Characterising and improving image quality in Optical ......Characterising and improving image...

Characterising and improving image quality in Optical Coherence Tomography

and Elastography by means of optical beam shaping and simulations

Andrea Curatolo

BSc, MSc

This thesis is presented for the degree of Doctor of Philosophy

of The University of Western Australia School of Electrical, Electronic & Computer Engineering

January, 2017

THE UNIVERSITY OF WESTERN AUSTRALIA

ii

Abstract

Most diseases manifest themselves when symptoms affect the patient at the organ level, but

they originate at the molecular and cellular level, with clear signatures, such as changes to

morphology or stiffness. Understanding disease at such a level may permit earlier, more

specific diagnosis, and may improve targeted treatment strategies; however, it requires

imaging techniques that are sensitive and specific to such small variations in cell and tissue

micro-environment properties. Unfortunately, medical imaging technologies, capable of

performing such high-resolution imaging at depths of up to several millimetres in tissue,

are limited and need improvement.

One path to understanding the genesis and progression of disease, and to early disease

detection, lies in the advancement of such a microscopy technique. Optical coherence

tomography (OCT) holds much promise towards this end, as an optical, three-dimensional,

non-invasive, high-resolution imaging technique, with the ability to penetrate tissue,

reaching 2-3 mm below the surface. Optical coherence elastography (OCE), an extension

of OCT to image a sample’s three-dimensional stiffness distribution, can aid in this

scientific and clinical effort, by providing complementary information on the tissue

mechanical properties.

Nevertheless, the penetration depth and image contrast of OCT are fundamentally

limited by light scattering in biological tissue. In addition, OCE currently lacks the

resolution to visualise mechanical interactions at the cellular scale.

Characterisation and improvement of image quality in OCT and OCE are fundamental

to the translation of these techniques into reliable, non-invasive providers of biological

tissues’ microscopic structural and functional information, and for their range of

applications.

The first part of this thesis describes the methods used: firstly, to alter and improve

OCT and OCE image quality; secondly, to compare it in realistic and controlled turbid

tissue scenarios; and, thirdly, to analyse it. We use energy-efficient Bessel beams to alter

image quality, as a viable alternative to conventional focussing schemes using Gaussian

beams. We design novel structured phantoms to compare image quality in turbid tissue,

providing fine control of the optical, mechanical and structural properties, and aiding in the

benchmarking effort. We implement and use novel simulations of beam propagation and

image formation in turbid samples to analyse image quality and guide system development.

In the second part of this thesis, we employ the previously described tools to quantify

the effect that using energy-efficient Bessel beams has on OCT image quality in turbid

iii

tissue. We demonstrate that the Bessel beam’s increased depth of field comes at the

expense of both reduced peak OCT sensitivity and contrast, when compared to a Gaussian

beam of identical transverse resolution and optical power. This is because the Bessel beam

does not reconstruct its amplitude profile any better than the Gaussian beam when

propagating through a turbid medium that exhibits distributed scattering, as is typical with

soft tissue. We also show, however, that for fixed focus beams, Bessel beams result in

contrast and resolution superior to that of Gaussian beams, at depths where, in free-space,

they possess a higher irradiance than the Gaussian beam. To improve OCT contrast

further, alternative adaptive approaches or dynamic focussing is required.

We then demonstrate an extended-focus optical coherence microscope and use it to

perform ultrahigh-resolution optical coherence elastography, which achieves an

unprecedented (𝑥𝑥 𝑦𝑦 𝑧𝑧) 2×2×15 μm strain resolution over a depth of field of nearly 100 μm.

This is obtained by combining Bessel beam illumination and Gaussian beam detection,

compromising between depth of field improvement and contrast reduction. Our

multiphysics phase-sensitive compression OCE simulation of the influence of the system

resolution and the applied load on the measured strain precision guided our acquisition

protocols to maximise the strain sensitivity. We demonstrate this record performance on a

structured phantom and freshly excised mouse aorta.

The tools developed for image quality analysis, the characterisation of the influence of

Bessel beams on contrast in OCT, and the resolution improvement demonstrated in

ultrahigh-resolution OCE, provide important contributions to the ability of OCT and

OCT-based techniques to provide better microscopic structural and functional information

on biological tissue.

iv

Contents

Abstract ii

Contents iv

List of figures viii

List of tables xiv

Acknowledgements xv

Statement of contribution xvii

List of publications xxiii

List of acronyms xxvi

Chapter 1 Introduction 1

1.1 Research objectives .................................................................................................. 4

1.2 Structure of the thesis .............................................................................................. 4

Chapter 2 Background of OCT and OCE 9

2.1 Optical coherence tomography .............................................................................. 9

2.1.1 Low coherence interferometry ....................................................................... 16

2.1.2 Time-domain optical coherence tomography .............................................. 21

2.1.3 Fourier-domain optical coherence tomography .......................................... 26

2.2 Optical coherence elastography ............................................................................ 29

2.2.1 Elasticity ............................................................................................................ 30

2.2.2 Principles of optical coherence elastography ............................................... 35

2.3 Image quality ........................................................................................................... 38

2.3.1 OCT resolution and depth of field ................................................................ 39

2.3.2 OCT signal-to-noise ratio, sensitivity and contrast ..................................... 43

2.3.3 Speckle ............................................................................................................... 45

2.3.4 Elastogram quality............................................................................................ 47

2.4 Conclusion ............................................................................................................... 49

Chapter 3 Energy-efficient Bessel beams: a tool to alter image quality 50

3.1 Optical beam shaping ............................................................................................ 50

3.1.1 Hermite-Gaussian beams, Laguerre-Gaussian beams and non-diffracting

beams ................................................................................................................. 52

3.1.2 Beam shaping with a passive component versus an active (reconfigurable)

optical element .................................................................................................. 55

3.1.3 Location of the beam shaper in the optical path ......................................... 57

v

3.1.4 Aberration characteristics with refractive versus diffractive optical

elements .............................................................................................................59

3.2 Energy-efficient low-Fresnel-number Bessel beams and their application in optical

coherence tomography [27] ...................................................................................64

3.3 Selected dual beam OCT setup .............................................................................71

3.4 Conclusion ...............................................................................................................72

Chapter 4 Structured phantoms: a tool to mimic turbid tissue and compare image

quality 75

4.1 Introduction .............................................................................................................75

4.2 Light-tissue interaction ...........................................................................................76

4.2.1 Optical properties .............................................................................................78

4.3 Controlling the optical properties of OCT phantoms .......................................85

4.3.1 OCT silicone phantoms with varying attenuation coefficients .................86

4.3.2 OCT silicone phantoms with varying anisotropy ........................................87

4.3.3 Scattering overlayers validation ......................................................................88

4.4 Mechanical properties.............................................................................................90

4.4.1 Controlling the mechanical properties of OCT silicone phantoms ..........92

4.5 Structured three-dimensional optical phantom for optical coherence tomography

[30] .............................................................................................................................95

4.5.1 Introduction ......................................................................................................95

4.5.2 Method ...............................................................................................................96

4.5.3 Results and discussion .....................................................................................99

4.5.4 Conclusions .................................................................................................... 102

4.5.5 Aknowledgements ......................................................................................... 102

4.6 Image quality test targets ..................................................................................... 103

4.6.1 Point spread function and contrast phantoms characteristics ................ 103

4.6.2 Phantom validation ....................................................................................... 105

4.7 Elastography structured phantoms.................................................................... 108

4.8 Conclusion ............................................................................................................ 110

Chapter 5 Simulation of beam propagation and image formation in turbid samples:

a tool to theoretically quantify image quality 111

5.1 Introduction .......................................................................................................... 111

5.2 OCT image simulation in the single-scattering regime ................................... 112

5.2.1 OCT image simulation as local sums of random phasors ....................... 115

5.2.2 OCT image simulation as superposition of linear system responses ..... 120

vi

5.3 A full wave 2-D model of image formation in optical coherence tomography

applicable to general samples [32] ................................................................... 125

5.3.1 Details of the model ...................................................................................... 128

5.3.2 Evaluation and analysis ................................................................................. 133

5.3.3 Discussion and conclusions .......................................................................... 144

5.3.4 Acknowledgments .......................................................................................... 144

5.4 3-D simulation of optical beam propagation in phantoms ............................ 145

5.4.1 Simulation of beam propagation in free-space .......................................... 145

5.4.2 Simulation of beam propagation in tissue phantoms ............................... 147

5.5 Conclusion ............................................................................................................. 149

Chapter 6 Quantifying OCT image quality in turbid samples using Gaussian and

Bessel beams 151

6.1 Introduction .......................................................................................................... 151

6.1.1 Testing arrangement: sample configuration with a Gaussian beam ....... 152

6.2 Quantifying the influence of Bessel beams on image quality in optical coherence

tomography [28].................................................................................................... 155

6.2.1 Introduction .................................................................................................... 155

6.2.2 Results .............................................................................................................. 158

6.2.3 Discussion ....................................................................................................... 164

6.2.4 Methods ........................................................................................................... 168

6.3 Conclusion ............................................................................................................. 171

Chapter 7 Improving OCE image quality: strain precision and resolution 173

7.1 Introduction .......................................................................................................... 173

7.1.1 Phase-sensitive OCT displacement measurement precision ................... 173

7.2 Analysis of image formation in optical coherence elastography using a multiphysics

approach [25] ......................................................................................................... 177

7.2.1 Introduction .................................................................................................... 177

7.2.2 Metrics of elastogram quality and precision ............................................... 179

7.2.3 Multiphysics model of optical coherence elastography ............................ 180

7.2.4 Experimental procedure ................................................................................ 184

7.2.5 Results .............................................................................................................. 187

7.2.6 Discussion ....................................................................................................... 190

7.2.7 Implications for the ultrahigh-resolution regime ...................................... 194

7.3 Ultrahigh-resolution optical coherence elastography [31] ............................. 196

7.4 Conclusion ............................................................................................................. 203

vii

Chapter 8 Conclusions 205

8.1 Research contributions and significance ........................................................... 206

8.2 Study limitations and future work ..................................................................... 211

8.2.1 Current limitations ......................................................................................... 211

8.2.2 Proposed future work ................................................................................... 213

8.3 Final remarks ........................................................................................................ 214

Appendix A Spatial light modulator characterisation 215

Bibliography 225

viii

List of figures

Figure 2-1. Schematic of terminology used when referring to OCT scan orientation.

