the tissue surface is 25mJ/cm2 which is still under the reported skin damage
threshold.69 A 50MHz focused transducer with 3mm activate area and 6mm focal length
is confocalled with the focused light beam. Combination of depth-resolved
photoacoustic waves with a 2D raster scanning along the x-y plane generates the
volumetric images.
24
2.3 In Vivo Animal Experiments Demonstration
To demonstrate the imaging abilities of our ORPAM system, we employed the
ORPAM to visualize the microvasculature in an ear of a nude mouse (body weight: 25g)
in vivo with excitation source at 532nm wavelength. All animal procedures were
conducted in conformity with the laboratory animal protocol approved by Animal Studies
Committee of University of Florida. Before experiments, the mice were anesthetized
with a mixture of Ketamine (85mg/kg) and Xylazine. An imaging area covering 1×1mm2
was scanned with a step size of 6µm without any signal average.
In Figure 2-2A, a typical microvasculature of mice ear is shown. From the result,
we can easily identify even the single capillary with size of less than 5µm. As indicated,
red blood cell (RBC) can be seen among these capillaries too. In Figure 2-2B, the top
panel shows the result of ORPAM and bottom panel shows the photograph taken with a
transmission optical microscope at a 4×magnification. These two blood vessels are
clearly identified in ORPAM; however they cannot be seen by a transmission optical
microscope.
2.4 Miniature Hybrid Probe Combining ORPAM and OCT for Endooscopic Vascular Visualization
2.4.1 Motivation
ORPAM is fundamentally sensitive to optical absorbers inside the tissue such as
hemoglobin and melanin. Hence, ORPAM is well suited for imaging blood vessels.
However, it is hard for ORPAM to image tissues with low absorption contrast.
Recently, studies from several research groups have shown the potential to
combine photoacoustic imaging with other modalities such as space fluorescent
microscopy32, OCT70 and ultrasound66. These combinations enable the hybrid system to
25
provide more information than each single modality by itself. The major drawback of
these combined systems is that the system is bulky and thus cannot be employed in
endoscopic or intravascular visualization. Several groups have proposed single-
modality or multi-modality miniature probes for endoscopic or intravascular purposes.
Yang et al. reported an integrated photoacoustic and ultrasound endoscopic probe with
outer diameter of 3.5mm. In this probe, they used a focused ultrasound transducer with
a hole in the center for optical illumination and a micromotor to achieve internal
scanning.67 Wang et al, showed the possibility of combining ultrasound with
photoacoustic imaging in an external rotation based intravascular probe. Due to the
limited resolving abilities of ultrasound, conventional PAT and ARPAM, it is hard for
these probes to obtain as high as optical resolution.65 In another study, Yang et al.
reported an endoscopic probe with a size of 5 mm in diameter for ovarian cancer
detection where three separated miniature probes for PAT, OCT and ultrasound,
respectively, were used, making the co-register of the images form the three modalities
difficult.68
To overcome the aforementioned limitations, we propose a miniature GRIN lens
based probe (2.3mm in diameter) integrating ORPAM with OCT. The common optical
path of both the ORPAM and OCT is built using a single-mode fiber, a miniature GRIN
lens and two microprisms, enabling these two modalities to scan the same tissue area.
The self-focusing ability of the GRIN lens results in a highly focused light beam for OCT
and ORPAM. As a result, both OCT and ORPAM yield a high lateral resolution of 15μm.
2.4.2 Materials and Methods
Figure 2-3A shows the schematic of the probe. In this probe, both illumination
beams for OCT and ORPAM are coupled into one tip of a single mode fiber (SMF-28e+,
26
Thorlabs) with 0.14 NA and 0.9mm outer diameter. The other tip of the fiber is cut with
an 8° angle to minimize back-reflection. Optical UV glue is used to connect the fiber tip
with the GRIN lens resulting in a 5mm working distance. Two microprisms (one without
coating and the other coated with aluminum) are glued together and a small unfocused
ultrasound transducer with 10MHz central frequency and 2mm aperture is mounted on
the top of the cubic prism group. The light beams are focused by the GRIN lens and
reflected by the thin aluminum film to the tissue surface. The generated ultrasound
waves transmit through the cubic prism group and are detected by the transducer. The
back-scattering photons from the tissue are reflected by the aluminum film and coupled
into the same fiber through GRIN lens. Figure 2-3B shows a photograph of the probe,
where both the cubic prism group and GRIN lens have 0.7mm in diameter. A stainless
steel tube with 1.0mm in diameter is used to protect the light path (i.e., the single mode
fiber, GRIN lens and cubic prism group). The probe is glued to the transducer and
protected by another bigger stainless steel tube (2.3mm in diameter).
The probe is mounted on a 3D linear stage shown in Figure 2-4. The light beam
for ORPAM is generated from a pulsed Nd:YAG laser with 532nm wavelength and a
10Hz repetition rate. Two neutral density filters are used to attenuate the light energy
and a small iris is utilized to select the homogeneous part of the light beam. The shaped
light beam is focused by a convex lens, which passes through a 50μm pinhole for
spatial filtering and is then coupled into a 2×1 beam coupler. The detected acoustic
waves are amplified by two wideband amplifiers and then digitized by a data acquisition
board (NI5152, National Instrument) at a sampling rate of 250MS/s.
27
A broadband light source (DenseLight, DL-BX9-CS3159A) with a center
wavelength of 1310nm is employed for OCT. The light source has a full width half
maximum (FWHM) of 75nm, providing an axial resolution of 10μm in air. The broadband
light is split into the reference arm and the sample arm by a beam splitter and coupled
into the same 2×1 beam coupler used for ORPAM. The depth scanning from 0 to
1.6mm at the reference arm is realized by a rapid scanning optical delay line (RSOD)
coupled with a galvanometer scanning at 1kHz. The OCT signal is then detected by a
balanced photodetector, whose output is acquired and stored by a DAQ card. The
sensitivity of the system is measured to be 74dB.
2.4.3 Results and Discussion
The lateral resolution is determined by imaging a selected part of an USAF 1951
resolution test target. Figure 2-5B shows the photograph of the resolution test target
and red dashed line shows the part selected for imaging. Figure 2-5A and C show one-
dimensional (1D) profile as marked in Figure 2-5B where the smallest resolvable bar
spacing is 15µm (group 6, element 1). The three bars can be clearly indentified by both
OCT and ORPAM, hence the lateral resolution is better than 15µm. In tissue imaging,
the scattering will reduce the spatial resolution of this probe; however, when the sample
is optically thin, the degradation of lateral resolution is not significant.
To demonstrate the microscopic imaging ability of this dual-mode probe, we chose
to image the ear of a mouse. Before starting the experiments, the hair on the ear was
gently removed using a human hair-removing lotion. The mouse was placed on a
homemade animal holder and was anesthetized with a mixture of Ketamine (85mg/kg)
and Xylazine. After the experiments, the mice were sacrificed using University of Florida
Institutional Animal Care and Use Committee (IACUC)-approved techniques. Strict
28
animal care procedures approved by the University of Florida IACUC and based on
guidelines from the NIH, guide for the Care and Use of Laboratory Animals were
followed.
The laser exposure was 25mJ/cm2 at the optical focus point which is higher than
the ANSI laser safety limit (20mJ/cm2)71, while it is still below the reported skin damage
threshold.69 The scanning step size is 6µm along X-Y plane.
Top rows of Figure 2-6A and B show the maximum amplitude projection (MAP)
images of ORPAM and OCT. OCT and ORPAM visualize different targets of the tissues,
where ORPAM clearly maps the microvasculature and OCT images the sebaceous
gland with high resolution. The bottom rows (cross-section of tissue) of Figure 2-6A and
B show the benefits of combining these two modalities more clearly. The ear's thickness
in the OCT image is from 400µm to 600µm. The dermal structure and the sebaceous
gland are clearly observed. In the cross-sectional OCT, we can identify epidermis,
dermis, and cartilage as indicated with yellow arrows. We have observed that ORPAM
is good at locating micro-vessels in the ear with limited surrounding tissue information.
Figure 2-6C shows the 3D reconstruction of co-registered OCT and ORPAM.
The signal to noise ratio (SNR) of the ORPAM is 25dB, which is lower than that
of conventional ORPAM. There are several reasons contributing to the reduced SNR: 1)
A 10MHz ultrasound transducer was used in this probe, while it is known that the
strongest generated acoustic signal lies between 30MHz to 70MHz; 2) The transducer
is flat resulting in reduced sensitivity compared with a focused transducer commonly
used in conventional ORPAM; 3) The size of the light spot at the focal point was around
20µm which is much larger than that used in conventional ORPAM (2µm to 5µm); and
29
4) During the experiments, we used an 8 bit resolution DAQ card which can only resolve
signals larger than 40mV. As a result, we lost some signals from small capillaries.
These limitations, however, can be easily overcome using a miniature focused high-
frequency transducer (>30MHz), which will in turn improve the axial resolution of
ORPAM and make the whole probe smaller as well. Using a high-resolution DAQ card
will also help solve the aforementioned problems. For future improvements, we also
need a suitable micro-motor to implement an internal scanning mechanism and a faster
laser to reduce the scanning time. Finally, we plan to replace the current slow time-
domain (TD) OCT with fast frequency domain (FD) OCT. These are necessary steps
towards the clinical evaluation of the ORPAM/OCT.
In this paragraph, we will introduce our ORPAM system and it potential application
for endoscopic cancer research combined with OCT. Due to penetration limitation of
highly focused light beam, it is difficult to apply this technique for cancer research if the
tumor lies deep under the surface. Hence, in the next paragraph we conduct a ARPAM
for broad study of cancer.
30
Figure 2-1. Schematic of ORPAM system
Figure 2-2. MAP photoacoustic images of mice ear. A) MAP photoacoustic image of microvascularization of blood vessels of a mouse ear. B) Comparison of ORPAM image (top panel) with conventional transmission microscope image (bottom panel).
31
Figure 2-3. Schematic and photograph of the hybrid probe.
Figure 2-4. Schematic of the integrated ORPAM and OCT system.
32
Figure 2-5. Resolution test for ORPAM and OCT. A) 1D profile of OCT. B) Photograph of USAF 1951 resolution test target. C) 1D profile ORPAM.
Figure 2-6. In vivo imaging of mouse ear by the hybrid probe. A) MAP image (top panel) and cross-section (bottom panel) of ORPAM. B) MAP image (top panel) and cross-section (bottom panel) of OCT. C) 3D rendering of the co-registered ORPAM and OCT images of the mouse ear.
33
CHAPTER 3 ACOUSTIC-RESOLUTION PHOTOACOUSTIC MICROSCOPY
3.1 Motivations
ARPAM employs a single mechanically translated or rotated focused transducer to
map the photoacoustic signals. Commonly, it comprises a focused transducer and
weakly focused light beam where excitation light is delivered through a conical lens in
order to eliminate the affect of tissue surface, termed as dark-field illumination.
However, to our experience, bright-field illumination, which is full field illumination
without a hollow hole, is fine for ARPAM too.
Although the lateral resolution of ARPAM from 50μm to 1mm is not as high as
ORPAM, it can detect target more deeper under surface than that of ORPAM. Hence it
breaks the penetration limitation of ORPAM and has been widely used for breast cancer
detection, lymph node detection and molecular imaging using targeted or non-targeted
contrast agents.
3.2 Materials and Methods
Figure 3-1A shows details of our PAM imaging system. In this system, a short-
pulsed laser beam of 6ns duration at 10Hz repetition rate generated by a 532nm
Nd:YAG pulsed laser was divided into two beams by a beam splitter and coupled into
two fiber bundles. Wideband photoacoustic waves induced as a result of thermoelastic
expansion of target was collected by a focused high-frequency ultrasound transducer
(50MHz) with 3mm aperture and 6mm focal length. The fiber bundles were well
adjusted to optimize the signal-to-noise ratio of the system as shown in Figure 3-1B. A
water tank with a window sealed with an optically and ultrasonically transparent
membrane was used. The imaging probe consisting of acoustic transducer and fiber
34
bundles was mounted on a two-dimensional (2D) moving stage with a scanning step
size of 60μm during experiments.
3.3 Quantitative ARPAM
When a short-pulsed laser light irradiates biological tissues, photoacoustic waves
are induced as a result of transient thermoelastic expansion, which is modeled by the
following Helmholtz-like photoacoustic wave equation72:
t
tJ
Ct
tp
vtp
p
,,1,
2
2
2
0
2 rrr
(3.1)
where p is the pressure wave, 0v is the acoustic velocity, is the absorbed
energy density of the medium and represents the product of optical absorption
coefficient and excitation light distribution, pC is the specific heat, is the thermal
expansion coefficient, and J is the time- and space-dependent light intensity. While
acoustic speed, 0v is practically a spatially varying parameter73-74
, in this work we
assume it a homogeneous constant for simplicity. The first-order absorbing boundary
conditions used are75-78
:
r
p
t
p
vnp
2
1ˆ
0
(3.2)
where n̂ is the unit normal vector.
In our PAM model, photoacoustic signal is first collected at each location of
transducer, resulting in 1D depth (defined as Z direction) resolved image. Thus for the
transducer which located on the surface plane, also named as X-Y plane, at the position
kr kk yx , , Mk ,2,1 , where M is the total number of the transducers, the acoustic
pressure tzyxp kk
c
k ,,, can be obtained by:
35
t
tJ
C
zyx
t
tzyxp
vz
tzyxp
p
kkkkkkkk
,,,,,1,,,2
2
2
0
2
2 (3.3)
The reconstruction approach we used to obtain the map of the absorbed energy
density is an iterative Newton method with combined Marquardt and Tikhonov
regularizations that can provide stable inverse solutions.73,79
The FE discretization of
equation(3.3) and the matrix equation capable of inverse solution can be stated as:
BpMpCKp kkk (3.4)
c
k
o
k
T
kkk
T
k ppΔχI (3.5)
where the elements of matrix K, C, and M are expressed as
02
1zji
l
jiij
rdl
zzk
,
0
0
1zjiij
vc ,
l
jiij dlv
m 2
0
1, respectively, and the
element of vector B can be expressed as
l n
nni
p
i dlt
J
Cb
. is the 1D FE
basis function in Z axis , o
kp and c
kp in equation(3.3) are vectors representing observed
and computed acoustic field data at the kth
transducer location, respectively. kΔχ is the
update vector for the absorbed energy density, k is the Jacobian matrix formed by
kkp at the boundary measurement sites; is the regularization parameter; and I is
the identity matrix. Thus here the image formation task is to update absorbed energy
density distribution zyx kkk ,, for each transducer located at the position kr kk yx , via
iterative solution of equations(3.3) and (3.4) so that an object function composed of a
weighted sum of the squared difference between computed and measured acoustic
data can be minimized.
36
A 3D image of absorbed energy density is produced by scanning the ultrasound
transducer in the X-Y plane and aggregated as:
M
k
kkk zyxzyx1
,,,, (3.6)
The second step is to recover the absorption coefficient a from the absorbed
energy density obtained earlier. We can obtain the distribution of the optical fluence
r through a model-based finite element solution to the light propagation equation. It
is generally believed that light propagation in tissues is best modeled by the radiative
transfer equation (RTE)80-82
:
,,,,
1rrr qd
nSsas , (3.7)
where s is the scattering coefficient; is the radiance; q is the source term;
1 nS
denotes a unit vector in the direction of interest. The kernel,
, is the
scattering phase function describing the probability density that a photon with an initial
direction
will have a direction
after a scattering event. In this study we assume
that the scattering phase function depends only on the angle between the incoming and
outgoing directions; thus
, . Here the commonly used Henyey-
Greenstein scattering function is applied [43]: 2 2(1 ) / 2 (1 2 cos )g g g
where is the angle between
and
, and 11 g . The optical fluence is related
to the radiance by
1,
nSd
rr . Thus the map of absorbed energy density
can be obtained by a .
37
Based on the FE solution to RTE83-84
, we can then determine the a distribution
by the iterative solution procedure.85
In this fitting procedure, s , and g are assumed as
constant, while the incident laser source strength and the absorbed energy density
are estimated in advance.
3.4 Phantom Validation
In the first three phantom experiments, we embedded a 0.8mm-diameter rounded
target in the background. The absorption coefficient of the target was 0.07mm-1, while
the depths are 0.0mm (beneath the surface), 5.0mm and 7.5mm, respectively. In the
fourth experiment, the absorption coefficient of the target was 0.035mm-1, and the target
was embedded at the depth of 2mm. In the fifth experiment, the absorption coefficient of
the target was 0.021mm-1, and the target was embedded at the depth of 1mm. We also
present a multi-target case in the last experiment. Three targets with different
absorption coefficients (0.07mm-1, 0.035mm-1 and 0.021mm-1, respectively) were
embedded at the depth of 2mm in the background. In all of above experiments, the
absorption coefficient of the background was 0.007mm-1, and the reduced scattering
coefficient of the background and targets were 1.0mm-1.
Figure 3-2 shows reconstructed absorption coefficient images from these
phantom experiments. We see that the object(s) are clearly detected by our finite-
element based quantitative PAM algorithm for all six cases. Based on these results, we
can make observations that the methodology outlined earlier is a feasible reconstruction
algorithm that can provide both qualitative and quantitative information about the targets
in terms of their location, size, shape and the optical property values. These conclusions
are further confirmed by the results provided in Figure 3-3, which presents
38
reconstructed absorption coefficient profiles along transects that across the target(s) in
these cases.
We also summarize the comparison of exact and reconstructed values (absorption
coefficient and size) in the target area for all these phantom experimental cases in
Table 3-1. Based on these quantitative results, we find that the relative errors for the
recovered absorption coefficient of the targets for all of the single target cases (case
I~V) are all less than 5%. It is worth noting that the values are better reconstructed for
the higher contrast cases (case I~III with the contrast level 10:1) than for the lower
contrast cases (case IV with the contrast level 5:1, and case V with the contrast level
3:1). It is also better reconstructed for the shallow target cases (case I) than for the
deeper target cases (case II~III). The results from the multi-target case (case VI) also
confirm this conclusion. Similar observations on the recovered size of the targets can be
made based on these results.
3.5 In Vivo Animal Experiments Validation
A rat ear experiment was chosen to further validate the algorithm. All experimental
animal procedures were forth by, IACUC of University of Florida, as based on the
guidelines from the NIH guide for the Care and Use of Laboratory Animals. In Figure 3-
4A we present the maximum amplitude projection (MAP) images projected along the
vertical direction (z axis) to the orthogonal plane, and seven orders of vessel branching
can be observed in the image, which indicated by numbers 1–7. The B-scan image and
the reconstructed absorption coefficient image obtained by the FE-based quantitative
PAM algorithm are shown in Figure 3-4B and C, respectively. These images are in the
vertical plane (x-z plane) at the location indicated by the dash line in Figure 3-4A. It can
be clearly seen that all of the vessel branching can be reconstructed by using our FE
39
based quantitative PAM algorithm. Although both methods provide the images with
almost the same high spatial resolution, we obtain the absolute optical absorption
coefficient values based on our quantitative PAM method.
In next paragraph, we will show three applications on cancer research using
ARPAM. Firstly, we used ARPAM to monitoring development of neovasculation of
breast tumor on an animal model with and without anti-angiogenesis drugs. Second,
ARPAM was used to image breast tumor with injection of targeted contrast agent which
is very important to study the procedure of drug delivery. Thirdly, it was combined with
fluorescence molecular tomography (FMT) to map the ovarian cancer with injection of
targeted contrast both for photoacoustic and fluorescent imaging.
40
Figure 3-1. Our experimental PAM system. A) Diagram of experimental installation. B)
The optical illumination area of the experimental system.
Figure 3-2. Reconstructed absorption coefficient images (mm-1) from all 6 experimental cases. A) Case I. B) Case II. C) Case III. D) Case IV. E) Case V. F) Case VI.
41
Figure 3-3. Reconstructed absorption coefficient profiles. A) y=0.0mm for image shown in Fig.2A. B) y=-4.8mm for image shown in Fig.2B. C) y=-7.6mm for image shown in Fig.2C. D) y=-2.1mm for image shown in Fig.2D. E) y=-1.1mm for image shown in Fig.2E. F) y=-1.9mm for image shown in Fig.2F.
42
Figure 3-4. In vivo imaging of the blood vessels in a rat ear. A) The MAP image of the photoacoustic signals projected on the orthogonal plane, with seven small vessels that can be observed in the image as indicated by numbers 1-7. B) B-scan image in the vertical plane at the location indicated by the dash line in A. C) Reconstructed absorption coefficient image in the vertical plane at the location indicated by the dash line in A.
43
Table 3-1. Comparison of exact and reconstructed values (absorption coefficient and size) in the target area for all 6 experimental cases.
Case #
Depth of
the target
(mm)
µa of target (mm-1) size of target (mm)
Exact Recovered Exact recovered
I 0
0.07
0.071
0.8
0.90
II 5.0 0.071 0.90
III 7.5 0.072 0.95
IV 2.0 0.035 0.034 0.95
V 1.0 0.021 0.020 1.0
VI 2.0
0.07 0.068 0.9
0.035 0.033 1.0
0.021 0.019 1.1
44
CHAPTER 4 APPLICATIONS OF ACOUSTIC-RESOLUTION PHOTOACOUSTIC MICROSCOPY
TO PRECLINICAL CANCER RESEARCH
4.1 Photoacoustic Imaging of Tumor Vasculature Development for Breast Cancer Research
4.1.1 Motivations
Tumors growth requires sustenance by nutrients and oxygen as well as an ability
to evacuate metabolic waste and carbon dioxide (CO2). To address these needs, an
“angiogenic switch” is permanently activated, causing new vessel growth from the
adjacent host vasculature. The vasculatures within tumors are typically aberrant,
controlled by a complex biological stress that involves both the cancer cells and stromal
microenvironment.86 Sustained aberrant tumor angiogenesis plays a central role in
breast cancer carcinogenesis and metastatic potential.87 Members of vascular
endothelial growth factor (VEGF) family are well-known angiogenesis activators.
Despite the promising activity of anti-angiogenic drugs in preclinical tumor models,
targeting VEGF signaling appears to be insufficient for permanently inhibiting tumor
angiogenesis in patients with breast cancers. The reasons for this are likely to be
multiple and complex. For example, animal studies revealed that some tumor vessels
induced by VEGF require continuous VEGF expression for their maintenance and
undergo apoptosis if VEGF levels fall below threshold level.88 Others, once induced by
VEGF, persist indefinitely in the absence of exogenous VEGF89, suggesting that tumor
vasculature is heterogeneous and an endogenous intracrine VEGF signaling may be
crucial for vascular homeostasis.90 Nevertheless, there remains an urgent and unmet
need for novel targeted therapies for patients with resistance to current anti-angiogenic
agents. An ideal anti-angiogenic model for cancer treatment should consist of three
45
elements: i) identification and validation of rationale-based anti-angiogenic targets; ii) an
adequate drug form and delivery route; iii) advanced imaging modalities allowing
noninvasive detection and monitoring tumor response to anti-angiogenic agents.
