Cette thèse s'intéresse au développement de méthodes automatiques pour la segmentation et la synthèse d'images par résonance magnétique de tumeurs cérébrales. La principale perspective clinique de la segmentation des gliomes est le suivi de la vitesse d'expansion diamétrique dans le but d'adapter les solutions thérapeutiques. A cette fin, la thèse formalise au moyen de modèles graphiques probabilistes des approches de segmentation multi-atlas fondées sur l'appariement de blocs de voxels. Un premier modèle probabiliste prolonge à la segmentation automatique de régions cérébrales pathologiques les approches multi-atlas classiques de segmentation de structures anatomiques. Une approximation de l'étape de marginalisation remplace la notion de fenêtre de recherche locale par un tamisage par atlas et par étiquette. Un modèle de détection de gliomes fondé sur un a priori spatial et des critères de pré-sélection de blocs de voxels permettent d'obtenir des temps de calcul compétitifs malgré un appariement non local. Ce travail est validé et comparé à l'état de l'art sur des bases de données publiques. Un second modèle probabiliste, symétrique au modèle de segmentation, simule des images par résonance magnétique de cas pathologiques, à partir d'une unique segmentation. Une heuristique permet d'estimer le maximum a posteriori et l'incertitude du modèle de synthèse d'image. Un appariement itératif des blocs de voxels renforce la cohérence spatiale des images simulées. Le réalisme des images simulées est évalué avec de vraies IRM et des simulations de l'état de l'art. Le raccordement d'un modèle de croissance de tumeur permet de créer des bases d'images annotées synthétiques. / This thesis focuses on the development of automatic methods for the segmentation and synthesis of brain tumor Magnetic Resonance images. The main clinical perspective of glioma segmentation is growth velocity monitoring for patient therapy management. To this end, the thesis builds on the formalization of multi-atlas patch-based segmentation with probabilistic graphical models. A probabilistic model first extends classical multi-atlas approaches used for the segmentation of healthy brains structures to the automatic segmentation of pathological cerebral regions. An approximation of the marginalization step replaces the concept of local search windows with a stratification with respect to both atlases and labels. A glioma detection model based on a spatially-varying prior and patch pre-selection criteria are introduced to obtain competitive running times despite patch matching being non local. This work is validated and compared to state-of-the-art algorithms on publicly available datasets. A second probabilistic model mirrors the segmentation model in order to synthesize realistic MRI of pathological cases, based on a single label map. A heuristic method allows to solve for the maximum a posteriori and to estimate uncertainty of the image synthesis model. Iterating patch matching reinforces the spatial coherence of synthetic images. The realism of our synthetic images is assessed against real MRI, and against outputs of the state-of-the-art method. The junction of a tumor growth model to the proposed synthesis approach allows to generate databases of annotated synthetic cases.

Insulin-like growth factor binding proteins (IGFBPs) modulate the bioavailability of IGFs in circulation. IGFBPs 1-6 bind IGFs with high affinity and can either potentiate or inhibit IGF signaling in a context dependent manner. IGFBP2 is a 36 kDa protein and the second most abundant IGFBP in serum. Numerous studies in the recent past have implied a pro-tumorigenic role of IGFBP2. Elevated expression of IGFBP2 has been observed in multiple malignancies, including glioblastoma multiforme (GBM), ovarian, pancreatic, gastric, prostate, colon, breast, thyroid cancer and leukemia. In addition, increased expression of IGFBP2 in both tissues and serum of patients has been correlated with poor prognosis in prostate, glioblastoma and colon cancers. Pro-tumorigenic actions of IGFBP2 have been supported by in vitro studies, where IGFBP2 increases the tumorigenic potential of adrenocortical tumor cells, epidermoid carcinoma cells, glioma cells and ovarian cancer cells. Further, using xenograft animal models, the role of IGFBP2 in the progression of glioma has been established. In breast cancer, IGFBP2 was found to be over expressed in ductal carcinoma in situ and invasive breast cancer samples. IGFBP2 over expression has been shown to confer drug resistance and an increased expression has been reported to correlate with lymph node metastasis in T1 breast carcinomas. These reports implicate IGFBP2 in