No description available.

Despite facing technical issues (e.g., overfitting, vanishing and exploding gradients), deep neural networks have the potential to capture complex patterns in data. Understanding how depth impacts neural networks performance is vital to the advancement of novel deep learning architectures. By varying hyperparameters on two sets of architectures with different depths, this thesis aims to examine if there are any potential benefits from developing deep networks for segmenting intracranial aneurysms from 3D TOF-MRA scans in the ADAM dataset. / Master of Science / With the technologies we have today, people are constantly generating data. In this pool of information, gaining insight into the data proves to be extremely valuable. Deep learning is one method that allows for automatic pattern recognition by iteratively improving the disparity between its prediction and the ground truth. Complex models can learn complex patterns, and such models introduce challenges. This thesis explores the potential benefits of deep neural networks whether they stand to gain improvement despite the challenges. The models will be trained to segment intracranial aneurysms from volumetric images.

Actin isoform specific mutations have been identified as causes for various human diseases. These include twelve missense mutations in γ-nonmuscle actin leading to early onset autosomal dominate nonsyndromic hearing loss and twenty two missense mutations in α-smooth muscle actin leading to thoracic aortic aneurysms and dissections (TAAD). The molecular mechanisms leading to these human pathologies are mainly unknown, principally due to the inability to isolate pure mutant γ-nonmuscle actin and α-smooth muscle actin in quantities required for biochemical analysis. To begin to address these problems, I have individually expressed the human nonmuscle actin isoforms, β– and γ– nonmuscle actin, in a baculovirus expression system and characterized their biochemical properties. Surprisingly, despite a conserved amino acid difference of only 4 residues at or near the N-terminus, Ca-γ-actin exhibits slower monomeric and filamentous biochemical properties than β-actin. In the Mg-form, the difference between the two is smaller. Mixing experiments with Ca-actins reveal the two will readily co-polymerize. Calcium bound in the high affinity binding site of γ-actin may cause a selective energy barrier relative to β-actin that retards the equilibration between G– and F-monomer conformations resulting in a slower polymerizing actin with greater filament stability. This difference may be particularly important in sites such as the γ-actin-rich cochlear hair cell stereocilium where local mM calcium concentrations may exist. In hair cells γ-nonmuscle actin seems to play a central role in stereocilia maintenance. To determine how the deafness causing D51N-γ-mutant actin mutation leads to deafness, I expressed and characterized it in the γ-actin background. The D51N mutation, lethal when cloned into yeast, displayed decreased filament stability and polymerization kinetics of an actin more dynamic than γ-actin. This result suggests that the hearing effects of the γ-actin mutations on the hearing apparatus are not simply caused by an inability to polymerize. The observed increased polymerization rates and decreased filament stability may have major implications for the human disease, as the mutation may alter the ability of the γ-actin to fulfill its maintenance functions. To address the basis by which TAAD mutations cause vascular dysfunction I introduced two of the know human mutations, N115T and R116Q, into yeast actin, 86% identical to human α-smooth muscle actin. I then generated yeast strains expressing each of these mutations as the sole actin in the cell to assess their effect on actin function in vivo and in vitro. Both mutant strains exhibited reduced ability to grow under a variety of stress conditions, although the N115T cells were more severely affected. In vitro the mutations caused exhibited altered thermostability and nucleotide exchange rates indicating effects on monomer conformation with R116Q the most severely affected. The N115T actin demonstrated a biphasic elongation phase during polymerization, while R116Q actin demonstrated a markedly extended nucleation phase. Allele-specific effects were also seen on critical concentration, rate of depolymerization and filament treadmilling. R116Q filaments were hypersensitive to severing by the actin-binding protein cofilin. In contrast, N115T filaments were hyposensitive to cofilin, despite near normal binding affinities of actin for cofilin. The mutant specific effects on actin behavior suggest that individual mechanisms may contribute to TAAD. Understanding the mechanisms of actin dependent human diseases requires elucidation of the effects of the mutations on the behavior of actin per se, its regulation, and the impact on actin mediated processes within the cell. The work provided in this thesis and future studies will provide the information required to understand the pathways involved in these diseases and form innovative treatments for deafness and TAAD.