In rapidly growing tumors, vascular endothelial cells face a hostile
microenvironments characterized by hypoxia, nutrient deprivation and acidosis. These
environmental stressors induce endoplasmic reticulum (ER) stress, and cells respond
by activating the unfolded protein response (UPR) pathway.91 The UPR relieves ER
stress by inducing genes i) to increase the protein-folding capacity of the ER, ii) to
enhance the clearance of unfolded proteins from the ER, and iii) to inhibit general
protein translation in the ER.92 Our recent study demonstrated that under the stress
conditions caused by tumor microenvironment or/and anti-VEGF therapies, tumor
endothelial cells adopt the up-regulation of IRE1α/XBP-1 and ATF697, which may
contribute to the up-regulation of molecular chaperone and in turn increase the protein-
folding capacity for misfolded/unfolded VEGF and maintain VEGF intracrine signaling
for endothelial cell survival.
In the past decades, remarkable progress has been made in using siRNA as a
promising new class of drugs by targeting the mutant or overexpression oncogenes.
Several siRNA cancer therapies have entered early clinical trials.98 In the present study,
we investigated whether IRE1α or XBP-1 or ATF6 can serve as novel antiangiogenic
therapeutic targets for breast cancer treatment. Also, we have been pursuing that
further development of alternative vector systems, such as self-complementary adeno-
associated virus (scAAV) vectors for their potential to transduce microvascular
46
endothelial cells with high efficiency, given that proven safety of AAV vectors in several
clinical trials.106-107
Several current non-invasive imaging modalities have differing limitations for
monitoring vasculature development. For instance, X-ray computed tomography (CT)
needs extrinsic contrast agent and exposures patients to ionization radiation108, positron
emission tomography (PET) screening often involves extrinsic contrast agents and
magnetic resonance imaging (MRI) is limited by its low temporal/spatial resolution.109
Pure high-resolution optical imaging modalities such as single-photon, multi-photon
fluorescence microscopy suffer from limited imaging depth (<1mm) and repeated
fluorescent dye injection.110 Adequate non-invasive imaging can help physicians to
determine whether to start and when to start anti-angiogenic therapies. In particular,
such imaging is essential for monitoring the tumor response to anti-angiogenic therapies
because tumor shrinkage may not occur within a short period of time even when anti-
angiogenic treatment is effective. One potential non-invasive imaging modality is
photoascoustic imaging (PA) consisting of the advantages of rich optical contrast in
optical imaging and high ratio of imaging depth to spatial resolution in ultrasound
imaging.
In the present study, we have identified that scAAV2 septuplet-mutant vector, in
which seven surface-exposed tyrosine residues of the capsid were changed to
phenylalanine, was able to infect mouse microvascular endothelial cells with high
efficiency. scAAV2 vectors encoding siRNAs against IRE1α, or XBP-1 or ATF6 were
effective in decreasing breast cancer-induced angiogenesis in vitro. Serial non-invasive
photoacoustic imaging further confirm that intratumoral delivering the siRNAs against
47
IRE1α, XBP-1 or ATF6 by scAA2 vectors resulted in a significant decreased in tumor
growth and tumor angiogenesis in breast cancer xenograft models. These data have
generated a proof-to-concept model with important implications for the development of
novel anti-angiogenic targeted therapies for patients with breast cancer.
4.1.2 Materials and Methods
4.1.2.1 Construction of scAAV2 vector for delivering siRNAs against the UPR proteins
Self-complementary AAV serotype 2 (scAAV2) and their corresponding single and
multiple tyrosine-to-phenylalanine (Y-F) mutants containing a ubiquitous, truncated
chimeric CMV-chicken β-actin (smCBA) promoter116 were generated and purified by
previous described methods.117 Vectors were titered by quantitative real-time PCR and
re-suspended in balanced salt solution (BSS: Alcon, Forth Worth, TX).118 siRNA has
been widely used to knock down target gene expression for a variety of purpose.
However, siRNAs can provoke unspecific degradation of all cellular RNAs through the
induction of the interferon response.119 We utilized three pre-validated sequences of the
siRNAs against mice IRE1α, XBP-1 and ATF6, respectively, with a significant
knockdown in our previous study120 (Figure 4-2D). The cDNAs encoding IRE1α (5’-
GGATGTAAGTGACCGAATA-3’), XBP-1 (5’-CAGCTTTTACGGGAGAAAA-3’), ATF6
(5’-GGATCATCAGCGGAACCAA-3’) and scramble (with the same nucleotide
composition of IRE1α cDNA) were ligated into scAAV2 backbones using Notl and Sall
(Figure 4-2B and C). The integrity of the siRNA coding region was confirmed by
sequence analysis (data not shown).
48
4.1.2.2 scAAV2 infection of cells in vitro
Infection of scAAV2 vectors were carried out as previous described.121 Briefly,
cells were plated in 14-mm microwells in a 35-mm Petri dish (MatTek, Ashland, MA)
and reached up to 60-70% confluence. scAAV2 vectors were diluted in serum-free
medium to achieve the desired multiplicity of infection (MOI). MOI was defined as the
number of genome containing vector particles per targeting cell. The cells were rinsed in
PBS and the virus dilutions were added. After 1 hour, complete medium was added to
the cells. Infected cells were maintained at 37°C in 5% CO2 for 3 days. All infections
were performed in triplicate. Each vector was used to infect 24 wells in total. Each
“sample” was pooled from eight wells, making a final sample count of three.
4.1.2.3 Microscopy and fluorescence activated cell sorting (FACS) analysis
Three days post infection, cells were observed using bright-field microscopy to
ensure 100% confluence. Cells were then analyzed by an OlympusI×81-DSU Spinning
Disk Confocal microscopy (Olympus America,Inc., Center Valley, PA) and digital
photomicrographs were captured by imaging software-SlideBook™ (Intelligent Imaging
Innovations, Denver, CO). For FACS analysis, the cells were dissociated with Accutase
solution (MP Biomedicals. Solon, OH) and collected by centrifugation (200×g for 5 min).
10,000 cells per sample were counted and analyzed using a BD LSR II flow cytometer.
Uninfected control cells were also counted and analyzed to establish transduction
efficiency of baselines.
4.1.2.4 Cell culture
Mice microvascular endothelial cells (MMECs) were isolated from mouse brain
tissue using modified methods as previous described.122 Briefly, freshly isolated mouse
brains were homogenized, and after trapping on an 83-μm nylon mesh, they were
49
transferred into an enzyme mixture including 500μg/ml collagenase, 200μg/ml Pronase,
and 200μg/ml DNase, at 37°C for 30min. The resultant vessel fragments were trapped
on a 53-μm mesh, washed, and pelleted, and cells were plated in microvascular
endothelial cell basal medium with growth supplement (Life Technologies™, Grand
Island, NY) at 37°C in 5% CO2 for 3 days. Purification of MMECs was achieved as
previous described.123 Briefly, the cells collected from primary culture T25cm2 flask by
trypsinization were reacted with rat anti-VE-cadherin antibody (Biolegend, Inc, San
Diego, CA) for 10min at 4°C. After washing, the cells were mixed with pan rat IgG beads
(Life Technologies™, Grand Island, NY) for 20min at 4°C with gentle tilting and rotation.
The samples were placed in a magnetic particle concentrator (DynaMag™-15, Life
Technologies™, Grand Island, NY) to immunoadsorbed cells for 2min at room
temperature. The bead-bound cells were then re-suspended in release buffer for 15min.
The purified cells were collected by centrifugation and seeded into T25cm2 flask in
complete medium. Cells were subculture at a ratio of 1:2 on reaching confluence. The
cells were used within three passages.
NeuT (a mice breast cancer cell line) was isolated from MMTV-c-neu transgene
mice as described.124 NeuT EMTCL2 served as a malignant variant of NeuT and was
developed by consecutive serial implantation of NeuT cells in athymtic nude mice. The
mouse breast cancer cells were grown in RPMI1640 supplemented with 10% fetal
bovine serum and gentamycin (Gibco-BRL, MD) to 70-80% confluence subjected to
sequential experiments or sub-cultured at ratio of 1:3 in fresh complete medium
supplemented with 10% FBS.
50
For in vitro co-culture systems, NeuT or NeuTEMTCL2 cells were seeded onto 6-
well Transwell inserts with 0.4μm pores (Corning Life Sciences, MA) in a 6-well plate for
72 hrs. MMECs were cultured in a separate 6-well plate. Confluent breast cancer cells
on Transwell inserts were then transferred on top of MMECs and placed at 37°C for 48
hr prior to sequential experiments.
4.1.2.5 In vitro angiogenesis assay
In vitro angiogenesis was measured by in vitro tube formation assay which reflects
a combination of proliferation, migration and tube formation of microvascular endothelial
cells. Briefly, MMECs were plated sparsely (2.5×104/well) on 24-well plates coated with
12.5% (v/v) Matrigel (BD, Franklin Lakes, NJ) and left overnight. The medium was then
aspirated and 250μl/well of 12.5% Matrigel was overlaid on the cells for 2 hrs to allow its
polymerization, followed by addition of 500μl/well of basal medium MCD131 with 10%
fetal calf serum (FCS) for 48 hrs. The culture plates were observed under a phase
contrast microscope and photographed at random in five fields (×10). The tubule length
(mm/mm2) per microscope field was quantified.
4.1.2.6 Apoptosis assay
Apoptosis was evaluated using FITC–conjugated annexin V/propidium iodide
assay kit (R&D System, Minneapolis, MN) based on annexin-V binding to
phosphatidylserine exposed on the outer leaflet of the plasma membrane lipid layer of
cells entering the apoptotic pathway. Briefly, MMECs were collected by EDTA
detachment and centrifuged (200×g for 5 min), washed in PBS and re-suspended in the
annexin V incubation reagent in the dark for 15 min before flow cytometric analysis. The
analysis of samples was performed by flow cytometric analysis on BD lSRII flow
cytometer (BD Biosciences, MD). An excitation wavelength of 488nm was used with
51
fluorescence emission measured at 530 ± 15nm through fluorescence channel one. A
minimum of 10,000 cells per sample were collected using log amplification for
fluorescence channel one and linear amplification for forward light scanner and 90° light
scatter before being analyzed using in-house software.
4.1.2.7 Proliferation assay
Crystal violet assay was conducted as previously described.125 Briefly, MMEC
suspensions (100 μl) were incubated in each well of 96-well plates at 1×105cell/ml. The
cells were fixed in 4% paraformaldehyde in PBS for 15min. After being washing with
distilled water, the plates were stained with 0.1% crystal violet solution for 20min. The
plates were washed with water and allowed to be air dry. Acetic acid (100μl of 33%)
was added to each well for extraction of dye. Absorbance of the staining was measured
by an automatic microtiter plate reader at 590nm.
4.1.2.8 Mice breast cancer xenograft models
All animals used in this study were maintained at the animal facility of the
University of Florida and handled in accordance with institutional guidelines. Athymic
female nude mice (nu/nu) at 5 to 8 weeks were purchased from Charles River
Laboratories (Charles River Laboratories, Inc., Wilmington, MA) and caged in groups of
5 or fewer. Mice breast cancer xenografts were established by subcutaneous injection
of 1×105 NeuT or NeuT EMTCL2 cells into the mammary fat pads of the mice. Tumor
volume was calculated from caliper measurements of the large (a) and smallest (b)
diameters of each tumor using formula a×b2×0.4. Three days after inoculation, most
tumors had grown to ~30mm3. All mice were euthanized when the tumor volume in the
non-treated group reached ~1000mm3.
52
For in vivo siRNA knockdown of the UPR proteins, mice with similarly sized tumors
were divided into two groups (siRNA treatment and scrambled control). Mice were
anesthetized with a mixture of ketamine (85mg/kg)/xylazine (4mg/kg) and were
intratumoral injected (10μl) with scAAV2 sept-mut vectors encoding siRNAs against
IRE1α or XBP-1 or ATF6 or scrambled siRNA at a titer of 2×1013 genome copies per ml.
4.1.2.9 Photoascoustic (PA) imaging systems and noninvasive monitoring in vivo tumor angiogenesis
In this study, photoacoustic microscopy (PAM) system was used to monitor
vasculature change in breast cancer xengraft models. The system is the same as
shown in Figure 3-1A. In this applications, a focused ultrasound transducer (50MHz,
3mm aperture and 6mm focal length) was used to receive induced photoacoustic
waves. PA signals amplified by two different amplifiers (one was 17dB from 100kHZ to
1GHz and the other was 20dB from 20MHz to 3GHz) were digitized by a 8-bit data
acquisition board (NI5152, National Instrument, Austin, TX) at a sample rate of
250MS/s. The two-dimensional transverse scanning combined with the depth-resolved
ultrasonic detection generated 3D PA images displayed in maximum amplitude
projection (MAP). 3D PA signals were processed by Hilbert transform, normalized to the
same scale (0-256) and applied with a threshold (-6dB level).126 The volumes of the
blood vessels were calculated by integrating the corresponding image voxels (1 for
blood vessel and background being set to 0). Entropy was calculated from normalized
MAP of each PA image at the same scale (0-256).
4.1.2.10 Statistics analysis
All experiments were repeated at least 3 times. Analysis of variance (ANOVA) was
used to assess the transduction efficiency of AAV2 vector infection. The unpaired
53
Student t-test was to assess statistic significant for the rest data obtained from in vitro
studies (in vitro angiogenesis assay, apoptosis assays and crystal violet assays) as well
as in vivo animal experiments including tumor volume, entropy and normalized vessel
volume. The data were expressed as mean±SEM. Statistical analysis was conducted by
GraphPad Prism (version 5.01; GraphPad Software, Inc., La Jolla, CA) with p<0.05
considered statistically significant.
4.1.3 Results
4.1.3.1 scAAV2 sept mut exhibits higher transduction efficiency in MMECs
AAV2 based vectors have, to date, exhibited higher transduction efficiency when
to targeting transgene expression to vascular endothelium than other native AAV.127 A
construct containing a truncated version of the hybrid chicken β-actin promoter/CMV
promoter (smCBA) driving green fluorescent protein (GFP) was packaged into scAAV2
containing combinations of up to 7 surface-exposed tyrosine residues (252, 272, 444,
500, 700, 704 and 730F). Combinations tested for transduction of mouse microvascular
endothelial cells (MMECs) included: triple (Y444+500+730F), quadruple
(Y272+444+500+730F) and septuplet (Y252+272+444+500+700+704+730F). The
transduction efficiency of each of the combination tyrosine-mutant vector at MOIs
ranging from 100 to 10,000 was analyzed (as measured by GFP expression) and
compared with unmodified scAAV2 in MMECs 72 hr later by flow cytometry. We present
data generated from cells infected only at an MOI of 10,000 as it is representative of
trends seen across all MOIs. We demonstrated that the transduction efficiency of the all
tyrosine-mutant vectors was significantly higher compared with the unmodified scAAV2
(Figure 4-1A). Specifically, the transduction efficiency of the septuplet-mutant vectors
was maximal, ~173-fold higher than the unmodified vector (Figure 4-1B). Similarly, the
54
triple-mutant and the quadruple mutant also had significant enhancement in GFP
expression (~2- and 6-fold, respectively) (Figure 4-1B).
4.1.3.2 Septuplet tyrosine-mutations improve the transduction efficiency of scAAV2-mediated siRNAs infection in MMECs
Our previous study showed a crucial role of the unfolded protein response
pathway in breast tumor-angiogenesis.97 Base on the observation (Figure 4-1) that
scAAV2 septuplet mutant (Y-F) exhibited the most transduction efficiency in MMECs,
we constructed a septuplet-mutant AAV2 vector containing three siRNA sequences
against IRE1α, XBP-1, ATF6, respectively and a scrambled siRNA (Figure 4-2A and B).
To verify that the inserted siRNA sequence transduced MMECs, scAAV2 sept-mut
containing the scrambled siRNA was used to infect MMECs at MOIs ranging from 400
to 12,000. As shown in Figure 4-1E, mCherry expression was detected by fluorescence
microscopy 72 hr post-infection. The transduction efficiency of sc AAV2 sept-mut vector
was not affected by siRNA insertion and elevated with the increased MOI of the vector
(Figure 4-2C).
4.1.3.3 siRNA knockdown of the UPR proteins decreased NeuT EMTCL2-induced in vitro angiogenic activity of endothelial cells
The balance between protein synthesis and protein proper folding in ER is
essential for cellular homeostasis and survival. Many disease conditions that affect
protein folding tip this balance and trigger an ER membrane-bound protein stress
pathway known as the unfolded protein response (UPR). Our recent study
demonstrated that malignant breast cancer cells significantly up-regulated three UPR
proteins including IRE1α, XBP-1 and ATF6 in the endothelial cells, implicating possible
pro-angiogenic roles for UPR.97 To address the possible involvements of the UPR
proteins for pro-angiogenic activity, MMECs were co-cultured with NeuT EMTCL2 as
55
described in Experimental Procedures. Since late stage of angiogenesis requires
morphologic alterations of endothelial cells, leading to lumen formation, NeuT EMTCL2
cell-induced angiogenesis in MMECs was studied by measuring a network of capillary-
like tubes in an in vitro 3-D Matrigel model with reflects a combination of proliferation,
migration and tubule formation of endothelial cells.125 MMECs pre-treated with the
scrambled siRNA exhibited a significant in vitro angiogenesis in contrast to non-
treatment (Figure 4-3A and B). As expected, the siRNAs against the UPR proteins
markedly reduced the angiogenic responses of MMECs to the NeuT EMTCL2 cells’
stimulation (Figure 4-3A and B). When compared with scrambled siRNA, the siRNA
against IRE1α or XBP-1 caused a ~50% reduction of in vitro angiogenesis wile ATF6
siRNA caused a ~30% reduction.
4.1.3.4 siRNA knockdown of the UPR proteins down-regulated the survival of endothelial cells
The effect of knockdown of IRE1α, XBP-1 or ATF6 on the survival of MMECs was
analyzed by apoptosis and crystal violet assays. The knockdown of XBP-1 elicited a
maximal apoptotic response in MMECs, as evidenced by a significant 8-fold increase in
the apoptotic cells compared to the scrambled siRNA (Figure 4-3C and D). siRNAs
against IRE1α or ATF6 caused a lesser, yet still significant apoptotic response in
MMECs (4-fold increase) (Figure 4-3C and D). Consistently, crystal violet assay showed
NeuT EMTCL2 stimulation significantly increased MMECs proliferation, which was
inhibited by the knockdown of XBP-1 at maximal reduction of ~70% (Figure 4-3E). Both
of siRNAs against IRE1α and ATF6 also caused a similar (~50%) reduction in NeuT
EMTCL2-induced survival of MMECs (Figure 4-3E).
56
4.1.3.5 Different malignant mice breast cancer cells induced differential agngiogenic responses in breast cancer xenograft models were monitored by serial photoacoustic (PA) imaging
To test the feasibility of noninvasively monitoring in vivo tumor vasculature
development, we performed PA imaging for two mice breast cancer xenograft models.
In one model, mice were injected with NeuT cells, while the other model was generated
via injection of NeuT EMTCL2 cells. Serial PA imaging were carried out on day 3, 5, 7
and 9 after tumor inoculation, respectively. The inoculation of NeuT cells only caused a
minimal tumor growth compared to a rapid tumor growth in NeuT EMTCL2 model which
tumor volume reached ~250mm3 by day 11 (Figure 4-4A and D). Additionally, NeuT
EMTCL2 cells induced a significant vasculature development evidenced as gradual
splitting large host blood vessels to small ones. This was noticeable on day 5 and
peaking on day 9. Using PA resolution and depth-section capability, we determined the
kinetics of the vessel density changes surrounding tumor mass using the Entropy
method.128 Entropy is a statistical measure of randomness that can be used to
characterize the texture of the input image. In NeuT model, Entropy did not reveal a
significant increase in vessel density. However, Entropy clearly indicated that NeuT
EMTCL2 inoculation resulted in a increased vessel density by 20 % compared with
NeuT implantation (Figure 4-4B). Another PA image extraction analysis was normalized
vessel volume changes that closely echo Entropy data. As shown (Figure 4-4C), NeuT
EMTCL2 implantation resulted in a steep increase in vessel volume, whereas there was
no significant change in tumor vessel volume detected in the NeuT model (Figure 4-4C).
57
4.1.3.6 Knockdown of the UPR proteins significantly inhibited Neut EMTCL2-induced in vivo angiogenesis
To confirm the purported anti-angiogenic activities of the siRNAs against the UPR
proteins (Figure 4-5A and B), mice of NeuT EMTCL2 xenograft models were divided
into four groups including one scrambled treatment, three siRNA treatments (IRE1α,
XBP-1 and ATF6) delivered by intratumoral injection of scAAV2 sept-mut vectors
encoding the siRNAs against the UPR proteins. The tumor volume reached ~250mm3
after 11 day of NeuT EMTCL2 inoculation in the scrambled siRNAs treatment group
(Figure 4-5D). Figure 4-5D also shows that the knockdown of IRE1α and XBP-1 both
exhibited significant decreased tumor growth to a similar extent (~45% on day 11) and
the ATF6 siRNA was even more effective on inhibition of tumor growth (~54% by day
11). Serial PA imaging was employed to monitor and evaluate the development of tumor
vasculature in the NeuT EMTCL2 xenograft models on day 3 after NeuT EMTCL2
inoculation followed on day 5, 7, 9 and 11. Entropy method delineated a significant
decrease in vessel density (25%) by day 9 and sustained at that level thereafter in the
IRE1α siRNA treatment group compared with the scrambled group (Figure 4-5B). The
NeuT EMTCL2 xenograft models treated with the siRNAs against XBP-1 or ATF6
evidenced a steady decrease in vessel density with statistic significant in comparison
with the scramble group (Figure 4-5B). All three siRNAs exhibited significantly steady
decrease in vessel volume compared with the scrambled siRNA treatment (Figure 4-
5C).