breast cancer biology. However, its role in breast cancer progression is not well defined. With this background, the following objectives were set for the current study: Functional characterization of IGFBP2 with respect to its possible role in breast cancer progression. Elucidation of the molecular mechanisms of IGFBP2 actions. Towards this, immunohistochemistry was performed on 132 invasive ductal carcinoma (IDC) grade III tumors using IGFBP2 specific antibody. It was observed that IGFBP2 expression was significantly higher in tumors in comparison to normal tissues that showed no detectable staining for IGFBP2. It was also observed that expression of IGFBP2 significantly correlated with the expression of ER. To understand the functional significance of IGFBP2 over expression in breast cancer, IGFBP2 was characterized with respect to proliferation, survival and tumor forming ability (in vitro and in vivo) in BT474 breast cancer cells. The knockdown of IGFBP2 expression resulted in suppression of colony formation (nearly 70%) in these breast cancer cells, which could be partially reversed upon exogenous addition of IGFBP2 protein. Proliferation assays using stable clones with knockdown of IGFBP2 in BT474 cells showed a significant decrease in proliferation as compared to vector transfected cells in the presence of serum. Culturing of IGFBP2 knockdown breast cancer cells in serum free medium resulted in their growth arrest in G0/G1 phase of cell cycle as compared to control cells, which progressed through the cell cycle. Prolonged culturing of IGFBP2 knockdown cells in serum free condition (up to 72 h) resulted in the increase of cells in sub G1 phase of the cell cycle. Prolonged depletion of growth factors (serum free conditions) could result in apoptosis of these G1 arrested IGFBP2 knockdown cells. When serum starved IGFBP2 knockdown cells were treated with IGFBP2 protein, the cells arrested in G0/G1 phase were able to progress through the cell cycle and concomitant decrease in sub G1 fraction was observed. Knockdown of IGFBP2 resulted in significantly decreased number and visibly smaller colonies in anchorage independent conditions in vitro. Consistent with this observation, in vivo tumor xenograft formation with IGFBP2 knockdown cells also showed significant reduction in tumor weight as compared to vector generated tumors. These results imply that IGFBP2 has potent growth promoting effects on breast cancer and acts as a mitogen/survival factor for breast cancer cells. To elucidate the molecular mechanisms underlying the pro-tumorigenic effects of IGFBP2, the transcriptome profile following IGFBP2 perturbation in breast cancer cells was determined. IGFBP2 knockdown resulted in significant changes in the expression of genes associated with cellular proliferation and tumorigenicity. The down regulated genes were found to be associated with several events, notably cell cycle, p53 and Wnt signaling, as revealed by Gene Set Enrichment Analysis (GSEA). To further validate these results in breast cancer tissues, whole genome expression analysis was performed in 19 breast tumor samples which were categorized as IGFBP2 positive or negative based on immunohistochemical staining pattern. In comparison to IGFBP2 negative tumors, IGFBP2 positive tumors showed increased expression of genes belonging to MAPK, focal adhesion and Wnt signaling pathway. In order to identify the genes commonly regulated by IGFBP2 in cell lines and tumors, the gene expression profiles of IGFBP2 positive versus IGFBP2 negative tumors and IGFBP2 knockdown breast cancer cells were compared. 347 genes were found to be common among IGFBP2 regulated genes in tumors and cell line. The most significant networks representing the web of interactions among these genes were found to be associated with cellular growth and proliferation, cellular movement and nucleic acid metabolism, indicating an association of IGFBP2 expression phenotype to the distinct changes in expression of genes associated with the regulation of cellular growth and migration. Silencing of IGFBP2 in BT474 cells resulted in a reduced IGF signaling as evidenced by the reduced phosphorylation of IGF1R and concomitantly that of ERK. This effect could be reversed upon addition of the IGFBP2 protein, implying that IGFBP2 potentiates IGF signaling in breast cancer cells. Besides IGF ligand and their receptors, regulation of proliferation associated genes like CENPF, TOP2A, CCND1 and FOXM1 by IGFBP2 was observed, thus providing a molecular basis for the pro-proliferative effects of IGFBP2 on breast cancer cells. Addition of IGFBP2 to immortal breast cells resulted in reduced IGF1R signaling and reduced pERK