PET (positron emission tomography) scans are still in the experimental phase, as one of the newest breast cancer diagnostic techniques. It is becoming the new standard in neurology, oncology and cardiology. PET, like other nuclear medicine diagnostic and treatment techniques, involves the use of radiation. Because of the negative impact of radioactivity to our bodies the radiation doses in PET should be small. The existing computing algorithms for calculating PET images can be divided into two broad categories: analytical and iterative methods. In the analytical approach the relation between the picture and its projections is expressed by a set of integral equations which are then solved analytically. The Fourier backprojection (FBP) algorithm is a numerical approximation of this analytical solution. Iterative approaches use deterministic (ART = Algebraic Reconstructed Technique) or stochastic (EM = Expectation Maximization) algorithms. My proposed kernel density estimation (KDE) algorithm also falls also into the category of iterative methods. However, in this approach each coincidence event is considered individually. The estimate location of the annihilation event that caused each coincidence event is based on the previously assigned location of events processed earlier. To accomplish this, we construct a probability distribution along each coincidence line. This is generated from previous annihilation points by density estimation. It is shown that this density estimation approach to PET can reconstruct an image of an existing tumor using significantly less data than the standard CT algorithms, such as FBP. Therefore, it might be very promising technique allowing reduced radiation dose for patients, while retaining or improving image quality. / October 2007

The development of a non-invasive method of visualizing gene expression in larger animals could revolutionize some aspects of gene research by opening up a wider variety of animal systems to explore; some of which may be better models of human systems. Presently, most gene expression studies employ Green Fluorescent Protein (GFP) transfected into the genome of the animal system. For larger animals, an x-ray equivalent of GFP would be desirable due to the high penetrating power of x-rays. A model gene modification system is to use the Sodium (Na) Iodide Symporter (NIS) which will cause the accumulation of iodine in cells which express the NIS. To non-invasively observe the dilute iodine accumulated by the cancer cells transfected with NIS in the head of small animals, such as a rat, two synchrotron-based imaging methods were studied: K-Edge Subtraction (KES) imaging and Fluorescence Subtraction Imaging (FSI).<p> KES needs wide monochromatic x-ray beams at two energies bracketing the K-edge of the contrast agent existing or injected in the tissues. The monochromatic beam in the synchrotron facility normally is prepared by a double crystal monochromator. The appearance of the azimuthal angle (tilt error) in the double crystal monochromator creates intensity variations across the imaging field. This misalignment was studied through another two synchrotron-based imaging methods, Diffraction Enhanced Imaging (DEI) and Multi-Image Radiography (MIR), which show this problem clearly in their processed images. The detailed analysis of the effect of the tilt error, how it affects the resulting images, and how to quantify such an error were presented in the thesis. A post processing method was implemented and the artifacts caused by the improper experimental settings were discussed.<p> With the wide monochromatic beam prepared by the double crystal monochromator, a sequence of KES experiments were done and the detection limit of KES was quantified at a projected amount of 17.5mM-cm iodine in a physical model of a rat head with a radiation dose of 2.65mGy. With the raster scan of the object relative to the monochromatic pencil beam, FSI was studied to obtain higher Signal to Noise Ratio (SNR) for local area and better detection limit compared to KES. The detection limit of FSI was measured as a projected amount of 2.5mM-cm iodine in the same physical rat head with a tolerable radiation dose of 24mGy. According to the comparison of these two imaging techniques with references to imaging time and area, radiation dose, spatial resolution, and SNR, it was concluded that these two imaging techniques can be used complementarily in imaging dilute contrast material. Due to the short imaging time and large imaging area, KES is used first to provide a global view of the object, locate the area of interest, do the preliminary diagnosis, and decide whether the further FSI is necessary. Due to its high SNR for the dilute sample, FSI can be used when the area of interest is known. The combination of these two imaging techniques will be very promising and powerful. To facilitate the comparison of KES and FSI, a quality factor was developed to evaluate the performance of the imaging system.<p> The measured detection limits in our experiments are far beyond the thyroidal iodine concentration of a rat (around 1mM). To further improve the performance of KES, a bent Laue crystal monochromator was designed to do the simultaneous iodine KES imaging which overcomes the artifacts in the iodine image caused by the temporal difference for a single set of images. The designed monochromator can provide two separated x-ray beams bracketing the K-edge of iodine at the same time with a very high spatial resolution which is only depends on the source size, a very high energy resolution which can almost compete with that of the double crystal monochromator, and an acceptable photon flux.