There is increasing interest in assessing the role of hemodynamics in aneurysm growth and rupture mechanism. Identification of the indicators of rupture risk can prove very valuable in the clinical management of patients. If rupture risk of aneurysms can be predicted, immediate preemptive treatments can be done for the high risk patients whereas others can avoid the risky intervention. Retrospective studies have been performed in the past to filter out indices that differentiate ruptured aneurysms from unruptured aneurysms. However, these differences may not necessarily translate to differences between aneurysms that present unruptured but fork towards growth/rupture and unruptured aneurysms that are invariably stable. The hypothesis of the present study is that hemodynamic indices of unruptured aneurysms when they first presented can be used to predict their longitudinal outcome. A prospective longitudinal cohort study was designed to test this hypothesis. Four clinical centers participated in this study and a total of 198 aneurysms were recruited. These aneurysms were chosen by the physicians to be kept under watchful waiting. Three-dimensional models of aneurysms and their contiguous vasculature generated using the initial scans of patients were used for computational fluid dynamic (CFD) simulations. Both pulsatile and steady flow analyses were performed for each patient. By collating all the prominent hemodynamic indices available in aneurysm literature and developing a few new indices, 25 hemodynamic indices were estimated for each subject. For statistical analysis, it was hypothesized a priori that low wall shear area is different between stable and unstable aneurysms. All other indices were tested in a post-hoc manner. The longitudinal outcome information of these patients was recorded at the clinical centers and the author was blinded until all analyses were complete. Aneurysms that grew during the follow up period were labeled as "grown" and otherwise they were called "stable" by the radiologists. After the hemodynamic analysis was complete, a non-parametric Mann Whitney U test was performed to determine if any index can statistically differentiate the two groups ("grown" versus "stable"). It was found that none of the indices distinguished the two groups with statistical significance. Comparison of the steady and pulsatile flow analysis suggested that the patient population is stratified in the same order by an index, irrespective of whether the index is computed using a steady or pulsatile flow simulation. Pearson correlation coefficient was obtained between basic geometric indices and hemodynamic indices of this population. No strong correlation was found in between morphology and hemodynamics, suggesting uniqueness of the hemodynamic indices. The hypothesis motivating the present study is that aneurysm blood flow based indices can be used as prognostic indicators of growth and/or rupture risk. This study is the first to analyze intracranial aneurysm hemodynamics of a large cohort in a longitudinal prospective manner. Results of the present study indicate that quantitative hemodynamics cannot be used to predict the longitudinal outcome of an aneurysm. Further studies are needed to gain additional clinical insights.

The success of endoluminal stent-graft treatment for abdominal aortic aneurysm relies on maintenance of an effective seal when the stent expands into the healthy artery. Clinical observation of aortic neck dilation following endoluminal grafting has led to the hypothesis that excessive stent expansion forces may cause remodelling and dilation of the artery to accommodate the strong forces. This may lead to failure of the seal, hence so-called endoleak. In this research, we analysed the force field generated by aortic stent-grafts and investigated in vitro approaches for studying the effects of these forces on cells within the vascular wall. The pressure-deformation behaviour of ovine arteries was examined experimentally and was found to vary with artery type. A finite element model of abdominal aorta (AA) characterised by Mooney-Rivlin hyperelastic material properties was validated. The property inputs were derived from the polynomial form of the strain energy density function proposed by Patel and Vaishnav. Stent-artery contact simulations revealed stresses 1.2-19 times higher than within a normal vessel at 120 mmHg when contacted by a zig-zag, square cross-section stent that expanded the AA by 3-16%. Streses 1.3-23 times normal were predicted for circular cross-section stents at the same range of expansions. The stress distribution was determined to be concentrated at the contacting surface and within the inner region of the aortic wall. These results confirmed that the forces within the vessel wall are likely to place unnatural physiological demands on the cells within. We then developed an in vitro system for studying the impact of this mechanical stress on cells within a three dimensional (3D) structure. A 20 wt% poly(vinyl alcohol) (PVA) - 5 wt% collagen tubular construct was developed to support cells, and was shown to sustain physiological blood pressures. Two cell-seeding techniques were examined, direct cell encapsulation and surface cell-seeding. Both demonstrated the capability of entrapping viable cells within the construct that remained viable for up to 4 days. In conclusion, stent contact does create abnormal stress concentrations within the vessel wall with a magnitude severely higher than physiological levels. A feasible tubular construct and an in vitro system were developed, enabling further assessments on the effects of these abnormality on the cells.

