Image-guided surgery (IGS) has emerged as a valuable tool for the interactive incorporation of pre-operative image data into the surgical setting. While a majority of research has focused on neurosurgical procedures, the feasibility of implementing IGS methods in hepatic procedures has become evident. However, a lack of robustness in the methods developed for performing image-guided liver surgery (IGLS) has impaired their utility. In order to improve the robustness of the algorithms used within IGLS, the incorporation of salient anatomical features that can be reliably identified on the hepatic anatomy is proposed within this work. More specifically, these anatomical features can be weighted within the performance of the image-to-physical space mapping, or registration, that is required for display of the intra-operative location of surgical instruments within the context of pre-operative image data. Additionally, the salient anatomical feature registration can be used to quantify the extent of soft tissue deformation that is known to compromise the intra-operative registration due to the use of rigid body assumptions in determining the mathematical mapping. The quantification and analysis of soft tissue deformation within the context of IGLS provides unique insight into the design of algorithms that can be used to compensate for the shift induced guidance errors. Based on the deformation studies, a novel atlas-based model updating method is proposed for the improvement of IGLS guidance accuracy. The atlas-based method relies on the utilization of pre-operatively computed model solutions to expedite the determination of a non-rigid image-to-physical space mapping. Further, the atlas-based method incorporates the hepatic salient anatomical features to improve the robustness of the method.

Non-specific antibody binding limits the sensitivity of antibody-based detection technologies. We explore the use of logical operations among DNA logic tags associated with antibodies to increase specificity and sensitivity. DNA sequences were developed to perform a logical NOT operation to subtract non-specific binding prior to PCR amplification. Antibody-associated tags A and A' were designed to be partially complementary and contain a restriction enzyme site. Tag A is associated with a specific antibody; tag A' is associated with an isotype matched control antibody. If the concentration of A is greater than A', hybridized AA' is enzymatically cleaved and remaining tag A is subsequently amplified during real-time PCR. Quartz crystal microbalance (QCM), DNA agarose gels, and PCR were used to experimentally characterize components of the NOT operation in respiratory syncytial virus (RSV) detection. QCM showed gold nanoparticles functionalized with both tag DNA and antibody can bind virus. Successful enzymatic cleavage of AA' was visualized on a DNA agarose gel. After cleavage, remaining tag A was amplified by the addition of primers and standard real-time PCR. In the presence of RSV, magnetic pulldown led to the delivery of both tag A and A'. When PCR was run after enzymatic cleavage, the PCR Ct was increased, indicating a decrease in initial tag A. Our results suggest that combining the careful design of DNA logic tags, their association with antibodies, and standard molecular biology techniques is a promising approach to increase the specificity and sensitivity of antibody-based detection methods.

BIOMEDICAL ENGINEERING QUANTUM DOT-BASED IN VIVO IMAGING OF INFLAMMATION ASHWATH JAYAGOPAL Thesis under the direction of Professor Frederick R. Haselton The recruitment of leukocytes to tissue and their specific interactions with adhesion molecules are essential processes which provide for a natural mechanism of localizing immune defense. However, inappropriate immune responses can manifest as harmful inflammation in a variety of diseases. Much information has been derived from immunohistochemistry and in vitro cell culture studies, which have identified various inflammatory mechanisms and mediators. A key challenge has been the in vivo investigation of detailed cellular and molecular events in real-time, such as leukocyte extravasation. A variety of cell adhesion molecules and cell types facilitate these activities, and the ability to monitor their recruitment and proliferation would likely have an impact on the development of diagnostic and therapeutic avenues. Current imaging techniques to characterize these events are limited by low signal to background ratios, invasiveness, and fading. In this study, we use fluorescence microscopy and quantum dot-antibody conjugates to specifically investigate the expression of the cell adhesion molecule VCAM-1 in diabetic rats. The retina was used to non-invasively probe inflammatory activity in the circulation, and to investigate the impact of diabetes-induced inflammation on ocular complications. We report elevated VCAM-1 levels in diabetic rats relative to untreated controls. In addition, we observed ex vivo and in vivo quantum dot-labeled leukocytes in the healthy and diseased retinal circulation. Features of this technique include stable, high intensity of labeled species and a narrow, size-tunable emission spectra. These studies demonstrate the promise of quantum dots in the in vivo visualization of molecular and cellular inflammatory mediators. Approved: Date: Frederick R. Haselton 7/29/05