Objective: The purpose of the research presented in this thesis is to explore new biomedical applications of synchrotron imaging in the field of neurosurgery.<p> Methods: Four different studies were performed, all using advanced biomedical synchrotron imaging techniques. In the first two experiments, diffraction enhanced imaging (DEI) and analyzer based imaging (ABI) were utilized to study the anatomy of the rat spine and a novel rat model of spinal fusion. In a third experiment, K-edge digital subtraction angiography (KEDSA) was used to study the cerebral vasculature in a rabbit model. In a fourth experiment, rapid scanning X-ray fluorescence spectroscopy (RS-XRF) was used to study stem cell migration in a rat stroke model.<p> Results: DEI had superior visualization of ligamentous and boney anatomy in a rat model. Analyzer based imaging was able to visualize physiologic amounts of bone graft material and progressive incorporation into the spine. Intravenous KEDSA showed excellent visualization of the cerebral vasculature in a rabbit model. Finally, RS-XRF was used to track iron labeled stem cells implanted in a rat stroke model. The technique was able to visualize the iron that represented the stem cell migration. This was correlated with histology and magnetic resonance imaging information.<p> Conclusions: 1) Diffraction enhanced imaging has excellent contrast for the study of boney and ligamentous anatomy. 2) Analyzer based imaging is an excellent tool to study animal models of boney fusion. 3) Intravenous KEDSA is able to clearly visualize the arterial vasculature in a rabbit model. 4) RS-XRF can be used to study the migration patterns of implanted iron labeled stem cells.

This research is primarily focused on the study of tumors in experimental animal models using functional imaging in the presence of various contrast agents. The study of malignant tumor angiogenesis, oxygen saturation, glucose and amino acid metabolisms will lead to better methods for cancer detection as well as diagnosing and managing cancer. Non invasive in vivo diagnostic imaging technique is an area of great clinical interest in present days. In this study, noninvasive in vivo photoacoustic tomography and conventional fluorescence imaging together with multiphoton microscopic tomography were implemented to study the malignant tumor morphology and physiology. Tumor structure and angiogenesis were successfully imaged by photoacoustic tomography and conventional fluorescence imaging. The important malignant tumor cellular parameters such as oxygen saturation and αvβ3 integrin concentration were measured in living small animals (rodents) using the novel photoacoustic tomography technique. By implementing multiphoton microscopy using Cy3.5 NHS ester contrast agent, tumor amino acid metabolism was successfully studied in cell culture. This method will at least give you a relative concentration map of amino acid in cells. Non invasive in vivo imaging can be achieved by modifying the current multiphoton imaging setup. A new method for studying amino acid and glucose metabolisms of tumor cells using multiphoton imaging was developed.