4.1.4 Discussion
The UPR is generally considered to involve 3 signaling pathways from the ER
membrane: IRE1α (inositol-requiring protein 1α), ATF6 (activating transcription factor 6)
58
and PERK (PKR-like ER kinase). Upon activation, IRE1α cleaves an intron of XBP-1 (X-
box-binding protein 1) mRNA, resulting in a frame shift and the translation of the spliced
form of XBP-1, a 41kDa basic leucine zipper (bZIP) family transcription factor that
induces genes involved in UPR response.93 IRE1α also cleaves many mRNAs, reducing
the load of proteins in the ER, in favor of restoration of ER homeostasis.94 During ER
stress, ATF6 is transported to the Golgi apparatus where it is cleaved and release a
50kDa bZIP to the nucleus to induced expression of ER chaperones.95 ATF6 also
supports ER biogenesis independently of XBP-1.96
IRE1α/XBP-1, ATF6 and PERK are aberrantly expressed in many types of
cancers. The IRE1α/XBP-1 pathway is important for tumor growth under deteriorating
conditions of microenvironment. The levels of XBP-1 correlate with glucose
starvation.129 XBP-1 splicing can be detected even in relatively small tumors in several
genetic models for breast cancer, implicating that ER stress may occurs from the early
stage of tumor development. Indeed, XBP-1 knockout caused cancer cell death and
XBP-1 knockout cells produced smaller tumors in xenograft models. Over-expression of
IRE1α/XBP-1 restored tumor growth under these conditions. The significance of ATF6
in tumor development is less well characterized. However, ATF6 was found to be over-
expressed and activated alone with increased XBP-1, suggesting a coordination and
interdependency between IRE1α/XBP-1-dependent and ATF6-dependent systems.130
ATF6 served as stress survival factor for dormant but not proliferative squamous
carcinoma cells via GTPase and mTOR.131 A role for ATF6 pathway is further supported
by an in vitro study in which activation of ATF6 resulted in apoptosis resistance of
melanoma cells.132 The UPR pathway is a general mediator of vascular endothelial cell
59
dysfunction in inflammatory disorders.133 Furthermore, our recent studies suggested
that activation of IRE1α/XBP-1 and ATF6 is closely related to angiogenic activity of
endothelial cells, whereas endothelial PERK functions were not drastically affected by
angiogenic stimulation.97 In the current study, we confirmed that selective knockdown of
IRE1α or XBP-1 or ATF6 can lead to inhibition of tumor angiogenesis and breast cancer
growth using in vitro and in vivo models.
A challenge for RNAi based-therapies is the efficiency of delivery for the siRNA to
the target cells. Delivery of siRNA into mammalian cells is usually achieved via the
transfection of double-strand oligonucleotides or plasmid encoding RNA polymerase III
promoter-driven small hairpin RNA.99 Recently, retroviral and lentiviral vectors have
been used for siRNA delivery, which have overcome some of the problems of poor
transfection efficiency seen with the plasmid-based systems.100 However, both retroviral
and lentiviral vectors undergo random integration into host chromosomal DNA, and
there is persistent risk for insertion mutagenesis.101-102 Non-integrating adenoviral vector
has also been tested for vascular endothelial gene transfer103, but many cell types
express the adenoviral receptor preventing selective endothelial cell infection and
precluding clinical use.104 Furthermore, adenoviral vectors are highly immunogenic and
are therefore unsuitable for long term gene expression in vivo.105 In this study, we
ultilized scAAV2 vectors containing seven surface exposed tyrosine to phenylanine
capsid mutations to deliver siRNAs against IRE1α, XBP-1 and ATF6 into endothelial
cells. AAV vectors have historically exhibited rather poor transduction efficiency in
endothelial cells in comparison with other cells. Key steps in AAV transduction have
been identified as cell surface receptor-and co-receptor-mediated viral binding,
60
intracellular trafficking, nuclear transport, uncoating, and viral second-strand DNA
synthesis.135-137 Of these, intracellular trafficking is considered a major impediment for
efficient transgene expression. Tyrosine phosphorylation of AAV2 can trigger
ubiquitination-dependent degradation of AAV2.138 Based on these observations, we
replaced tyrosine residues with phenylalanine to generation several multiple mutations
of the seven surface-exposed tyrosine residues on scAAV2 capsids.121 In this study, we
identified scAAV2 sept-mut vector that can transduce mouse microvascular endothelial
cell ~173-fold more efficiently than wild-type scAAV2. Intriguingly, our current data is
different from our previous published data which showed that the quadruple-mutant was
the most efficacious in bovine retinal microvascular endothelial cells139, suggesting that
tyrosine-mutants of scAAV2 may have different levels of efficacy in different cell/tissue
types. It is noteworthy that another group has also reported a similar discrepancy for
scAAV2 transduction efficiency in the other cell types, possibly due to the differences in
the intracellular milieus in different species and/or conformation-dependent binding and
accessibility to host cellular factors.140
The chimeric CMV-chicken β-actin promoter (CBA) has been utilized extensively
as a promoter that supports expression in a wide variety of cells due to a broad tropism.
However, CBA is ~1700 base pairs in length, too large in most cases to be used in
conjunction with AAV vectors. As the result, we selected a truncated CBA promoter
(smCBA), in which the hybrid chicken β-actin/rabbit β-globin intron is greatly shortened.
The most significant advantage of the modification is that smCBA can initiate transgene
expression sooner and for a longer time period, compared to expression using other
61
promoters. Because smCBA is truncated, it permits packaging of more polynucleotides
into the vector, which is particularly relevant for double-stranded viral vectosr.
Our results demonstrated that knockdown of IRE1α or XBP-1 or ATF6 exerts
robust anti-angiogenic activities and suppresses tumor growth via local intratumorally
administration of scAAV2 sept-mut vectors encoding the corresponding siRNAs.
Although there are controversies surrounding the approaching of local intratumoral gene
therapy, this strategy may offer several advantages for anti-angiogenic therapy: 1)
intratumoral gene therapy can reduce the risk of widespread anti-angiogenesis resulting
from systemic administration of an anti-angiogenic agent; 2) gene transfer can lead to a
local accumulation of the anti-angiogenic agents; 3) anti-angiogenic gene therapy does
not require that genes be transferred to all target cells.
Photoacoustic microscopy (PAM), a hybrid imaging technique combining high
optical contrast and high ratio of imaging depth to spatial resolution of ultrasound, has
multiple advantages.111 First, it can identify red blood cell perfused microvasculature
supplying oxygen for tissues through the endogenous hemoglobin contrast. Second,
unlike conventional optical microscopic, PAM is 100 time less acoustic scattering in
biological tissues, thus greatly improving tissue transparency. Thirdly, the high contrast
of hemoglobin at 532nm (>100:1) enables using of low-level laser exposure. Recently, a
study showed noninvasively label-free serial imaging of red blood cell perfused
vasculature in ear of small mice by using ORPAM.112 The authors demonstrated that the
spatial resolution is high enough (<2μm) to image a single capillary and even a single
red blood cell. However, the imaging depth of 400-700μm may limit ORPAM application
for monitoring neovascularization during tumor growth. In contrast, acoustic-resolution
62
PAM with imaging depth reaching 3mm is more adequate for monitoring tumor vascular
development along with tumor growth.141 Using our established acoustic-resolution PAM
systems, we were able to serially image dynamics of tumor neovascular network
development in mice breast cancer xenograft models. More importantly, it was first time
that we successfully monitored and determined differential tumor response to siRNAs
against IRE1α, XBP-1 and ATF6. These were delivered using AAV2 septuplet-mutant
vectors by intratumoral injection in mice breast cancer xenograft models. The acoustic-
resolution PAM systems system used here equips a lower repetition rate (10Hz) pulse-
laser resulting in relative longer scanning time (20min) in comparison with some
commercial high repetition rate (>100Hz) pulsed laser enabling completing each
scanning within 2min. To apply PA imaging in the clinic in the future, it will be greatly
beneficial if we can develop quantitative PAM systems enabling extraction of more
functional information, such as oxygen saturation, concentration of hemoglobin
correlating to tumor growth and anti-angiogenic therapy.
4.2 uPAR Targeted Magnetic Iron Oxide as a Contrast Agent for In Vivo Molecular Photoacoustic Tomography of Breast Cancer
4.2.1 Motivations
Breast cancer affects 1 in 4 women in the U.S.143and in 2011 there were 230480
new breast cancer cases and 39520 breast cancer deaths.143-144 Conventional breast
imaging techniques such as X-ray, ultrasound and magnetic resonance imaging (MRI)
have limitations. X-ray mammography is the most widely used and remains the imaging
standard for breast cancer diagnosis, but exposes patients to ionization radiation, is not
ideal for imaging dense breasts, which are typically associated with younger female
patients (<50 yr).145 Although ultrasound is very useful in characterizing breast lesions in
63
dense breasts, it is not used as a primary screening method due to its low contrast level
and specificity. Breast MRI shows high sensitivity but low specificity producing a high
false-positive rate.146 Therefore, there is an urgent need to develop noninvasive, and
highly sensitive specific techniques with a high resolution for early detection, accurate
diagnosis and precise resection of breast cancer.
Molecular imaging techniques using target-specific probes for positron emission
tomography (PET) or single photon emission computed tomography (SPECT) have
been demonstrated for breast cancer diagnosis and treatment monitoring. These
imaging methods have shown improved specificity and sensitivity in cancer detection,
compared with conventional mammography. However, PET and SPECT have low
spatial resolution in determining anatomic location of the tumor. Additionally, their
dynamic and time-resolved imaging ability is limited because of the long half-life of the
radiotracers. PET and SPECT both involve ionization radiation.147-148
PAT is an imaging technique that detects wide-band acoustic waves that are
generated when biological tissue absorbs short laser pulses. This imaging method has
the advantages of high optical contrast as well as increased ratio of imaging depth to
spatial resolution. The image resolution and maximum imaging depth can be adjusted
with the ultrasonic frequency and the penetration of diffuse photons. The optical-to-
acoustic conversion efficiency represents how many incident photons will be absorbed
and converted to heat and how fast this heat can diffuse from the target during
thermoelastic expansion and wave generation. As such, this conversion efficiency will
determine the contrast intensity of photoacousic imaging. Although an increase in
hemoglobin (Hb) content in breast tumors results in enhanced absorption of photon
64
energy that leads to 2 to 4 times higher photoacoustic signals, compared to the healthy
breast tissue, the signal is not strong enough to detect breast tumors located deep
below the skin. To enhance the photoacoustic signals, contrast agents are used to
generate acoustic transients. Several recent studies have reported the use of targeted
and non-targeted contrast agents in imaging lymph nodes, melanomas, angiogenesis,
the cerebral cortex and brain tumors. IONP has been widely used to enhance the image
contrast for MRI in mice experiments and has recently been approved for pilot studies
on humans. uPAR targeted nanoparticles (ATF-IONP) have been used successfully in
in vivo magnetic resonance imaging of mouse mammary tumors.149 In this study, tumor
cells selectively bound and internalized the ATF-IONP and provided high contrast for
MRI. Ekaterina et al. also demonstrated the use of IONP as a contrast agent to detect
circulating tumor cells.150 In our study, we used near-infrared dye labeled amino-
terminal fragments of uPA conjugated to iron oxide nanoparticles (NIR830-ATF-IONP)
to specifically bind to uPAR, a cellular receptor highly expressed in many types of
human cancer tissues, including breast cancer.
4.2.2 Material and Methods
4.2.2.1 Cell line
Mouse mammary carcinoma cell line 4T1 was used. Cells were cultured at 37°C
and 5% CO2 in a humidified incubator in Dulbecco’s Modified Eagle’s Medium
supplemented with 10% fetal bovine serum and antibiotics. Cells were harvested at
80% confluence.
4.2.2.2 Preparation of NIR830-ATF-IONP
Recombinant amino-terminal fragments (ATF) were generated using from
pET101/D-TOPO expression vectors containing a mouse ATF of the receptor binding
65
domain of uPA cDNA sequence and expressed in E. coli BL21 (Invitrogen, Carlsbad,
CA). ATF peptides were purified from bacterial extracts with Ni2+ nitrilotri-acetic acid
(NTA)-agarose columns (Qiagen, Valencia, CA) using manufacturer’s suggested
instructions. IR-783 (Sigma-Aldrich, St. Louis, MO) was used to synthesize near-
infrared dye (NIR-830) as described by Lipowska.151 The schematic and spectral
characterization (excitation wavelength: 800nm and emission wavelength: 825nm) of
NIR-830 dye were shown in Figures 4-6A and C. Free thiol groups on the ATF’s were
labeled with NIR-830 dye and then conjugated to amphiphilic polymer-coated, 10nm
magnetic iron oxide nanoparticles (IONPs) via cross-linking of caboxyl groups of the
amphilphilic polymer to the animo side groups of the peptides (Figure 4-6B).
Unconjugated peptides were removed by washing with 100k spin columns for three
times.149
4.2.2.3 Animal preparation
Mouse mammary tumor 4T1 cells (2×106) were implanted into mammary fat pads
of 6-8 week-old BALB/C mice. Tumors were allowed to grow for 6-10 days to a size of
0.5-0.8cm. Animals were anesthetized with a mixture of Ketamine (85mg/kg) and
Xylazine, and were sacrificed using University of Florida Institutional Animal Care and
Use Committee (IACUC)-approved techniques. Strict animal care procedures approved
by the University of Florida IACUC and based on guidelines from the NIH, guide for the
Care and Use of Laboratory Animals were followed.
4.2.2.4 Photoacoustic microscopy imaging system
The schematic of photoacoustic microscopy system is shown in Figure 3-1A. Two
pulsed lasers were used in this study: 1) a tunable Ti:Sapphire laser (LT-2211A, LOTIS
66
TII) with 8-30nanosecond (ns) pulse duration and 10Hz repetition rate for macroscopic
imaging of tumors; and 2) a Nd:YAG laser (NL 303HT from EKSPLA, Lithuania) with
6ns pulse duration and 10Hz repetition rate for microscopic imaging of the blood
vessels. The laser beam was spit and coupled into two optical fiber bundles separately
which were both mounted and adjusted to allow optimal illumination in the imaging area.
Induced photoacoustic waves were collected by a focused ultrasound transducer
(50MHz or 3.5MHz). The 50MHz transducer with 3mm aperture and 6mm focal length
yields 15µm resolution in axial and 60µm resolution in lateral at the focal point. The
3.5MHz transducer (V383, Olympus) with 15mm aperture and 35mm focal length yields
axial and lateral resolution of 200μm and 820μm. The imaging probe, transducer and
optical fiber bundles were mounted on a two-dimension (2D) moving stage. One-
dimensional depth-resolved images (A-line) at each transducer location were acquired
and additional scanning along a transverse direction produced the 2D images referred
to as B-scans. Further raster scanning along the other transverse direction enabled the
reconstruction of 3D images. All photoacoustic images were displayed in maximum
amplitude projection (MAP) form.
4.2.2.5 Near-infrared planar fluorescence imaging system
As shown in Figure 4-7 two laser beams from separate 785nm CW lasers (M5-
785-0080, Thorlabs) were coupled into optical fiber bundle III and optical fiber bundle IV
which were both fixed and adjusted for homogeneous illumination. A high-performance
fluorescent band-pass filter (NT86-381, Edmund Optics) was mounted onto the front of
a fast charge-coupled device camera (CoolSNAP EZ, Photometrics) which was used to
collect the fluorescence signal. All experiments were conducted using the same power
illumination and camera exposure times.
67
4.2.2.6 Image processing
Photoacoustics were reconstructed by a program implemented in Matlab 7.0 and
merged with Amira 5.3.3. Fluorescent images were collected by RS Image (Roper
Scientific, Inc) provided by the manufacturer and processed by Matlab 7.0 and fused
through Amira 5.3.3.
4.2.2.7 Histologic analysis
Tumors collected from mice were preserved in 10% neutral buffered formalin for
10 hours at room temperature. Histological sections were stained with Prussian blue
staining and analyzed using standard procedures to confirm the presence of iron oxide
nanoparticles.
4.2.2.8 Statistical analysis
All data obtained from the experiments were summarized using means ± standard
error of the mean (SEM).
4.2.3 Results
The absorbance of light by IONPs is much stronger compared to blood at
wavelength longer than 650nm. The absorbance of IONPs decreases with increasing
wavelength (Figure 4-8B). Although the strongest absorption for IONPs lies near
650nm, the appropriate wavelength for medical imaging is within the transparent
window of 700-1000nm, which increases the penetration depth. We measured the
photoacoustic signal of NIR830-ATF-IONP between wavelengths of 730-870nm (Figure
4-8B) and our findings agreed well with IONP absorption spectra presented by
Galanzha.150 Tumor-bearing mice were imaged at 24 hours after injection of NIR830-
ATF-IONP and 730nm was chosen as the best wavelength for in vivo photoacoustic
imaging. From the MAP photoacoustic images shown in Figure 4-8A, NIR830-ATF-
68
IONP-targeted tumors displayed the highest absorption at 730nm compared to 800nm
and 850nm. From the quantitative plot shown in Figure 4-8C, the in vivo photoacoustic
signal at 730nm is 21% and 49% greater, compared to 800nm and 870nm, respectively.
Similarly, the absorption of NIR830-ATF-IONP at 730nm was 18% and 47% higher,
compared to 800nm and 870nm, respectively. This in vivo data is consistent with in vitro
signal enhancement studies.
To investigate the feasibility of photoacoustic contrast enhancement with targeted
NIR830-ATF-IONP, we performed in vivo PA imaging of 4T1 mouse mammary tumors
in the following three groups of mice: 1) one group (n=3) received an uPAR targeted
IONP targeting agent (100 pmol NIR830-ATF-IONP, 2) one group (n=2) received a non-
targeted control that conjugated with bovine serum albumin (BSA) IONP (100pmol
NIR830-BSA-IONP) and 3) the third control group didn’t receive IONP injection (n=2).
Nanoparticle probes were injected via the tail vein. As shown in Figure 4-9B, F, J, the
contrast between tumor to normal tissue is low before injection. At 24 hours post
NIR830-ATF-IONP injection the contrast increased significantly while there was no
significant contrast increase using the non-targeted NIR830-BSA-IONP or in animals
without injection (Figure 4-9D, H, L). To verify the delivery of NIR830-ATF-IONP to
tumor cells, NIR fluorescent images were collected before each photoaocustic
experiment. We observed the strongest NIR signal in the tumor site at 24 hours post-
injection for NIR830-ATF-IONP (Figure 4-9A) and the strongest NIR signal for NIR830-
BSA-IONP was in the spleen (Figure 4-9E).
We plotted the photoacoustic enhancement in the tumor as a function of time in
Figure 5A. The photoacoustic signals with NIR830-ATF-IONP increased 3 and 10 times
69
at 5 and 24 hours post injection compared to non-injection control mice. This result
indicates that NIR830-ATF-IONP accumulated in the tumors, as expected. Figure 4-10B
shows the quantitative comparisons of photoacoustic signals in the tumor using different
concentrations of NIR830-ATF-IONP. Compared with non-injection control mice, the
photoacoustic contrast enhancement for 50pmol, 100pmol and 170pmol is 300%,
1000% and 1200%, respectively (Figure 4-10B). The trend of photoacoustic signal
enhancement over concentration was significantly different. Specifically, the
photoacoustic signal increased steeply from 50pM to 100pM and then slowly from
100pM to 170pM. Tumors were resected and imaged by the NIR fluorescent system
(Figure 4-10C). The NIR signal inside the tumor is strong and no NIR signals from the
surrounding tissue or organs.
To evaluate the clinical utility of this technique, the imaging depth was investigated
by adding biological tissues (chicken breast) to the top of the mice skin. Figure 4-11A
(left) shows the MAP of the tumor in mice injected with NIR830-ATF-IONP at 24 hours
post-injection without adding chicken breast. The dashed red line represents the
position of B-scan images shown in Figure 4-11C. Even after adding 31mm of chicken
breast (Figure 4-11A, right), the tumor is clearly seen in Figure 4-11C (d6). As the
imaging depth increased, the signal to noise ratio (SNR) decreased from 40dB to 18dB
as shown in Figure 4-11B.
4.2.4 Discussion and Summary
This work represents the first report on in vivo molecular photoacoustic
tomography of breast cancer with receptor-targeted magnetic iron oxide in a tumor-
bearing animal model. The photoacoustic signal enhancement for NIR830-ATF-IONP
70
was 3 times higher compared to that for non-targeted NIR830-BSA-IONP and 10 times
higher than the non-injection controls.
The uPAR-targeted IONP nanoparticle probe used in these studies address some
of the challenges for the use of targeting-specific tumor imaging agents. Our
methodology utilizes stable and high-affinity targeting ligands and produces strong
imaging contrast for multi-image modalities. From prior extensive studies, human breast
cancer and tumor stromal cells have a higher level of uPA receptor compared with other
normal breast tissues.152-153 In addition, the highest level of uPA receptor expression is
detected in the invasive edge of the tumor region usually enriched in blood vessels
making it accessible for uPA receptor-targeted IONP in this area.154-155 The high-quality
and uniformly sized IONPs used in this study were coated with a thin ampiphilic
copolymer and have a relatively small particle complex (18nm) which is more suitable
for in vivo delivery of the imaging probe compared with other receptor-targeted or non-
targeted nanoprobes for photoacoustic imaging. It also has been shown that polymer-
coated IONPs have more than 8 hours of plasma retention time compared with other
molecular imaging agents for PAT, which are usually less than 2 hours. The delivery of
NIR830-ATF-IONP to the tumor was verified by planar NIR fluorescent imaging. This
shows the potential for us to combine fluorescence molecular tomography (FMT) with
photoacoustic tomography. While FMT has lower spatial resolution than PAT, it has
higher specificity than photoacoustic imaging techniques and the combination of PAT
and FMT would provide complementary information for breast cancer detection.
The demonstrated imaging depth of 31mm indicates the potential of our method
for tumor imaging in humans. While the photoaocustic system as described here is not
71
yet suitable for clinical applications, it can be improved to be a clinical prototype using a
commercial ultrasound array and/or a high frame rate of laser pulses with a high-speed
scanning system.
By using high-frequency focused ultrasound transducer (>50MHz), the
microvasculature inside and around the tumor can be imaged as shown in Figure 4-9,
indicating the potential of understanding the interplay between the tumor
microvasculature and delivery of targeted contrast agents. We are also investigating
multi-wavelength and quantitative PAT reconstruction methods to calculate functional
parameters and nanoparticle concentrations in tissue. This will allow us to quantitatively
study the delivery mechanism of NIR830-ATF-IONP through microvasculature to tumor
tissue.
4.3 Photoacoustic and Fluorescence Tomography of HER-2/Neu Positive Ovarian Cancers Using Receptor-Targeted Nanoprobes In An Orthotopic Human Ovarian
Cancer Xenograft Model
4.3.1 Motivations
Ovarian cancer has a low survival rate due to the lack of specific early
symptoms.156-157 Although prophylactic oophorectomy (PO) can reduce the risk of
ovarian cancer by more than 50% in BRCA mutation carriers, PO also increases the
mortality rate in women undergoing this procedure prior to the age of 45.158 Several
imaging diagnostic technologies developed in the past decades for early detection of
ovarian cancers has some limitations due to their low specificity and sensitivity.159-161
More recently, the use of imaging probes that specifically target ovarian cancer showed
promises for enhanced sensitivity and specificity in tumor imaging. Positron emission
tomography (PET) and single photon emission tomography (SPECT) using targeted
cancer probes with high specificity and sensitivity have been used clinically to detect
72
cancers. However, their resolutions are generally poor and they are not able to
accurately locate tumors.162-163
Fluorescence molecular tomography (FMT) is an inexpensive and fast optical
imaging modality that has been widely used in molecular imaging. FMT is able to
sensitively detect low molecular concentration within the tissue. In contrast to planar
fluorescence molecular imaging, it provides depth information. However, similar to PET
and SPECT, FMT has poor spatial resolution.164
Photoacoustic tomography (PAT) is a hybrid imaging method that combines the
rich optical contrast with the high resolution of ultrasound imaging. PAT detects wide-
band acoustic waves that are generated through transient thermoelastic expansions
due to the absorption of short pulses by biological tissue. PAT has been used in
vascular biology, ophthalmology, dermatology, gastroenterology, and breast imaging.
