Traditional information-processing accounts of the reasoning process in radiology assume that humans process the details of the input image in order to compute judgments. In these accounts, the development of expertise involves the acquisition of increasingly precise and complex internal problem representations that are based on a normal anatomy prototype. Fuzzy-trace theory predicts that accurate judgments rely on the reasoners ability to ignore irrelevant detail, to retrieve relevant gist memories and to accurately instantiate the image information with respect to the internal representation. Fuzzy-trace theory predicts that the development of expertise involves the ability to access and process less precise and complex internal representations (i.e., gist). The purpose of this study was to examine the internal representations used to make judgments in radiology and to quantify the changes in complexity of the internal representations, as well as the differences in time, accuracy and confidence that might be associated with experience. Thirty-five subjects from general and specialized expertise samples participated. Each subject was presented with 32 chest films including normal films, films with precise disease patterns (mass category) and diffuse disease patterns (interstitial and airspace category). For each film, the participant made a series of judgments (normal/abnormal; category; specific diagnosis) and then sketched the features that were essential to the judgments. The information content and complexity of the representations were calculated using an approach that considered the underlying meaning of the sketches rather than the surface form. The sketches were converted to propositions and the information in the propositions was evaluated in terms of possible world semantics. Time, accuracy, confidence and content measures supported the prediction of fuzzy-trace theory that the internal representations are abnormality-based. Consistent with predictions regarding the acquisition and use of gist representations expertise was associated with greater improvements in accuracy for interstitial (as opposed to mass) films; accuracy was higher with interstitial films when judgments were less reliant on surface detail (normal/abnormal); and accuracy was higher for mass films when judgments were more reliant on surface detail (specific diagnosis). Complexity measures showed that the overall representations did not get more complex with the development of expertise.

<p> The motivation for this work is a vision of widespread adoption of a priori quantitative epidemiological information for clinical decision-making, and can be seen as a quantitative large-scale extension of evidence-based medicine (EBM). Medical images can be seen as a spatially encoded map of physiological measurements that can be used to predict prognosis and to drive treatment plans. This paradigm can be very powerful and is driven by the recent big data revolution in computer science as well as the increasing availability of medical imaging modalities due to decreases in manufacturing costs. In order to achieve this overarching goal, three practical requirements must be reached and correspond to the parts of this thesis: Part A: Developing IT infrastructure and technology that enables the dataset to be properly collected and organized for analysis. Part B &amp; C: Generation of functional (Part B) and structural (Part C) medical imaging contrast that are optimized for analysis. Part D: Pattern recognition techniques (including both image processing and machine learning techniques) to mine information from the large imaging datasets generated. As part of the thesis, I discuss my contribution to IT infrastructure (Part A) by developing a Short Message Service (SMS)-based system to control the clinically used Picture Archival and Communication System (PACS) (Ch.2) as well as an imaging study tool that categorizes patient imaging data for use in retrospective studies(Ch.3). I then go on to detail my work with functional neuroimaging of obesity using functional magnetic resonance imaging (fMRI)(Ch.4) and (Ch.5). Chapters 6-9 details my efforts at studying abnormal aging versus normal aging using diffusion MRI as well as applications of diffusion MRI to surgical planning. Chapters 10 discusses my work integrating diffusion MR with FLAIR MRI to investigate the properties of white matter lesions and how it can be used in the clinical setting. Chapter 11 then moves on to talk about my work modifying standard brain parcellation techniques to allow them to work with aged brains with large infarcts. Chapters 6-11 altogether represent my efforts in structural neuroimaging using MRI (Part C). The thesis then closes with capstone work in development staging using hand x-rays using fuzzy logic (Ch. 12 &amp; 13). To close the work with Alzheimer's Disease (AD) and aging, we used machine learning techniques to predict disease progression based on a baseline MRI scan as well as higher order analysis of our diffusion MRI dataset by integrating MRI information with other clinical information such as neuropsychological tests, cardiovascular status. This is all in an effort to computationally explore the relationship between MRI measurements and clinical presentation of disease as measured by neuropsychological scores. Similarly with the Obesity work, we related fMRI activation differences between high and low calorie foods with non-imaging information such as insulin resistance (Ch. 16).</p>

The purpose of this study was to determine the cumulative doses of x-ray radiation to the thyroid gland and female breast from spinal radiographs for Adolescent idiopathic scoliosis (AIS) and to estimate the number of cancers at these sites attributable to x-rays. / Subjects for this study were patients referred from 1960 to 1979 for AIS to Hopital Ste-Justine, Montreal. / About 85 percent of 2,181 subjects were first referred for scoliosis between the ages of eleven and seventeen and the average time under observation was about three years. The mean number of radiographs was about twelve. The mean cumulative dose to the thyroid gland and to the female breast was about three cGy. Seven excess breast cancer and thyroid cancer cases were projected to occur over the lifetime of the women; among these, two excess deaths from cancers were projected. Approximately one in every 250 women in this cohort would be expected to develop breast or thyroid cancer over their lifetime, and one in every 900 women would be expected to develop a fatal cancer. (Abstract shortened by UMI.)