Corneal wound healing is a complex, and highly coordinated process of cell-cell signaling, cell spreading, migration, proliferation, and cell-extracellular matrix interaction. The main limitation in natural wound healing is that cells can only fill in the wound void from the outside wound margin. We hypothesize that epithelial cells applied to the interior of a wound will attach to the exposed extracellular matrix of the wound, and accelerate the healing process. In order to deliver epithelial cells to wounds, we have developed a cell transfer contact lens. Our objective was to produce and evaluate transfer contact lens designs that allow for primary epithelial cell attachment and in vitro culture growth, and in addition after application result in the transfer of a portion of these cells to wound areas to aid in re-epithelialization of corneal injuries. In the process of meeting these design criteria, we explored the use of surface topographies and surface coatings to enhance cell transfer. We found that cell transfer from Polydimethylsiloxane contact lenses to extracellular matrix was a function of both cell motility and adhesion. In order to enhance cell transfer through the modulation cell motility and adhesion, we explored some of the underlying biological mechanisms of these two inter-related cellular behaviors. Though there were many potential molecular targets to augment cellular adhesion and motility, we specifically explored the regulation of cell-cell contacts, which have a profound impact on cell phenotype and migration. Normal corneal epithelial cells are motile cells that move slowly as a sheet. In contrast, fibroblasts move more quickly as individual cells. When epithelial cells exist as individual cells at very low cell density, they are more motile and more closely resemble fibroblast cells. Individual epithelial cells do not exhibit cell junction proteins. In this investigation we altered Bves, a protein we have shown is associated with tight junctions, in an effort to make epithelial cells become more like fibroblasts regardless of cell density. This transition from an epithelial cell type to a fibroblastic cell type is called epithelial-mesenchymal transition, and has broad reaching implications for other fields such as metastasis in cancer biology.

Dealing with a cancer diagnosis and cancer treatment involves communication among clinicians, patients, families, friends and others affected by the illness. The hypothesis for this research is that an informatics system can effectively support the communication needs of cancer patients and their formal and informal caregivers. How can an online communication system be made accessible and desirable for all of the different players involved in the patients care and support? The initial design of a novel informatics-based system for cancer communication was conducted in three phases (The Discovering Phase, The Developing Phase and The Testing Phase). Five types of clinical and supportive relationships were identified and supported by in-depth interviews with cancer patients and their informal caregivers. A prototype web-based communication system was created with an emphasis on the interpersonal relationships between patients, families, and clinicians. The system was used over the course of two months by a group of clinicians and head and neck cancer patients in the Vanderbilt Ingram Cancer Center. A case study is presented of one patients detailed feedback and use of the system with the clinicians and with family and friends. A larger study of the systems effect on the clinical and social outcomes will be conducted during the PhD phase of this research. The current phase of this research concludes with the initial user feedback and iterative design of the system prototype.

Cancer is a complex and adaptable disease, and knowledge of the mechanisms that cause its progression is vital to creating and monitoring anti-cancer therapies. As a tumor grows beyond a few mm3, blood vessels are recruited to provide additional nutrients in a process called angiogenesis. Some novel drug therapies specifically target this process, and the efficacy of these drugs can in principle be monitored by a technique called angiography. 3D angiography, a method of imaging the vasculature, can be performed by several imaging modalities typically with the use of a contrast agent. A technique is introduced which uses high resolution ultrasound in conjunction with an ultrasound contrast agent to produce 3D images of the vasculature. This method offers a faster, more accessible, and cheaper alternative to assess the efficacy of anti-angiogenic drugs in preclinical cancer models. In addition to vessel imaging, modeling the kinetic behavior of the contrast agent in the vasculature can elucidate parameters such as blood flow, which can also serve as an indicator of drug treatment efficacy. The repeatability of a commonly used mono-exponential model is assessed in order to determine thresholds for inter/intra-subject error. The 3D vessel imaging technique presented in this thesis correlated with other measures of blood flow (r = 0.55 ± 0.04, p < 0.01) and shows an increased sensitivity to microvasculature within tumors. Also, preliminary repeatability analysis (n = 6) on the modeling parameter which is proportional to contrast agent velocity shows a mean difference of 0.061 ± 0.298 between independent measurements, and the limits of agreement range from -0.536 and 0.656. The developments exhibited provide additional methods for monitoring longitudinal anti-angiogenic cancer treatments in preclinical models.