A variety of PAT contrast agents have been used to image brain tumors,
angiogenic processes, melanomas, lymph nodes, and the cerebral cortex. Recently,
Ekaterina et al. demonstrated the use of targeted IONP as a contrast agent to detect
circulating tumor cells using the photoacoustic effect. Here, we firstly report the use of a
new near-infrared dye labeled HER-2/neu-specific affibody conjugated IONP (NIR830-
ZHER2:342-IONP) for PAT and FMT imaging of orthotopic ovarian cancers in a human
ovarian cancer xenograft model in nude mice. This nanoparticle probe provided
photoacoustic signal enhancement and fluorescent signal as well as tumor-targeting
capabilities.
73
4.3.2 Materials and Methods
4.3.2.1 Cell lines
The human ovarian cancer cell line SKOV3, stably expressing a firefly luciferase
gene (SKOV3-Luc) was kindly provided by Dr. D. Matei at Indiana University Purdue
University at Indianapolis (IUPUI). SKOV3-luc cells were cultured in McCoy's 5A (Cat #
10-050-CV, Cellgro, Mediatech Inc. VA, USA) supplemented with 10% fetal bovine
serum (FBS-Cat # SH30396.03, Hyclone from Thermo scientific) and 1% penicillin
streptomycin (Cat # SV30010, Hyclone, Logan, Utah). Cultured cells were maintained at
37◦C in a humidified atmosphere containing 5% CO2. Growth medium was replaced
every other day. When appropriate, cells were detached with trypsin (Cat # 25-050-C1,
Cellgro, Mediatech Inc. VA, USA) and transferred to single cell suspensions.
4.3.2.2 HER-2/neu specific affibody conjugation to IONP
A histidine-tagged HER-2-specific affibody (His6-ZHER2:342-Cys), was provided by
Dr. J. Capala at the National Institute of Biomedical Imaging and Bioengineering
(NIBIB), Bethesda, Maryland. The NIR dye, NIR830 maleimide, was treated with Tris (2-
carboxyethyl) phosphine (TCEP) for 30 minutes and then conjugated with HER-2/neu-
specific affibody. Amphiphilic polymer-coated IONPs were activated by treatment with
ethyl-3-dimethyl amino propyl carbodiimide (EDAC) and sulfo-N-hydroxysuccinimide
(sulfoNHS). Activated IONP were then incubated with the NIR-830-conjugated ZHER2:342-
afiibody The resultant ZHER2:342 -NIR830-IONP were purified using Nanosep 100Kgel
filtration columns (Pall Corp, Ann Arbor, MI).
4.3.2.3 Orthotopic human ovarian cancer xenograft model
The ovarian tumor model was established by injecting 2 x 107 SKOV3 luciferase
gene positive human ovarian cancer cells into the ovaries of 6-8 week old female
74
athymic nude mice (Harlan). All mouse surgical and imaging procedures were approved
by University Institutional Animal Care and Use Committee. Mice were imaged when
orthotopically xenografted tumors reached around 5mm in size, usually 4-6 weeks after
tumor cell injection. Tumor growth was monitored and quantified weekly using a
bioluminescence imaging system (Caliper Life Sciences, Hopkinton, MA).
4.3.2.4 Fluorescence molecular tomography imaging System.
The photoacoustic system is the same as that used in Figure 3-1A. The handheld
optical fiber based FMT system is shown in Figure 4-12D. In this system, a continuous-
wave (CW) 785nm laser (M5-785-0080, Thorlabs) was mounted in a 2D linear stage
(Model 17AMA045, CVI MellesGriot). A convex lens was used to focus the laser into
each source fiber in the array. For each source illumination, a 1024×1024 pixel CCD
camera (Princeton Instruments, Trenton NJ) equipped with a high-performance band-
pass filter (Thin Film Imaging Technologies, Greenfield, MA) was used to collect the
emission light from the detection interface. Optimal exposure time was determined for
each experiment and binning of 4×4 pixels was used to improve the signal-to-noise
ratio (SNR). A photograph of the imaging probe with 10 sources and 15 detectors is
shown in Figure 4-12D. Fluorescence images were reconstructed using an iterative
finite element-based algorithm, described by Zhao et al..165
4.3.2.5 In vivo and ex vivo planar near infrared fluorescence imaging
Planar NIR fluorescence imaging was performed using a Kodak in vivo imaging
system Fx (Carestream Health Inc., New Haven, CT) to demonstrate specific
accumulation of the targeted IONPs in the ovarian tumors and anatomic localization of
the tumors. All images were captured using an 800nm excitation and 850nm emission
75
filter set with a 180s exposure time and a value of 0.2. For each optical image, a
corresponding radiographic image was taken to establish the anatomical location of the
tumor.
4.3.2.6 Histological analysis
After imaging, mice were sacrificed and tumors were excised and fixed with 10%
buffered formalin. Paraffin tissue sections were stained with Prussian blue or
hematoxylin and eosin (H&E). Images were acquired at 20Xmagnification using a Zeiss
Axioplan 2 upright microscope.
4.3.2.7 Statistical analysis
All data obtained from the experiments were summarized using means ± standard
error of the mean (SEM).
4.3.3 Results
To evaluate our latest PAT or FMT tomographic systems for imaging tumors,
human ovarian tumor xenografts were generated orthotopically in the nude mice.
Specifically, SKOV3-luc cells were injected into the mouse ovaries (Figure 4-13A) and
tumor growth was monitored by bioluminescence imaging (Figure 4-13B). Planar NIR
fluorescence images paired with X-ray were taken 24 and 48 hours after tail vein
injection of 400 picomoles of NIR830-ZHER2:342-IONPs to show the optical signal in the
luciferase positive tumor location. It seems that the fluorescence intensity reached to
the strongest level in the ovarian cancer at 48 hrs following systemic delivery (Figure 4-
13C), indicating that the nanoprobes specifically accumulated in the ovarian tumor.
Prussian blue staining of paraffin tumor tissue sections revealed the presence of tumor
cells with blue iron staining, indicating the internalization of IONPs in the tumor cells
(Figure 4-13D).
76
The absorption spectra of IONP for photoacoustic enhancement is shown in
Figure 4-12B. The IONP absorption decreases as the wavelength increases, with the
highest absorption occurring at 650nm. For clinical applications, the optimal operational
wavelength lies within the tissue transparent window of 700-1000nm. The absorption of
nanoprobes in photoacoustic phantom studies (black dots, Figure 4-12B) agrees well
with the absorption spectra of IONP. Three groups of mice were imaged at 730nm,
800nm and 870nm and HER-2/neu-targeted ovarian cancer nanoprobes (NIR830-
ZHER2:342-IONP) had the highest photoacoustic signal (left image in Figure 4-14A), in
contrast to surrounding normal tissues. In contrary, there was no significant
photoacoustic signal enhancement with non-target bovine serum albumin- NIR-IONP
(NIR830-BSA-IONP) (middle image in Figure 4-14A) or in non-injection controls (right
image in Figure 4-14A). Data from quantitative plots shown in Figure 4-14B revealed
that the strongest photoacoustic signal of IONPs accumulated in the ovarian tumors
was at 730nm, which was 500% higher than non-injection controls and 300% greater
than signals from the mice injected with non-targeted BSA.
In clinical applications, ovarian tumors are often located deeply inside the low
peritoneal cavity. To evaluate the use of this technique in tumor imaging in humans, we
examined the capability of detection of tumors in the deep tissue using PAMT by adding
different thicknesses of chicken breast tissue placed on the top of mouse back. The top
row in Figure 4-15A-D shows the B-scan images with the addition of different
thicknesses of chicken breast. The bottom row (Figure 4-15A-D) shows the MAP
images with chicken breast additions. As shown, the ovarian tumors are clearly
imageable even after the addition of 19mm thick of exogenous tissues. We also
77
investigated the imaging depth of FMT (Third and fourth rows in Figure 4-16A-C) and
found an imaging depth of up to 10mm.
We compared the spatial resolution of FMT and PAMT imaging at different depths
(Figure 4-16). The 3D and 2D images of PAMT with different thicknesses of chicken
breast are shown, as well as results with FMT. As shown, the axial and lateral
resolutions of FMT are relatively poor. Although the position of image center lies in the
right depth, the axial and lateral sizes are much larger than actual sizes. In contrast, the
lateral and axial resolutions of PAMT are 820µm and 200µm in our system. Therefore,
for PMAT, the reconstructed lateral and axial sizes are very close to real sizes. The
lateral and axial reconstructed sizes of PAMT and FMT were plotted (Figure 4-17) and
the average lateral and axial FWHM of the tumor in PAMT at three different depths (3, 6
and 10mm) were 5mm and 3mm, which are comparable to the size of the resected
tumor. Using FMT, the lateral FWHM was 5 to 13mm and axial FWHM was 3mm to 14
mm, as the depth increased.
4.3.4 Discussion
In this report, we demonstrated the feasibility of using PAMT and FMT to detect
ovarian tumors in mice after systemic delivery of HER-2/neu targeted NIR830-ZHER2:342-
IONP. Unlike gold nanoparticles, non-targeted IONP has been approved for pilot studies
on humans and is already used as a contrast agent for magnetic resonance imaging
(MRI). NIR830-ZHER2:342-IONP is a novel receptor-targeted nanoprobe that specifically
binds to ovarian tumors in vivo, as detected by planar NIR fluorescence imaging and
histology.
Planar NIR fluorescence imaging can monitor the delivery of nanoprobes and relay
accurate tumor position and size information. However, it cannot map deeply positioned
78
ovarian tumors because of poor spatial resolution and a lack of depth information. In
this study, our handheld FMT probe showed the potential to noninvasively map ovarian
tumors in 3D at depths up to 10mm. The method was fast, taking less than 2 minutes to
cover a 20×20mm imaging area. However, the poor axial and lateral resolutions affect
the accuracy, which limits the future applications in human patients. As a
complementary method to FMT, PAMT was employed to detect ovarian tumors with
contrast enhancement from IONP. PAMT had high axial and lateral resolutions of
200µm and 820µm, respectively, as well as imaging depth capabilities up to 19mm.
However, the system is not portable and image acquisition is slow (20 minutes to cover
18×18mm2 area). For these reasons, PAMT is not yet suitable for imaging large areas.
Here, we show a fast handheld FMT system can be used to mark areas possibly
containing ovarian tumors, and then use PAMT to image the marked area and
accurately map the position of ovarian tumors.
Our study has some limitations. The slow and bulky photoacoustic imaging system
is not suitable for clinical application. However, this is not a fundamental problem of
PAMT and could be overcome using a commercial ultrasound array and/or a high frame
rate of laser pulses with a high-speed scanning system. Although our handheld FMT
system is fast and portable, it offers limited imaging depth capabilities (10mm). For
tumors positioned more than 10mm under the skin surface, our FMT system is not
applicable. However we are currently investigating the use of high compact optical fiber
bundles to make a small FMT probe containing more sources and detectors. From our
simulation and phantom tumor detection experiments, we were easily able to image up
79
to 25mm away from the detector. A goal is to integrate these two systems into one
handheld probe.
In summary, noninvasive in vivo FMT and PAMT mapping of ovarian tumors after
systemic delivery of NIR830-ZHER2:342-IONPs was successfully accomplished in an
orthotopic human ovarian tumor xenograft model in the mice.
In this Chapter, we discussed about three applications of cancer research using
ARPAM. All these three applications are cancer treatment related biological research.
From our experience, it is difficult to translate ARPAM to clinical application such as
image-guided surgery that is very important for surgeons to improve the success rate of
surgeries. In next Chapter, I will introduce a MEMS-based intraoperative photoacoustic
probe for image-guided surgery.
80
Figure 4-1. Comparative analysis of scAAV2-mediated transduction of MMECs. Cells were infected with the WT or the triple-, quadruple- or septuplete-tyrosine mutant scAAV2-hGFP vectors at a of multiplicity infection (MOI) of 10,000 vgs/cell. Transgene expression was detected by fluorescence microscopy, 72 hr post-infection. A) Representative images of hGFP expression in MMECs infected with different tyrosine-mutant scAAV2-hGFP. B) Quantitative analysis of AAV2 transduction efficiency in MMECs. hGFP expression is shown in arbitrary units calculated by multiplying the percentage of positive cells by the mean fluorescence intensity in each sample. Each value represents the average of three samples based on 10,000 counted cells.
81
Figure 4-2. scAAV2 encoding siRNAs against UPR proteins. A) shows a scAAV2
construct with smCBA driving expression of siRNAs against mice UPR proteins. B) shows the sequences of the oligonucleotides encoding siRNAs against mice IRE1α, XBP-1 and ATF6. C) Transduction efficiency of scAAV2 sept mut vectors in MMECs. Representative images of mCherry expression in MMECs infected with scAAV2 sept mut vectors at three days post infection (left column) and a comparison of mCherry expression at different MOIs as indicated shown in arbitrary units calculated by multiplying the percentage of positive cells by the mean fluorescence intensity (right column). Each value represents the average of three samples (eight pooled wells of a 24-well plates/sample), based on 10,000 counted cells. UnST=unstimulated; scr=scrambed siRNA scr=scrambed siRNA
82
Figure 4-3. Pro-angiogenic and survival role of the UPR proteins on endothelial cells. MMCEs were co-cultured with NeuT EMTCL2 for 48 hr. A) MMECs were cultured between two layers of Matrigel for 48 hr. Morphometric analysis of the degree of tubule formation was then performed. B) Representative photomicrographs of microscope fields showing tubule formation. C) Annexin V-propidium iodide-positive cells were shown (top right, late stage apoptosis; bottom right, early apoptosis). D) Cell apoptosis was expressed as a percentage of apoptotic cells in the total cell population. E) The cell proliferation was assessed by crystal violet staining.
83
Figure 4-4. Serial photoacoustic imaging of the developing tumor vasculature and quantitative
analysis for mice breast cancer xenograft models. Mice breast cancer xenografts were established by subcutaneous injection of 1×105 NeuT or NeuT EMTCL2 cells into the mammary fat pads of the mice. Serial PA imaging was performed on the same tumor inoculation site on day 3, 5, 7 and 9 post tumor inoculations. A) Representative serial PA images of NeuT model (top panel) and NeuT EMTCL2 model (bottom panel). B) Entropy extraction for change in the vessel density over different time points as indicated. C) Comparative analysis of normalized vessel volumes at different time point as indicated. D) Tumor volume was calculated from daily caliper measurements of the large A and smallest B diameters of each tumor using formula a×b2×0.4. Representative images of H&E staining for breast cancer tissue sections of NeuT (top) and NeuT EMTCL2 (bottom) xenograft models.
84
Figure 4-5. Knockdown of the siRNAs against the UPR protein resulted in decreased tumor growth and
tumor vasculature in mice breast cancer xenografts. Mice breast cancer xenografts received intratumoral scAAV2 sept-mutant vector-delivered siRNAs against IRE1α or XBP-1 or ATF6. PA imaging was performed on the same tumor inoculation site on day 3, 5, 7 and 9 post tumor inoculations. A) Representative PA images of different treatments: the scrambled siRNA (top panel), IRE1α siRNA (second panel), XBP-1 siRNA (third panel) and ATF6 siRNA (bottom panel). B) Entropy extraction for change in the vessel density over different time points as indicated. C) Comparative analysis of normalized vessel volumes at different time point as indicated. D) Tumor volume was calculated from daily caliper measurements of the large A and smallest B diameters of each tumor using formula a×b
2×0.4.
85
Figure 4-6. Illustration of NIR-830 dye and ATF-conjugated IONP probe. A) Schematic
of NIR-830 dye. B) NIR 830 dye is conjugated to mouse ATF peptide through a bond between maleimide esters and free thiol groups of cystidine residues of the the peptide. Dye-labeled peptides were then conjugated to carboxyl group of the polymer coating on the IONPs. C) Resulting optical probe has an Ex 800nm and Em 825nm.
Figure 4-7. Schematic of NIR fluorescence imaging system.
86
Figure 4-8. In vivo and in vitro test of targeted nanoprobe using different wavelengths. A)
In vivo MAP photoacoustic images of tumors using signals of 730nm, 800nm and 870nm, 24 hours after injection of NIR830-ATF-IO. B) Absorption characteristics of NIR830-ATF-IO, IO and blood. C) PA signals using different wavelength inputs.
87
Figure 4-9. In vivo photoacoustic MAP and fluorescence images before and after
injection. Micrographs were merged with fluorescence images taken 24 hours post injection with indicated agent (A, E, I). Panels B thru L. Photoacoustic MAP images were merged with images of blood vessels before injection (B, F, J), and at 5 hours (C, G, K) and 24 hours post injection (D, H, L).
88
Figure 4-10. Quantitative plot and comparison of photoacoustic and fluorescent signals.
A) NIR fluorescence imaging of mice with resected tumor (top panel), cross-section of tumor along black dashed line (middle panel) and section stained with Prussian blue (bottom panel). B) Quantitative plot of fluorescence signals before and after injection of NIR830-ATF-IO. C) Quantitative plot of photoacoustic signals before and after injection of NIR830-ATF-IO. D) Comparison of fluorescence signals with different concentrations of injected NIR830-ATF-IO. E) Comparison of photoacoustic signals with different concentrations of injected NIR830-ATF-IO.
89
Figure 4-11. In vivo photoacoustic images with adding chicken breast between detector and tumor. A) MAP image of tumor at 24 hours after injection without adding chicken breast (left) and photograph of 31mm thick chicken breast (right). B) SNR versus depth. C) B-scan images by adding different thickness of chicken breast
90
Figure 4-12. Schematic and spectrum of targeted nanoprobe and FMT system
description. A) Schematic representation of HER-2 targeted ZHER2:342-NIR-830 dye-IONP: NIR-830 dye was conjugated to cysteine residues of HER-2 specific affibody . The ZHER2:342-NIR-830 complex was conjugated to the C-terminal ends of amphiphilic polymer-coated IONP’s (10nm in diameter). B) Photoacoustic absorption spectra of IONP, NIR830-IONP and blood from 650nm to 900nm are shown. C) Absorption and emission spectra of NIR830. The highest absorbing wavelength was 800nm and the strongest emission wavelength was 825nm. D) Schematic of the hand-held FMT system.
91
Figure 4-13. Multi-mode in vivo imaging and histological validation. A) SKOV3-luc cells
were injected orthotopically into mouse ovaries. B) Bioluminescence imaging showed the tumor location. C) Examination of target specificity of ZHER2:342-NIR-830 dye-IONPs to ovarian tumors in vivo. In these experiments, SKOV3-luc tumor-bearing mice were injected with NIR830-ZHER2:342-IO conjugates via tail veins and optical imaging was performed 24h and 48h post-injection. The NIR signal scale was generated using the Kodak molecular imaging software. Anatomical X-ray images were simultaneously captured to validate tissue position. D) Pathological analysis of xenograft tumors. All tumors were formalin-fixed, paraffin-embedded, sectioned, and stained for analysis and photography. Representative images demonstrate the tumor tissue and presence of IO nanoparticles, stained by Prussian blue inside the tumor tissue (yellow arrows).
92
Figure 4-14. Quantitative analysis based on different wavelengths. A) MAP
photoacoustic images of ovarian tumors injected with actively-targeted NIR830-ZHER2:342-IONP (left), non-targeted NIR830-BSA-IONP (middle) at 730nm, 48 hours after injection and without injection/control case (right). B) Quantitative comparison of PA signals of actively-targeted (HER-2-affibody), non-targeted (BSA) and non-injection cases at 730nm, 800nm and 850nm.
93
Figure 4-15. Image depth ability evaluation. B-scan (top row) and MAP (bottom row) PA
images of actively-targeted tumors with A) 0mm, B) 4mm, C) 10mm, D) 19mm exogenous chicken breast tissues added between tumors and the detector. Scale bar 1mm, white arrows indicate the chicken breast surface.
94
Figure 4-16. 3D results for both PAMT and FMT. 3D PAMT images (top row), B-scan
PAMT images along white dashed lines (second row) and 2D slices along white dashed lines (third row) in 3D FMT images(bottom row) with actively- targeted imaging agent and additional A) 3mm, B) 6mm, C) 10mm chicken breast tissue.
95
Figure 4-17. Quantitative comparison of spatial resolution between photoacoustic and fluorescence molecular tomography images with increased imaging depth. A) Graph shows results for three 1D profiles taken from along the brown dashed lines in the second row in Fig. 5 A-C. B) Graph shows results for three 1D profiles taken from along the yellow dashed lines in the third row in Fig. 5a-c. C) 1D PA profiles along Z-axis through tumor center at depths of 3mm, 6mm and 10mm. D) 1D FL profiles along Z-axis through tumor center at depths of 3mm, 6mm and 10mm.
96
CHAPTER 5 INTRAOPERATIVE PHOTOACOUSTIC TOMOGRAPHY
5.1 Motivations
In a lumpectomy, the surgeon removes the tumor along with some healthy, non-
cancerous surrounding breast tissue. The outside of the tumor, or margin, is typically
examined using pathological methods to determine if the border of a tumor mass is
more than 2mm away from the surrounding normal breast tissues; this is considered to
be a negative margin. However, if the tumor cells come to the edge of the resected
tissue, it is considered to be a positive margin.166-167
Methods currently available for intraoperative margin assessment, such as gross
examination, frozen section, ultrasound, and touch-prep, have various limitations with
false-negative diagnoses in 20 to 50% of the patients. Therefore, there is an urgent
need to develop sensitive and accurate intraoperative methods for detecting tumor
margins in order to reduce tumor recurrence and to improve the survival rate of breast
cancer patients.168-169
Optical-based technologies appear to be ideal for intraoperative imaging as they
can be miniaturized, and are inexpensive, fast and sensitive. In recent years, an
increasing number of optical studies using elastic scattering spectroscopy (ESS),
Raman spectroscopy, Optical coherence tomography (OCT) or diffuse reflectance
spectroscopy (DRS) have been carried out for tumor margin assessment. However,
these methods are surface weighted which cannot provide depth information and their
maximum sensing depth is limited from 300μm to 2mm. Laser-induced photoacoustic
imaging, where a single short-pulsed light beam illuminates an object and the
photoacoustic waves excited by thermoelastic expansion are measured using wideband
97
ultrasound transducer(s), retains the desired high optical contrast/sensitivity while
providing much improved spatial resolution (30~50μm, adjustable with ultrasound
frequency) than pure optical tomographic methods by detecting less-scattering
ultrasonic waves. The penetration depth for our iPAI system is ≥ 2mm which is excellent
for tumor margin assessment.