Fluorescence enhanced optical imaging and tomography studies involve the detection of weak fluorescent signals emanating from nano- to picomolar concentrations of exogenous or endogenously produced fluorophore concurrent with the rejection of an overwhelmingly large component of backscattered excitation light. The elimination of the back-reflected excitation light of the collected signal remains a major and often unrecognized challenge for further reducing the noise floor and increasing sensitivity of small animal fluorescence imaging. In this dissertation, we adapted collimating and gradient index (GRIN) lenses in an existing frequency-domain system to improve excitation light rejection and enhance planar and tomographic imaging. To achieve this goal, we developed planar and tomographic imaging systems based upon ray tracing calculations for improved rejection of excitation light. The “out-of-band (S (λx))” to “in-band (S (λm) - S (λx))” signal ratio assessing excitation leakage was acquired with and without collimating optics. The addition of collimating optics resulted in a 51 to 75% reduction in the transmission ratio of (S (λx))/ (S (λm) - S (λx)) for the phantom studies and an increase of target to background ratio (TBR) from 11% to 31% in animal studies. Additionally, we presented results demonstrating the improvement of model match between experiments and forward simulation models by adaptation of GRIN lens optics to a breast phantom study. In particular, 128 GRIN lenses on the fiber bundle face were employed to align the collected excitation and emission light normal to the filter surface in an existing frequency-domain system. As a result of GRIN lens collimation, we reduced the transmission ratio between 10 and 86 % and improved the model match for tomographic reconstruction of one (1 cm3) and two (0.1 cm3) targets in a 1087 cm3 of breast phantom. Ultimately, this work improves the sensitivity of NIR fluorescence imaging by enhancing the rejection of excitation light and shows that the current sensitivity challenges for translating fluorescence-enhanced optical imaging into the clinic can be overcome.

Seismic imaging of sub-basalt sedimentary layers is difficult due to high impedance of the basalt layer, the roughness of the top and bottom of the basalt layer and sometimes the heterogeneities within the basalt layer. In this thesis we identify specific problems within the modern imaging technology which limit sub-basalt imaging. The basic framework for the identification of this limitation is that we are able to group most basalt layers into the following four categories: A basalt layer having smooth top and bottom surfaces. A basalt layer having rough top and bottom surfaces. Small-scale heterogeneities within the basalt layer. Intra-basalt velocity variation due to different basalt flows. All the above models of basalt layers obviously have high impedance with respect to the surrounding sedimentary layers. These four models encapsulate all the possible heterogeneities of basalt layers seen in areas like the Voring and More basins off mid- Norway, basins in the Faroes, W. Greenland, Angola and Brazil margins, and the Deccan Traps of India. In this work, problems in seismic processing and imaging specific to these models have been presented. For instance, we have found that the application of the multiple attenuation technique, which first predicts the multiples and then subtracts them from the data, using least-squares criteria, can be effective for all the models except for the model, which has intra-bedded layers within the basalt. The failure in the second case is due to the destructive interference of multiple scattering from the intra-bedded layers within the basalt and the multiples located below the primary associated with the top of the basalt layer. This interference degrades the signal-to-noise (S/N) ratio of the multiples contained in the data, whereas the predicted multiples, which are constructed from the reflectors above the basalt, have a much higher signal-to-noise ratio. Our recommendation is to subtract the predicted multiples from the data using either leastabsolute- value criteria or any other higher-order-statistics-based criteria.

Novel microscale velocity and temperature measurement techniques were studied based on confocal laser scanning microscopy (CLSM) and optical serial sectioning microscopy (OSSM). Two microscopic measurement systems were developed, 1) a CLSM micro particle image velocimetry (PIV) system with a dual Nipkow disk confocal unit (CSU-10), a CW argon-ion laser and an upright microscope, and 2) an OSSM micro- particle tracking velocimetry (PTV) system with an epi-fluorescence microscope and a non-designed specimen to make a three-dimensional (3-D) diffraction particle image. The CLSM micro-PIV system shows a unique optical slicing capability allowing true depth-wise resolved vector field mapping. A comparative study is presented between the CLSM micro-PIV and a conventional epi-fluorescence micro-PIV. Both have been applied to the creeping Poiseuille flows in two different microtubes of 99-µm (Re = 0.00275) and 516-µm ID diameters (Re = 0.021). The CLSM micro-PIV consistently shows significantly improved particle image contrasts, the definition of "optical slicing" and measured flow vector fields more accurately agreeing with predictions based on the Poiseuille flow fields, compared to the conventional micro-PIV. The OSSM micro-PTV technique is applied for a 3-D vector field mapping in a microscopic flow and a Brownian motion tracking of nanoparticles. This technique modifies OSSM system for a micro-fluidic experiment, and the imaging system captures a diffracted particle image having numerous circular fringes instead of an in-focus particle image. The 3-D particle tracking is based on a correlation between the 3-D diffraction pattern of a particle and the defocus distance from a focal plane. A computational program is invented for the OSSM micro-PTV, and provides a 3-D velocity vector field with a spatial resolution of 5.16 µm. In addition, a concept of nonintrusive thermometry is presented based on the correlation of the Brownian motion of suspended nanoparticles with the surrounding fluid temperature. Detection of fully three-dimensional Brownian motion is possible by the use of the OSSM, and the measured value of mean square displacement (MSD) is compared fairly well with Einstein's predictions.