......................................................................................................................... 10

Figure 2-2. Spectral absorption for a range of tissue chromophores in the diagnostic

window in the near-infrared. ....................................................................... 11

Figure 2-3. Comparison of OCT resolution and imaging depths to other biomedical

imaging methods. .......................................................................................... 12

Figure 2-4. Bar chart of a snapshot of results from a Scopus search of selected OCT

biological applications from 2005 to 2015, inclusive. .............................. 13

Figure 2-5. Main applications of OCT. ............................................................................. 13

Figure 2-6. Video showcasing representative examples of the main clinical and

laboratory research applications of OCT. ................................................. 14

Figure 2-7. Video showcasing areas of fundamental research aimed at image quality

improvement in OCT. .................................................................................. 14

Figure 2-8. Optical coherence tomography (OCT) setups. ............................................ 15

Figure 2-9. Transform-limited pulse generation by addition of weighted spectral

components with their phases locked to each other. ............................... 17

Figure 2-10. Temporal evolution of polychromatic wave generated by addition of

weighted spectral components with their phases randomly assigned. .. 18

Figure 2-11. Temporal evolution of monochromatic (left) wave and a polychromatic

CW (right) wave. ........................................................................................... 19

Figure 2-12. Spectra (in green) of the monochromatic (left) and polychromatic (right)

waves shown in Figure 2-11 after 143 ps. ................................................. 19

Figure 2-13. Autocorrelation of the monochromatic (left) and polychromatic (right)

waves shown in Figure 2-12. ....................................................................... 20

Figure 2-14. Equivalence of temporal autocorrelation of polychromatic waves

generated by a transform-limited pulse (left) and by a continuous wave

(CW) (right). ................................................................................................... 21

Figure 2-15. Interferometry with a monochromatic wave. ............................................. 23

Figure 2-16. TD-OCT photocurrent over time. .............................................................. 24

Figure 2-17. Field amplitudes (left panel), electric charge generated at the

photodetector (middle panel) and detector photocurrent as a function

of reference arm group delay. ...................................................................... 25

ix

Figure 2-18. An example of the spectral interference generated by a discrete sample

reflectivity function (top), and the corresponding A-scan captured by

FD-OCT (bottom). .......................................................................................28

Figure 2-19. Reported values and ranges of Young’s modulus for selected tissues and

tissue constituents. .........................................................................................34

Figure 2-20. Illustration of phase-sensitive compression optical coherence

elastography on a structured phantom. ......................................................37

Figure 2-21. Optical coherence elastography of a malignant breast tumour with

quantitative elasticity estimation. .................................................................38

Figure 2-22. Schematic illustrating diffraction-limited resolution in microscopy. .......40

Figure 2-23. Schematic illustrating the determinants of resolution in OCT. ...............42

Figure 2-24. Calculated sensitivity for a typical spectrometer-based SD-OCT system.

..........................................................................................................................44

Figure 2-25. OCT image of a human fingertip with high contrast speckle pattern

plotted on a logarithmic grayscale. ..............................................................45

Figure 2-26. Marginal probability density function 𝑝𝑝𝑝𝑝𝑝𝑝 for the OCT signal phase. .48

Figure 3-1. Trade-off between the DOF and transverse resolution in OCT, at

wavelength of 0.84 µm in air. ......................................................................51

Figure 3-2. Collimated Gaussian beam shaped and focussed by different lenses. ......54

Figure 3-3. Examples of shaped beams. ............................................................................55

Figure 3-4. Static optical elements: lenses. .........................................................................56

Figure 3-5. Reconfigurable beam shapers. .........................................................................56

Figure 3-6. Bessel beam and its typical spatial frequency spectrum. .............................58

Figure 3-7. Typical experimental configurations for the generation of Bessel beams.

..........................................................................................................................58

Figure 3-8. Refractive and diffractive optical elements. ..................................................60

Figure 3-9. Simulations of the beam profile of a Bessel beam, using VirtualLab. .......61

Figure 3-10. Graphical illustration of the cause of the achromatic characteristics of a

DOE-generated Bessel beam. ......................................................................61

Figure 3-11. Simulation results for the diffraction efficiency of an SLM-generated

Bessel beam. ...................................................................................................62

Figure 3-12. Energy efficiency of a Bessel beam compared to a Gaussian beam of

same NA. .........................................................................................................66

Figure 3-13. Experimental setup for OCT imaging using low-Fresnel-number Bessel

beams. ..............................................................................................................68

x

Figure 3-14. Experimental results comparing the imaging performance of low-

Fresnel-number Bessel beams and a Gaussian beam of the same

transverse resolution. .................................................................................... 69

Figure 3-15. Dual-beam OCT system schematic. ............................................................ 72

Figure 3-16. Self reconstruction property of a Bessel beam........................................... 73

Figure 4-1. Scattering interaction in tissue classified into three categories. ................. 78

Figure 4-2. Mie theory of scattering of a plane wave by a sperical dielectric particle. 81

Figure 4-3. Typical average OCT signal, on a logarithmic scale, as a function of

depth. .............................................................................................................. 83

Figure 4-4. OCT signal characteristics as a function of increasing scatterer

concentration. ................................................................................................ 86

Figure 4-5. Scattering overlayers. ........................................................................................ 88

Figure 4-6. Measurements of elasticity and viscoelasticity of materials and apparatus.

......................................................................................................................... 92

Figure 4-7. Comparison of range of Young’s moduli of phantom materials and soft

tissue. ............................................................................................................... 93

Figure 4-8. 3-D structured phantom design and characterisation. ................................ 97

Figure 4-9. OCT images of Phantom I and II. ............................................................... 100

Figure 4-10. Video of sequential multiplanar views and a fly-through of the 3-D solid

rendering of the phantom. ......................................................................... 100

Figure 4-11. Speckle reduction performed on Phantom II. ......................................... 101

Figure 4-12. Imaging targets. ............................................................................................. 104

Figure 4-13. OCT characterisation of the imaging targets with Gaussian and Bessel

beams. ........................................................................................................... 106

Figure 4-14. Ultrahigh-resolution OCE inclusion phantom. ....................................... 109

Figure 5-1. Simulation of the axial distortion of an OCT A-scan caused by speckle as

a result of the coherent detection process. .............................................. 113

Figure 5-2. Simulation of the one-dimensional convolution operation resulting in the

single-scattering OCT A-scan signal phasor. .......................................... 114

Figure 5-3. Comparison between images of a structured phantom acquired using a

coherent (OCT) system (a) and a bright field microscope, i.e., an

incoherent system (b). ................................................................................ 119

Figure 5-4. Simulation of (a) a hypothetical incoherent and (b) a coherent (OCT)

image formation process, according to the simple model presented in

this section.................................................................................................... 120

xi

Figure 5-5. Illustration of the model of image formation in OCT, and comparison

with experimental results. .......................................................................... 123

Figure 5-6. A schematic of the modelled optical system. ............................................. 129

Figure 5-7. Schematic diagram of where the source field is introduced 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 and

where the scattered field is recorded 𝑆𝑆𝑆𝑆𝑆𝑆 in FDTD simulations. ........ 132

Figure 5-8. Numerical dispersion. .................................................................................... 135

Figure 5-9. Simulation of the OCT image of a stratified medium. ............................. 136

Figure 5-10. Images and plots which demonstrate the depth-dependent PSFs

employed in OCT by simulating images of 24 scatterers arranged

equidistantly along the optical axis. .......................................................... 137

Figure 5-11. 3-D structured phantom for OCT. ........................................................... 140

Figure 5-12. Diagram showing how a collection of spherical scatterers in a three-

dimensional beam (left) can be approximated in a two-dimensional

system (right). .............................................................................................. 140

Figure 5-13. Simulation of OCT image formation for a structured phantom. ......... 141

Figure 5-14. Optical beams: Gaussian and Bessel beams of equal resolution and

power. ........................................................................................................... 146

Figure 5-15. Illustration of the EM simulation. ............................................................. 147

Figure 5-16. (a) Simulated and (b) experimental transverse PSFs for Gaussian and

Bessel beams, top and bottom, respectively. .......................................... 149