The purpose of this study is to evaluate in vivo high resolution imaging of tumor
margins and resection inspection by using iPAI system. We demonstrate this technique
using a tumor-bearing animal model. The results obtained show accurate correlation
with the histology.
5.2 Materials and Methods
5.2.1 Animal Protocol
Approval for animal studies was obtained from, and strict animal care procedures
followed for all experiments that were set forth by, the Institutional Animal Care and Use
Committee (IACUC), as based on the guidelines from the NIH guide for the Care and
Use of Laboratory Animals. Human breast carcinoma cells (MDA-MB-231-luc-D3H2LN,
Caliper Life Sciences) were implanted into the mammary fat pad of 6 week-old female
nu/nu mice (Harlan Laboratories). Tumors were allowed to grow for 6 days prior to
imaging experiments to a size of ~0.5cm. Animals were anesthetized via intraperitoneal
injection for imaging and surgical procedures, and were subsequently sacrificed using
IACUC approved techniques.
5.2.2 Intraoperative Photoacoustic Tomography System
In our microelectromechanical system (MEMS) based photoacoutc imaging
system (Figure 5-1), short laser pulses of 6ns duration are generated from a Nd:YAG
532nm laser (NL 303HT from EKSPLA, Lithuania), and a reflection mirror is employed
98
to deliver light into a miniaturized probe (Figure 5-2A). Inside the probe, there are an
aperture, a wedge-shaped mirror (Figure 5-2C) and a MEMS mirror (Figure 5-2B). A
custom designed ultrasound transducer (unfocused ring-shape, 5.5MHz central
frequency, Resource Center for Medical Ultrasonic Transducer Technology, University
of South California) is attached to the end of the probe to detect photoacoustic waves.
The structural design, dimensions and performance of the transducer are presented in
Figure 5-3. Through the ultrasound transducer, the photoacoustic waves are converted
to electrical signals which are amplified, filtered, and fed into a computer through a data
acquisition card (Figure 5-1).
A MEMS mirror based on electrothermal bimorph actuation is utilized to realize
two-dimensional light scanning. The moveable mirror plate is as large as
0.8mm0.8mm in a 2mm 2mm device footprint. The MEMS mirror is fixed on a ramp
with an angle of 22.5° (Figure 5-4A), allowing light to scan from position 3 to position 4
(Figure 5-4B), while the MEMS mirror moves from position 1 to position 2. The
measured vertical resonant frequency is 500Hz. In Video 1 online, four channels of
voltage signals are utilized to drive four actuators of the MEMS mirror.
The probe is miniaturized through the integration of MEMS mirror and transducer.
In the current design, the probe diameter is 12mm but further size reductions are
achievable. Our miniaturized probe works effectively and stably in the front-view and
reflection mode, which is easier for a surgeon to operate during surgery than a probe in
the side-view or transmission mode.
To determine the axial and transverse resolution versus target position, we imaged
a 100µm diameter tissue mimicking target (absorption coefficient a =0.049mm 1 ,
99
reduced scattering coefficient '
s =0.5mm 1 ) embedded in turbid media
( a =0.012mm 1 , '
s =0.35mm 1 ). The target can be treated as an ideal line target
because the central frequency of the transducer is much longer than the diameter of the
target. Chicken breast with thickness of up to 2.3mm was added to investigate the
resolution and signal-noise ratio versus target position.
5.2.3 Imaging Procedure
The capability of our iPAI system was qualitatively and quantitatively validated by
mouse experiments involving tumor extirpation (n=5). The mouse was anesthetized
during the experiments. The laser beam illuminated the tumor surface along Z axis at an
energy density of 8mJ/cm2 that is lower than the American National Standards Institute
safety limit of 20mJ/cm2. Two 0~4V ramp signals with repetition frequency of 0.1Hz
were used to drive two actuators moving along Y axis and the other two signals with the
same magnitude but different repetition frequency of 0.002Hz were used to drive the
actuators moving along X axis. The photoacoustic signals were recorded without
averaging at each scanning point, amplified and converted to a one-dimensional (1D)
image assuming a constant ultrasound velocity in soft tissue (1.54mm/μs). A square
area (10mm10mm) in XY plane was imaged to produce a 3D image of the tissue
volume.
5.2.4 Histology
Each tumor was cut into two pieces by a sharp blade along Z axis for H&E stain in
XY plane and along X or Y axis for H&E stain in XZ or YZ plane. Tumors were
harvested and preserved in 10% neutral buffered formalin for 10 hours at room
temperature. Histological sections were stained with hematoxylin-and-eosin and
100
prepared by standard procedures. Aperio ImageScope V10.2.2.2352 was used to
microscopically image the slices and measure the actual size.
5.3 Spatial Resolution of the System
Figure 5-5A shows the Hilbert transform of a typical A-line acoustic signal and
Figure 5-5C shows the target’s transverse point spread function. The -6dB width of the
signals shown in Figure 5-5A and Figure 5-5C present the axial and transverse
resolutions. In Figure 5-5B, the system’s axial and transverse resolutions are plotted
versus the target’s position by blue (0.36mm~0.4mm) and red (0.24mm~0.96mm)
dashed lines. The signal-to-noise ratio (SNR) measured in turbid media is shown in
Figure 5-5D, decreasing from 32dB to 18dB within 2.3mm.
5.4 Phantom Experiments
As shown in Figure 5-6A, the pencil lead was embedded 1.2mm below the surface
of a 50mm-diameter cylindrical solid phantom, consisting of Intralipid as the scatterer
and India ink as the absorber. The phantom had10.007a mm (absorption coefficient)
and 11.0s mm (reduced scattering coefficient). Agar powder (2%) was used to
solidify the Intralipid/ink suspensions. Figure 5-6B presents a typical A-line signal
collected (blue line), along with its amplitude after the Hilbert transform (red line), while
Figure 5-6C shows the 2D image of the object. Using the criteria defined by the FWHM
of multiple A-line signals, the size of the recovered object is estimated to be 0.72mm
compared to the actual object size of 0.70mm.
Figure 5-7A and B gives the photograph and recovered image of another phantom
where multiple targets were embedded at different positions/depths. From the image
shown in Figure 5-7B, we see that the targets are all detected.
101
The results given in Figure 5-8 show the three dimensional imaging ability of our
system. In this experiment, a pencil lead with a diameter of 0.7mm was embedded in
one piece of chicken (Figure 5-8A) at a depth of 1.5mm from the surface. The recovered
3D images are presented in Figure 5-8B-D (at coronal, sagittal- and cross-section
views) and in Figure 5-8E (3D rendering of the recovered image). Again, the object is
accurately imaged in terms of its size, shape and location.
5.5 Human Blood Vessels Experiments
Here we demonstrate the in vivo imaging ability of our MEMS-based photoacoustic
imaging system for visualizing blood vessels in a human hand. The photograph of the
hand and the barely visible blood vessels under the skin are shown in Figure 5-9A,
while the recovered 3D image of the blood vessels is given in Figure 5-9B. We can see
that the blood vessels are clearly imaged with correct shape and size for most part. We
also note that some middle portions of the blood vessels (indicated by arrow in Figure 5-
9B) are notably distorted or missing. This is most likely attributed to the fact that these
portions of the blood vessels are physically located deeper over other parts of the blood
vessels of interest and that the sensitivity of the transducer responsible for this part of
the imaging area is lower due to the limited directivity of the transducer used. It is noted
that the shape of the vessels is not round likely due to the non-linear effect of the MEMS
mirror. We are seeking ways to resolve this issue by using a proper calibration method.
5.6 Preclinical Evaluation of Intraoperative Photoacoustic Tomography
The volumetric shape, size and position of the tumor are shown in Figure 5-10B.
The surgeon removed the tumor marked with “T” in Figure 5-10B. In these experiments,
the line of surgical resection of tumor was intentionally carried out beyond the margin of
the imageable tumor, consistent with clinical practice. After the surgical procedure, the
102
surgical area shown in Figure 5-10C was inspected again by the same imaging method
to ensure the complete resection of tumor. The image shown in Figure 5-10D confirmed
the completed resection of tumor. The signal (indicated by the arrow in Figure 5-10D)
was from the inner surface of the ring-shaped transducer. This signal disappeared in
Figure 5-10B due to strong signals from the tumor and normalized image procedure
was used.
Following tumor removal, the tumor size as determined by iPAT was compared
with the actual tumor size as measured histologically. H&E stained sections along XZ
plane were obtained after the photoacoustic experiments. A 2D PAT slice along the red
dashed line in Figure 5-10B as selected and compared with the corresponding H&E
stained section (the black dashed line in Figure 5-10B) as shown in Figure 5-11A and B.
Figure 5-11C shows the photograph of the tumor in comparison with the H&E and PAT
slices. Also, the red dashed line along the top margin of the tumor shown in Figure 5-
11A matches well with the top margin (black dashed line in Figure 5-11B) of the H&E
section and photograph (red dashed line in Figure 5-11C). The size along two typical
directions (Lines 1 and 2 in Figure 5-11A) was estimated to be 2.1 and 1.2mm,
respectively, in excellent agreement with the actual dimensions of 2.2 and 1.1mm
measured from the histology slice (Lines 1′ and 2′ in Figure 5-11B). Such comparisons
were applied to other two H&E staining sections along XZ plane for this mouse, and we
found that the measured error was less than 9.1% for all three slices.
In another independent mouse tumor experiment, photoacoustic imaging
demonstrated tumor nodules as shown in Figure 5-11D. The tumor size imaged by PAT
along lines 1, and 2 (Figure 5-11D) was 2.5mm, 1.9mm, showing an accurate
103
measurement of the actual size of 2.5mm, 2.1mm, measured by histology along lines 1′
and 2′ (Figure 5-11E). In this case, the largest observed error for the remaining two H&E
sections was 9.5%.
Photoacoustic slices with histological correlations from the third mouse are shown
in Figure 5-12A-C. The structures indicated by the arrow in Figure 5-12B are consistent
with each other. After quantitative analysis, the largest observed error was 6.5%. Figure
5-12D-F show the PAT slices and histological correlations for the fourth mouse where
we added 1.1mm-thick chicken breast on top of the tumor to simulate the real specimen
resected from the human breast. The 3D top margins of the tumor were clearly seen in
the PAT images and consistent with the H&E sections. The white dashed line shows the
surface of the chicken breast. The error between the PAT image and H&E sections was
found to be less than 16.5%. A 1.8mm-thick chicken breast was added to the top of the
fifth mouse and the PAT images with the associated H&E section are shown in Figure
5-12G-I. The tumor shape imaged by PAT was close to that from the H&E sections but
the error increased to 21.5% in this case.
5.7 Discussion
We have demonstrated that our iPAT technique with miniaturized MEMS probe
can be effectively used for mapping tumors three-dimensionally and for inspecting
completeness of tumor resection in small animals. The qualitative and quantitative
results obtained are significant for several reasons. First, iPAT does not use any
radioactive tracer or intravenous contrast. Second, the existing PAT techniques are not
directly applicable to intraoperative imaging due to their bulky size or inflexibility of the
imaging probe. Our MEMS-based probe breaks the limitation with its novel compact
104
design. Third, after tumor resection, a surgeon can inspect the tumor cavity wall to
ensure completeness of the procedure. Finally, iPAT has the potential to realize real-
time image guidance and allow acquisition of functional parameters of tumor when a
multispectral laser with high repetition frequency is used.
While our iPAT system has been demonstrated to be practical for intraoperative
tumor imaging, several challenges still remain. First, a handheld probe is necessary for
a surgeon to operate during surgery. In the current study, our probe was fixed in
position using a holder and an aperture was used to generate a small light spot. This
set-up would not be applicable to real operative procedures. In the future, we will use an
optic fiber coupled with a micro-optical lens to replace the aperture and make the probe
handheld. Second, imaging speed and area of our current iPAT system is limited by the
10Hz repetition frequency of the laser source currently available in our lab.
Considerably faster lasers (up to 5kHz) are now commercially available. Since the
resonant frequency of the MEMS mirror is up to 500Hz, the imaging speed will be
reduced to 5 s when a laser having a repetition frequency of 500Hz is used. With this
improvement, the probe can be moved easily to get a 70mm 70mm within 5 minutes.
Third, the probe was submerged in a tank filled with water which is not applicable to
clinical procedures. We plan to make a MEMS probe with a water- or ultrasound gel-
filled balloon attached to the surface of transducer which will then be used for clinical
intraoperative tumor imaging. (Figure 5-13). Finally, the bleeding problem during a
clinical surgery should be considered because the absorption of blood on 532nm is very
strong. During the mouse experiments, we used cautery to stop the bleeding and
normal saline to wash out the surgery area after tumor resection. For translating this
105
technique to clinical applications, near-infrared pulsed laser should be used to avoid the
strong affect from the blood and to increase the imaging depth.
The ability of visualizing the tumor in three-dimension permits the translation of
this technology to real clinical image-guided surgeries for breast cancer patients. Tumor
margin and inspection of completed tumor resection allows the clinicians to diagnose
the resected specimen and inspect the surgical area timely in order to reduce tumor
recurrence and improve the survival rate of breast cancer patients.
Beside employing photoacoustic imaging for image-guided surgery, it also
necessary to develop a photoacoustic imaging system to detect tumor noninvasively,
especially for breast cancer that is the second most important public health problem. In
next Chapter, we proposed a new circular-array based PAT system combined with DOT
for noninvasive clinical breast cancer detection.
106
Figure 5-1. System description of the experimental platform. The mouse was fixed into position on the holder with the head exposed to air. An aquatic heater was placed in the tank to maintain a constant body temperature for mouse. The probe was immersed in the water. The output trigger signal of the laser was used to synchronize the data acquisition and function generators (AFG3022B, Tektronics). The magnitude of ultrasonic signals were amplified by a pre-amplifier with a gain of 17dB and further amplified by an amplifier with a changeable gain from 5 dB to 20dB. The data acquisition sampling rate was 50MHz and a high-pass filter (Panametrics, Waltham, Massachusetts) with 1 MHz cutoff frequency was used to cut off the low frequency noises.
107
Figure 5-2. Schematic representation of MEMS-based photoacoustic imaging system.
A) The illuminating beam from a Nd:YAG laser passes through a mirror mounted in a 45° plane, a miniaturized probe mounted on a three-dimensional (3D) linear stage and irradiates the sample surface. B) 3D rendering of the probe in a; an aperture is used to form a small light spot of 0.2mm in diameter and plastic film is used to protect the MEMS mirror from water or ultrasound gel. C) One surface of wedge-shaped silicon base plated with aluminum redirects the light beam to the MEMS mirror. D) Photograph of the MEMS mirror with four actuators and five pads bonded with gold wire.
108
Figure 5-3. Schematic and performance of a ring-shaped ultrasound transducer. A) non-
focused ring-shape ultrasonic transducer was fabricated. To obtain high detection sensitivity, high performance PZT (DL-53HD, DeL Piezo Specialties, LLC, Wellington, FL) ceramic was used to fabricate the ring ultrasonic transducer. The inner and outer diameters of the transducer are 9mm and 11mm, respectively. B) The central frequency of the transducer was measured as 5.5MHz and the -6dB bandwidth of the transducer was larger than 50%.
109
Figure 5-4. Schematic of the miniaturized probe. A wedge-shaped mirror plated with
aluminum is fixed in position to redirect the light beam to the MEMS mirror. The MEMS mirror is fixed on a ramp with an angle of 22.5° allowing light to scan from position 3 to position 4 while the MEMS mirror moves from position 1 to position 2.
110
Figure 5-5. Performance evaluation of the system. A) Hilbert transform of a typical A-
line photoacoustic signal from the target. B) Axial and transverse resolutions versus target depth. C) Transverse point spread function of the target. D) Signal-to-noise ratio (SNR) versus target depth.
111
Figure 5-6. Phantom experiment with single target. A) Photograph of phantom. B) Raw
signal (blue) and signal after Hilbert transform (red). C) Image result of the phantom
Figure 5-7. Diagram of the phantom and imaging result with multiple targets
112
Figure 5-8. Ex vivo experiment with single target. A) Diagram of the sample. B) 2D and
3D result of the sample
Figure 5-9. In vivo experimental evaluation using human hand. A) Photograph of the
image area. B) 3D image result of the blood vessels
113
Figure 5-10. In vivo three-dimensional (3D) tumor mapping in a mouse model. A)
Photograph of the mouse with tumor implanted in abdomen before surgery. B) In vivo 3D photoacoustic image. The distance shown along Z axis includes the tumor depth and the distance between the surface of transducer and mouse skin. C) Photograph of the mouse after tumor resection. D) 3D photoacoustic image of the tumor cavity after surgery to examine the completeness of the tumor resection. E) Photograph of the tumor resection guided by 3D photoacoustic image shown in B.
114
Figure 5-11. Quantitative analysis of the photoacoustic slices and H&E stained sections.
A) Photoacoustic slice. Lines 1 and line 2 indicate the imaged dimensions along two different directions. B) H&E stained section in the same position. Lines 1′ and line 2′ denote the measured dimensions in the H&E stained section. C) Photograph of the tumor. All lines indicated by arrows in Fig. 5-11A, Fig. 5-11B and Fig. 5-11C show the top margins of the tumor. D). Transverse photoacoustic slice from another mouse. Lines 1 and line 2 indicate the imaged dimensions along two different directions. E) H&E stained section in the same position. Lines 1′ and line 2′ present the actual dimensions in the H&E stained section. F) Photograph of the tumor. All lines indicated by arrows in Fig. 5-11D, Fig. 5-11E and Fig. 5-11F present the margins of the tumor in transverse plane. Scale bar, 500µm.
115
Figure 5-12. Quantitative analysis of photoacoustic images in correlation with H&E
sections with increasing image depth. (A,B,C) PAT slices and H&E sections for a tumor beneath the skin of mouse (0.3mm depth) . (D,E,F) PAT slices and H&E sections for a tumor beneath the mouse skin and a 1mm-thick chicken breast. (G,H,I) PAT slices and H&E sections for a tumor beneath the mouse skin and a 1.8mm-thick chicken breast. White dashed lines show the surface of chicken breast. Scale bar, 500um.
116
Figure 5-13. Design of future imaging probe. A transparent and deformable balloon filled
with water or ultrasound gel is attached to the ultrasound transducer. The balloon serves as the ultrasound coupling medium while protecting the transducer from direct contact with blood.
117
CHAPTER 6 CIRCULAR ARRAY-BASED PHOTOACOUSTIC TOMOGRAPHY AND ITS
APPLICATION TO BREAST CANCER DETECTION
6.1 Motivations
To date several PAT systems for breast cancer detection have been reported.
Oraevsky et al. developed a laser-based optoacoustic imaging system (LOIS) with an
arc-shaped 64-element acoustic array capable of providing a spatial resolution of 1mm.
Haisch et al. reported a combined photoacoustic and ultrasound imaging system.
Manohar et al. built a three-dimensional PAT system that could cover a 90mm field of
view. In their system, a planar array of 590 PVDF transducers was used, offering a
spatial resolution of 2.3-3.9mm given an imaging depth of up to 32mm. Pramanik et al.
described a hybrid PAT and thermoacoustic tomography system. Kruger et al. recently
reported an interesting PAT study of breast angiography using a 64×64×50mm field of
view given a 40mm imaging depth.
While the results from these studies are promising, PAT has limitations. For
example, PAT can provide high quality images only for certain size of targets due to the
limited detection band of an ultrasound transducer. It is difficult to accurately detect a
relatively large size target (e.g., ≥ 6mm) using a typical ≥ 1MHz transducer. In addition,
the limited directivity of a transducer may lead to imbalanced image quality for the entire
imaging area/volume. Yet, it still remains a major challenge for PAT to recover tissue
scattering coefficient, an important parameter for breast cancer detection.
DOT is capable of overcoming the above mentioned limitations associated PAT,
although it has relatively low spatial resolution. Thus, a combination of PAT and DOT
appears to be an ideal approach for breast imaging. Towards this, we recently reported
for the first time a hybrid system that integrated PAT and DOT in a single platform.13
118
The goal of the current work is to develop an improved G2 PAT/DOT system using a
ring-shaped 64-element array with improved sensitivity and directivity over the previous
32-element array used in our first prototype. The G2 PAT/DOT is validated using
phantom and ex vivo experiments.
6.2 Materials and Methods
6.2.1 System Description
Figure 6-1A presents the schematic of our PAT/DOT hybrid system. Detailed
description of the DOT part has been reported previously. Briefly, light generated from a
diode laser is delivered sequentially by an optical switch and optical fiber bundles to 16
source positions. For each source position, diffused light is detected by 16 detection
fiber bundles coupled with detection units and data acquisition board and used for DOT
image reconstruction.
For the PAT part, a pulsed light from a Nd:YAG or Ti:Sapphire laser with a 6ns
pulse duration and 10Hz repetition rate is sent to the object through a light delivery
system. The laser-generated ultrasound signals are collected by 64 transducers which
are connected via a mechanic switching system to a 16-channel pre-amplifier and 16-
channel data acquisition (DAQ) board. The sampling rate of the DAQ board triggered by
the laser is 50MHz.
Figure 6-1B, the finished optical fibers/transducers/object interface contains a
home-made 64-elements acoustic transducer array. We used commercial PVDF film
(110µm) with silver electrode in both sides. The film was shaped into 32 rectangle units
with a size of 5mm×30mm. For each unit, a hole in the center was drilled for the
placement of a fiber bundle of DOT. The positive electrode was divided into two parts
with an equal size of 2.3mm×30mm and the PVDF film was attached to backing material
119
with the same size as the film by transparent epoxy resin. The positive and negative
electrodes of a coaxial cable were connected to the film using silver epoxy. A copper
housing was used to shield electromagnetic noise for each unit to improve the signal-to-
noise ratio.
We initially tested a fiber bundle based light delivery approach and found it was
not feasible since the damage threshold for a typical fiber bundle was 20mJ which was
much lower than the at least 60mJ needed for us to illuminate an object area of 3cm2
(calculated based on the maximum permissible exposure to light for human tissue
surface). Therefore, we developed a light delivery system by mounting three prisms to a
two-dimensional step motor (Figure 6-2A). And a concave lens and ground glass
indicated by arrow 3 in Figure 6-2A were attached to the front of output end of the light
delivery system to extend the light beam to be 3cm2. The output laser beam was
delivered from bottom of the exam table to the phantom. During an imaging experiment,
the light beam is scanned in two-dimension (arrows 1 and 2 in Figure 6-2A) via two step
motors to realize the light illumination to cover a large area.