and pAKT signaling. Additionally, the genes involved in cellular proliferation were down regulated upon IGFBP2 treatment in immortal cells. IGFBP2 knockdown clones had reduced expression of FOXM1, a key regulator of cell cycle for G1/S and G2/M transition, and M phase progression. The regulation of CENPF and CCND1 genes was established following over expression of FOXM1 in IGFBP2 knockdown cells. One of the important and novel finding of this study is the regulation of Wnt signaling pathway genes such as CCND1, MMP7, FGF18, MYCBP, FN1 and survivin by IGFBP2. In support of this, β-catenin protein was found to be regulated by IGFBP2 in breast cancer and GBM cells, as evidenced by knockdown and over expression studies. Furthermore, regulation of β-catenin by IGFBP2 was found to involve integrin-FAK and IGF1R signaling. Another important finding of this study is the correlation of IGFBP2 over expression with elevated β-catenin levels in breast tumors. When expression of both IGFBP2 and β-catenin was correlated with the lymph node status of breast cancers, a significant association of IGFBP2 and β-catenin staining with increased lymph node metastasis was observed in comparison with tumors that did not show staining for either protein. Altogether, in this study employing genomic, cellular and molecular approaches, a pro- tumorigenic role for IGFBP2 in breast cancer has been established. Furthermore, this study provides novel insights into the molecular mechanisms employed by IGFBP2 involving IGF1R, FAK and Wnt signaling pathways during breast cancer progression.

Proton MR spectroscopy is a non-invasive tool for measuring in vivo concentrations of several metabolites. The aim of this thesis was to test its applicability and reliability in neurosurgical praxis. In the first part of the study multiple MR spectroscopy methods were applied in a group of patients after surgery and oncologic treatment for high-grade glioma to test which method performed best in discriminating recurrent tumor from radionecrosis in the presence of a new enhancing lesion. The best diagnostic yield was achieved by comparison of choline, creatine and lactate between lesion and contralateral side (sensitivity 93.3%, specificity 78.6%). Creatine was significantly decreased in patients compared to controls. The inhibiting effect of ongoing oncologic treatment on cerebral and tumoral metabolism makes differential diagnosis trickier. Therefore, a diagnosis of radionecrosis assessed during ongoing radio- and chemotherapy should be confirmed after its completion. In the second part of the study MR spectroscopy data was compared with MR hippocampal volumetry and transcranial doppler examination in a cohort of patients with unilateral occlusion of the internal carotid artery. The N-acetylaspartate/choline ratio and hippocampal volume were significantly lower in both hemispheres of patients...

Patients and methods: Malignant supratentorial glioma (anaplastic astrocytoma, oligoastrocytoma, oligodendroglioma and glioblastoma incl. gliosarcoma), age 16-70y, KPS 50-100. Postoperative randomization to chemotherapy with either BCNU (B) (80 mg/m2 x 3 every 6 weeks) alone or additional VM 26 (V) (50 mg/m2 x 3 every 6 weeks) starting concomitant with radiotherapy. Central histopathological review was required. Primary endpoints were survival time (ST) and progression free survival (PFS) . In addition confirmative analysis of prognostic factors and their interaction with therapy was performed. Results: Eligible: 501 of 522 randomized pts: 82% WHO grade IV gliomas, 18% grade III gliomas. 57% male, mean KPS 74, mean age 50.9 years. The high incidence of lung toxicity – with a cumulative risk of 19% during the first year - was alarming. Survival was not significantly different ( median 50.3 (B) versus 52.4 (V) (weeks), but an increase in long term survivors was observed (18 months: 29% B, 34% V, 5 years 5% B, 12% V) and PFS showed a significant difference with a median of 31.4 (B) versus 34.3 (V) weeks. Qualitative interaction between KPS and therapy (p < 0.01) was demonstrated: pts with a KPS ≥ 70 benefited from additional VM26, those with reduced KPS < 70 did better with BCNU-monotherapy. Conclusion: Adding VM26 to BCNU is effective in the chemotherapy of malignant gliomas. Because of the demonstrated interaction with therapy performance status, not tumor grade is the crucial factor to determine application and aggressiveness of chemotherapy. With risk adapted therapy a significant proportion of patients even with glioblastoma survive for years in good general condition. BCNU should be replaced by an equipotent alkylans to avoid the unacceptable high rate of lung toxicity.

Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG-geförderten) Allianz- bzw. Nationallizenz frei zugänglich.

Brain tumors represent a diverse spectrum of cancer types which can induce grave complications and lead to poor life expectancy. Amongst the various brain tumor types, gliomas are primary brain tumors that compose about 30% of adult brain tumors. They are graded according to the World Health Organization into Grades 1 to 4 (G1-G4), where G4 is the highest grade with the highest malignancy and poor prognosis. Early diagnosis and classification of brain tumor grade is very important since it can improve the treatment procedure and (potentially) prolong a patient's life, since life expectancy largely depends on the level of malignancy and the tumor's histological characteristics. While clinicians have diagnostic tools they use as a gold standard, such as biopsies these are either invasive or costly. A widely used example of a non-invasive technique is magnetic resonance imaging, due to its ability to produce images with different soft-tissue contrast and high spatial resolution thanks to multiple imaging sequences. However, the examination of such images can be overwhelming for radiologists due to the overall large amount of data. Deep learning approaches, on the other hand, have shown great potential in brain tumor diagnosis and can assist radiologists in the decision-making process. In this thesis, brain tumor grade classification in MR images is performed using deep learning. Two popular pre-trained CNN models (VGG-19, ResNet50) were employed using single MR modalities and combinations of them to classify gliomas into three grades. All models were trained using data augmentation on 2D images from the TCGA dataset, which consisted of 3D volumes from 142 anonymized patients. The models were evaluated based on accuracy, precision, recall, F1-score, AUC score, as well as the Wilcoxon Signed-Rank test to establish if one classifier was statistically significantly better than the other. Since deep learning models are typically 'black box' models and can be difficult to interpret by non-experts, Gradient-weighted Class Activation Mapping (Grad-CAM) was used in order to address model explainability. For single modalities, VGG-19 displayed the highest performance with a test accuracy of 77.86%, whilst for combinations of two and three modalities T1ce, FLAIR and T2, T1ce, FLAIR were the best performing ones for VGG-19 with a test accuracy of 74.48%, 75.78%, respectively. Statistical comparisons indicated that for single MR modalities and combinations of two MR modalities, there was not a statistically significant difference between the two classifiers, whilst for combination of three modalities, one model was better than the other. However, given the small size of the test population, these comparisons have low statistical power. The use of Grad-CAM for model explainability indicated that ResNet50 was able to localize the tumor region better than VGG-19.

Le cancer du cerveau est associé à une morbidité importante et à un fardeau économique considérable pour les systèmes de santé, les patients et leur famille. Malheureusement, on en sait toujours très peu sur l’étiologie de cette maladie. Les métaux, les métalloïdes et les fumées de soudures constituent une grande famille de cancérogènes professionnels potentiels à laquelle des millions de travailleurs sont exposés. La littérature scientifique fournit certains éléments de preuve que l’exposition professionnelle à quelques composés métalliques pourrait augmenter le risque de cancer du cerveau, mais la plupart des études publiées étaient limitées dans leur taille d’échantillons et en leurs capacités de mesurer efficacement l’exposition professionnelle à vie. Cette thèse a pour objectif de fournir de nouveaux éléments de preuve concernant l’association entre l’exposition professionnelle à certains composés métalliques et les deux principaux sous-types histologiques du cancer du cerveau, le gliome et le méningiome. Deux projets existants constituent la base de cette thèse: INTEROCC, une grande étude internationale cas-témoins sur l’association entre l’exposition professionnelle et le cancer du cerveau, incluant 2 054 cas de gliome, 1 924 cas de méningiome et 5 601 témoins, ainsi que CANJEM, une nouvelle matrice emplois-exposition basée sur plus de 30 000 emplois. CANJEM est un tableau croisé de trois axes: un