<p> Diffuse optical tomography (DOT) is an optical imaging modality that uses near infrared light to recover functional information of tissue. In this thesis we focus on breast imaging where estimation of the optical properties of the breast can assist in detecting cancerous tumors and in judging overall breast health. </p><p> To this end we explore the application of a parametric level set method (PaLS) for image reconstruction for hyperspectral DOT. Chromophore concentrations and diffusion amplitude are recovered using a linearized Born approximation model and employing data from over 100 wavelengths. The images to be recovered are taken to be piecewise constant and a newly introduced, shape-based model is used as the foundation for reconstruction. The PaLS method significantly reduces the number of unknowns relative to more traditional level-set reconstruction methods and has been shown to be particularly well suited for ill-posed inverse problems such as the one of interest here. We extend the PaLS method to imaging problems by considering a redundant dictionary matrix for basis functions allowing for recovery of a wide array of shapes. </p><p> Additionally we explore the ability of diffuse optical tomography (DOT) to recover 3D tubular shapes representing vascular structures in breast tissue. Using the PaLS method, we incorporate the connectedness of vascular structures in breast tissue to reconstruct shape and absorption values from severely limited data sets. The approach is based on a decomposition of the unknown structure into a series of two dimensional slices. Using a simplified physical model that ignores 3D effects of the complete structure, we develop a novel inter-slice regularization strategy to obtain global regularity. We report on simulated and experimental reconstructions using realistic optical contrasts where our method provides a more accurate estimation compared to an unregularized approach and a pixel based reconstruction.</p>

Magnetic Resonance (MR) imaging is a 3-D, multi-slice, radiological technique that acquires multiple intensities corresponding to each voxel. The transverse relaxation time, T$\sb1$, and the axial relaxation time, T$\sb2$, are two commonly obtained intensities that tend to be orthogonal. Automated segmentation of 3-D regions is very difficult because some borders may be delineated only in T$\sb1$ images, while others are delineated only in T$\sb2$ images. Classical segmentation techniques based on either global histogram segmentation or local edge detection often fail due to the non-unique and random nature of MR intensities. A 3-D, neighborhood based, segmentation method was developed based on both spatial and intensity criteria. The spatial criterion requires that only voxels connected by an edge or face to a voxel known to be in the region be considered for inclusion. Therefore, the region "grows" outward from an initial voxel. An intensity criterion that tries to balance local and global properties must also be satisfied. It determines the vector distance between the intensity of the voxel in question and a characteristic intensity for the neighboring voxels known to be in the region. Voxel intensities within a 95% confidence interval of the characteristic intensity are considered part of the region. The kernel size used to determine the characteristic intensity determines the balance between global and local properties. The segmentation terminates when no additional voxels satisfy both spatial and error criteria. Some regions, such as the brain compartments, are highly convoluted, resulting in a large number of border voxels containing a mixture of adjoining tissues. A sub-voxel estimate of the fractional composition is necessary for accurate quantification. A least-squares estimator was derived for the fractional composition of each voxel. Additionally, a maximum likelihood estimator was derived to globally estimate the fraction for all mixture voxels. This estimator is a minimum variance estimator in contrast to the least-squares estimator. The estimation methods in conjunction with the 3-D, neighborhood based, segmentation method resulted in an automated, highly accurate, quantification technique shown to be successful even for the brain compartments. Widespread applicability of these methods was further demonstrated by segmentation of kidneys in CT images.

The use of contrast media in positron emission tomography (PET)/computed tomography (CT) dual modality imaging has been shown to cause artifacts in the PET image. These artifacts are attributed to an overestimation of the PET attenuation coefficients, which are obtained from contrast-enhanced CT numbers. This dissertation evaluates three algorithms, which segment intravenous contrast-enhanced tissue from CT images, so as to minimize this bias. The algorithms evaluated are the template matching; 3D region growing, and snake-based methods, and they were tested using 5 patient studies. Segmentation results for each method were compared to corresponding manually segmented images on a pixel-wise basis. The snake-based technique was judged to be most suitable for efficiently segmenting the contrast-enhanced CT images. This technique can lead to a more efficient acquisition of high quality PET/CT data, by enabling the use of contrast media without introducing related artifacts.