Ultrasound signal generated by photoacoustic affect on a sphere absorber is
distributed in bipolar shape and the Fourier spectrum of this signal reveals the major
frequency components the acoustic energy contains. Therefore, the frequency
bandwidth of the transducer determines the detectable range of target size. We used an
impulsive wave method described by Manohar et al. 8 to measure the frequency
response of each element. Figure 6-2B shows that the -6dB bandwidth of the
transducer is from 380kHz to 1.48MHz and maximum frequency response is up to
2MHz.
120
Figure 6-2C presents the directivity of a transducer measured in the plane parallel
to the short side of the transducer. It was estimated to be 30。 based on the -6dB
level. The methodology we used for the estimation was described by Ermilov et al.7 The
normalized sensitivity for the 64 elements is shown in Figure 6-2D where we see that
the sensitivity lied in the range of 0.7-1.0. We used this set of data to calibrate the
measurements from each element for image reconstruction.
6.2.2 Quantitative Reconstruction Methods of PAT and DOT
Various quantitative reconstruction methods have been explored for PAT and
DOT. The reconstruction methods for both PAT and DOT in our PAT/DOT system are
finite element (FE) based.
For PAT, the FE-based dual-meshing reconstruction algorithm described by
Yao and Jiang14
was used to recover absorption coefficient from the photoacoustic
measurements. The algorithm includes two steps. The first is to obtain the map of the
absorbed optical energy density. The second step is to recover the distribution of the
absorption coefficient from the absorbed energy density. The core procedure of our PAT
reconstruction algorithm can be described by the following two equations:
2 2 00 0
( )( , ) ( , ) ,
p
c rp r k p r ik
C
(6.1)
0( ) ( ),T T cI p p (6.2)
where p is the pressure wave; 0 0/k c is the wave number described by the
angular frequency, , and the speed of the acoustic wave in the medium, 0c ; is the
thermal expansion coefficient; pC is the specific heat; is the absorbed energy density
121
that is the product of the absorption coefficient, a , and optical fluence, (i.e., a );
0
1 2( , ,..., ) ,c c c T
Mp p p p and 0
ip and c
ip are the observed and computed complex acoustic field
data for 1,2,...,i M boundary locations; is the update vector for the absorbed optical
energy density; is the Jacobian matrix formed by /p at the boundary measurement
sites; is the regularization parameter determined by combined Marquardt and
Tikhonov regularization schemes; and I is the identity matrix. Thus here the image
formation task is to update the absorbed optical energy density distribution via the
iterative solution of equations(6.1) and (6.2) so that an object function composed of a
weighted sum of the squared difference between the computed and measured acoustic
data can be minimized. The second step is based on the iterative solution to the
following radiation transport equation (RTE):
1( ) ( , ) ( , ) ( , ) ( , ),
ns a sS
r r d q r
(6.3)
where s is the scattering coefficient; ( , )r is the radiance; ( , )q r is the source
term; 1nS denotes a unit vector in the direction of interest. The kernel ( , ) is the
scattering phase function describing the probability density that a photon with an initial
direction will have a direction after a scattering event. If the incident laser source
strength and the absorbed energy density are estimated in advance, absorption
coefficient distribution can be determined by iterative solution procedure using the finite
element method. In the PAT reconstruction, s are assumed constant in this study. A
fine mesh of 6285 nodes and a coarse mesh of 1604 nodes were applied in the
reconstruction, and the images were converged within 50 iterations using a parallel
computer.
122
For DOT, both the absorption and scattering coefficient images were recovered
from the algorithms described in detail by Li and Jiang and Jiang et al.. In short, the
algorithms use a regularized Newton’s method to update an initial optical property
distribution iteratively in order to minimize an object function composed of a weighted
sum of the squared difference between computed and measured optical data at the
medium surface. The computed optical data (i.e., photon density) is obtained by solving
the photon diffusion equation with a finite element method. The core procedure in our
reconstruction algorithms is to iteratively solve the following regularized matrix equation:
( ) ( )( ) ( ),T T m cI q (6.4)
where is the photon density, I is the identity matrix, and can be a scalar or a
diagonal matrix. 1 2 ,1 ,2 ,( , ,..., , , ,..., )T
N a a a Nq D D D is the update vector for the optical
property profiles, where N is the total number of nodes in the finite element mesh used
and D is the diffusion coefficient. ( ) ( ) ( ) ( )
1 2( , ,..., )m m m m
M and ( ) ( ) ( ) ( )
1 2( , ,..., )c c c c
M , where ( )m
i
and ( )c
i , respectively, are measured and calculated data for 1,2,...,i M boundary
locations. is the Jacobian matrix that is formed by / D and / a at the boundary
measurement sites. In DOT, the goal is to update the a and D or s distributions
through the iterative solution of equation(6.4) so that a weighted sum of the squared
difference between computed and measured data can be minimized. A single mesh of
700 nodes was used, and the images were converged within 15 iterations in a 3GHz PC
with 1GB memory.
6.3 Performance Evaluation and Phantom Experiments
Since detailed system performance for DOT has been evaluated previously, 5 here
we focus on evaluating the PAT part of the PAT/DOT system on the imaging depth,
123
active imaging area and multi-target imaging ability using the light delivery system, and
conduct a comparison between PAT and DOT in these areas. We also test our
PAT/DOT system using ex vivo tumor tissue embedded in a tissue-mimicking phantom.
6.3.1 Performance of PAT
Spatial resolution of a PAT system is determined by several factors including the
number of transducers, inter-element spacing, and sensitivity and frequency response
of the transducer as well as the reconstruction methods used. Two experiments were
performed to estimate the best spatial resolution of our system. In the first experiment,
we put three thin metal wires (0.15mm in diameter each) in the center of a phantom
background. The distance was 0.5mm and 0.7mm, respectively, between wires 1 and 2
and between wires 2 and 3. In the second experiment, we placed two metal wires into
the phantom background and the distance between the two wires was 0.3mm. The
reconstructed PAT images from the two experiments are shown in Figure 6-3A and B,
respectively, where we note that the three targets can be clearly distinguished (Figure
6-3A), while the two targets are not resolvable (Figure 6-3B). Hence, we determine that
the best spatial resolution of our system is 0.5mm.
The detail parameters of the tissue-mimicking phantom (Intralipid + India ink) and
5 targets used for evaluating the PAT performance are listed in Table 6-1. These
experiments were performed using the 532nm Nd:YAG pulsed laser. The light beam
was guided and extended by our light delivery system to be an area source of 3cm2 and
the PAT images were reconstructed by a delay and sum method. The light energy
density at the surface of the phantom was 20mJ/cm2.
Figure 6-4A and B show the PAT images of target 1 located at 0 and 22mm
below the phantom surface, respectively. While the target is clearly better imaged with
124
0mm depth, the target with 22mm depth is still detectable. To detect a 2mm-radius
centrally located target, we found that the maximum depth was 7mm and 22mm for the
G1 and G2 systems, respectively. In addition, due to the use of electromagnetic
shielding cases for the transducers in the G2 system, the signal needed not to be
averaged while it needed to be averaged 100 times for the G1 system. The active
imaging area for PAT is validated by the result shown in Figure 6-4C where target 3
(30mm off center positioned) is detected. Due to the limited directivity of transducers,
we note that the shape of reconstructed target was distorted and stronger artifacts were
seen around the target. We also tested and found that when a target was placed at a
>30mm off center position, the image quality became unacceptable. For the single
target experiments, the light beam was directly delivered to the target area without
scanning. The image shown in Figure 6-4D demonstrated the ability of imaging multiple
targets using the developed light delivery system. Three targets (2, 4, and 5) separated
with a large distance were imaged by scanning the laser beam 9 (3×3) times to cover
the whole phantom surface. The 9 images obtained from the 9 laser beam positions
were then fused into a single image shown in Figure 6-4D.
6.3.2 Phantom Experiments
Figure 6-5 presents the reconstructed quantitative absorption coefficient images
by PAT and absorption and scattering coefficient images by DOT for some of targets 1-
5 and for a relatively large target, target 6 positioned 20mm off center (radius=5mm,
absorption coefficient=0.028mm-1 and scattering coefficient=4.0mm-1). These
quantitative images were recovered using our finite element based PAT and DOT.
Figure 6-5 also gives the reconstructed optical property profiles depicted along one cut
125
line crossing through the center of the target (red dashed line) for each case in
comparison with the exact property values (black solid line).
From Figure 6-5A, we see that target 5 (radius=1mm) is accurately recovered by
PAT in terms of its size, position and absorption coefficient value, while it is not
detectable by DOT in both the absorption and scattering images. For target 2
(radius=2mm), again PAT can accurately reconstruct its size, position and absorption
coefficient value (Figure 6-5B). In this case, DOT can detect the target from both the
absorption and scattering images; however, the values of both absorption and
scattering coefficients are significantly underestimated. Figure 6-5C shows that target 3
(radius=3mm) is quantitatively recovered by both PAT and DOT while we note an
overestimated target size by DOT.
Interestingly, the largest target 6 (radius=5mm) is poorly recovered by PAT (Figure
6-5D), due to the loss of low frequency signals given the limited frequency response of
the transducers. Here we see that only the edge of the target is recovered by PAT and
the reconstructed absorption coefficient value is considerably underestimated. DOT in
this case provides accurate recovery of the target in terms of its size, position, and
absorption and scattering coefficient values.
Overall from the results shown in Figure 6-5 and the summary in Table 6-2, PAT
can provide both qualitatively and quantitatively better images than DOT when the
target size is equal to or small than 3mm in radius, while DOT can offer better image
quality when the target size is larger than 3mm in radius. We also notice that the
smallest detectable target size is 2mm in radius for DOT given the experimental
conditions used in this study.
126
6.4 Ex Vivo Experiment
For the ex vivo experiment, a tumor was removed from a rat bearing a 4T1 tumor
of approximately 3.5mm in radius and embedded in the phantom background. In this
case, 730 nm pulsed light from a Ti:Sapphire laser was employed and provided an
energy density of 20mJ/cm2 at the phantom surface. Figure 6-6 shows the recovered
absorption and/or scattering images where we see that the tumor is detected by both
PAT and DOT. The absorption coefficient of the tumor recovered by PAT is consistent
with that by DOT. Through quantitative analysis of ex vivo PAT and DOT images, the
size of tumor (3mm in radius) estimated by the absorption image of PAT and by the
scattering image of DOT is consistent with the actual tumor size, while it is
overestimated by the absorption image of DOT to be 5mm in radius.
6.5 Multilayer Ultrasound Transducer with Improved Sensitivity and Bandwidth
In our PAT&DOT system, the circular array were made of PVDF because PVDF
owns some immediate advantages compared with other materials (PZT, PMN-PT).
PVDF has a high mechanical damping ability and a unique permittivity, resulting in a
high bandwidth. In addition, the acoustic impedance (Z0) of PVDF is 2.7 MRays, which
is much lower than that of PZT (>30MRays). Hence it provides a good coupling
efficiency to human tissue or commonly used ultrasound gels. PVDF is flexible and can
be easily fabricated in convex shapes replacing the acoustic lens commonly used in
PZT or PMN-PT. We note, however, that the sensitivity of the PVDF transducers
currently used for PAI is not comparable to the PZT and PMN-PT based transducers,
leading to limited SNR especially when targets are deeply located within tissues. In
order to improve the sensitivity of the PVDF transducers, several transducer designs
based on the same-thickness PVDF layers including folded transducer (FT), barker
127
coded transducer (BCT) and switchable Barker coded transducer (SBCT) have been
suggested in the field of ultrasound imaging.170-172 While the sensitivity is improved,
there is no significant improvement in bandwidth from these designs. Here we propose
a design of PVDF transducer with variable-thickness layers to provide improvement in
both the sensitivity and bandwidth for photoacoustic imaging.
The design of our multilayered transducer is shown in Figure 6-7A where the
PVDF films (Mettalized Film Sheets, Measurement Specialties) having three different
thicknesses (28µm, 52µm and 110µm) are sandwiched by electrodes. The fabrication
procedure of this transducer is as follows: (1) shaping the lateral dimension of the three
PVDF films to 10mm×3mm; (2) removing the unneeded silver coating from both sides of
each film to ensure electric insulation between the positive and negative electrodes; (3)
stacking/gluing the films/backing material together using ultrasound transparent epoxy
(EPO-TEK 301-2, Epoxy Technology); and (4) connecting coaxial cables to the
electrodes.
The multilayered transducer fabricated is tested using a circular scanning
photoacoustic imaging system (Figure 6-7B). A laser beam generated from a pulsed
Nd:YAG laser (532nm) was redirected through the center of the rotator and expanded
by a lens to illuminate an imaging area of 6.25cm2. A two-dimensional linear stage
(NLS4 Series, Newmark Systems) was used to optimize the position of transducer. A
water-tank with a through hole in the bottom sealed with transparent membrane was
used to provide coupling for ultrasound transmission. The signal received by the PVDF
transducer was amplified by two low-noise amplifiers and digitalized/stored by a data
acquisition card at 50MHz sampling rate. A pulse/echo test system (Figure 6-7C) was
128
used to evaluate the bandwidth of the multilayered transducer. The transducer was
driven by a pulser/receiver (5073PR, Panametrics-NDT) and a piece of metal with a
thickness of 5mm was placed in front of the transducer to act as a reflector. The
reflected pulse was received by the transducer, amplified by the internal amplifier of the
pulser/receiver and recorded using an oscilloscope.
Before the performance evaluation and in vivo experiments, the transducer was
calibrated through an experiment with a hair embedded solid phantom. The solid
phantom was a mixture of Intralipid (scatter), India ink (absorber), and 2% Agar powder
(solidifier). The absorption and reduced scattering coefficients of the phantom were
0.007mm-1 and 1.0mm-1. A human hair was inserted in the solid phantom, which was
illuminated by the light beam. Figure 6-8A, B and C show the acoustic signals from
different layers of the transducer where we note a time shift between these signals,
caused by the different distance of these three layers from the target. Hence, calibration
of this time shift is needed before further evaluating the performance of the transducer.
Assuming that the arrival times of the signals are T1 , T2 and T3 for the layers of 28, 52
and 110µm, respectively, and setting the arrival time for the 28µm-thick layer as the
baseline, we can then calculate the time shift of the other two layers relative to T1 by
the following relationships:
1 2 1t T T (6.5)
2 3 1t T T (6.6)
Thus, the photoacoustic (PA) signal from the multilayered transducer can be
obtained as
28 52 1 110 2( ) ( ) 1 ( ) 2 ( ) 3m m mPA t PA t P PA t t P PA t t P (6.7)
129
where P1, P2, and P3 are the phase of the PA signals which are 1, -1, 1 in our
case for the 28µm-, 53µm- and 110µm-thick layer, respectively. Figure 6-8D shows the
calibrated signal (red dashed line) and non-calibrated signal (blue dashed line). We see
that the SNR of the calibrated signal is improved significantly (by 2.5 times) compared
with that without the calibration.
Figure 6-9B, C and D show the frequency characteristics by fast Fourier
transforming the reflected acoustic waves detected by each of the three layers using the
pulse/echo system. For the 110µm-, 52µm- and 28µm-thick single-layer transducer, the
central frequency and the bandwidth (-6dB level) are, respectively, 5MHz/85%,
10MHz/78% and 15MHz/59%, while the central frequency of the multilayered
transducer is 10MHz and the -6 dB level bandwidth is improved to be 140%, as shown
in Figure 6-9A.
In vivo animal experiments were performed to evaluate the merits of the improved
bandwidth and sensitivity for photoacoustic imaging. The brain of adult BALB/C mice
(~25g) was imaged. Before imaging, the hair was removed with a hair remover lotion.
The mice were anesthetized by a mixture of Ketamine (85mg/kg) and Xylazine during
the imaging, and were sacrificed afterwards using the University of Florida Institutional
Animal Care and Use Committee (IACUC)-approved techniques. The laser beam
provided an incident energy density of 9 mJ/cm2 at the skin of the mice, which is much
lower than the safety standard provided by ANSI (20mJ/cm2 at 532nm). For each
imaging experiment, we collected PA signals at 120 positions by scanning the
transducer around the mouse brain with a scanning step of 3°. The PA image was
reconstructed by a delay and sum algorithm.
130
Figure 6-10A shows a typical signal from in vivo mouse brain measured by the
multilayered transducer without (top panel) and with (bottom panel) calibration. Without
the calibration, the SNR is low and the image formed with such signals has relatively
poor quality. As we see from Figure 6-10B (left), in such case, some blood vessels
inside the brain are not visualized and almost every blood vessel imaged is not
continuous, as compared to the photograph of the brain after removal of the scalp
(Figure 6-10B (right)). With the calibration, the SNR is significantly improved as shown
in the bottom panel of Figure 6-10A. From the image formed with the calibration (Figure
6-10B (middle)), we see that the eyes, blood vessels and other tissues are identified
clearly in comparison with the photograph.
In a separated in vivo experiment, we compared the sensitivity of the multilayered
transducer with the single-layer transducers and the PA images obtained are presented
in Figure 6-11. We immediately note the clearly higher SNR shown by the image using
the multilayered transducer (Figure 6-11A) compared to that using the single-layer
transducer (Figure 6-11B, C and D). In particular, we can identify the tissues, organs
and blood vessels inside the brain from Figure 6-11A much more clearly than that from
Figure 6-11B, C and D as indicated by red arrows.
6.6 Discussion and Future Directions.
We have developed a second generation prototype PAT/DOT system and
evaluated its validity using phantom and ex vivo tumor experiments. This system takes
full advantages of both PAT and DOT to provide high resolution absorption image and
scattering image. While we plan to test this system in breast cancer patients in the near
future, we are aware that there are still several improvements that need to be
undertaken before this testing. First, multi-spectral PAT imaging will be considered to
131
provide tissue functional information. Second, we will use the absorption coefficient
recovered by PAT as a priori knowledge to improve the scattering image reconstruction
by DOT, or use the scattering coefficient provided by DOT to enhance the absorption
recovery of PAT. Third, the data acquisition for PAT will be improved by using a fast
electronic switch with in-board pre-amplifiers that connects between the 64 transducers
and the 16-channel data acquisition board. Finally, we note that when the light beam
was delivered to the region near the boundary where the transducers are located, the
acoustic signal generated by light delivered to the surface of PVDF film resulted in
notable artifacts in the reconstructed photoacoustic images. We are currently
investigating some ways to minimize this effect including attaching a thin layer to the
transducer to absorb the light and preprocessing the photoacoustic signals using
wavelet transform before reconstruction.
We have shown that the novel design of variable-thickness multilayered PVDF
transducer provides significant improvement in both the sensitivity and bandwidth
compared to the conventional single-layer transducer. However, we note several
limitations associated with the current design, which should be overcome in future. First,
the sensitivity of the current multilayered transducer is still not comparable with that of a
commercial PZT unit. This, however, can be overcome by stacking more layers (>3)
together to make a more sensitive transducer. Second, the PVDF films used in the
current work are commercial units with electrodes on both sides of each film. After
stacking them together, there are two layers of electrodes between two adjacent films.
As a result, the improvement of the sensitivity is not linearly proportional to the number
of the layers stacked due to the attenuation from the electrodes. This problem can be
132
solved by using raw film liquid, which will allow us to use only one much thinner
electrode (just a few µm) between two adjacent films during depositing the films. Third,
no matching layer and electromagnetic shield (EMS) housing were used in the current
design. Hence we had to average 50 times on the signals collected to reduce the
electromagnetic noise, resulting in a relatively long data acquisition time.
Figure 6-1. System description. A) Schematic of the PAT/DOT system. B) Photograph of the interface and schematic of a single transducer.
133
Figure 6-2. Performance evaluation of the hybrid system. A) 3D schematic of the light
delivery system. B) Frequency response of a transducer. Red dashed line shows the -6dB level. C) Directivity of a transducer. Red dashed line shows the -6dB level. D) Normalized sensitivity for all the elements in the transducer array.
134
Figure 6-3. System resolution evaluation. A) PAT image of three separated metal wires
(0.15mm in diameter each) embedded in the phantom background. B) PAT image of two metal wires (0.15mm in diameter each) placed in the phantom background.
135
Figure 6-4. System performance evaluation by phantom experiments. A) PAT image of
target 1 without any depth. B) PAT image of target 1 positioned at 22mm beneath the surface. C) PAT image of target 3 (30mm off center positioned). D) Fused image of multiple targets.
136
Figure 6-5. Quantitative PAT and DOT images of targets. A) Target 5. B) target 2. C) target 3. D) target 6.
137
Figure 6-5. Continued
138
Figure 6-5. Continued
139
Figure 6-5. Continued
140
Figure 6-6. Reconstructed optical properties of ex vivo tumor. A) Absorption image by PAT. B) Absorption image by DOT. C) Scattering image by DOT.
141
Figure 6-7. Schematic of the multi-layered transducer and performance evaluation systems. A) Configuration of the multilayered PVDF transducer. B) Circular-scanning based photoacoutic imaging system. C) Description of pulse/echo evaluation system.
Figure 6-8. Calibration of multi-layered transducer. Signals of 28μm (A), 52μm (B) and 110μm (C) layer transducer from hair. D) Signals with (red) and without (blue) calibration.
142
Figure 6-9. Frequency response of multilayered and single-layered PVDF transducer. A) Frequency response of multilayered PVDF transducer. B) 110μm PVDF transducer. C) 52μm PVDF transducer. D) 28μm PVDF transducer.
143
Figure 6-10. In vivo experimental evaluation of multi-layered transducer. A) In vivo phtoacoustic signals of multilayer transducer with (bottom panel) and without (top panel) calibration. B) Reconstructed brain images using signals without (left panel) and with (middle panel)calibration. Photograph of the brain after removing the scalp (right panel).
144
Figure 6-11. In vivo photoacoutic imaging of mouse head using multilayered and single-layered transducer. A) Multilayered transducer. B) 110μm single-layered transducer. C) 52μm single-layered transducer. D) 28μm single-layered transducer.