Figure 6-1. Effect of increasing degree of wide-angle multiple scattering

(Category III) signal on OCT images of the two different structured

phantoms, described in Section 4.5. ........................................................ 153

Figure 6-2. Effect of increasing degree of low-angle forward scattering (Category II)

signal on OCT images of the Structured Phantom I overlaid by

phantoms containing different diameter polystyrene beads. ................ 154

Figure 6-3. Beam and sample configuration. (a) Schematic diagram of the beam,

overlayer and imaging target, for transparent (top) and scattering

overlayers (bottom). ................................................................................... 157

Figure 6-4. Simulated beam profiles after propagation through Overlayers 1–3 with

increasing anisotropy from left to right compared to (left) propagation

through a scattering-free (SF) overlayer. ................................................. 159

Figure 6-5. Simulated transverse beam profiles. ............................................................ 160

Figure 6-6. Simulated total beam irradiance and image-carrying and -degrading

components of the beam versus depth in the PSF phantom after each

overlayer. ...................................................................................................... 161

xii

Figure 6-7. On-axis signal-to-background ratio (SBR) versus depth for Overlayers 1–

3. .................................................................................................................... 162

Figure 6-8. OCT contrast assessment. ............................................................................. 163

Figure 6-9. Angular spectra intensity of the beams. ...................................................... 166

Figure 7-1. Origin of translation-induced phase decorrelation in one dimension. ... 175

Figure 7-2. Origin of strain-induced phase decorrelation in one dimension. ............ 176

Figure 7-3. FEM simulation of a sample containing a stiff inclusion under quasi-static

compression. ................................................................................................ 181

Figure 7-4. Computing the new location of the scattering potentials under an applied

load. ............................................................................................................... 183

Figure 7-5. Flowchart of the multiphysics simulation of OCE. .................................. 183

Figure 7-6. Inclusion phantom and phase-sensitive compression OCE signal

signatures. ..................................................................................................... 186

Figure 7-7. Experimental scans of a silicone inclusion phantom compared to results

of the multiphysics simulation of phase-sensitive compression OCE.

....................................................................................................................... 187

Figure 7-8. Regions used for comparing experiment to simulation, shown on (a) the

simulated OCT image, and (b) the simulated strain elastogram. ......... 188

Figure 7-9. Displacement sensitivity (𝑆𝑆𝑠𝑠) vs. (a) local strain at various depths in the

sample, and (b) vs. depth at selected values of strain in the sample. ... 189

Figure 7-10. Strain SNR (SNR𝜀𝜀) (a) vs. local strain at various depths in the sample,

and (b) vs. depth at selected values of strain in the sample. ................. 190

Figure 7-11. Effect of the OCT axial resolution on the precision of the sample axial

displacement measurement. ....................................................................... 195

Figure 7-12. UHROCE system. ........................................................................................ 197

Figure 7-13. OCT images and strain elastograms of an inclusion phantom taken with

two systems. ................................................................................................. 200

Figure 7-14. OCT images and strain elastograms of a mouse aorta taken with the two

systems: OCE and UHROCE, compared with histology. .................... 202

Figure A-1. Illustration of the connection diagram of the SLM, with a picture of one

module, a close-up of the LCoS screen, with a schematic of the pixel

structure and the operating principle. ...................................................... 216

Figure A-2. Installed and connected SLM. ..................................................................... 217

Figure A-3. Testing setup for the SLM characterisation. .............................................. 217

Figure A-4. Setup to test the effect of phase droop on the Gaussian beam focus. .. 219

xiii

Figure A-5. Time sequence over 8.34 ms of 8 images of the SLM screen and

corresponding transverse focal intensity. ................................................ 219

Figure A-6. Recurrent defocus. ........................................................................................ 220

Figure A-7. Time-resolved (over 50 ms) powers in the 0th and 1st diffractive orders as

a function of the modulation depth discretization and grating period for

the two SLM modules loaded with different configurations. .............. 221

Figure A-8. Phase droop as a function of the corresponding grey level for the two

SLM modules and various configurations. ............................................. 222

Figure A-9. SLM phase calibration to obtain a linear conversion from grey scale to

phase retardation. A 256 grey level offset corresponds to a 2π phase

shift for a given wavelength. ..................................................................... 223

Figure A-10. Characterisation of the linearity of the phase retardation with grey level.

....................................................................................................................... 224

xiv

List of tables

Table 2-1. Imaging parameters for OCE displacement estimation methods. Adapted

from [18]. ........................................................................................................ 35

Table 2-2. Experimental and theoretical imaging parameters for OCE loading

methods. Adapted from [18]. ...................................................................... 35

Table 3-1. Experimental comparison Bessel vs. Gaussian ............................................. 70

Table 4-1. Specified and calculated scattering overlayer characteristics ....................... 87

Table 4-2. Young’s modulus of various mixing ratios (Cross-linker: Catalyst) for

Wacker Elastosil 601 silicone phantoms ................................................... 92

Table 4-3. Young’s modulus of various mixing ratios (Cross-linker: Catalyst: PDMS

oil) for Wacker Elastosil 601 silicone and Wacker AK50 PDMS oil

phantoms ........................................................................................................ 93

Table 4-4. Contrast phantom nominal characteristics ................................................... 105

Table 5-1. Simulated and experimental beam characteristics ....................................... 145

Table 5-2. Simulation characteristics ................................................................................ 148

Table 6-1. Reduction in contrast ...................................................................................... 163

Table 6-2. Simulated and experimental beam characteristics ....................................... 169

Table 6-3. Specified and calculated scattering overlayer characteristics ..................... 169

Table 6-4. Contrast phantom nominal characteristics ................................................... 170

Table 7-1. OCE simulation inputs and parameters ....................................................... 184

Table 7-2. Numerical comparison of experimentally acquired vs. simulated

elastograms for the regions marked in Figure 7-8. ................................. 188

xv

Acknowledgements

Many people have helped, directly or indirectly, to shape this thesis, and I personally want

to acknowledge them for their contribution and support.

First, David Sampson, for giving me the opportunity to join the Optical + Biomedical

Engineering Laboratory (OBEL) in 2008 and, after several years of research appointments,

encouraging me to pursue a PhD degree to capitalise on my experience. Thanks David for

mentoring me, providing and leading an excellent research facility, with high-profile

international collaborations and a very healthy work environment that promotes work-life

balance. I also feel very grateful to my other supervisors, Peter Munro and Dirk Lorenser,

who have taught me a great deal of knowledge and helped me manage this research project

effectively, navigating through a lot of unknowns and difficulties, technical or otherwise. I

am also very thankful to The University of Western Australia for supporting me through an

Australian Postgraduate Award scholarship.

Special thanks goes to my newfound home of Western Australia and Perth. It helped

me connect my work with the places and nature around me through countless and

beautiful examples of the ubiquity of the topics being studied in this thesis. These awe-

inspiring optical and physical phenomena might take a different form from those

researched, but they are all related: light scattering by the west coast pristine sky in

enchanting sunsets, refraction by water droplets in Perth’s winter rainbows, reduced

visibility by stormy and foggy clouds over the river, wave interference patterns in

Freshwater bay, summer light attenuation by sunscreen at Floreat Beach.

My professional development and research creativity have greatly benefitted from in-

depth discussions and guidance by colleagues and high-calibre researchers Timothy

Hillman, Brendan Kennedy, Maciej Wojtkowski and Martin Villiger. Without their input

and our exchange of ideas, the richness of this research would have suffered.

I want to thank Blake Klyen and Loretta Scolaro for having been, not only peers

providing the biggest continuing support and having shared their research experience with

me, but also for being true friends.

Other colleagues and fellow PhD candidates have been a source of knowledge and

inspiration throughout the duration of this research: Bryden with his fearless “can do”

attitude, Philip with his humbleness and wit, Kelsey with her positive work ethics, Lixin

with his competence and diligence, Robert with his focus and drive and the rest of the

OBEL team and collaborators with their unique characteristics.

xvi

I cannot forget the moral support and keen interest from Cibele in my research and my

career development, which strengthened my confidence in the pursuit of this achievement.

Roberta, Laurent, Monica, Seon, Danila, Uros, Cris, Valerio and Jana, you are all good

friends and I am fortunate to have you in my life, and I am sure you are all delighted by this

accomplishment. Through your smiles, our laughs and the good times together, you have

helped me strike the right balance, and tackle the intricacies of research with a positive

attitude. Thanks to my housemates and friends, especially Kat, Lana, Dario and Jon who

have been supportive and understanding at all times.

Thanks to all the UWA OSA student chapter members, and Danka in particular, for

having created something fun and valuable in Perth for the young and upcoming optics

and photonics community, through professional development and scientific outreach

efforts.

Mens sana in corpore sano. A very true Latin statement, and for me a healthy body comes

from the sporting activity that I enjoy the most: volleyball. So thanks to all my volleyball

and beach volleyball mates.

Heather and Chris, you have been like family to me here in Perth. From the very

beginning, you helped me settle in the very best of ways, and I will always be grateful to

you for that and the warmth you gave me.