The distribution of neuropeptide Y (NPY) receptor sites was investigated in certain mammalian species using receptor autoradiography and membrane binding assay techniques. NPY receptor sites are discretely distributed throughout the central nervous systems of hamster, rat, guinea pig and monkey but are particularly concentrated in cortex and hippocampus. NPY receptor autoradiography techniques also revealed important labelling of numerous thalamic nuclei while most hypothalamic nuclei showed surprisingly low densities of NPY receptor sites, considering the high levels of NPY-like immunoreactivity in this brain area. The widespread distribution of NPY receptor sites suggests that this peptidergic system must have important roles in mammalian central nervous system. In particular, the possible involvement of NPY in cognitive function deserves further investigation since we observed that ($ sp3$H) NPY receptor site densities are decreased in temporal cortex and hippocampus of individuals decreased with Alzheimer's disease. / The possible existence of NPY/peptide YY (PYY) receptor subtypes was investigated in the rat brain. Overall, the similar autoradiographic distribution of ($ sp{125}$I) BH-NPY and ($ sp{125}$I) PYY in most areas suggests that these two receptor probes most likely interact with the same population of NPY/PYY receptor sites. ($ sp{125}$I) PYY may recognize both a high and low affinity state/subtype of NPY/PYY receptors while ($ sp{125}$I) BH-NPY recognize a single affinity state of receptor having the binding characteristics of the low affinity ($ sp{125}$I) PYY receptor state/subtype. The exact nature of this high affinity receptor state/subtype remains to be established. / Finally, the structural requirements of NPY receptors in central (CNS) and peripheral (PNS) nervous systems were also examined with a binding assay on rat brain membrane preparation and with the rat vas deferens bioassay preparation. The amino acid residues responsible for the activation of NPY receptors in the rat vas deferens preparation lie in the C-terminal half of the NPY molecule as revealed by the loss of potency, but not of biological activity with C-terminal fragments up to NPY$ sb{18-36}$. The N-terminal portion of the NPY molecule appears to be mostly important to insure adequate affinity for central and peripheral NPY receptors. Moreover, a series of NPY analogs revealed that the two tyrosine residues in position 20 and 21 are not directly involved in NPY receptor activation in this bioassay preparation, although they appear relatively important for the maintenance of adequate affinity for the receptors. Interestingly, modifications of the tyrosine residue in position 20 led to the development of two analogs demonstrating a certain degree of selectivity for the PNS receptors while other modifications of tyrosine residues in position 21 may provide some selectivity for CNS receptor sites. (Abstract shortened by UMI.)

Using Monte Carlo techniques, the 18 MV accelerators CL2300 and CL21EX, 8 physical wedges and the enhanced dynamic wedge have been modeled and validated. The simulated PDDs, profiles, output factors and wedge factors are in good agreement with measured values. These models have been applied in the dose calculation of lung cancer patients to investigate the effect of tissue heterogeneity by comparing with the CADplan treatment planning system with and without heterogeneity correction. Plan I and plan II for 8 patients have been re-calculated with the Monte Carlo method. On average, the mean PTV dose calculated with the Monte Carlo method is 2.4% higher than that calculated with CADplan without correction but 2.5% lower than that with EqTAR correction. The Monte Carlo method predicts 13% higher mean lung dose than CADplan without heterogeneity correction. Because of this difference, the risk of radiation pneumonitis based on available NTCP models will be underestimated. A relationship of mean lung dose between two types of dose algorithms can be used to correct this underestimation.

This study involves the measurement of the neutron equivalent dose ( NED) and the induced activity produced from medical linear accelerators. For the NED, various parameters such as the profile, field effects and energy responses were studied. The NED in a Solid Water(TM) phantom was measured and a new quantity, the neutron equivalent dose tissue-air ratio (NTAR), was defined and determined. Neutron production for electron beams was also measured. For the induced activity, comparisons were carried out between different linacs, fields and dose rates. The half life and activation saturation were also studied. A mathematical model of induced activity was developed to explain the experimental results. Room surveys of NED and induced activity were performed in and around a high energy linear accelerator room. Unwanted doses from photoneutrons and induced activity to the high energy linear accelerator radiotherapy staff and patient were estimated.

A Cisplatin-resistant subline of a Human Ovarian cancer cell line, A2780 CP also exhibits cross resistance to 60Co gamma-radiation. Cellular Glutathione (GSH) levels were measured by the Tietze Spectrophotometric method. GSH levels in A2780 CP cells were approximately three time that of the parental A2780 cells. In cells depleted of GSH by L-Buthionine Sulfoximine (BSO), Cisplatin-resistance was significantly decreased in both cell lines. BSO had no effect on the radiation resistance of either cell lines. However treatment of cells with a combination of BSO and Dimethyl Fumarate (DMF) significantly decreased the radio-resistance in A2780 CP cells. The combination treatment had no effect on the radio-response of the parental cells. Ethacrynic acid (EA), an inhibitor of Glutathione S-Transferase (GST) activity had no effect on cellular GSH levels nor did it affect the Cisplatin or radiation resistance in either cell line. These results suggest that GSH levels are involved in Cisplatin and radiation resistance in these cells. However the mechanism of GSH involvement in radiation and Cisplatin-resistance appears to be different in the cell lines studied.