In Magnetic Resonance Imaging, the time required to generate an image is proportional to the number of steps used to encode the spatial information. In rapid imaging, an array of coil elements and receivers are used to reduce the number of encoding steps required to generate an image. This is done using knowledge of the spatial sensitivity of the array and receiver channels. Recently, these arrays have begun to include a large number of coil elements. Ideally, each coil element would have its own receiver channel to acquire the image data. In practice, this is not always possible due to economic or other constraints. In this dissertation, methods are explored to combine a large array to a limited number of receivers so as to optimize the performance for parallel imaging; this dissertation focuses on SENSE in particular. Simple combinations that represent larger coils that might be constructed are discussed. More complex solutions form current sheets. One solution uses Roemer'€™s method to optimize image SNR at a set of points. In this dissertation, Roemer's solution is generalized to give the weighting coefficients that optimize SNR over regions. Also, solutions fitted to ideal profiles that minimize noise amplification are shown. These fitted profiles can allow the SENSE algorithm to function at optimal reduction factors. Finally, a description of how to build the combiner in hardware is discussed.

To reduce the image blurring which originates from contributions of a cylindrical array of photolysis events in a photo-induced experiment, a variant of fluorescence imaging techniques has been developed to study photodissociation dynamics and collisional relaxation processes in the bulk. The experimental arrangement utilizes sliced imaging techniques of photofragments by the laser-induced fluorescence detection scheme. An unconventional procedure is employed to guide the photolysis laser in the viewing direction of the imaging detector with a proper obstruction. The sliced image in the direction perpendicular to the photolysis laser is equivalent to a two-dimensional projection of the fluorescence image of photoproducts from a single photolysis center. Experimental images of state-selected CN photofragments from the ICN photodissociation are presented to illustrate the versatility of the present method.

Computer vision encompasses a host of computational techniques to process visual information. Medical imagery is one particular area of application where data comes in various forms: X-rays, ultrasound probes, MRI volumes, EEG recordings, NMR spectroscopy, etc. This dissertation is concerned with techniques for accurate reconstruction of neural pathways from diffusion magnetic resonance imagery (dMRI). This dissertation describes a filtered approach to neural tractography. Existing methods independently estimate the diffusion model at each voxel so there is no running knowledge of confidence in the estimation process. We propose using tractography to drive estimation of the local diffusion model. Toward this end, we formulate fiber tracking as recursive estimation: at each step of tracing the fiber, the current estimate is guided by those previous. We argue that this approach is more accurate than conventional techniques. Experiments demonstrate that this filtered approach significantly improves the angular resolution at crossings and branchings. Further, we confirm its ability to trace through regions known to contain such crossing and branching while providing inherent path regularization. We also argue that this approach is flexible. Experiments demonstrate using various models in the estimation process, specifically combinations of Watson directional functions and rank-2 tensors. Further, this dissertation includes an extension of the technique to weighted mixtures using a constrained filter.