axe de codes professionnels, un axe de périodes de temps et un axe d’agents chimiques. CANJEM fournit diverses mesures d’exposition à des agents professionnels sélectionnés en fonction d’un titre occupationnel et d’une période de temps. CANJEM étant un outil complexe conçu pour offrir une flexibilité considérable à l’utilisateur, les deux premiers volets de cette thèse ont été consacrés à l’examen de certaines des considérations méthodologiques associées à l’utilisation de CANJEM dans le cadre d’une étude épidémiologique. Premièrement, nous avons examiné comment la modification de la résolution des axes de codes professionnels et de périodes de temps influençait la proportion d’emplois pouvant être liés à CANJEM dans l’étude INTEROCC. Nous avons ensuite comparé l’accord de paires de versions de CANJEM pour la probabilité d’exposition et la concentration pondérée par la fréquence d’exposition de 19 composés métalliques en utilisant le coefficient d’accord de Gwet (AC2). Nous avons observé que, selon la résolution utilisée, CANJEM pouvait lier entre 70,7% et 98,1% de l’ensemble des emplois disponibles dans l’étude INTEROCC. De plus, la modification de l’axe de code professionnel avait un impact plus important que la modification de l’axe de période de temps sur les mesures d’expositions. Deuxièmement, l’évaluation par des experts est généralement considérée comme l’étalon-or dans l’évaluation rétrospective de l’exposition professionnelle. Différents seuils peuvent être appliqués à la probabilité d’exposition fournie par CANJEM afin de distinguer «exposé» de «non exposé». Nous avons comparé les rapports de cotes (RC) obtenus à l’aide de plusieurs versions de variables d’exposition binaire et cumulative pour neuf cancérogènes potentiels du poumon avec des RC obtenus à l’aide de l’évaluation par des experts. Des modèles de régression logistique inconditionnels ont été utilisés pour examiner l’association entre chaque variable d’exposition et le cancer du poumon chez 1 200 cas de cancer du poumon et 1 505 témoin issus d’une étude cas-témoin basée à Montréal. La sensibilité de l’évaluation dérivée de CANJEM par rapport à l’évaluation par experts variait de 0,12 à 0,78, tandis que la spécificité variait de 0,84 à 0,99. Dans l’ensemble, CANJEM a été capable reproduire les associations obtenues avec l’évaluation par experts, l’utilisation de seuils de probabilité de 25% ou 50% fournissant généralement les meilleurs résultats. Finalement, nous avons examiné le lien entre l’exposition professionnelle à 21 composés métalliques et le gliome ainsi que le méningiome dans l’étude INTEROCC à l’aide de régressions logistiques conditionnelles. La stratégie analytique était basée sur les observations faites dans les deux premiers volets. Nous n’avons observé aucune preuve de la présence d’association entre les agents sélectionnés et le gliome, mais la présence d’associations positives entre ces agents et le méningiome a été suggérée. Des associations statistiquement significatives ont également été observées entre le méningiome et une exposition inférieure à 15 ans aux fumées de plomb (RC (intervalle de confiance de 95%)) (1,67 (1,02-2,74)), aux composés du zinc (2,14 (1,02-3,89)), aux fumées de soudure (1,80 (1,17-2,77)), aux fumées d’oxydes métalliques (1,51 (1,03-2,21)) et entre une faible exposition cumulée au chrome VI (1,99 (1,03-3,84)) et aux fumées de brasage (1,83 (1,17-2,87)). L’évaluation rétrospective de l’exposition constitue l’un des principaux défis de l’épidémiologie professionnelle. Dans cette thèse, nous avons constaté que CANJEM, bien qu’imparfaite, était une approche appropriée pour l’évaluation de l’exposition professionnelle dans les études épidémiologiques. Bien qu’il soit difficile de déterminer le rôle exact joué par chacun des agents examinés, nos résultats supportent la présence d’une association positive entre les composés métalliques et plus particulièrement les fumées métalliques et le méningiome. / Brain cancer is associated with substantial lifelong morbidity and considerable economic burden for public health systems, patients, and their families. Very little is known regarding the etiology of this disease. Metals, metalloids, and welding fumes are a large family of potential occupational carcinogens to which millions of workers are exposed. The literature provides some evidence that occupational exposure to a few metallic compounds could increase