Abdominal Aortic Aneurysms (AAA) are amongst the top ten most common cause of death in those over 55 years of age. The disease is usually asymptomatic, often being diagnosed incidentally. Once diagnosed, elective repair of an AAA results in excellent long-term survival with a 3-5% operative mortality. However, up to one half of patients present with ruptured aneurysms, a complication that carries an 80% mortality in the community, and of those reaching hospital, a 50% mortality. Clearly early diagnosis and treatment results in improved survival. Screening for AAA, with ultrasound, would detect aneurysms early, prior to rupture. However, debate continues over the cost effectiveness of population based screening programmes. The identification of a sub-population at a higher risk of developing AAA would increase the yield of a screening prograrmne. A number of populations have been examined, none of which have received international acceptance. About 20% of patients with an AAA have a family history of an aneurysm. The disease is also considered to be a disease of Caucasians, both facts suggesting a strong genetic component to the disease. Perhaps a genetically identified sub-population at a high risk of developing an AAA would prove to be an ideal population for screening. This thesis examines the incidence of aneurysms and the family histories of patients with AAA in the Otago region of New Zealand. Almost twenty percent of the population has a family history of AAA. DNA was collected from each of these patients for genetic analysis. The population was divided into familial AAA and non-familial AAA for the purpose of genetic analysis and compared to a control population. AAA is believed to be a disease of Caucasians; a non-Caucasian population with a low incidence of AAA may prove to be a good control population for genetic studies. A literature review demonstrated a higher incidence of AAA in Caucasians than other ethnic groups and within Caucasians a higher incidence in patients of Northern European origin. The incidence was low in Asian communities, even in studies involving of migrant Asian populations. The New Zealand Maori are believed to have originated from South East Asia, therefore could be expected to have a low incidence of AAA and would make an ideal control population for genetic studies. A pilot study was undertaken to examine the incidence of AAA in the New Zealand Maori. The age standardised incidence of AAA proved to be at least equal in Maori to non-Maori, with a more aggressive form of the disease in Maori, manifesting with a younger age at presentation and a higher incidence of ruptured aneurysms at diagnosis. It is well known that at the time of surgery, an AAA is at the end stage in its life. At this time, inflammation and matrix metalloproteinases (MMP) enzymes are prevalent within the aneurysm wall and have destroyed the wall of the aorta. One of the most important genetic pathways regulating these enzymes is the plasminogen activator inhibiter 1-Tissue plasminogen activator-plasmin pathway. Genetic analysis of this pathway demonstrated an association of the 4G5G polymorphism in the promoter of the PAl-1 gene with familial AAA. In this insertion:deletion polymorphism, the 5G variant binds an activator and repressor, resulting in reduced PAI-1 expression and ultimately increased MMP activation. This allele was associated with familial aneurysms, 47% versus 62% non-familial AAA and 61% controls (p=0.024). A polymorphism within the tissue plasminogen activator gene was also examined and no association was found with AAA. Another way the MMPs expression could be increased is from mutations or polymorphisms in their own genetic structure. Stromelysin 3 is itself a MMP capable of destroying the aortic wall and it has a role in activating other MMPs. A 5A6A insertion:deletion polymorphism exists in the promoter of this gene. The 5A allele variant results in increased stromelysin expression and is associated with AAA 46% versus 33% in controls p=0. 0006. The actions of the MMPs are themselves inhibited by the tissue inhibitors of matrix metalloproteinases. The TIMP genes have been sequenced; two polymorphisms have been identified in the non-coding promoter area of the TIMP 1 gene. Further studies are necessary to examine the effect of these polymorphisms. Inflammation has been implicated in aneurysm progression. One of the roles of the inflammatory cells found in an aneurysm is to deliver the MMP�s to the AAA. The HLA system is integral in controlling this inflammation and was therefore examined. From this series of studies it is concluded that there is a genetic component to AAA. This thesis presents the first genetic polymorphism associated with familial AAA and explores the role of a genetic pathway in the formation of AAA.

Thesis (Ph. D.)--University of Washington, 2001. / Vita. Includes bibliographical references (leaves 158-162).

A retrospective review was made of the case records, angiograms and computed tomography (CT) relating to a total of 263 patients with subarachnoid haemorrhage (SAH) due to ruptured berry aneurysms who were admitted to the Department of Neurosurgery, Wentworth Hospital during the four years 1983-1986. The part of the thesis concerning vasospasm (VS) includes two independent studies on calcium blocker Nimodipine (NO) in the prevention and treatment of VS done by the author. The aim of the thesis is to analyse the management problems of aneurysmal SAH, and investigate factors influencing outcome in order to establish the best possible management policy. The results are discussed and related to the recent data from literature. The main factors influencing outcome were: clinical condition of the patient, the timing of admission and surgery, hypertension and hyperglycaemia on admission, presence of vasospasm and related CT appearance of a thick layer of blood or clot in subarachnoid haemorrhage (CT-Fisher 3). The systemic administration of the calcium blocker nimodipine did not reverse or prevent delayed vasospasm and caused serious adverse effects i.e. hypotension and hyperglycaemia. The results of the thesis suggest a change in management policy and timing of surgery should depend. on clinical condition of the patient on admission (Hunt & Hess grading)(HH I/II grade (HH as possible regardless of timing of admission and results of radiological investigations (CT, angiography). Early surgery (1-3 days) should be the aim of the effort including referral, transport and hospital organisation. III grade (HH surgery should be performed soon after day 10 post-SAH. Particular attention should be paid to the careful preparation and selection of patients for angiography. IV/V grade (HH in specialised units as s000n as possible, preferably neurological or neurosurgical wards, and operated on as soon as their grade improves or, in selected (by surgeon, radiologist and anaesthetist) cases by delayed surgery ( after day 10 post-SAH). / Thesis (M.Med.)-University of Natal, Durban, 1988.