Last, but not least, thanks to my friends and family back home in Italy. Thanks to

Francesca, Chino and Andrea. Your affection has travelled through continents. Grazie to

my close relatives, my cousins, my late grandparents, my brother Marco and my parents

Lucia and Roberto. Your unconditional encouragement and love has provided, at the same

time, a safe haven to reflect upon my values and interests, and a springboard that

strengthened my motivation to look ahead and to reach new grounds.

xvii

Statement of contribution

This thesis contains the results of the research that I, Andrea Curatolo (AC), performed

within the School of Electrical, Electronic, and Computer Engineering at The University of

Western Australia, between 2011 and 2016. Except where indicated below and throughout

the text, all work and writing are my own. The chapters of this thesis are primarily derived

from nine published works: four first-authored, journal articles (three full-length and one

letter); three co-authored, journal articles (two full-length and one letter); one co-authored

review article; and one first-authored book chapter1. I am the sole author of the remainder

of the document. Sections 3.2, 4.5, 5.3, 6.2, and 7.3 are reproductions of publications, as

reported below, included without change in content, but modified only in formatting for

consistency with the remainder of this document. I am the sole author of Chapters 1, 8,

Appendix A and the vast majority of Chapter 2.

Listed below are the contributions of each author to each publication. My contributions

are given as a percentage next to my name, and in the descriptions.

1. B. F. Kennedy, R. A. McLaughlin, K. M. Kennedy, L. Chin, A. Curatolo

(15%), A. Tien, B. Latham, C. M. Saunders, and D. D. Sampson, "Optical

coherence micro-elastography: mechanical-contrast imaging of tissue

microstructure," Biomedical Optics Express 5, 2113-2124 (2014).

Publication 1 features in parts in Section 2.2 and in Section 4.7.

B. F. Kennedy (BFK) was the principle author of this article. BFK led the design of the

system, and all experiments, assisted with the design of the signal processing code, and led

the writing of the manuscript, which was reviewed and edited by all co-authors. R. A.

McLaughlin (RAM) assisted with analysis of results, provided software for, and assistance

with, co-registration with histology, organised access to tissue and clinical collaboration,

and helped develop the signal processing code. K. M. Kennedy (KMK) fabricated

phantoms, helped performed phantom imaging, tissue imaging, and interpretation of

results. L. Chin (LC) helped develop the signal processing chain, performed data

processing, helped perform tissue imaging and generated schematic figures. AC assisted

with early technology development, phantom and tissue imaging, and generated schematic

figures. A. Tien (AT) obtained patient consent, facilitated access to tissue, and in some

1 This thesis was prepared according to The University of Western Australia’s guidelines on “Thesis ss a series of papers”

(http://www.postgraduate.uwa.edu.au/students/thesis/series. Accessed 27 June 2016).

http://www.postgraduate.uwa.edu.au/students/thesis/series

xviii

cases performed surgery. B. Latham (BL) interpreted histopathology for diagnosis, and

assisted co-registration to OCT and OCE data. C. M. Saunders (CMS) performed surgeries,

aided in the conceptual development of the research, and facilitated access to tissue. D. D.

Sampson (DDS) conceptualised the work, provided overall guidance, facilitated

collaboration, and supervised drafting of the manuscript.

2. L. Chin, A. Curatolo (40%), B. F. Kennedy, B. J. Doyle, P. R. T. Munro, R. A.

McLaughlin, and D. D. Sampson, “Analysis of image formation in optical

coherence elastography using a multiphysics approach,” Biomedical Optics

Express 5, 2913–2930 (2014).

Publication 2 features in parts in Section 2.2.2, in Section 4.7 and in Section 5.2.2. Section

7.2 includes it from the methods section until the conclusion.

LC and AC were the joint principle authors of this article. AC fabricated the inclusion

and homogeneous phantoms, performed the phantom imaging experiments, led the design

of the optical simulation, performed the OCE simulation of the homogeneous phantoms,

led the data analysis and generated the first draft of the figures. LC led the design of the

mechanical simulation, and the combination of the optical and mechanical models,

optimised the simulation code, performed the OCE simulation of the inclusion phantom,

refined the figures, and led the writing of the manuscript, which was reviewed and edited

by all co-authors. BFK provided guidance and supervision, and facilitated collaboration. B.

J. Doyle (BJD) provided advice and training on the mechanical simulation. P. R. T. Munro

(PRTM) provided advice and guidance on the optical simulation. RAM provided advice on

the simulation code. DDS provided overall guidance and supervision of the research.

3. A. Curatolo (60%), B. F. Kennedy, D. D. Sampson, and T. R. Hillman,

"Speckle in Optical Coherence Tomography," in Advanced Biophotonics: Tissue

Optical Sectioning (V. V. Tuchin, and R. Wang, eds), Taylor & Francis: 211-277

(2013).

Publication 3 features in parts in Section 2.3.3, in Section 4.2 and in Section 5.2.1.

AC was the principle author of this book chapter. AC prepared an extensive literature

review, proposed the chapter structure, organised the material, collected and updated

existing figures, generated new figures, and led the writing of the chapter, which was

reviewed and edited by all co-authors. BFK led the writing of Section 6.7 about mitigation

of OCT speckle. DDS conceptualised the work, provided overall guidance, facilitated the

discussion and timeline management with the editors, and supervised drafting of the

xix

manuscript. T. R. Hillman (TRH) provided published and unpublished material for the

chapter and strong technical and writing guidance.

4. D. Lorenser, C. C. Singe, A. Curatolo (20%), and D. D. Sampson, "Energy-

efficient low-Fresnel-number Bessel beams and their application in optical

coherence tomography," Optics Letters 39, 548-551 (2014).

Publication 4 appears in full as published in Section 3.2.

D. Lorenser (DL) was the principle author of this article. DL led the engineering and

development of the theory of low-Fresnel Bessel beams and their energy efficiency in

comparison to equal resolution Gaussian beams. DL designed the dual beam OCT system

and the software to extract the DOF improvement and OCT sensitivity penalty from OCT

images of a nanoparticle embedded phantom. He also led the writing of the manuscript,

which was reviewed and edited by all co-authors. C. C. Singe (CCS) contributed to the

system design through simulations, built parts of the OCT system, aligned it and help

characterise it. AC chose, characterised and calibrated the beam shaper. He built parts of

the OCT system, measured and verified its specifications. He also acquired and processed

the images of phantoms and biological tissue, and helped prepare some of the manuscript

figures. DDS conceptualised the work, provided overall guidance and supervision of the

research.

5. A. Curatolo (65%), P. R. T. Munro, D. Lorenser, P. Sreekumar, C. C. Singe, B.

F. Kennedy, and D. D. Sampson, "Quantifying the influence of Bessel beams

on image quality in optical coherence tomography," Scientific Reports 6, 23483

(2016).

Publication 5 features in parts of its supplementary information in Section 3.3, in Section

4.3, in Section 4.6 and in Section 5.4. The main manuscript also appears in full as published

in Section 6.2.

AC was the principle author of this article. AC prepared an extensive literature review

and proposed the image quality benchmarking setup. AC also rebuilt, aligned and

characterised the selected dual beam OCT setup. He designed and validated the scattering

overlayers, i.e., the tissue-mimicking phantoms with varying scattering anisotropy. He

designed and sourced the 3-D structured phantom used as a contrast target. AC wrote code

to process the OCT images, to verify phantom specifications, to model the numerical

phantoms and to measure and analyse OCT contrast. AC performed the experiments and

processed the simulation results. AC generated the manuscript figures and led the writing

xx

of the article, which was reviewed and edited by all co-authors. PRTM designed, provided

and run the simulation code to analyse the beam propagation through the scattering

numerical phantoms using a rigorous full-wave solver of Maxwell’s equations. He also

provided supervision of the project, manuscript overview, guidance and insight into

scattering phenomena interpretation. DL supervised the dual beam OCT setup redesign

and characterisation, provided supervision of the project, and useful discussion about the

validation of the imaging test target. P. Sreekumar (PS) manufactured the scattering

overlayers and helped in their validation and verification of their specifications. CCS helped

in the alignment and characterisation of the dual beam OCT setup and in the initial stages

of the scattering overlayers validation. BFK provided help and insight in the design of the

contrast target. DDS conceptualised the work, provided overall guidance and supervision

of the research with insightful discussions.

6. G. Lamouche, B. F. Kennedy, K. M. Kennedy, C.-E. Bisaillon, A. Curatolo

(10%), G. Campbell, V. Pazos, and D. D. Sampson, "Review of tissue

simulating phantoms with controllable optical, mechanical and structural

properties for use in optical coherence tomography," Biomedical Optics

Express 3, 1381-1398 (2012).

Publication 6 features in parts from Section 4.1 to Section 4.4.

Guy Lamouche (GL) and BFK were the joint principle authors of this article. GL led

the writing of the manuscript, which was reviewed and edited by all co-authors. BFK

supervised the writing of all sections relating to mechanical properties, fibrin phantoms,

and complex structures. KMK led the writing of the mechanical properties section of the

article. C.-E. Bisaillon (CEB) led the experimental work on PVA phantoms and

characterisation of optical properties. AC led the writing of the first part of the complex

structures section of the article. G. Campbell (GC) consulted on the artery phantom work.

V. Pazos (VP) consulted on sections regarding PVA phantoms. DDS supervised the

drafting of the manuscript.

7. A. Curatolo (70%), B. F. Kennedy, and D. D. Sampson, "Structured three

dimensional optical phantom for optical coherence tomography," Optics

Express 19, 19480-19485 (2011).


