About 50% of cancer patients are treated with X-ray radiation therapy; however, with current treatment feedback, the effects and the efficacy of the treatment are generally detected several weeks/months after treatment completion. This makes the adjustment of the treatment based on early response, and identification of non-responding patients, nearly impossible. In this thesis a novel method combining optical coherence tomography and a gamut of image analysis methods is explored as a potential approach to detecting tissue variability. Applying texture analysis to the optical coherence tomography images may allow for the tracking of radiation therapy induced cell microstructural changes in cancer patients and help in the adjustment of treatment based on early response.

Optical imaging has been developed quickly in the past decades because it has become an important research tool in biology, biochemistry, and biomedical sciences. Coherence domain optical imaging is one of the well developed optical imaging modalities, as it provides high resolution and long penetration depth. In this dissertation, we will report our work on development of molecular contrast in coherence domain optical imaging. In order to image important molecules which are poor fluorophores, we developed a high resolution molecular imaging technique, pump-probe optical coherence microscopy (PPOCM), which does not rely on fluorescent tags. PPOCM is the fusion of Pump-Probe spectroscopy and optical coherence microscopy (OCM). We have demonstrated the prototype system on a fixed human skin sample containing a nodular melanoma. The results indicate that PPOCM can clearly provide strong contrast between the melanotic and amelanotic regions. This technique can be applied to early diagnosis of melanoma and the mapping of tumor margins during excision. It also can be extended to any biological chromophore with a known absorption spectrum and sufficient concentration. In order to differentiate further multiple chromophores, we developed a spectrally resolved two color pump-probe Optical Coherence Microscopy (SRPPOCM). We showed the prototype system on a red hair and a black hair. Our preliminary results show that the SRPPOCM technique could provide a contrast between pheomelanin and eumelanin. This technique could be used potentially as a clinical tool for diagnosing different progression stages of melanoma. This technique could also be applied to differentiate other mixed chromophores. Second harmonic optical coherence tomography (SHOCT) is non-linear high resolution optical molecular imaging modality which is widely used in non-centrosymmetric material. However, depth ambiguity is associated with SHOCT in tissue sample because forward generated second harmonic signal does not correctly report where the second harmonic signal is generated. We studied the feasibility of collecting the backward generated second harmonic signal from nanocrystals through a Second Harmonic Optical Coherence Tomography in Fourier domain. The preliminary result shows that we can collect backward generated second harmonic signal from nanocrystals which indicates that this technique could suppress the depth ambiguity.

<p>Clinical imaging within ophthalmology has had transformative effects on ocular health over the last century. Imaging has guided clinicians in their pharmaceutical and surgical treatments of macular degeneration, glaucoma, cataracts and numerous other pathologies. Many of the imaging techniques currently used are photography based and are limited to imaging the surface of ocular structures. This limitation forces clinicians to make assumptions about the underlying tissue which may reduce the efficacy of their diagnoses. </p><p>Optical coherence tomography (OCT) is a non-invasive, non-ionizing imaging modality that has been widely adopted within the field of ophthalmology in the last 15 years. As an optical imaging technique, OCT utilizes low-coherence interferometry to produce micron-scale three-dimensional datasets of a tissue's structure. Much of the human body consists of tissues that significantly scatter and attenuate optical signals limiting the imaging depth of OCT in those tissues to only 1-2mm. However, the ocular anterior segment is unique among human tissue in that it is primarily transparent or translucent. This allows for relatively deep imaging of tissue structure with OCT and is no longer limited by the optical scattering properties of the tissue. </p><p>This goal of this work is to develop methods utilizing OCT that offer the potential to reduce the assumptions made by clinicians in their evaluations of their patients' ocular anterior segments. We achieved this by first developing a method to reduce the effects of patient motion during OCT volume acquisitions allowing for accurate, three dimensional measurements of corneal shape. Having accurate corneal shape measurements then allowed us to determine corneal spherical and astigmatic refractive contribution in a given individual. This was then validated in a clinical study that showed OCT better measured refractive change due to surgery than other clinical devices. Additionally, a method was developed to combine the clinical evaluation of the iridocorneal angle through gonioscopy with OCT.</p> / Dissertation