Abdominal Aortic Aneurysms (AAA) are amongst the top ten most common cause of death in those over 55 years of age. The disease is usually asymptomatic, often being diagnosed incidentally. Once diagnosed, elective repair of an AAA results in excellent long-term survival with a 3-5% operative mortality. However, up to one half of patients present with ruptured aneurysms, a complication that carries an 80% mortality in the community, and of those reaching hospital, a 50% mortality. Clearly early diagnosis and treatment results in improved survival. Screening for AAA, with ultrasound, would detect aneurysms early, prior to rupture. However, debate continues over the cost effectiveness of population based screening programmes. The identification of a sub-population at a higher risk of developing AAA would increase the yield of a screening prograrmne. A number of populations have been examined, none of which have received international acceptance. About 20% of patients with an AAA have a family history of an aneurysm. The disease is also considered to be a disease of Caucasians, both facts suggesting a strong genetic component to the disease. Perhaps a genetically identified sub-population at a high risk of developing an AAA would prove to be an ideal population for screening. This thesis examines the incidence of aneurysms and the family histories of patients with AAA in the Otago region of New Zealand. Almost twenty percent of the population has a family history of AAA. DNA was collected from each of these patients for genetic analysis. The population was divided into familial AAA and non-familial AAA for the purpose of genetic analysis and compared to a control population. AAA is believed to be a disease of Caucasians; a non-Caucasian population with a low incidence of AAA may prove to be a good control population for genetic studies. A literature review demonstrated a higher incidence of AAA in Caucasians than other ethnic groups and within Caucasians a higher incidence in patients of Northern European origin. The incidence was low in Asian communities, even in studies involving of migrant Asian populations. The New Zealand Maori are believed to have originated from South East Asia, therefore could be expected to have a low incidence of AAA and would make an ideal control population for genetic studies. A pilot study was undertaken to examine the incidence of AAA in the New Zealand Maori. The age standardised incidence of AAA proved to be at least equal in Maori to non-Maori, with a more aggressive form of the disease in Maori, manifesting with a younger age at presentation and a higher incidence of ruptured aneurysms at diagnosis. It is well known that at the time of surgery, an AAA is at the end stage in its life. At this time, inflammation and matrix metalloproteinases (MMP) enzymes are prevalent within the aneurysm wall and have destroyed the wall of the aorta. One of the most important genetic pathways regulating these enzymes is the plasminogen activator inhibiter 1-Tissue plasminogen activator-plasmin pathway. Genetic analysis of this pathway demonstrated an association of the 4G5G polymorphism in the promoter of the PAl-1 gene with familial AAA. In this insertion:deletion polymorphism, the 5G variant binds an activator and repressor, resulting in reduced PAI-1 expression and ultimately increased MMP activation. This allele was associated with familial aneurysms, 47% versus 62% non-familial AAA and 61% controls (p=0.024). A polymorphism within the tissue plasminogen activator gene was also examined and no association was found with AAA. Another way the MMPs expression could be increased is from mutations or polymorphisms in their own genetic structure. Stromelysin 3 is itself a MMP capable of destroying the aortic wall and it has a role in activating other MMPs. A 5A6A insertion:deletion polymorphism exists in the promoter of this gene. The 5A allele variant results in increased stromelysin expression and is associated with AAA 46% versus 33% in controls p=0. 0006. The actions of the MMPs are themselves inhibited by the tissue inhibitors of matrix metalloproteinases. The TIMP genes have been sequenced; two polymorphisms have been identified in the non-coding promoter area of the TIMP 1 gene. Further studies are necessary to examine the effect of these polymorphisms. Inflammation has been implicated in aneurysm progression. One of the roles of the inflammatory cells found in an aneurysm is to deliver the MMP�s to the AAA. The HLA system is integral in controlling this inflammation and was therefore examined. From this series of studies it is concluded that there is a genetic component to AAA. This thesis presents the first genetic polymorphism associated with familial AAA and explores the role of a genetic pathway in the formation of AAA.

Mémoire (M.Sc.)-- Université du Québec à Chicoutimi, 1992. / Ce mémoire a été réalisé à l'UQAC dans le cadre du programme de maîtrise en médecine expérimentale (volet génétique) extensionné de l'Université Laval à l'UQAC. CaQCU Bibliogr.: f. 71-75. Document électronique également accessible en format PDF. CaQCU