AC was the principle author of this article. AC contracted the Australian National

Fabrication Facility (ANFF) node with equipment and expertise in advanced micro-fluidic

xxi

device fabrication, with the goal of manufacturing a structured OCT phantom. AC

designed the 3-D structured phantom, and pursued its validation through several

interactions with the ANFF by overcoming manufacturing constraints. He then verified the

specifications of the four different manufactured types. He performed the experiments and

the image and data processing. He also generated the manuscript figures and led the writing

of the article, which was reviewed and edited by all co-authors. BFK supervised and helped

in the discussions with the ANFF and provided valuable feedback on the design

adjustments to suit manufacturing constraints. DDS advocated and inspired the start of

this research and provided overall guidance and supervision.

8. P. R. T. Munro, A. Curatolo (15%), and D. D. Sampson, "Full wave model of

image formation in optical coherence tomography applicable to general

samples," Optics Express 23, 2541-2556 (2015).


PRTM was the principle author of this article. PRTM conceptualised this research,

designed the OCT image simulator and its various modules, including the rigorous

full-wave solver of Maxwell’s equations to calculate the spatial distribution of the scattered

light field form the sample. He wrote the code, and performed the simulations. He also

generated the majority of the manuscript figures and led the writing of the article, which

was reviewed and edited by all co-authors. AC provided feedback on the way to implement

the OCT working principles in the simulation. AC performed the experiments with the

structured phantom, processed the images and helped in generating the related figure. He

also provided information for the input simulation parameters and feedback on the

simulation results verification. DDS conceptualised the work, provided overall guidance

and supervision.

9. A. Curatolo (50%), M. Villiger, D. Lorenser, P. Wijesinghe, A. Fritz, B. F.

Kennedy, and D. D. Sampson, "Ultrahigh-resolution optical coherence

elastography," Optics Letters 41, 21-24 (2016).


AC was the principle author of this article. AC contributed to the hardware design of

the extended-focus optical coherence microscope (xf-FDOCM) and to building it. He

aligned and characterised the system. He devised acquisition strategies to acquire 3-D

elastograms with it and to process them. He designed and manufactured suitable inclusion

phantoms. He performed the experiments and processed the images and data. He liaised

xxii

with CellCentral and provided the tissue for histology. He also generated the manuscript

figures and led the writing of the article, which was reviewed and edited by all co-authors.

M. Villiger (MV) participated in the project conceptualisation. He also contributed to the

hardware design, and to its characterisation. He helped with the spectrometer calibration

and provided useful feedback throughout the project duration. DL participated in the

project conceptualisation. He designed and chose the components for the xf-FDOCM

system. He closely supervised the building of it. PW provided support in the discussions

about the acquisition strategies. He also refined the image reconstruction and processing

code. He helped in the image acquisition effort and some of the experiments. He also

scanned the histology sections, which formed part of Figure 3. A. Fritz (AF) helped in

building and aligning the xf-FDOCM system. BFK participated in the project

conceptualisation and provided close supervision throughout the whole project. He gave

insightful feedback on all elastography aspects of the research and sourced the mouse aorta

for tissue imaging. DDS conceptualised the work, and provided overall guidance and

supervision.

xxiii

List of publications

The following is a chronological list of publications arising during the duration of this

thesis. PDFs and online multimedia material of some of these publications are available on

the website of the Optical + Biomedical Engineering Lab (OBEL):

http://obel.ee.uwa.edu.au/publications.

Fully refereed journal articles

Key: † These authors contributed equally to this work

1. A. Curatolo, B. F. Kennedy, and D. D. Sampson, "Structured three dimensional

optical phantom for optical coherence tomography," Optics Express 19, 19480-

19485 (2011).

2. R. A. McLaughlin, B. C. Quirk, A. Curatolo, R. W. Kirk, L. Scolaro, D. Lorenser, P.

D. Robbins, B. A. Wood, C. M. Saunders, and D. D. Sampson, "Imaging of Breast

Cancer With Optical Coherence Tomography Needle Probes: Feasibility and Initial

Results," IEEE Journal of Selected Topics in Quantum Electronics 18, 1184-1191

(2012).

3. A. Curatolo†, R. A. McLaughlin†, B. C. Quirk, R. W. Kirk, A. G. Bourke, B. A.

Wood, P. D. Robbins, C. M. Saunders, and D. D. Sampson, "Ultrasound-Guided

Optical Coherence Tomography Needle Probe for the Assessment of Breast

Cancer Tumor Margins," American Journal of Roentgenology 199, W520-W522

(2012).

4. G. Lamouche, B. F. Kennedy, K. M. Kennedy, C.-E. Bisaillon, A. Curatolo, G.

Campbell, V. Pazos, and D. D. Sampson, "Review of tissue simulating phantoms

with controllable optical, mechanical and structural properties for use in optical

coherence tomography," Biomedical Optics Express 3, 1381-1398 (2012).

5. D. Lorenser, C. C. Singe, A. Curatolo, and D. D. Sampson, "Energy-efficient low-

Fresnel-number Bessel beams and their application in optical coherence

tomography," Optics Letters 39, 548-551 (2014).

6. B. F. Kennedy, R. A. McLaughlin, K. M. Kennedy, L. Chin, A. Curatolo, A. Tien,

B. Latham, C. M. Saunders, and D. D. Sampson, "Optical coherence micro-

elastography: mechanical-contrast imaging of tissue microstructure," Biomedical

Optics Express 5, 2113-2124 (2014).

http://obel.ee.uwa.edu.au/publications

xxiv

7. L. Chin†, A. Curatolo†, B. F. Kennedy, B. J. Doyle, P. R. T. Munro, R. A.

McLaughlin, and D. D. Sampson, "Analysis of image formation in optical

coherence elastography using a multiphysics approach," Biomedical Optics Express

5, 2913-2930 (2014).

8. P. R. T. Munro, A. Curatolo, and D. D. Sampson, "Full wave model of image

formation in optical coherence tomography applicable to general samples," Optics

Express 23, 2541-2556 (2015).

9. B. F. Kennedy, R. A. McLaughlin, K. M. Kennedy, L. Chin, P. Wijesinghe, A.

Curatolo, A. Tien, M. Ronald, B. Latham, C. M. Saunders, and D. D. Sampson,

"Investigation of optical coherence micro-elastography as a method to visualize

cancers in human breast tissue," Cancer Research 75, 3236-3245 (2015).

10. A. Curatolo, M. Villiger, D. Lorenser, P. Wijesinghe, A. Fritz, B. F. Kennedy, and

D. D. Sampson, "Ultrahigh-resolution optical coherence elastography," Optics

Letters 41, 21-24 (2016).

11. A. Curatolo, P. R. T. Munro, D. Lorenser, P. Sreekumar, C. C. Singe, B. F.

Kennedy, and D. D. Sampson, "Quantifying the influence of Bessel beams on

image quality in optical coherence tomography," Scientific Reports 6, 23483 (2016).

Selected conference papers

Key: * International | ^ Domestic | † Full paper| § Abstract|

1. *§ A. Curatolo, B. F. Kennedy, K. M. Kennedy, R. A. McLaughlin, and D. D.

Sampson, "Portable 3D optical coherence elastography for applying microscopic

imaging of tissue mechanical properties in the clinic," in Optics Within Life Science

(Genoa, Italy, 2012).

2. *§ A. Curatolo, L. Chin, C. Stynes, B. F. Kennedy, and D. D. Sampson, "Simulation

of image formation process in phase-sensitive optical coherence elastography," in

Opto, Meeting for Young Researchers (Torun, Poland, 2013).

3. ^§ A. Curatolo, C. C. Singe, D. Lorenser, P. R. T. Munro, and D. D. Sampson,

"Improving OCT image quality in turbid structured phantoms by beam shaping," in

ANZCOP (Fremantle, Australia, 2013).

4. *§ A. Curatolo, D. Lorenser, P. R. T. Munro, P. Sreekumar, C. C. Singe, B. F.

Kennedy, and D. D. Sampson, "Analysis of beam shaping in optical coherence

tomography in the presence of sample-induced aberrations and scattering," in SPIE

Photonics West, BiOS: Optical Coherence Tomography and Coherence Domain Optical Methods

in Biomedicine XIX (San Francisco, USA, 2015).

xxv

5. *§ A. Curatolo, M. L. Villiger, D. Lorenser, A. Fritz, B. F. Kennedy, and D. D.

Sampson, "Cellular resolution optical elastography using phase-sensitive extended

depth-of-field optical coherence microscopy," in SPIE Photonics West, BiOS: Optical

Elastography and Tissue Biomechanics II (San Francisco, USA, 2015).

6. *§ A. Curatolo, P. R. T. Munro, P. Sreekumar, C. C. Singe, B. F. Kennedy, D.

Lorenser, and D. D. Sampson, "Effect on optical coherence tomography image

quality of turbid tissue scattering using Gaussian or Bessel beams," in SPIE ECBO

2015, European Conference in Biomedical Optics (Munich, Germany, 2015).

7. *† A. Curatolo, M. L. Villiger, D. Lorenser, A. Fritz, B. F. Kennedy, and D. D.