Optical coherence tomography (OCT) imaging is used in an increasing number of biomedical and industrial applications. A massively parallel simulator of OCT of inhomogeneous turbid media, e.g., biological tissue, could be used as a practical tool to expedite and expand the study of the physical phenomena involving such imaging technique, as well as, to design OCT systems with enhanced performance. Our work presents the open-source implementation of this massively parallel simulator of OCT to satisfy the ever-increasing need for prompt computation of OCT signals with accuracy and flexibility. Our Monte Carlo-based simulator uses graphic processing units (GPUs) to accelerate the intensive computation of processing tens of millions of photon packets undergoing a random walk through a sample. It provides computation of both Class I diffusive reflectance due to ballistic and quasi-ballistic scattered photons and Class II diffusive reflectance due to multiple scattered photons. Our implementation was tested by comparing results with previously validated OCT simulators in multilayered and inhomogeneous (arbitrary spatial distributions) turbid media configurations. It models the objects as a tetrahedron-based mesh and implements and advanced importance sampling technique. Our massively parallel simulator of OCT speeds up the simulation of OCT signals by a factor of 40 times when compared to it central processing unit (CPU)-based sequential implementation.

Optical Coherence Tomography (OCT) is a subsurface imaging technique with many biomedical and industrial applications. In this thesis, we describe our design and implementation of a time domain OCT system. We used this system to obtain OCT images of objects that are important in different applications. We also used an existing quadrature OCT system to obtain both real and imaginary parts of an OCT image. We introduced a new interpretation of OCT images as the 2nd derivative of the scattering potential of an object. To obtain this scattering potential from its 2nd derivative, we implemented a method of definite integration in the spectral-domain. The obtained scattering potential was used to separate the scattering profile from the absorption profile of an object. We applied this new spectroscopic imaging method to quadrature OCT images of different objects.

The objective of this study is to apply digital image processing (DIP) techniques to optical coherence tomography (OCT) images and develop computer-based non-subjective quantitative analysis, which can be used as diagnostic aids in early detection of dental caries. This study first compares speckle reduction effects on raw OCT image data by implementing spatial-domain and transform-domain speckle filtering. Then region-based contour search and global thresholding techniques examine digital OCT images with possible lesions to identify and highlight the presence of features indicating early stage dental caries. The outputs of these processes, which explore the combination of image restoration and segmentation, can be used to distinguish lesion from normal tissue and determine the characteristics prior to, during, and following treatments. The combination of image processing and analysis techniques in this thesis shows potential of detecting early stage caries lesion successfully.

Obtaining higher depth of imaging is an important goal in Optical Coherence Tomography (OCT) systems. One of the main factors that affect the depth of OCT imaging is the presence of noise. That’s why the study of noise statistics is an important problem. In the first part of this thesis we obtain an empirical estimate of the second order statistics of noise by using a sequence of Time domain (TD) OCT images. These estimates confirm the non-stationary nature of noise in TD-OCT. In the second part of the thesis these estimates are used to design a depth-dependent matched filter to maximize the Signal-to-Noise Ratio (SNR) and increase the Contrast-to-Noise Ratio (CNR) in TD-OCT. By applying our filter to TD-OCT images of both vascular rabbit tissue and a human tooth, both SNR and CNR were increased and a higher imaging depth was achieved.

This project initiated with the aim for improving the ultrahigh resolution optical coherence tomography (UHR-OCT) system performance by considering the limitations to the axial OCT resolution for in vivo imaging of human and animal retina. To this end, a computational model was developed to simulate the effect of wavelength-dependant water absorption on the detected spectral shape of the broad-bandwidth light source used in UHR-OCT at 1060nm wavelength region, which effectively determines the axial OCT resolution in the retina. For experimental verification of the computational model, a custom built light source with a re-shaped spectrum (Superlum Inc.) was interfaced to the state-of-the-art UHR-OCT system. About 30% improvement of the axial OCT resolution in the rat retina and ~12% improvement of the axial OCT resolution in the human retina was achieved compared to the case of the almost Gaussian shaped spectrum of the standard, commercially available SLD. Although water absorption in the 1060nm spectral region strongly affects the sample beam, selecting a suitable light source with specific spectral shape can compensate for the undesired water absorption effect and thus result in significantly improved axial resolution in in vivo OCT retinal images. To demonstrate the advantages of the state-of-the-art OCT technology for non invasive retinal imaging, an established animal model of outer retina degeneration (sodium iodate (NaIO3)-induced retina degeneration) was employed for longitudinal monitoring of the degeneration and investigation of possible early and dynamic signs of damage undetected by other imaging modalities. The long-term (up to 3 months) and short-term (up to 12 hours) effect of sodium iodate toxicity on the layered structure of retina was monitored longitudinally and in vivo for the first time using OCT. An initial acute swelling of the retina, followed by progressive disruption and degeneration of outer retina was observed as a result of sodium iodate-induced damage. Changes in the thickness and optical reflectivity of individual retinal layers were extracted from the OCT images to quantify the changes occurring at different stages of the disease model. Results from this project present the theoretical and practical limits to the highest axial OCT resolution achievable for retina imaging in the 1060nm spectral range both in small animals and humans, and provided a framework for future development of novel light sources. Furthermore, UHR-OCT imaging was shown to be an effective and valuable modality for in vivo, non invasive investigation of retina degenerative disease.