the risk of brain cancer, but most published studies were limited in sample size and ability to effectively measure lifetime occupational exposure. In this thesis, we aimed to provide new evidence concerning the association between occupational exposure to selected metallic compounds and glioma and meningioma, the two major histological subtypes of brain cancer. Two existing projects provided the basis for the thesis: INTEROCC, a large international pooled case-control study on the association between occupational exposures and brain cancer, including 2,054 glioma cases, 1,924 meningioma cases, and 5,601 controls; CANJEM a new job exposure matrix based on the expert assessment of > 30,000 jobs. CANJEM is a cross-tabulation of three axes: an occupation code axis, a time period axis, and a chemical agent axis that provides various metrics of exposure to selected occupational agents based on a job title and a time period. However, CANJEM is also a complex tool designed to offer considerable flexibility to the user. The first two components of this thesis focused on the examination of some of the methodological considerations associated with the use of CANJEM in the context of an epidemiological study. First we examined how changing the resolution of the occupational code and time period axes, affected the proportion of jobs in the INTEROCC study that could be linked to CANJEM. We then compared the agreement among pairs of versions of CANJEM for the probability and frequency weighted concentration of exposure to 19 metallic compounds using Gwet’s agreement coefficient (AC2). We observed that, depending on the resolution used, CANJEM could be linked to 70.7% to 98.1% of all jobs available in the INTEROCC study. Furthermore, we observed that varying the occupation code axis had a greater impact than varying the time period axis. Neither the metrics of exposure nor the linkage rate were strongly affected by other aspects of CANJEM examined. Second, expert assessment is usually considered the gold standard in retrospective occupational exposure assessment. Different cutpoints can be applied to the probability of exposure provided by CANJEM to distinguish “exposed” from “unexposed”. We compared odds ratios (ORs) obtained using multiple versions of a binary ever and a cumulative exposure variable for nine potential and known lung carcinogens with ORs obtained using expert assessment. Unconditional logistic regression models adjusted for potential confounders were used to examine the association between each exposure variable and lung cancer in 1,200 lung cancer cases and 1,505 controls from a Montreal based case-control study. Sensitivity of the CANJEM-derived assessment vs. the expert assessment ranged from 0.12 to 0.78 while Specificity ranged from 0.84 to 0.99. Overall, CANJEM was fairly successful in reproducing the associations obtained with the expert assessment method, with the use of probability thresholds of 25% or 50% generally providing the best results for both exposure variables. Finally, we examined the association between occupational exposure to 21 metallic compound and glioma and meningioma in the INTEROCC study using conditional logistic regression adjusted for potential confounders. The analytical strategy was based on the observations made in the two previous components. We observed no evidence of association between the selected agents and glioma, but there was evidence of positive associations between some of the agents and meningioma. Statistically significant associations with OR (95% confidence interval) were also observed between < 15 years of exposure to lead fumes (1.67 (1.02-2.74)), zinc compounds (2.14 (1.02-3.89)), soldering fumes (1.80 (1.17-2.77)), and metal oxide fumes (1.51 (1.03-2.21)) and low cumulative exposure to chromium VI (1.99 (1.03-3.84)) and soldering fumes (1.83 (1.17-2.87)) and meningioma. One of the main challenges in occupational cancer epidemiology is retrospective exposure assessment. In this thesis we found that, while imperfect, CANJEM was a cost-efficient approach to occupational exposure in epidemiological studies. Although it is difficult to determine the exact role played by individual agents examined, our results provide some support for the presence of a positive association between metallic compounds, and more particularly metallic fumes, and meningioma.

No description available.

No description available.