Sampson, "Ultrahigh-resolution optical coherence elastography using a Bessel beam

for extended depth of field," in SPIE Photonics West, BiOS: Optical Coherence

Tomography and Coherence Domain Optical Methods in Biomedicine XX (San Francisco,

USA, 2016).

Book chapters

A. Curatolo, B. F. Kennedy, D. D. Sampson, and T. R. Hillman, "Speckle in Optical

Coherence Tomography," in Advanced Biophotonics: Tissue Optical Sectioning (V. V. Tuchin, and

R. Wang, eds), Taylor & Francis: 211-277 (2013).

Other publications

R. A. McLaughlin, B. C. Quirk, D. Lorenser, X. Yang, B. Y. Yeo, A. Curatolo, K. M.

Kennedy, L. Scolaro, R. W. Kirk, and D. D. Sampson, "A microscope in a needle", Optics

and Photonics News, vol. 23, no. 12, p. 40, December 2012. (This article was selected for

the 2012 special issue "Optics in 2012". It also featured in a video highlight of the issue,

https://www.osapublishing.org/opn/abstract.cfm?uri=opn-23-12-40).

B.F. Kennedy, L. Chin, K.M. Kennedy, P. Wijesinghe, A. Curatolo, S. Es’haghian, P.R.T.

Munro, R.A. McLaughlin, and D.D. Sampson, "Optical elastography: a new window into

disease, " Optics & Photonics News, vol. 25, no. 12, December 2014. (This article was

selected for the 2014 special issue "Optics in 2014". http://www.osa-

opn.org/home/articles/volume_25/december_2014/extras/a_new_window_into_disease)

https://www.osapublishing.org/opn/abstract.cfm?uri=opn-23-12-40

http://www.osa-opn.org/home/articles/volume_25/december_2014/extras/a_new_window_into_disease/#.VfaPqBGqqko

http://www.osa-opn.org/home/articles/volume_25/december_2014/extras/a_new_window_into_disease/#.VfaPqBGqqko

xxvi

List of acronyms

Acronym Definition

1-D One-dimensional

2-D Two-dimensional

3-D Three-dimensional

AFM Atomic force microscopy

ASP Angular spectrum propagation

CPU Central processing unit

CT Computed tomography

CW Continuous wave

DOE Diffractive optical element

DOF Depth of field

DVC Digital volume correlation

EM Electromagnetic

FD-OCT Fourier-domain optical coherence tomography

FDODL Frequency-domain optical delay line

FDTD Finite-difference in the time-domain

FEM Finite-element method

FFT Fast Fourier transform

FOV Field of view

FWHM Full-width at half maximum

xxvii

GPU Graphical processing unit

H&E Haematoxylin and eosin

HMDS Hexamethyldisilazane

ISAM Interferometric synthetic aperture microscopy

LCoS Liquid crystal on Silicon

MEMS Micro-electrical mechanical system

MIP Maximum-intensity projection

MRI Magnetic resonance imaging

NA Numerical aperture

OCE Optical coherence elastography

OCM Optical coherence microscopy

OCT Optical coherence tomography

OFDI Optical frequency domain imaging

ONH Optic nerve head

PDMS Polydimethylsiloxane

PF Phase Fresnel (lens)

PMMA Poly-methyl methacrylate

PSTD Pseudospectral time-domain

PSF Point spread function

PVA Polyvinyl alcohol

PWM Pulse-width modulation

RAM Random access memory

xxviii

ROI Region of interest

RTE Radiative transport equation

RTV Room-temperature vulcanising

SAW Surface acoustic wave

SD-OCT Spectral-domain optical coherence tomography

SF Scattering-free

SLD Superluminescent diode

SLM Spatial light modulator

SBR Signal-to-background ratio

SNR Signal-to-noise ratio

SS-OCT Swept-source optical coherence tomography

SW Shear wave

TD-OCT Time-domain optical coherence tomography

TE Transverse electric

TM Transverse magnetic

UHROCE Ultrahigh-resolution optical coherence elastography

VVG Verhoeff-Van Gieson

WLS Weighted least squares

xf-FDOCM Extended-focus Fourier-domain optical coherence microscopy

1

Chapter 1

Introduction 1

A major challenge in improving patient health and treatment lies in understanding the

origin of disease and how it affects living tissue across different length scales, from the

molecules inside a cell, to the whole organs within the body. Another challenge lies in early

detection of disease, in order to contain its progression and minimise damage.

Medical imaging has enabled great progress in these two areas. No technique, however,

can produce images with resolution sufficient to see cells and reveal the structural and

functional information at depths of more than a few hundred micrometres below the tissue

surface of a living subject, a task that is necessary to provide a viable substitute for

histopathological analysis.

For a pathology laboratory to provide this information, costly tissue excisions or

biopsies are required, followed by extensive and time-consuming preparation for slicing

and staining before performing microscopy investigation. In certain cases, biopsies are not

even feasible, where tissue excision would compromise organ functionality, e.g., in the brain

or in the eye.

Let us consider the eye, perhaps the most demanding example, where microscopic

functional and structural information is vital in the understanding, early diagnosis and

management of a condition like glaucoma. Even though this thesis does not concern

ophthalmology, the following example serves to highlight the need for improved image

quality and performance in high-resolution bio-imaging to be able to clearly identify tissue

types and constituents, and study and unequivocally diagnose disease.

Glaucoma is the second most common cause of blindness worldwide [1] and leads to

vision loss by damaging retinal ganglion cell axons in and around the optic nerve head

(ONH) [2]. The onset of glaucoma is influenced by many factors [3], with ONH

biomechanics being an important driving mechanism of this disorder, as demonstrated by a

large body of research [4, 5]. However, quantifying ONH biomechanics is complex, and so

far investigators have used general analytical [6, 7] or computational models [8, 9],

untailored to individual patients. To understand the influence of ONH biomechanics on

glaucoma, and then achieve clinical utility, in vivo measurements of the geometry and the

2 Chapter 1 Introduction

mechanical properties of all tissues within the ONH, as well as the load (i.e., intraocular

pressure) that acts on them, are required. This information can then be used in patient-

specific biomechanical models of the ONH. Optical coherence tomography (OCT) has the

potential to be a powerful tool for quantification of the in vivo biomechanics of the ONH.

OCT is a three-dimensional structural biomedical imaging technique that provides

higher spatial resolution (1-20 μm) than modalities such as ultrasound, magnetic resonance

imaging (MRI), and X-ray computed tomography (CT) to depths of several millimetres in

tissue [10]. This penetration depth exceeds that of other high-resolution optical imaging

techniques, such as confocal microscopy or non-linear microscopy, as near-infrared light

beams and coherent detection of backscattered light are used.

The newest generation of OCT devices can acquire serial cross-sectional 2-D images

rapidly (50,000+ depth lines/second), yielding near-real time 3-D volumes of tissue with an

axial resolution of approximately 4 μm for commercial devices, improving to 1 μm for

experimental ultrahigh-resolution devices [11]. The structures of the lamina cribrosa and the

adjacent peripapillary sclera [12, 13], both of which are believed to strongly influence ONH

biomechanics [14, 15] are so deep in the eye that light penetration of commercially available

OCT scanners is only sufficient to start visualizing and identifying their upper boundaries.

The principal reason for failing to visualise these deep structures is the influence on the

OCT image quality of blood vessels, in particular those arising from the central retinal

vessel trunk, as they cast signal shadows and alter structural and tissue displacement

measurements. Furthermore, as incident light travels through the ONH, it attenuates with

depth, such that reflected signals from deep structures may be too weak or the image

quality too poor for them to be reliably detected.

Image quality is defined by descriptors, such as resolution, depth of field (DOF),

signal-to-noise ratio (SNR), sensitivity, contrast, penetration depth and speckle contrast.

These parameters can be severely affected by the alteration of the optical beam (phase and

amplitude) as it propagates through turbid media, resulting in a degradation of the image

quality. These alterations are mainly caused by sample refractive index inhomogeneities

present in closely packed tissue constituents or at tissue boundaries. Degradation can take

the form of tissue-induced aberrations, caused by large-scale refractive index fluctuations,

or diffused haze on the image brought upon by either small-angle multiply-scattered light

wavefronts or wide-angle multiply-scattered light wavefronts that carry little or no

structural information about the sample under investigation.

The intrinsic contrast provided by scattering from the tissue reveals many

morphological features, but is, on occasion, insufficient. For example, it does not readily

reveal microvasculature, and it can be difficult to distinguish tissue types. Mechanical

1.1 Research objectives 3

properties are important to measure in their own right, but they also represent an

alternative form of contrast to optical properties, which provides new opportunities in

imaging tissues. Elastography is a medical imaging technique based on measuring the

spatially resolved response of tissue to mechanical loading and can provide a map of

mechanical properties. Better resolution than manual palpation can be achieved with

elastography based on magnetic resonance imaging or ultrasound imaging [16]. These types

of elastography are emerging as a clinical tool in the diagnosis liver fibrosis and breast

cancer [17]. With resolution of tens of micrometres, optical coherence elastography (OCE)

[18], a form of elastography based on OCT, shows promise in visualizing mechanical

contrast in tissue on a scale intermediate between that of cells and organs.