People with Down’s syndrome (DS) are known to experience premature ageing and have a high propensity for clinical diagnosis of dementia due to Alzheimer’s disease (AD). In DS there is a unique and natural model of over-expression of amyloid-beta (Aβ) protein, the accumulation of which is proposed to be the central early event in the pathogenesis of AD. In DS, AD neuropathology is universally seen in the brain from the fourth decade. Identifying biomarkers are essential to the evaluation of future treatment trials. The retina has been shown to experience changes in patients with AD, such as retinal thinning, compared to age-matched controls. As an extension of the brain, the retina can be quickly and non-invasively imaged and may provide a proxy measure of brain changes in AD. Using optical coherence tomography (OCT), cross-sectional retinal examinations were completed in 50 people with DS aged 18 years and over. Comparisons between retinal thickness of the DS and control groups were examined, as well as the effect of age on thickness in both groups. For the DS group, further investigations were made into the relationships between retinal thickness and (i) cognitive performance, (ii) diagnosis of dementia, (iii) cortical thickness and, (iv) presence of Aβ binding in the brain. Contrary to expectations, people with DS had thicker retina compared to age-matched controls. In addition, normal age-related retinal thinning was not seen in the DS group. People with DS have a life-long overproduction of Aβ, deposits of which have been previously imaged in the retina. Aβ may be responsible both directly, through physical mass, and indirectly through inflammation as a response to Aβ, for increased retinal thickness in people with DS. Consequently, retinal thickness in DS may be a proxy measure of Aβ deposition in the retina. As part of a collaborative study, brain Aβ binding was measured using positron emission tomography neuroimaging in a subset of the DS group. Individuals with positive Pittsburgh compound [11C]-PIB (PIB) binding to Aβ displayed a trend towards having thinner retina than those with negative PIB binding. These results indicate that a shift towards thinning retina in people DS may reflect changes in brain pathology. Future studies are discussed which aim to investigate Aβ and Aβ driven pathology in the retina.

Cardiovascular disease is one of the leading causes of the mortality in the western world. Many imaging modalities have been used to diagnose cardiovascular diseases. However, each has different forms of noise and artifacts that make the medical image analysis field important and challenging. This thesis is concerned with developing fully automatic segmentation methods for cross-sectional coronary arterial imaging in particular, intra-vascular ultrasound and optical coherence tomography, by incorporating prior and tracking information without any user intervention, to effectively overcome various image artifacts and occlusions. Combinatorial optimisation methods are proposed to solve the segmentation problem in polynomial time. A node-weighted directed graph is constructed so that the vessel border delineation is considered as computing a minimum closed set. A set of complementary edge and texture features is extracted. Single and double interface segmentation methods are introduced. Novel optimisation of the boundary energy function is proposed based on a supervised classification method. Shape prior model is incorporated into the segmentation framework based on global and local information through the energy function design and graph construction. A combination of cross-sectional segmentation and longitudinal tracking is proposed using the Kalman filter and the hidden Markov model. The border is parameterised using the radial basis functions. The Kalman filter is used to adapt the inter-frame constraints between every two consecutive frames to obtain coherent temporal segmentation. An HMM-based border tracking method is also proposed in which the emission probability is derived from both the classification-based cost function and the shape prior model. The optimal sequence of the hidden states is computed using the Viterbi algorithm. Both qualitative and quantitative results on thousands of images show superior performance of the proposed methods compared to a number of state-of-the-art segmentation methods.