Tesis por compendio / [ES] El cáncer constituye la segunda causa de muerte en España. El tromboembolismo venoso (TEV), una complicación del cáncer, conlleva gran gasto del presupuesto sanitario y representa la segunda causa de muerte en estos pacientes. Sin embargo, las herramientas actuales disponibles para la identificación de pacientes oncológicos con elevado riesgo trombótico son limitadas. Adicionalmente, no existen métodos simples, mínimamente invasivos y económicos de diagnóstico de cáncer vesical. Por este motivo, se utilizan técnicas dañinas como la tomografía computarizada la cual implica una elevada dosis de exposición a radiación y procedimientos invasivos como la cistoscopia. Además, un estado hipercoagulable parece tener una relación directa con una mayor carga tumoral y un peor pronóstico. El objetivo principal de la presente Tesis Doctoral es explorar la utilidad clínica de nuevos métodos diagnósticos y pronósticos para el cáncer y sus complicaciones trombóticas. En la primera parte de la Tesis, nos hemos centrado en el papel de miRNAs en orina como biomarcadores de cáncer vesical. Hemos identificado al miR-29c-3p como el miRNA más estable por lo que fue utilizado como normalizador. Hemos ajustado un modelo de regresión logística ordinal para el diagnóstico y estratificación de cáncer vesical utilizando la expresión de miRNAs en orina de pacientes y controles. Este modelo incluyó la expresión de 7 miRNAs: miR-221-3p, miR-93-5p, miR-362-3p, miR-191-5p, miR-200c-3p, miR-192-5p y miR-21-5p. En la segunda parte de la Tesis, nos centramos en el estudio de nuevos biomarcadores para la trombosis asociada a cáncer. Analizamos el potencial predictivo de los miRNAs y de marcadores de activación de neutrófilos en pacientes con cáncer pancreático y pacientes con glioma y meningioma. En cáncer pancreático, obtuvimos un perfil de 7 miRNAs (miR-486-5p, miR-106b-5p, let-7i-5p, let-7g-5p, miR-144-3p, miR-19a-3p y miR-103a-3p) capaz de estimar el riesgo de TEV al diagnóstico con dianas incluidas en las rutas pancreatic cancer y complement and coagulation cascades. En el estudio de los marcadores de activación de neutrófilos, obtuvimos un nuevo modelo predictivo de TEV con la calprotectina como variable predictora. Respecto al estudio de trombosis asociada a cáncer en tumores intracraneales, en pacientes con glioma, ajustamos y validamos un modelo predictivo de embolismo pulmonar (EP) postquirúrgico con 6 miRNAs: miR-363-3p, miR-93-3p, miR-22-5p, miR-451a, miR-222-3p y miR-140-3p y otro con cfDNA y mieloperoxidasa como predictores. Además, hemos combinado los dos tipos de marcadores y hemos obtenido un modelo con mayor capacidad predictiva que incluye a miR-140-3p y a la mieloperoxidasa como predictores. En pacientes con meningioma, ajustamos y validamos un modelo predictivo de EP postquirúrgico con 6 miRNAs: miR-29a-3p, miR-660-5p, miR-331-3p, miR-126-5p, miR-23a-3p y miR-23b-3p. En conclusión, proponemos diferentes perfiles de biomarcadores para el diagnóstico de cáncer de vejiga y para la identificación de pacientes oncológicos con elevado riesgo de trombosis. / [CA] El càncer constitueix la segona causa de mort a Espanya. El tromboembolisme venós (TEV), una complicació del càncer, representa la segona causa de mort en aquests pacients i comporta una gran despesa sanitària. No obstant això, les eines disponibles actualment per a la identificació de pacients oncològics amb elevat risc trombòtic són limitades. Actualment, no existeixen mètodes diagnòstics per al càncer de bufeta senzills, mínimament invasius i econòmics. Per aquest motiu, s'utilitzen tècniques nocives com la tomografia computada la qual implica una elevada dosi d'exposició a radiació i procediments invasius com la cistoscòpia. A més, un estat hipercoagulable sembla tindre una relació directa amb una major càrrega tumoral i un pitjor pronòstic. L'objectiu principal de la present Tesi Doctoral fou explorar la utilitat clínica de nous mètodes diagnòstics i pronòstics per al càncer i les seues complicacions trombòtiques. En la primera part de la Tesi, ens hem centrat en el paper dels microRNAs (miRNAs) en orina com biomarcadors de càncer de bufeta. Hem identificat al miR-29c-3p com el miRNA més estable per la qual cosa va ser utilitzat com a normalitzador. Hem ajustat un model de regressió logística ordinal per al diagnòstic i estratificació de càncer de bufeta utilitzant l'expressió de miRNAs en orina de pacients i controls. Aquest model va incloure l'expressió de 7 miRNAs: miR-221-3p, miR-93-5p, miR-362-3p, miR-191-5p, miR-200c-3p, miR-192-5p i miR-21-5p. En