Magnetic resonance imaging (MRI) is unique in its sensitivity to a wide array of contrast mechanisms in soft tissue. Unfortunately, the physiological and/or microanatomical characteristics that give rise to this contrast can show significant heterogeneity on the scale of a typical voxel, resulting in an observed nuclear magnetic resonance (NMR) signal that is a summation of these spatially varying characteristics. Multicomponent analysis, which allows one to separate the observed NMR signal into components that represent underlying sub-voxel tissue compartments, can be used to deal with this limitation. For example, this approach has been applied to myelinated tissue (e.g., white matter, peripheral nerve) to measure the so-called myelin water fraction, which has been shown to correlate with myelin content. Despite this promise, several fundamental issues regarding the compartmental models used to describe myelinated tissue exist. These include: 1) quantifying the effect of intercompartmental exchange and 2) the inability to resolve axonal water (water within myelinated axons) from interaxonal water (water outside myelinated axons) in the central nervous system. This dissertation presents a series of studies aimed to address these fundamental issues. To address the effect of exchange, a novel method for quantifying exchange rates, which allows for a significant reduction in scan time relative to existing methods, was developed. This method was tested and compared to existing methods via simulations studies, validated via phantom studies, and applied in excised myelinated tissue samples. To investigate methods for resolving axonal and interaxonal water, compartmental relaxation measurements were performed in myelinated tissue samples before and after administration of contrast reagents as a means to characterize the compartmental enhancement pattern associated with each reagent. The results of these studies suggest that administration of potassium dichromate in white matter and optic nerve ex vivo might allow one to resolve axonal and interaxonal water in these tissues. Though not directly related, an additional study was included, which showed that novel information about tumor microenvironment may be available in vivo using the multicomponent analysis techniques presented throughout this work.

MRI provides an excellent way of visualizing muscle inflammation; however, there are few techniques that serve to quantitatively assess edematous muscle. In this thesis, integrated relaxation measurements were made in vivo on edematous rat muscle with varying degrees of swelling. To investigate the effect of exchange on the observed relaxation parameters a two pool model was created and the Bloch-McConnell equations were solved for the varying amounts of swelling. Results from the simulation of the exchange model were compared to the observed data to extract fitted parameters for the compartmental relaxation times. These simulations also provided a comparison for the observed changes in the long-lived apparent T1. Edematous muscle was found to display both multiexponential T1 and multiexponential T2. Normal muscle, however, was found to exhibit only a single T1-T2 component. It was shown that the apparent T1 of the long-lived signal component in edematous muscle increased monotonically with an increase in the amount of edema. Knowledge of changes in T1 and the exchange kinetics in edematous muscle might help in further characterizing the micro-anatomy of muscle tissue in various stages of injury.

Functional magnetic resonance imaging (fMRI) is a noninvasive imaging technique capable of mapping cognitive networks in the human brain that has become widely used for both research and clinical applications. The activity of focal regions within the brain can be inferred based on their blood oxygen level dependent (BOLD) signal changes through time by recording a series of images while subjects perform carefully designed tasks. Steady state functional connectivity methods generate similar maps of neural networks through analyses of underlying activity. Resting state functional connectivity measurements reduce the demands placed on patient/subject compliance and can potentially extend fMRI to a wider range of applications. However, the factors that affect measurements of steady state functional connectivity remain unclear, and methods of making these measurements may be further improved. This dissertation addresses three different aspects of functional connectivity measured with fMRI. First, we study the effects of cognitive load on measurements of functional connectivity in the working memory and default mode networks, building on previous work in the motor network. We report increases in functional connectivity within both networks, and present evidence of changing connectivity between them. Second, we apply functional connectivity analyses to electroencephalography and fMRI data recorded simultaneously to investigate the regions underlying variance in frontal theta power. Our results show both positive and negative correlations to theta power across working memory loads, and changing correlations in the parahippocampal gyrus, among other regions. Third, we demonstrate two methodological improvements to the measurement of functional connectivity using data acquired at ultra high field (7T) and applying nonlinear measurements of connectivity. We show that detection and significance of functional connectivity can be improved by decreasing voxel volumes to sizes that are not practically achievable without ultra-high magnetic fields, presumably due to decreases in partial volume effects. In addition, we show that mutual information can be used to detect functional connectivity between regions that are commonly left unidentified by measurements based on linear correlation coefficients.

Tandem mass spectrometry-based shotgun proteomics has become a widespread technology for analyzing complex protein mixtures. A number of database searching algorithms have been developed to assign peptide sequences to tandem mass spectra. Assembling the peptide identifications to proteins, however, is a challenging issue because many peptides are shared among multiple proteins. IDPicker is an open-source protein assembly tool that derives a minimum protein list from peptide identifications filtered to a specified False Discovery Rate. Here, we update IDPicker to increase confident peptide identifications by combining multiple scores produced by database search tools. By segregating peptide identifications for thresholding using both the precursor charge state and the number of tryptic termini, IDPicker retrieves more peptides for protein assembly. The new version is more robust against false positive proteins, especially in searches using multispecies databases, by requiring additional novel peptides in the parsimony process. IDPicker has been designed for incorporation in many identification workflows by the addition of a graphical user interface and the ability to read identifications from the pepXML format. These advances position IDPicker for high peptide discrimination and reliable protein assembly in large-scale proteomics studies. The source code and binaries for the latest version of IDPicker are available from http://fenchurch.mc.vanderbilt.edu/.