This resolution is yet too coarse to probe changes to the mechanical properties of tissue

on the cellular scale (< 25 µm), as is ideally required to study the onset and development of

disease [19], preventing OCE from realising its potential as a promising and unique

instrument in the field of cell mechanics. The field of cell mechanics, in part, focuses on

pathogenesis studies at the cellular scale and on the characterisation of the mechanical

signatures of healthy and diseased cellular tissue constituents [20]. Ideally, such studies

would include the ability to provide in situ images at cellular resolution in live tissues in their

native environment.

However, despite the efforts of the scientific community to improve the system

resolution and increase the signal depth penetration, there are severe limitations when using

OCT to image deep tissue structures and when using OCE to resolve tissue mechanical

interaction on the cellular scale. Such limitations lead to artefacts, which may result in

clinical misinterpretation and morphometric errors, or low sensitivity (high rate of false

negative detection of tissue that is actually diseased) and low specificity (high rate of false

positive detection of tissue that is actually healthy). Strongly scattering and attenuating

structures (e.g., pigment and blood, generally highly forward scattering) adversely affect

biomedical OCT applications, by reducing SNR, contrast and resolution. OCT phase

decorrelation noise, limited depth of field, low scan rates or coarse displacement estimation

methods have so far prevented the use of OCE in an ultrahigh-resolution regime.

Since, for the study and diagnosis of diseases, the ultimate goal is to measure

morphometric and biomechanical parameters, it is crucial to obtain high-quality images of

the structures of interest at depth below the surface. The challenges previously mentioned

can be overcome using particular approaches. Amongst these, one is using software post-

processing techniques and algorithms to improve the quality of OCT images [21]. A

potentially more effective one is using hardware techniques, and the use of optical imaging

simulations to guide the hardware development. This is because, in general, a model and


the results of a simulation inform the experiments and help advance understanding of the

technique, which remains relatively weak in OCT, and is embryonic in OCE.

The aim of this thesis is to address these issues with hardware techniques, in order to

maximise the potential of OCT and OCE imaging at high resolution in turbid tissues. We

shall not focus on a particular application, as we want to keep a general approach, suited to

most imaging applications of OCT or OCE in turbid biomedical tissues. Therefore, we

want to concentrate on the technological push within these boundaries, with the goal of

significantly improving the microscopic detection of diseased tissue.

1.1 Research objectives

The focus of this PhD thesis is to understand, characterise and improve image quality in

OCT and OCE, performed at millimetre depths in biological tissue, by means of both

experimental and theoretical approaches: optical beam shaping and optical simulations.

We plan to do so, by addressing fundamental limitations of coherent light-tissue

interaction and technical constraints, affecting OCT image quality, with the objective of

opening up new avenues for applications. The research required to achieve this goal can be

broken down into three parts:

1. Providing the tools needed to improve and quantify OCT and OCE image

quality in turbid tissue, i.e., the tools to alter it (beam shaping), measure it (tissue

phantoms), and analyse it (simulations);

2. Analysing OCT image quality in turbid tissue, altered using Bessel and Gaussian

beams, and measuring the relative improvements;

3. Improving OCE image resolution without compromising other image quality

descriptors (e.g., depth of field, strain sensitivity), in order to perform OCE at

the cellular level, with applications in cell mechanics in situ.

1.2 Structure of the thesis

The content of each chapter is briefly summarised below. Where journal papers are

included in the chapters, they are reproduced as published, as noted in this introduction.

Where chapter sections are only partially based on journal papers or other sources, proper

referencing is also listed in the following breakdown.

1.2 Structure of the thesis 5

Chapter 2 - Background of OCT and OCE

This chapter provides a brief overview of OCT technology, concepts and characteristics

that we exploit in our research effort to improve image quality and diagnostic ability in

biological tissue. We review OCT, its working principles and image quality descriptors,

such as resolution, DOF, SNR, sensitivity, contrast, speckle and present an introduction to

OCE, an OCT-based modality for imaging tissue elasticity [18, 22, 23]. Section 2.2 is based

on the introduction of [24]. Section 2.2.2 is based on Section 2 of [25]. Section 2.3.3 is

based on Section 6.1 of [26].

Chapter 3 - Energy-efficient Bessel beams: a tool to alter image quality

This chapter explores the characteristics of differently shaped optical beams used to

interrogate biological tissue and form an image. We present different ways of shaping an

optical beam, with refractive and diffractive optical elements, and with the use of

reconfigurable beam shapers, such as a phase-only liquid crystal-on-silicon spatial light

modulator (SLM). With appropriate characterisation, programming and placement in the

sample arm optical path, we obtained phase-stable, achromatic and diffraction-efficient

Bessel-like beams, and characterised their use in OCT, including their energy efficiency.

Section 3.2 reports the journal paper [27] as published. In this section, we show that,

for quasi non-diffracting Bessel beams, the Fresnel number is the key parameter

determining the trade-off between DOF extension and OCT sensitivity loss when

compared to a Gaussian beam of equal resolution. Section 3.3 is based on the experiment

section of the supplementary information of [28]. We conclude this chapter by discussing

the implication of the so-called self-reconstructing property of the Bessel beam in turbid

tissue imaging.

Chapter 4 - Structured phantoms: a tool to mimic turbid tissue and compare image

quality

This chapter introduces tissue-mimicking targets, called phantoms, that reproduce the

optical, mechanical and structural properties of biological tissue, and can be used to

characterise OCT image quality. Section 4.1 is based on the introduction of [29]. Section

4.2 is based on Section 6.5 of [26], and it reviews the basics of light-tissue interaction, the

different categories of OCT signal contributions from different scattering regimes and their

implications for image quality. We then review Mie theory, describing the main tissue

optical properties, scattering coefficient 𝜇𝜇𝑠𝑠, and scattering anisotropy, 𝑔𝑔. Those properties


influence the OCT signal, with respect to both its attenuation with depth and its contrast.

This section is partly based on Sections 2.1.2 and 2.1.3 of [29]. Section 4.3 describes and

validates the fabrication of phantoms with controlled attenuation coefficients 𝜇𝜇𝑡𝑡, and

scattering anisotropy, 𝑔𝑔 and is partly based on Section 2.2 of [29] and the beam and sample

configuration section of the supplementary information of [28]. Section 4.4, based on

Section 3 of [29], discusses and validates fabrication of phantoms with controlled

mechanical properties, especially stiffness. Section 4.5 reports the journal paper [30] as

published, describing novel phantoms containing three-dimensional structure, suitable for

mimicking the complexity of tissue structures on a scale intermediate between the OCT

system resolution and the field of view, and their potential application in the assessment of

speckle. In Section 4.6, based on the beam and sample configuration and validation

sections of the supplementary information of [28], we look more specifically at the use of

phantoms to test OCT image quality descriptors, including, resolution and contrast, with

both a nanoparticle-embedded phantom and an advanced version of a 3-D-structured

phantom. In Section 4.7, based on Section 2.2. of [24], Section 4.1 of [25] and partly [31],

we describe the fabrication and use of structured phantoms for OCE. We conclude the

chapter with a discussion of the use and limitations of silicone phantoms as tissue-

mimicking objects and image test targets.

Chapter 5 - Simulation of beam propagation and image formation in turbid

samples: a tool to theoretically quantify image quality

This chapter presents a suite of OCT image simulation and optical beam propagation tools

for the study and understanding of speckle phenomena in OCT, phase-sensitive OCT

measurements, and assessment of the quality of images of scattering turbid media.

We proceed through increasing complexity and realism of the model on which the

simulations are based, and increasing dimensionality from one to three dimensions. Section

5.2 deals with image formation in the single-scattering regime. In Section 5.2.1, based on

Section 6.2 of [26], we start by simulating the OCT signal from a multitude of scatterers in

one dimension and in two dimensions as the result of local sums of random phasors, and

in Section 5.2.2, based on Section 3.1 of [25], we explore a more advanced linear systems

model providing realistic OCT amplitude and phase images from numerical scattering

phantoms. Section 5.3 presents our journal paper [32], as published, on a two-dimensional

full wave model of image formation in OCT applicable to general samples. Section 5.4,

based on the supplementary information of [28], extends the simulation of the beam

propagation into the sample to three dimensions.

1.2 Structure of the thesis 7

Chapter 6 - Quantifying OCT image quality in turbid samples using Gaussian and

Bessel beams

In this chapter, we answer the question of how energy-efficient Bessel beams perform in

turbid tissue OCT imaging, and whether they hold an advantage in terms of image quality

compared to imaging with Gaussian beams.

Light interaction with turbid tissue affects the image quality indicators. In fact, the

presence and relative contribution of the different categories of image-carrying and

image-degrading light in the detected OCT signal determines the precision and ability to

localise and discriminate the position and intensity of the backscattering events generated

by the tissue refractive index distribution.

In our effort to benchmark image quality and to provide strategies for improvement,

we present our work done in quantifying the influence of Bessel beams on image quality in

OCT. This analysis will bring together all the tools developed in the previous chapters.

Firstly, the beam shaping platform used to produce Gaussian and Bessel beams of equal

transverse resolution. Secondly, the overlayers and structured phantoms to test different

realistic scattering conditions and to quantify the image quality. Finally, the 3-D beam

propagation simulation to verify the different image-carrying and image-degrading light

contributions to the OCT signal. These tools will help us answer the question of which

beam type attains better OCT image quality in turbid tissue under general tissue-like

scattering conditions. Section 6.2 reports our recent journal paper [28], as published.