la segona part de la Tesi, ens centràrem en l'estudi de nous biomarcadors per a la trombosi associada a càncer. Analitzàrem el potencial predictiu dels miRNAs i de marcadors d'activació de neutròfils en pacients amb càncer pancreàtic i pacients amb glioma i meningioma. En càncer pancreàtic, vàrem obtindre un perfil de 7 miRNAs (miR-486-5p, miR-106b-5p, let-7i-5p, let-7g-5p, miR-144-3p, miR-19a-3p i miR-103a-3p) capaç d'estimar el risc de TEV al diagnostic dels pacients els quals tenen dianes incloses en les rutes biològiques pancreatic cancer y complement and coagulation cascades. En el estudi dels marcadors d¿activació de neutròfils, vàrem obtenir un altre model predictiu de TEV amb la calprotectina com a variable predictora. Respecte a l'estudi de trombosi associada a càncer en tumors intracranials, en pacients amb glioma, ajustàrem i validàrem un model predictiu d'embolisme pulmonar (EP) incidental postquirúrgic amb 6 miRNAs (miR-363-3p, miR-93-3p, miR-22-5p, miR-451a, miR-222-3p i miR-140-3p) i un altre amb cfDNA i mieloperoxidasa com a predictors. A més, vàrem combinar els dos tipus de marcadors i vàrem obtenir un model amb major capacitat predictiva que inclou al miR-140-3p i la mieloperoxidasa com a predictors. En pacients amb meningioma, ajustàrem i validàrem un model predictiu d¿EP incidental postquirúrgic amb 6 miRNAs: miR-29a-3p, miR-660-5p, miR-331-3p, miR-126-5p, miR-23a-3p i miR-23b-3p. En conclusió, proposem diferents perfils de biomarcadors per al diagnòstic de càncer de bufeta i per a la identificació de pacients oncològics amb elevat risc de trombosi. / [EN] Cancer is the second leading cause of death in Spain. Collaterally, venous thromboembolism (VTE), as a complication of cancer, consumes a great part of its healthcare budget and, more importantly, it is the second cause of death in these patients. However, limited tools are available to identify high risk patients. Additionally, a simple, minimally invasive and economical diagnostic methods for bladder cancer are also lacking. For that aim, harmful techniques are used like CT scan with high radiation exposure and invasive procedures like cystoscopy. Moreover, a hypercoagulable state seems directly related to a large tumor burden and poor prognosis. The overall aim of this Doctoral Thesis is to explore the clinical utility of novel diagnostic and prognostic methods for cancer and its thrombotic complications. In the first part of this Doctoral Thesis, we focused on the role of urine miRNAs as bladder cancer biomarkers. We identified miR-29c-3p as the most stable miRNA and was therefore used as normalizer. We adjusted an ordinal logistic regression model for the diagnosis and stratification of BC using the urine miRNA expression levels of patients and controls. This model included 7 miRNAs: miR-221-3p, miR-93-5p, miR-362-3p, miR-191-5p, miR-200c-3p, miR-192-5p and miR-21-5p. In the second part of this Doctoral Thesis, we focused on the study of novel biomarkers for cancer-associated thrombosis. We analyzed the predictive potential of miRNAs and neutrophil activation markers of thrombotic events in patients with pancreatic cancer and patients with glioma and meningioma. In pancreatic cancer, we obtained a profile of 7 miRNAs (miR-486-5p, miR-106b-5p, let-7i-5p, let-7g-5p, miR-144-3p, miR-19a-3p and miR-103a-3p) able to estimate the risk of potential VTE at diagnosis with targets involved in the pancreatic cancer and complement and coagulation cascades pathways. In the study of the neutrophil activation makers, we obtained a new predictive model of VTE with calprotectin as predictor. Regarding the study of cancer-associated thrombosis in intracranial tumors, in glioma patients, we adjusted and validated a predictive model for post-surgical pulmonary embolism (PE) with 6 miRNAs: miR-363-3p, miR-93-3p, miR-22-5p, miR-451a, miR-222-3p and miR-140-3p, and another with cfDNA and myeloperoxidase as predictors. Furthermore, we combined both types of biomarkers and obtained an improved model using myeloperoxidase and miR-140-3p as predictors. In meningioma patients we fitted and validated a predictive model with 6 miRNAs: miR-29a-3p, miR-660-5p, miR-331-3p, miR-126-5p, miR-23a-3p and miR-23b-3p. In conclusion, we propose several profiles of biomarkers for the diagnosis of bladder cancer and for the identification of oncologic patients at high risk of suffering a thrombotic event. / Oto Martínez, AJ. (2022). Novel Diagnostic and Prognostic Methods for Cancer and Cancer Associated Thrombosis [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/181510 / Compendio