Table 6-1. Exact size (diameter), location (off center), depth, and absorption and reduced scattering coefficients of the five targets
Size (mm) Location (mm) Depth* (mm) a (mm-1) '
s (mm-1)
Target 1 4 0 0 (22) 0.028 2
Target 2 4 20 5 (15) 0.028 2
Target 3 6 30 5 (10) 0.028 2
Target 4 6 20 5 (15) 0.028 2
Target 5 2 20 5 (15) 0.028 2
145
Table 6-2. Exact and reconstructed target size (diameter), location (off center) and absorption and reduced scattering coefficients for the phantom experiments
Size (mm) a (mm-1) s (mm-1) Location (mm)
Case 1
Exact 2 0.028 2 20
PAT 2 0.026 NA 20
DOT NA NA NA 20
Case 2
Exact 4 0.028 2 20
PAT 4 0.025 NA 20
DOT 10 0.017 1.1 20
Case 3
Exact 6 0.028 2 20
PAT 6 0.024 NA 20
DOT 10 0.024 1.8 20
Case 4
Exact 10 0.028 4 20
PAT 8 0.015 NA 20
DOT 12 0.028 4 20
146
CHAPTER 7 CONCLUSION AND FUTURE WORK
In this dissertation, four modalities of photoacoustic imaging covering from macro-
scale to micro-scale are described. For each modality, we show the potential
applications for different cancers studies. However, there are still several aspects need
to be improved for each modality. Firstly, for ORPAM, more applications should be
excavated. Secondly, for ARPAM, one improvement should be focused on employing
quantitative photoacoustic microscopy to exhume functional parameters such as
hemoglobin concentration, blood flow, oxygen consumption et al. with multi wavelengths.
Quantitative photoacoustic microscopy is also very helpful to explore the binding rate.
The other aspect is to investigate a real time handheld photoacoustic probe which is a
basic requirement for clinical applications. For intraoperative photoacoustic tomography,
and integrated PAT and DOT system, we are focusing on making it suitable for clinical
tries.
147
LIST OF REFERENCES
1. A.G. Bell, "On the production and reproduction of sound by light," Am. J. of Sci. 20, 305-324 (1880).
2. A. G. Bell, "The production of sound by radiant energy " Science 2, 242-253
(1881).
3. A. Tam, "Applications of photoacoustic sensing techniques," Rev. Mod. Phys. 58, 381-431 (1986).
4. R. Kruger, P. Liu and C. Appledorn, "Photoacoustic ultrasound (PAUS)-
reconstruction tomography," Med. Phys. 22, 1605-1609 (1995).
5. R. O. Esenaliev, A. A. Karabutov, F. K. Tittel, B. D. Fornage, S. L. Thomsen, C. Stelling and A. A. Oraevsky, "Laser optoacoustic imaging for breast cancer diagnostics: limit of detection and comparison with X-ray and ultrasound imaging," Proc. SPIE 2979, 78-82 (1997).
6. C. G. Hoelen, de F. F. Mul, R. Pongers and A. Dekker, "Three-dimensional
photoacoustic imaging of blood vessels in tissue," Opt. Lett. 23, 648-650 (1998).
7. A. A. Oraevsky, V. A. Andreev, A. A. Karabutor, K. R. Declan Fleming, Z. Gatalica, H. Singh and R. O. Esenalier, "Laser optoacoustic imaging of the breast: detection of cancer angiogenesis," Proc. SPIE 3597, 352-363 (1999).
8. R. A. Kruger, K. D. Miller, H. E. Reyolds, W. L. Kiser, D. R. Reinecke and G. A.
Kruger, " Breast cancer in vivo: contrast enhancement with thermoacoustic CT at 434 MHz-feasilibility staudy," Radiology 216, 279-283 (2000).
9. D. Feng, Y. Xu, G. Ku and L. V. Wang, "Microwave-induced thermoacoustic
tomography: reconstruction by synthetic aperture," Med. Phys. 28, 2427-2431 (2001).
10. Z. Yuan and H. Jiang, "Quantitative photoacoustic tomography: Recovery of
optical absorption coefficient of heterogeneous media," Appl. Phys. Lett. 88, 213301-1-3 (2006)
11. Z. Yuan, Q. Wang and H. Jiang, "Reconstruction of optical absorption coefficient
maps of heterogeneous media by photoacoustic tomography coupled with diffusion equation based regularized Newton method, " Opt. Express 15, 18076-18081 (2007).
12. L. Yao, Y. Sun and H. Jiang, "Quantitative photoacoustic tomography based the
radiative transfer equation," Opt. Lett. 34, 1765-1767 (2009).
148
13. Z. Yuan and H. Jiang, "Quantitative photoacoustic tomography," Phil. Trans. R. Soc. A 367, 3043-3054 (2009).
14. X. Wang, X. Xie, G. Ku, G. Stoica, and L. V. Wang, "Non-invasive imaging of
hemoglobin concentration and oxygenation in the rat brain using high-resolution photoacoustic tomography," Journal of Biomedical Optics 11, 024015 (2006).
15. H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, "Functional photoacoustic
microscopy for high-resolution and noninvasive in vivo imaging," Nature Biotechnology 24, 848–851 (2006).
16. X. M. Yang, S. E. Skrabalak, Z. Li, Y. Xia, and L. V. Wang, "Photoacoustic
tomography of a rat cerebral cortex in vivo with Au nanocages as an optical contrast agent," Nano Letters 7, 3798–3802 (2007).
17. L. Li, R. Zemp, G. Lungu, G. Stoica, and L. V. Wang, "Photoacoustic imaging of
lacZ gene expression in vivo," Journal of Biomedical Optics 12, 020504 (2007).
18. G. Ku, B. D. Fornage, X. Jin, M. Xu, K. K. Hunt, and L. V. Wang, "Thermoacoustic and photoacoustic tomography of thick biological tissues toward breast imaging," Technology in Cancer Research and Treatment 4, 559-566 (2005).
19. Y. Xu and L. V. Wang, "Effects of acoustic heterogeneity on thermoacoustic
tomography in the breast," IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control 50, 1134-1146 (2003).
20. M. Pramanik, G. Ku, C. H. Li, and L. V. Wang, "Design and evaluation of a novel
breast cancer detection system combining both thermoacoustic (TA) and photoacoustic (PA) tomography," Medical Physics 35, 2218–2223 (2008).
21. X. Li, L. Xi, R. Jiang, L. Yao, and H. Jiang, "Integrated diffuse optical tomography
and photoacoustic tomography: phantom validations," Biomed. Opt. Express 2, 2348-2353 (2011).
22. L. Xi, X. Li, L. Yao, S. Grobmyer and H. Jiang, "Design and evaluation of hybrid
photoacoustic tomography and diffuse optical tomography system for breast cancer detection," Med. Phys. 39, 2584-2595 (2012).
23. A. E. Cerussi, A. Berger, F. Bevilacqua, N. Shah, D. Jakubowski, J. Butler, R.
Holcombe, and B. Tromberg, "Sources of absorption and scattering contrast for near infrared optical mammography," Acad. Radiol. 8, 211–218 (2001).
24. Y. Yamashita, A. Maki, and H. Koizumi, "Wavelength dependence of the
precision of noninvasive optical measurement of oxy-, deoxy-, and total-hemoglobin concentration," Med. Phys. 28, 1108–1114 (2001).
149
25. S. Manohar, A. Kharine, J. C. G. van Hespen, W. Steenbergen, and T. G. van Leeuwen, "Photoacoustic mammography laboratory prototype: Imaging of breast tissue phantoms," J. Biomed. Opt. 9, 1172-1181 (2004).
26. R. A. Kruger, K. D. Miller, H. E. Reynolds, W. L. Kiser, Jr., D. R. Reinecke, and G.
A. Kruger, "Contrast enhancement of breast cancer in vivo using thermoacoustic CT at 434 MHz," Radiology 216, 279-283 (2000).
27. J. T. Oh, M. L. Li, H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, "Three-
dimensional imaging of skin melanoma in vivo by dual-wavelength photoacoustic microscopy," Journal of Biomedical Optics 11, 034032 (2006).
28. Y. Zhang, X. Cai, S.-W. Choi, C. Kim, L. V. Wang, and Y. Xia, "Chronic label-free
volumetric photoacoustic microscopy of melanoma cells in three-dimensional porous scaffolds," Biomaterials 31, 8651-8658 (2010).
29. Z. Xu, C. H. Li, and L. V. Wang, "Photoacoustic tomography of water in phantoms
and tissue," Journal of Biomedical Optics 15, 036019 (2010).
30. K. Maslov, H. F. Zhang, and L. V. Wang, "Effects of wavelength-dependent fluence attenuation on the noninvasive photoacoustic imaging of hemoglobin oxygen saturation in subcutaneous vasculature in vivo," Inverse Problems 23 S113–S122 (2007).
31. H. F. Zhang, K. Maslov, M. Sivaramakrishnan, G. Stoica, and L. V. Wang,
"Imaging of hemoglobin oxygen saturation variations in single vessels in vivo using photoacoustic microscopy," Applied Physics Letters 90, 053901 (2007).
32. Y. Wang, S. Hu, K. Maslov, Y. Zhang, Y. Xia, and L. V. Wang, "In vivo integrated
photoacoustic and confocal microscopy of hemoglobin oxygen saturation and oxygen partial pressure," Optics Letters 36, 1029–1031 (2011).
33. J. Yao, K. I. Maslov, Y. Shi, L. A. Taber, and L. V. Wang, "In vivo photoacoustic
imaging of transverse blood flow by using Doppler broadening of bandwidth," Optics Letters 35, 1419-142 (2010).
34. J. Yao, K. I. Maslov, and L. V. Wang, "In vivo photoacoustic tomography of total
blood flow and potential imaging of cancer angiogenesis and hypermetabolism," Technology in Cancer Research and Treatment 11, 301-307 (2012).
35. K. J. Rowland, J. Yao, L. Wang, C. R. Erwin, K. I. Maslov, L. V. Wang, and B. W.
Warner, "Immediate alterations in intestinal oxygen saturation and blood flow after massive small bowel resection as measured by photoacoustic microscopy,”"Journal of Pediatric Surgery 47, 1143-1149 (2012).
150
36. K. H. Song, C. H. Kim, C. M. Cobley, Y. N. Xia, and L. V. Wang, "Near-infrared gold nanocages as a new class of tracers for photoacoustic sentinel lymph node mapping on a rat model," Nano Letters 9, 183–188 (2009).
37. D. Pan, M. Pramanik, A. Senpan, S. Ghosh, S. A. Wickline, L. V. Wang, and G.
M. Lanza, "Near infrared photoacoustic detection of sentinel lymph nodes with gold nanobeacons," Biomaterials 31, 4088–4093 (2010).
38. W. Lu, Q. Huang, G. Ku, X. Wen, M. Zhou, D. Guzatov, P. Brecht, R. Su, A.
Oraevsky, L. V. Wang, and C. Li, "Photoacoustic imaging of living mouse brain vasculature using hollow gold nanospheres," Journal of Biomaterials 31, 2617-26 (2010).
39. X. Cai, W. Li, C. Kim, Y. Yuan, L. V. Wang, and Y. Xia, "In vivo quantitative
evaluation of the transport kinetics of gold nanocages in a lymphatic system by noninvasive photoacoustic tomography," ACS Nano 5, 9658-9667 (2011).
40. C. Kim, K. H. Song, F. Gao, and L. V. Wang, "Sentinel Lymph Nodes and
Lymphatic Vessels: Noninvasive Dual-Modality in Vivo Mapping by Using Indocyanine Green in Rats-Volumetric Spectroscopic Photoacoustic Imaging and Planar Fluorescence Imaging," Radiology 255, 442-450 (2010).
41. T. N. Erpelding, C. Kim, M. Pramanik, L. Jankovic, K. Maslov, Z. Guo, J. A.
Margenthaler, M. D. Pashley, and L. V. Wang, "Sentinel Lymph Nodes in the Rat: Noninvasive Photoacoustic and US Imaging with a Clinical US System," Radiology 256, 102–110 (2010).
42. W. Denk, J. H. Strickler and W. W. Webb, "Two-photon laser scanning
fluorescence microscopy," Science 248, 73–76 (1990).
43. D. Huang, E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W. Chang, M. R, Hee, T. Flotte, K. Gregory, C. A. Pullaftto, J. G. Fujimoto, "Optical Coherence tomography," Science 254, 1178-1181 (1991).
44. R. Henriques, C. Griffiths, E. Hesper, M. Mhlanga, "PALM and STORM: unlocking
live-cell super-resolution," Biopolymers 95, 322-331 (2011).
45. K. Maslov, H. F. Zhang, S. Hu and L. V. Wang, "Optical-resolution photoacoustic microscopy for in vivo imaging of single capillaries," Optics Letters 33, 929–931 (2008).
46. H. F. Zhang, K. Maslov, and L. V. Wang, "In vivo imaging of subcutaneous
structures using functional photoacoustic microscopy," Nature Protocols 2, 797–804 (2007).
151
47. L. V. Wang and S. Hu, "Photoacoustic tomography: in vivo imaging from organelles to organs," Science 335, 1458–1462 (2012).
48. S. Hu, K. Maslov and L. V. Wang, "Noninvasive label-free imaging of
microhemodynamics by optical-resolution photoacoustic microscopy," Optics Express 17, 7688-7693 (2009).
49. S. Hu, K. Maslov and L. V. Wang, "In vivo functional chronic imaging of a small
animal model using optical-resolution photoacoustic microscopy," Med. Phys. 36, 2320-2323 (2009).
50. S. Hu, K. Maslov, V. Tsytsarev, and L. V. Wang, "Functional transcranial brain
imaging by optical-resolution photoacoustic microscopy," Journal of Biomedical Optics 14, 04050301-03 (2009).
51. S. Hu, P. Yan, K. Maslov, J.-M. Lee, and L. V. Wang, “Intravital imaging of
amyloid plaques in a transgenic mouse model using optical-resolution photoacoustic microscopy,” Optics Letters 34, 3899–3901 (2009).
52. S. Hu, B. Rao, K. Maslov, and L. V. Wang, "Label-free photoacoustic ophthalmic
angiography," Optics Letters 35, 1–3 (2010).
53. V. Tsytsarev, S. Hu, J. Yao, K. Maslov, D. L. Barbour, and L. V. Wang, "Photoacoustic microscopy of microvascular responses to cortical electrical stimulation," Journal of Biomedical Optics 16, 076002 (2011).
54. S. Oladipupo, S. Hu, J. Kovalski, J. Yao, A. C. Santeford, R. Sohn, R. V. Shohet,
K. Maslov, L. V. Wang, and J. M. Arbeit, "VEGF is essential for hypoxia-inducible factor mediated neovascularization but dispensable for endothelial sprouting," Proceedings of the National Academy of Sciences 108,13264-13269 (2011).
55. S. Oladipupo, S. Hu, A. C. Santeford, J. Yao, J. Kovalski, R. V. Shohet, K.
Maslov, L. V. Wang, and J. M. Arbeit, "Conditional HIF-1 induction produces multistage neovascularization with stage-specific sensitivity to VEGFR inhibitors and myeloid cell independence," Blood 117, 4142-4153 (2011).
56. J. Laufer, P. Johnson, E. Zhang, B. Treeby, B. Cox, B. Pedley and P. Beard, "In
vivo preclinical photoacoustic imaging of tumor vasculature development and therapy," Journal of Biomedical Optics 17, 056016 (2012).
57. J. Laufer, F. Norris, J. Cleary, E. Zhang, B. Treeby, B. Cox, P. Johnson, P.
Scambler, M. Lythgoe and P. Beard, "In vivo photoacoustic imaging of mouse embryos," Journal of Biomedical Optics 17, 061220 (2012).
152
58. Q. Ruan, L. Xi, S. Boye, W. W. Hauswirth, S. Han, Z. Chen, M. E. Boulton, B. Law, W. G Jiang, H. Jiang and J. Cai, "Development of anti-angiogenic therapeutic model by combining scAAV2-delivered siRNAs and noninvasive photoacoustic imaging of tumor vasculature development, " Cancer letters (in press).
59. L. Xi, L. Zhou and H. Jiang "C-scan photoacoustic microscopy for in vivo imaging
of Drosophila pupae," Applied Physics Letter. 101, 013702 (2012).
60. H. F. Zhang, K. Maslov, G. Stoica, and L. V. Wang, "Imaging acute thermal burns by photoacoustic microscopy," Journal of Biomedical Optics 11, 054033 (2006).
61. H. F. Zhang, K. Maslov, M.-L. Li, G. Stoica, and L. V. Wang, "In vivo volumetric
imaging of subcutaneous microvasculature by photoacoustic microscopy," Optics Express 14, 9317–9323 (2006).
62. L. Xi, S. Grobmyer, G. Zhou, W. Qian, L. Yang and H. Jiang, "In vivo molecular
photoacoustic tomography of breast cancer with receptor-targeted magnetic iron oxide," Journal of Biophotonics (in press).
63. D. Finch, S. K. Patch and S. J. Rakesh, "Determining a function from its mean
values over a family of spheres," SIAM J. Math. Anal. 35, 1213–1240 (2004).
64. B. Cox, J. G. Laufer, S. R. Arridge and P. Beard, "Quantitative spectroscopic photoacoustic imaging: a review," Journal of Biomedical Optics 17, 061202 (2012).
65. P. Beard, "Biomedical photoacoustic imaging," Interface Focus 1, 602-631
(2012).
66. B. Wang, A. Karpiouk, D. Yeager, J. Amirian, S. Litovsky, R. Smalling, and S. Emelianov, "Intravascular photoacoustic imaging of lipid in atherosclerotic plaques in the presence of luminal blood." Opt. Lett. 37, 1244 (2012).
67. J. M. Yang, K. Maslov, H. C. Yang, Q. Zhou, K. K. Shung, and L. V. Wang,
"Photoacoustic endoscopy." Opt. Lett. 34, 1591 (2009).
68. Y. Yang, X. Li, T. Wang, P. D. Kumavor, A. Aguirre, K. K. Shung, Q. Zhou, M. Sanders, M. Brewer, and Q. Zhu, " Integrated optical coherence tomography, ultrasound and photoacoustic imaging for ovarian tissue characterization. " Biomed. Opt. Express 2, 2551 (2011) .
69. V. P. Zharov, E. I. Galanzha, E. V. Shashkov, N. G. Khlebtsov, and V. V.
Tuchin, "In vivo photoacoustic flow cytometry for monitoring of circulating single cancer cells and contrast agents." Opt. Lett. 31, 3623 (2006).
153
70. S. Jiao, M. Jiang, J. Hu, A. Fawzi, Q. Zhou, K. K. Shung, C. A. Puliafito, and H. F.
Zhang, "Photoacoustic ophthalmoscopy for in vivo retinal imaging " Opt. Express 18, 3967 (2010).
71. Laser Institute of America, American National Standard for Safe Use of Lasers
ANSI Z136.1-2007 (American National Standards Institute, Inc., 2007).
72. A. C. Tam, "Applications of photoacoustic sensing techniques," Rev. Mod. Phys. 58, 381-430 (1986).
73. H. Jiang, Z. Yuan, and X. Gu, “Spatially varying optical and acoustic property
reconstruction using finite-element-based photoacoustic tomography,” J. Opt. Soc. Am. A 23, 878-888 (2006).
74. Z. Yuan, Q. Zhang and H. Jiang, “Simultaneous reconstruction of acoustic and
optical properties of heterogeneous media by quantitative photoacoustic tomography,” Opt. Express 14, 6749-6754 (2006).
75. A. Bayliss and E. Turkel, “Radiation Boundary Conditions for Wave-Like
Equations,” Comm. Pure. Appl. Math. 33, 707-725 (1980).
76. R. Mittra and O. Ramahi, “Absorbing boundary conditions for the direct solution of partial differential equations arising in electromagnetic scattering problems,” in Differential Methods in Electromagnetic Scattering, ed J. A. Kong and M. A. Morgan, Elsevier, 133-173 (1989).
77. H. Gan, P. L. Levin and R. Ludwig, “Finite element formulation of acoustic
scattering phenomena with absorbing boundary condition in the frequency domain,” J. Acoust. Soc. Am. 94, 1651-1662 (1993).
78. T. Van and A. Wood, “A time-domain finite element method for Helmholtz
Equations,” J. Comput. Phys. 183, 486-507 (2002).
79. L. Yao and H. Jiang, “Finite-element-based photoacoustic tomography in time domain”, Journal of Optics A: Pure and Applied Optics, 11, 085301 (2009).
80. K. Ren, G. Bal, and A. H. Hielscher, “Transport- and diffusion-based optical
tomography in small domains: a comparative study,” Appl. Opt. 46, 6669-6679 (2007).
81. A. D. Klose, U. Netz, J. Beuthan, and A. H. Hielscher, “Optical tomography using
the time-independent equation of radiative transfer. Part 1: Forward model,” J. Quan. Spec. Radi. Tran. 72, 691-713 (2002).
154
82. Lei Yao, Yao Sun and Huabei Jiang, “Transport-based quantitative photoacoustic
tomography: simulations and experiments”, Physics in Medicine and Biology 55, 1917-1934 (2010).
83. J. Heino, S. R. Arridge, J. Sikora, and E. Somersalo, “Anisotropic effects in highly
scattering media,” Phys. Rev. E 68, 031908: 1-8 (2003).
84. B. Cox, S. Arridge, K. Kostli, and P. Beard, “Two-dimensional quantitative photoacoustic image reconstruction of absorption distributions in scattering media by use of a simple iterative method,” Appl. Opt. 45, 1866-1875 (2005).
85. H. Jiang, Y. Xu, and N. Iftimia, “Experimental three-dimensional optical image
reconstruction of heterogeneous turbid media from continuous-wave data,” Opt. Express, 7, 204-209 (2000).
86. J. A. Nagy, S. H. Chang, S. C. Shih, A. M. Dvorak and H. F. Dvorak,
"Heterogeneity of the tumor vasculature," Semin Thromb Hemost 36, 321-331(2010).
87. H. M. Jensen, I. Chen, M. R. DeVault and A. E. Lewis, "Angiogenesis induced by
normal human breast tissue: a probable marker for precancer," Science 218, 293-295 (1982).
88. C. Sundberg, J. A. Nagy, L. F. Brown, D. Feng, I. A. Eckelhoefer, E. J. Manseau,
A. M. Dvorak and H. F. Dvorak, "Glomeruloid microvascular proliferation follows adenoviral vascular permeability factor/vascular endothelial growth factor-164 gene delivery," Am J Pathol 158, 1145-1160 (2001).
89. J. A. Nagy, E. Vasile, D. Feng, C. Sundberg, L. F. Brown, M. J. Detmar, J. A.
Lawitts, L. Benjamin, X. Tan, E. J. Manseau, A. M. Dvorak and H. F. Dvorak, "Vascular permeability factor/vascular endothelial growth factor induces lymphangiogenesis as well as angiogenesis," J Exp Med 196, 1497-1506(2002).
90. S. Lee, T. T. Chen, C. L. Barber, M. C. Jordan, J. Murdock, S. Desai, N. Ferrara,
A. Nagy, K. P. Roos and M. L. Iruela-Arispe, "Autocrine VEGF signaling is required for vascular homeostasis," Cell 130, 691-703 (2007).
91. D. E. Feldman, V. Chauhan and A. C. Koong, "The unfolded protein response: a
novel component of the hypoxic stress response in tumors," Mol Cancer Res 3, 597-605 (2005).