Chapter 7 - Improving OCE image quality: strain precision and resolution

This chapter focuses on the characterisation and improvement of optical coherence

elastography image quality. We present the results of the optical simulations introduced in

Chapter 5, and incorporate those in a multiphysics simulation, combining optical and

mechanical models, to highlight the influence of acquisition parameters, such as the loading

conditions (compression amplitude), on the elastogram image quality. We do so with

specific reference to elastogram precision, i.e., the repeatability of the tissue displacement

and strain measurement, as determined by indicators such as displacement and strain

sensitivity and strain SNR. Section 7.2 includes most of the journal paper [25] on this

subject.

With the goal of expanding the capabilities of OCE to serve the field of cellular

biomechanics in situ, we then concentrate on improving the elastogram resolution, without

compromising its precision. We do so, by designing an ultrahigh-resolution optical


coherence microscopy system with an advanced beam shaping setup, combining Bessel

beam illumination and Gaussian beam detection, and devising an image acquisition strategy

that preserves elastogram precision. This highly novel system and its results are reported in

Section 7.3, where the journal paper [31] is presented as published.

We demonstrate this improvement on both tissue-mimicking phantoms and freshly

excised mouse aorta, revealing the mechanical heterogeneities of vascular smooth muscle

and elastin sheets in the aorta wall in exceptional detail.

Chapter 8 - Conclusions

This chapter summarizes the significance and limitations of the research presented in this

thesis, and makes recommendations for future work. The thesis concludes with a summary

of the key contributions and some final remarks.

9

Chapter 2

Background of OCT and OCE 2

2.1 Optical coherence tomography

This chapter provides a brief overview of the imaging technology, concepts and

characteristics required to understand the context and our effort to improve image quality

and diagnostic ability in biological tissue. We will review OCT, its working principles and

imaging properties, such as resolution, depth of field, signal-to-noise ratio, sensitivity,

contrast, speckle and presents an introduction to optical coherence elastography (OCE), an

OCT-based modality for imaging tissue elasticity [18, 22, 23].

OCT is an optical imaging technique acquiring three-dimensional (3-D), high-

resolution, near real-time images of samples and their sub-surface structure, suitable for

biomedical diagnostics in vivo [33]. It is non-ionizing and non-invasive as it uses infrared

light without requiring direct contact between probe and tissue.

OCT is analogous to ultrasound imaging in its working principle, as it gates the

detected backscattered radiation from the sample and determines the associated distance

from the measured echo-time delay, but uses light waves instead of sound waves. Direct

time-of-flight measurement of the reflected waves is infeasible, since the speed of light is

roughly six orders of magnitude faster than the speed of sound. Instead, low-coherence

interferometry [34, 35], explained in detail in Section 2.1.1, is used to make depth-resolved

measurements of the sample’s reflectance. Light from a spectrally broadband source is split

into two optical paths (‘arms’), and the interference of light reflected or backscattered from

the sample with light reflected from a reference mirror is detected. The use of

low-coherence interferometry for in vivo measurements of tissue has been demonstrated

since at least the mid 1970s [36]. It wasn’t until 1991 that Huang et al. presented the first

demonstration of OCT, producing images by raster scanning the sample beam of a low

coherence interferometer over ex vivo tissue samples [37].

In OCT, a 3-D image is usually formed by scanning in the following order: a one-

dimensional (1-D) axial scan in depth, 𝑧𝑧, is performed by low-coherence interferometry.

10 Chapter 2 Background of OCT and OCE

Acquiring multiple 1-D scans at different transverse locations along 𝑥𝑥 leads to a two-

dimensional (2-D), cross-sectional scan. A 3-D volume image is then formed by scanning

laterally in 𝑦𝑦, and acquiring multiple 2-D scans at different transverse locations. Using the

same terminology as used in ultrasound imaging, the 1-D depth scan is called an A-scan,

the 2-D scan is called a B-scan, and the 3-D scan is called a C-scan. Once acquired, the data

volume can then be digitally sliced in different orientations, similarly to multiplanar

reformatting in CT. Slices in the 𝑥𝑥𝑦𝑦-plane, thus at some depth 𝑧𝑧, are referred to as the

en-face images. Figure 2-1 shows a schematic of this geometry and terminology.

Figure 2-1. Schematic of terminology used when referring to OCT scan orientation. (a) A-scan: a 1-D axial scan formed from the irradiance of light vs. depth, z. (b) B-scan: a 2-D cross-sectional scan formed by scanning laterally (x) and acquiring multiple A-scans. (c) C-scan: a 3-D volume formed by scanning laterally (y) in the orthogonal direction to (b), and acquiring multiple B-scans. Adapted from [11].The resolutions (in the axial and transverse

directions) of a typical OCT system (at focus and in low-scattering media) are around 5–20

μm [11], but systems with ultrahigh resolutions of 1–3 μm have been demonstrated [38].

Unlike in most other microscopy techniques, the mechanisms determining the axial and

transverse resolution in OCT are decoupled, as only the transverse resolution is determined

by the sample optics, while the axial resolution is determined by the coherence property of

the light source. Section 2.3.1 deals in detail with the factors influencing OCT resolution.

Light extinction from tissue is caused by absorption and scattering processes. The

wavelengths used for OCT are usually chosen to lie within the so-called diagnostic window

(650–1,350 nm) to minimise the absorption of light by tissue constituents [39, 40]. Figure

2-2 shows a diagram of the spectral molar extinction by selected tissue chromophores.

2.1 Optical coherence tomography 11

Figure 2-2. Spectral absorption for a range of tissue chromophores in the diagnostic window in the near-infrared. For wavelengths in the range 650–1350 nm, the extinction of light due to absorption and scattering is relatively low. Based on data from [41-44].

The optical absorption of any given chromophore in tissue is determined by the molar

extinction coefficient multiplied by the concentration of that chromophore. Tissue

absorption is dominated by the spectral response of water since water is the most

prominent constituent of most soft tissues. Thus, water limits light penetration in tissue on

the long wavelength side, and melanin and blood, where present, on the short wavelength

side. Within the diagnostic window, light scattering is two or more orders of magnitude

stronger than absorption, making it very suitable for OCT, which relies on the elastic

backscattering of light to form an image. The most commonly used wavelengths for OCT

are centred around 800 nm, which provides higher resolution (for a given bandwidth, see

Section 2.3.1), and 1,300 nm, which provides greater depth penetration in tissue. The

typical depth penetration of OCT is in the range 0.5–1.5 mm, determined primarily by the

interplay between scattering, system sensitivity and numerical aperture.

In terms of resolution and penetration depth, OCT occupies a niche between

laboratory imaging techniques, such as confocal microscopy, and medical imaging

techniques, such as ultrasound and magnetic resonance imaging, as illustrated in Figure 2-3.

12 Chapter 2 Background of OCT and OCE

Figure 2-3. Comparison of OCT resolution and imaging depths to other biomedical imaging methods. The “pendulum” length represents imaging depth, and the “sphere” size represent resolution. Reproduced from [45].

OCT is clinically used primarily in ophthalmology [46, 47], especially for imaging the

retina, but increasingly also the anterior segments of the eye, such as the cornea. OCT is

also being used clinically, in order of decreasing importance, for cardiology [48, 49],

dermatology [50], gastroenterology [51], dentistry [52], and pulmonology [53, 54], as shown

in Figure 2-4. Examples of the three most published research application areas are

schematically shown in Figure 2-5. Figure 2-6 shows movies of representative OCT images

from those clinical applications as well as laboratory applications.

The advances into fibre-optics technology brought by the telecommunications industry

had fostered the growth of OCT imaging and helped overcome its short penetration depth

by delivering light through flexible probes, allowing for in vivo and in situ imaging of internal

tissues.

Fibre-optics and miniaturised optics are at the core of compact endoscope [55],

catheter [56] and needle [57] probes. In addition, OCT has been used as the basis for

several derived modalities, including Doppler OCT for measuring fluid velocities,

particularly of blood flow in vasculature [58-60], and spectroscopic OCT for measuring

depth-resolved, wavelength-dependent attenuation [61].

2.1 Optical coherence tomography 13

Figure 2-4. Bar chart of a snapshot of results from a Scopus search of selected OCT biological applications from 2005 to 2015, inclusive. Searches were conducted on the publication title and only included English results. Search terms used were as follows: “optical coherence tomography” AND terms relevant for each heading, e.g. "optical coherence tomography" AND (breast OR mamm*), where the use of a * indicates a search for all words containing the truncated term. (http://www.scopus.com).

Figure 2-5. Main applications of OCT. (a) Ophthalmology. (b) Cardiology. (c) Dermatology. Reproduced from [62-72].

13152135365682103110153

5612736

0 500 1000 1500 2000 2500 3000

GynaecologyLaryngology

MuscleBreast

UrologyGastroenterology (inc. pancreas and biliary tract)

Pulmonary medicineDentistry

Neurology (exc. ophthalmology)Dermatology

CardiologyOphthalmology

# of publications between 2005 and 2015

# of publications

http://www.scopus.com/

Characterising and improving image quality in Optical ......Characterising and improving image...

Documents

Transcript of Characterising and improving image quality in Optical ......Characterising and improving image...