92. D. Ron, and P. Walter, "Signal integration in the endoplasmic reticulum unfolded
protein response," Nat Rev Mol Cell Biol 8, 519-529 (2007).
155
93. M. Calfon, H. Zeng, F. Urano, J. H. Till, S. R. Hubbard, H. P. Harding, S. G. Clark and D. Ron, "IRE1 couples endoplasmic reticulum load to secretory capacity by processing the XBP-1 mRNA," Nature 415, 92-96 (2002).
94. J. Hollien and J. S. Weissman, "Decay of endoplasmic reticulum-localized
mRNAs during the unfolded protein response," Science 313, 104-107 (2006).
95. J. Wu, D. T. Rutkowski, M. Dubois, J. Swathirajan, T. Saunders, J. Wang, B. Song, G. D. Yau and R. J. Kaufman, "ATF6alpha optimizes long-term endoplasmic reticulum function to protect cells from chronic stress," Dev Cell 13, 351-364 (2007).
96. K. Yamamoto, H. Yoshida, K. Kokame, R. J. Kaufman and K. Mori, "Differential
contributions of ATF6 and XBP1 to the activation of endoplasmic reticulum stress-responsive cis-acting elements ERSE, UPRE and ERSE-II," J Bio Chem 136, 343-350 (2004).
97. Q. Ruan, S. Han, W. G. Jiang, M. E. Boulton, Z. J. Chen, B. K. Law and J. Cai,
"alphaB-crystallin, an effector of unfolded protein response, confers anti-VEGF resistance to breast cancer via maintenance of intracrine VEGF in endothelial cells," Mol Cancer Res 9, 1632-1643 (2011).
98. D. H. Kim and J. J. Rossi, "Strategies for silencing human disease using RNA
interference," Nat Rev Genet 8, 173-184 (2007).
99. P. J. Paddison, A. A. Caudy and G. J. Hannon, "Stable suppression of gene expression by RNAi in mammalian cells," Proc Natl Acad Sci 99, 1443-1148 (2002).
100. T. R. Brummelkamp, R. Bernards and R. Agami, "Stable suppression of
tumorigenicity by virus-mediated RNA interference," Cancer Cell 2, 243-247 (2002).
101. A. S. Cockrell and T. Kafri, "Gene delivery by lentivirus vectors," Mol Biotechnol
36, 184-204 (2007).
102. V. Nair, "Retrovirus-induced oncogenesis and safety of retroviral vectors," Curr Opin Mol Ther 10, 431-438 (2008).
103. H. C. Champion, T. J. Bivalacqua, S. S. Greenberg, T. D. Giles, A. L. Hyman and
P. J. Kadowitz, "Adenoviral gene transfer of endothelial nitric-oxide synthase (eNOS) partially restores normal pulmonary arterial pressure in eNOS-deficient mice," Proc Natl Acad Sci 99, 13248-53 (2002).
156
104. M. D. Martin-Martinez, M. Stoenoiu, C. Verkaeren, O. Devuyst and C. Delporte, "Recombinant adenovirus administration in rat peritoneum: endothelial expression and safety concerns," Nephrol Dial Transplant 19, 1293-1297 (2004).
105. A. Pandey, N. Singh, S. V. Vemula, L. Couetil, J. M. Katz, R. Donis, S. Sambhara
and S. K. Mittal, "Impact of preexisting adenovirus vector immunity on immunogenicity and protection conferred with an adenovirus-based H5N1 influenza vaccine," PLoS One 7, e33428 (2012).
106. W. W. Hauswirth, T. S. Aleman, S. Kaushal, A. V. Cideciyan, S. B. Schwartz, L.
Wang, T. J. Conlon, S. L. Boye, T. R. Flotte, B. J. Byrne and S. G. Jacobson, "Treatment of leber congenital amaurosis due to RPE65 mutations by ocular subretinal injection of adeno-associated virus gene vector: short-term results of a phase I trial," Hum Gene Ther 19, 979-990 (2008).
107. A. C. Nathwani, E. G. Tuddenham, S. Rangarajan, C. Rosales, J. McIntosh, D. C.
Linch, P. Chowdary, A. Riddell, A. J. Pie, C. Harrington, J. O'Beirne, K. Smith, J. Pasi, B. Glader, P. Rustagi, C. Y. Ng, M. A. Kay, J. Zhou, Y. Spence, C. L. Morton, J. Allay, J. Coleman, S. Sleep, J. M. Cunningham, D. Srivastava, E. Basner-Tschakarjan, F. Mingozzi, K. A. High, J. T. Gray, U. M. Reiss, A. W. Nienhuis and A. M. Davidoff, "Adenovirus-associated virus vector-mediated gene transfer in hemophilia B," N Engl J Med 365 2357-2365 (2011).
108. H. Anderson, P. Price, M. Blomley, M. O. Leach and P. Workman, "Measuring
changes in human tumour vasculature in response to therapy using functional imaging techniques," Br J Cancer 85, 1085-1093 (2001).
109. Q. Zhu, M. Huang, N. Chen, K. Zarfos, B. Jagjivan, M. Kane, P. Hedge and S. H.
Kurtzman, "Ultrasound-guided optical tomographic imaging of malignant and benign breast lesions: initial clinical results of 19 cases," Neoplasia 5, 379-388 (2003).
110. D. Fukumura and R. K. Jain, "Imaging angiogenesis and the microenvironment.
APMIS 116, 695-715 (2008).
111. S. Hu, K. Maslov and L. V. Wang, "Second-generation optical-resolution photoacoustic microscopy with improved sensitivity and speed," Optics Letters 36, 1134–1136 (2011).
112. B. Rao, K. Maslov, A. Danielli, R. Chen, K. K. Shung, Q. Zhou and L. V. Wang,
"Real-time four-dimensional optical-resolution photoacoustic microscopy with Au nanoparticle-assisted subdiffraction-limit resolution," Optics Letters 36, 1137-1139 (2011).
113. L. Song, K. Maslov and L. V. Wang, "Multifocal optical-resolution photoacoustic
microscopy in vivo," Optics Letters 36, 1236–1238 (2011).
157
114. E. Z. Zhang, B. Povazay, J. Laufer, A. Alex, B. Hofer, B. Pedley, C. Glittenberg, B. Treeby, B. Cox, P. Beard and W. Drexler, "Multimodal photoacoustic and optical coherence tomography scanner using an all optical detection scheme for 3D morphological skin imaging," Biomed Opt Express 2, 2202-2215 (2011).
115. B. Wang, E. Yantsen, T. Larson, A. B. Karpiouk, S. Sethuraman, J. L. Su, K.
Sokolov and S.Y. Emelianov, "Plasmonic intravascular photoacoustic imaging for detection of macrophages in atherosclerotic plaques," Nano Lett 9, 2212-2217 (2009).
116. S. E. Haire, J. Pang, S. L. Boye, I. Sokal, C. M. Craft, K. Palczewski, W. W.
Hauswirth and S. L. Semple-Rowland, "Light-driven cone arrestin translocation in cones of postnatal guanylate cyclase-1 knockout mouse retina treated with AAV-GC1," Invest Ophthalmol Vis Sci 47, 3745-3753 (2006).
117. S. Zolotukhin, M. Potter, I. Zolotukhin, Y. Sakai, S. Loiler, T. J. Fraites, Jr., V. A.
Chiodo, T. Phillipsberg, N. Muzyczka, W. W. Hauswirth, T. R. Flotte, B. J. Byrne and R. O. Snyder, "Production and purification of serotype 1, 2, and 5 recombinant adeno-associated viral vectors," Methods 28, 158-167 (2002).
118. S. G. Jacobson, G. M. Acland, G. D. Aguirre, T. S. Aleman, S. B. Schwartz, A. V.
Cideciyan, C. J. Zeiss, A. M. Komaromy, S. Kaushal, A. J. Roman, E. A. Windsor, A. Sumaroka, S. E. Pearce-Kelling, T. J. Conlon, V. A. Chiodo, S. L. Boye, T. R. Flotte, A. M. Maguire, J. Bennett and W. W. Hauswirth, "Safety of recombinant adeno-associated virus type 2-RPE65 vector delivered by ocular subretinal injection," Mol Ther 13, 1074-1084 (2006).
119. A. J. Bridge, S. Pebernard, A. Ducraux, A. L. Nicoulaz and R. Iggo, "Induction of
an interferon response by RNAi vectors in mammalian cells," Nat Genet 34, 263-264 (2003).
120. L. Liu, Z. J. Chen, L. Shaw, J. Cai, X. Qi, L. Smith, M. Grant and M. Boulton,
"Targeting the IRE1α/XBP1 and ATF6 arms of the unfolded protein response as an adjunct therapy to prevent retinal and choroidal neovascularization," Am. J. Pathol. (under review).
121. R. C. Ryals, S. L. Boye, A. Dinculescu, W. W. Hauswirth and S. E. Boye,
"Quantifying transduction efficiencies of unmodified and tyrosine capsid mutant AAV vectors in vitro using two ocular cell lines," Mol Vis 17, 1090-1102 (2011).
122. J. Cai, W. G. Jiang, A. Ahmed and M. Boulton, "Vascular endothelial growth
factor-induced endothelial cell proliferation is regulated by interaction between VEGFR-2, SH-PTP1 and eNOS," Microvasc Res 71, 20-31 (2006).
158
123. J. Cai, O. Kehoe, G.M. Smith, P. Hykin and M. E. Boulton, "The angiopoietin/Tie-2 system regulates pericyte survival and recruitment in diabetic retinopathy," Invest Ophthalmol Vis Sci 49, 2163-2171 (2008).
124. P. E. Corsino, B. J. Davis, P. H. Norgaard, N. N. Parker, M. Law, W. Dunn and B.
K. Law, "Mammary tumors initiated by constitutive Cdk2 activation contain an invasive basal-like component," Neoplasia 10, 1240-1252 (2008).
125. J. Cai, W. G. Jiang, M. B. Grant and M. Boulton, "Pigment epithelium-derived
factor inhibits angiogenesis via regulated intracellular proteolysis of vascular endothelial growth factor receptor 1," J Biol Chem 281, 3604-3613 (2006).
126. X. Xu, H. Liu and L. V. Wang, "Time-reversed ultrasonically encoded optical
focusing into scattering media," Nature Photonics 5, 154–157 (2011).
127. L. M. Work, H. Buning, E. Hunt, S. A. Nicklin, L. Denby, N. Britton, K. Leike, M. Odenthal, U. Drebber, M. Hallek and A. H. Baker, "Vascular bed-targeted in vivo gene delivery using tropism-modified adeno-associated viruses," Mol Ther 13, 683-693 (2006).
128. M. Sabuncu, "Entropy-based image registration," The Department of Electrical
Engineering, Princeton University, pp. 152 (2006).
129. M. T. Spiotto, A. Banh, I. Papandreou, H. Cao, M. G. Galvez, G. C. Gurtner, N. C. Denko, Q. T. Le and A. C. Koong, "Imaging the unfolded protein response in primary tumors reveals microenvironments with metabolic variations that predict tumor growth," Cancer Res 70, 78-88 (2010).
130. H. Yoshida, T. Matsui, A. Yamamoto, T. Okada, and K. Mori, "XBP1 mRNA is
induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor," Cell 107, 881-891 (2001).
131. D. M. Schewe, and J. A. Aguirre-Ghiso, "ATF6alpha-Rheb-mTOR signaling
promotes survival of dormant tumor cells in vivo," Proc Natl Acad Sci 105, 10519-10524 (2008).
132. C. C. Jiang, L. H. Chen, S. Gillespie, K. A. Kiejda, N. Mhaidat, Y. F. Wang, R.
Thorne, X. D. Zhang and P. Hersey, "Tunicamycin sensitizes human melanoma cells to tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis by up-regulation of TRAIL-R2 via the unfolded protein response," Cancer Res 67, 5880-5888 (2007).
133. P. S. Gargalovic, N. M. Gharavi, M. J. Clark, J. Pagnon, W. P. Yang, A. He, A.
Truong, T. Baruch-Oren, J. A. Berliner, T. G. Kirchgessner and A. J. Lusis, "The unfolded protein response is an important regulator of inflammatory genes in endothelial cells," Arterioscler Thromb Vasc Biol 26, 2490-2496 (2006).
159
134. K. Pajusola, M. Gruchala, H. Joch, T. F. Luscher, S. Yla-Herttuala and H. Bueler, "Cell-type-specific characteristics modulate the transduction efficiency of adeno-associated virus type 2 and restrain infection of endothelial cells," J Virol 76, 11530-11540 (2002).
135. J. Hansen, K. Qing, H. J. Kwon, C. Mah and A. Srivastava, "Impaired intracellular
trafficking of adeno-associated virus type 2 vectors limits efficient transduction of murine fibroblasts," J Virol 74, 992-926 (2000).
136. J. Hansen, K. Qing and A. Srivastava, "Adeno-associated virus type 2-mediated
gene transfer: altered endocytic processing enhances transduction efficiency in murine fibroblasts," J Virol 75, 4080-4090 (2001).
137. W. Zhao, L. Zhong, J. Wu, L. Chen, K. Qing, K. A. Weigel-Kelley, S. H. Larsen,
W. Shou, K. H. Warrington and A. Srivastava, "Role of cellular FKBP52 protein in intracellular trafficking of recombinant adeno-associated virus 2 vectors," Virology 353, 283-293 (2006).
138. L. Zhong, B. Li, G. Jayandharan, C. S. Mah, L. Govindasamy, M. Agbandje-
McKenna, R. W. Herzog, K. A. Weigel-Van Aken, J. A. Hobbs, S. Zolotukhin, N. Muzyczka and A. Srivastava, "Tyrosine-phosphorylation of AAV2 vectors and its consequences on viral intracellular trafficking and transgene expression," Virology 381, 194-202 (2008).
139. X. Qi, J. Cai, Q. Ruan, L. Liu, S. L. Boye, Z. Chen, W. W. Hauswirth, R. C. Ryals,
L. Shaw, S. Caballero, M. B. Grant and M. E. Boulton, "gamma-Secretase inhibition of murine choroidal neovascularization is associated with reduction of superoxide and proinflammatory cytokines," Invest Ophthalmol Vis Sci 53, 574-585 (2012).
140. M. Li, G. R. Jayandharan, B. Li, C. Ling, W. Ma, A. Srivastava, and L. Zhong,
"High-efficiency transduction of fibroblasts and mesenchymal stem cells by tyrosine-mutant AAV2 vectors for their potential use in cellular therapy," Hum Gene Ther 21, 1527-1543 (2010).
141. J. Yao, K. I. Maslov and L. V. Wang, "In vivo photoacoustic tomography of total
blood flow and potential imaging of cancer angiogenesis and hypermetabolism," Technology in Cancer Research and Treatment 11, 301-307 (2012).
142. A. Jemal, R. Siegel, E. Ward, Y. Hao, J. Xu, T. Murray and M. J. Thun., " Cancer
statistics 2008,"CA Cancer. J. Clin. 58, 71-96 (2008).
143. American Cancer Society. American Cancer Society, Atlanta (2010).
144. American Cancer Society. American Cancer Society, Atlanta (2011).
160
145. P. A. Carney, D. L. Miglioretti, B. C. Yankaskas, K. Kerlikowske, R. Rosenberg, C. M. Rutter, B. M. Geller, L. A. Abraham, S. H. Taplin, M. Dignan, G. Cutter and R. Ballard-Barbash, " Individual and combined effects of age, breast density, and hormone replacement therapy use on the accuracy of screening mammography," Ann. Intern. Med.138, 168-175 (2003).
146. D. Saslow, C. Boetes, W. Burke, S. Harms, M. O. Leach, C. D. Lehman, E.
Morris, E. Pisano, M. Schnall, S. Sener, R. A. Smith, E. Warner, M. Yaffe, K. S. Andrews and C. A. Russell, "American Cancer Society guidelines for breast screening with MRI as an adjunct to mammography," Ca-Cancer J. Clin. 57,75-89 CA (2007).
147. M. Hofmann, " From scinti-mammography and metabolic imaging to receptor targeted PETGnew principles of breast cancer detection," Phys. Med. 21, 11 (2006).
148. F. Benard and E. Turcotte, " Imaging in breast cancer: singlephoton computed
tomography and positron-emission tomography," Breast Cancer Res. 7,153-162 (2005).
149. L. Yang, X. H. Peng, Y. A. Wang, X. Wang, Z. Cao, C. Ni, P. Karna, X. Zhang, W.
C. Wood, X. Gao, S. Nie and H. Mao, "Receptor-targeted nanoparticles for in vivo imaging of breast cancer," Clin. Cancer Res. 15, 4722-4732 (2009).
150. E. I. Galanzha, E. V. Shashkov, T. Kelly, J. W. Kim, L. Yang and V. P. Zharov, "In
vivo magnetic enrichment and multiplex photoacoustic detection of circulating tumor cells," Nat. Nanotech. 4, 8558-8560 (2009).
151. M. Lipowska, G. Patonay and L. Strekowski, "New near-infrared cyanine dyes,
e.g. (1), for labeling of proteins," Synth. Commum. 23, 3087-3094 (1993).
152. S. DelVecchio, M. P. Stoppelli, M. V. Carriero, R. Fonti, O. Massa, P. Y. Li, G. Botti, M. Cerra, G. D'Aiuto, G. Esposito and M. Salvatore, "Human urokinase receptor concentration in malignant and benign breast tumors by in vitro quantitative autoradiography: comparison with urokinase levels," Cancer Res. 53,3198-3206 (1993).
153. Y. Li, N. Wood, D. Yellowlees and P. K. Donnelly, "Cell surface expression of
urokinase receptor in normal mammary epithelial cells and breast cancer cell lines," Anticancer Res. 19, 1223-1228 (1999).
154. E. Dublin, A. Hanby, N. K. Patel, R. Liebman and D. Barnes,
"Immunohistochemical expression of uPA, uPAR, and PAI-1in breast carcinoma. Fibroblastic expression has strong associations with tumor pathology," Am. J. Pathol. 157, 1219 -1227 (2000).
161
155. A. Hemsen, L. Riethdorf, N. Brunner, J. Berger, S. Ebel, C. Thomassen, F. Jänicke and K. Pantel, "Comparative evaluation of urokinase-type plasminogen activator receptor expression in primary breast carcinomas and on metastatic tumor cells," Int. J. Cancer 107, 903-909 (2003).
156. W. J. Hoskins, "Prospective on ovarian cancer: why prevent?" J. Cell Biochem
Suppl. 23,189-199 (1995).
157. American Cancer Society. Cancer facts &figures 2011. American Cancer Society, Atlanta, GA.
158. W. A. Rocca, B. R. Grossardt, de M. Andrade, G. D. Malkasian and L. J. Melton,
"Survival patterns after oophorectomy in premenopausal women: a population-based cohort study," Lancet Oncol. 7, 821-828 (2006).
159. Ovarian cancer: screening, treatment, and followup. NIH Consensus Statement 12, 1-30 (1994).
160. J. Tammela and S. Lele, "New modalities in detection of recurrent ovarian
cancer," Current Opinion in Obstetrics and Gynecology 16,5-9 (2004).
161. A. Shaaban and M. Rezvani, "Ovarian cancer: detection and radiologic staging," Clinical Obstetrics and Gynecology 52, 73-93 (2009).
162. H. S. Pan, S. L. Lee, L. W. Huang and Y. K. Chen, "Combined positron emission
tomography and tumor markers for detecting recurrent ovarian cancer," Archives of Gynecology and Obstetrics 283, 335-341 (2011).
163. von G. K. Schulthess, H. C. Steinert and T. F. Hany, "Integrated PET/CT: current
applications and future directions," Radiology 238, 405-422 (2006).
164. V. Ntziachristos, "Fluorescence molecular imaging," Annual Review of Biomedical Engineering 8, 1-33 (2006).
165. Q. Zhao, H. Jiang, Z. Cao, L. Yang, H. Mao and M. Lipowska, "A handheld
fluorescence molecular tomography system for intraoperative optical imaging of tumor margins," Med. Phys. 38, 5873-5878 (2011).
166. A. Taghian, M. Mohiuddin, R. Jagsi, S. Goldberg, E. Ceilley and S. Powell,
"Current perceptions regarding surgical margin status after breast-conserving therapy: results of a survey," Ann Surg. 241, 629–639 (2005).
167. L. Assersohn, T. J. Powles, S. Ashley, A. G. Nash, A. J. Neal, N. Sacks, J.
Chang, U. Querci della Rovere and N. Naziri, "Local relapse in primary breast cancer patients with unexcised positive surgical margins after lumpectomy, radiotherapy and chemoendocrine therapy," Ann Oncol. 10, 1451–1455 (1999).
162
168. G. C. Balch, S. K. Mithani, J. F. Simpson and M. C. Kelley, "Accuracy of intraoperative gross examination of surgical margin status in women undergoing partial mastectomy for breast malignancy" Am Surg. 71, 22–27 (2005).
169. B. Sigal-Zafrani, J. S. Lewis, K. B. Clough, A. Vincent-Salomon, A. Fourquet, M.
Meunier, M. C. Falcou and X. Sastre-Guarau, "Histological margin assessment for breast ductal carcinoma in situ: precision and implications," Mod Pathol. 17, 81–88 (2004).
170. Q. Zhang and P. A. Lewin, "Enhanced bandwidth multilayer transducer for
imaging applications," Archives of Acoustics 20, 77-90 (1995).
171. M. Fukuda, M. Nishihara and K. Imano, "Real time extraction system using double-layered piezoelectric transducer for second-harmonic ultrasound pulse waves," Jpn. J. Appl. Phys. 45, 4556-4559 (2006).
172. M. Nakazawa, M, Tararu, K. Nakamura, S. Ueha and A. Maezawa, "Multilayered
transducer using polyurea film,"Jpn. J. Appl. Phys. 46 (7B), 4466 (2007).
163
BIOGRAPHICAL SKETCH
Lei Xi received his B.S. in the College of Optoelectronic Science and Engineering
from Huazhong University of Science and Technology, China, in 2007. In Fall 2008, he
started to pursue his Ph.D. in the Department of Biomedical Engineering at the
University of Florida. His research focused on photoacoustic tomography for clinical
breast detection, intraoperative photoacoustic imaging and photoacoustic microscopy
for biological cancer research.
![Nonlinear quantitative photoacoustic tomography with two …kr2002/publication_files/Ren-Zhang-TP-PAT-2016.pdf · Two-photon photoacoustic tomography (TP-PAT) [35,36,51,53,56,57,58,60,59]](https://img.pdfslide.net/doc/110x75/5e26be0daa2e5d594541a49c/nonlinear-quantitative-photoacoustic-tomography-with-two-kr2002publicationfilesren-zhang-tp-pat-2016pdf.jpg)
