Glioblastoma multiforme (GBM) is a debilitating malignant brain tumor with expected patient survival of less than a year and limited responsiveness to most treatments, often requiring biopsy for diagnosis and invasive surgery for treatment. We propose a DNA computer system, consisting of input, computation, and output components, for diagnosis and treatment. The input component will detect the presence of three GBM biomarkers: vascular endothelial growth factor (VEGF), caveolin-1α (CAV), and B2 receptors. The computation component will include indicator segments for each of these genes, and ensure that output is only released if all the biomarkers are present. The output component will consist of the therapeutic agent interleukin-12 (IL-12). This study will designate four groups of animals: untreated tumor-free (control), tumor-inoculated (RG2), treated and tumor-free (DNA), and treated and tumor-inoculated (RG2/DNA). In the RG2 and RG2/DNA groups, we will inoculate adult male Fischer rats with RG2 cells into the striatum to induce tumor growth. Rats in the DNA and RG2/DNA groups will be implanted with the DNA system at the same location via recombinant adeno- associated viral vectors. The effectiveness of the DNA system will be evaluated through tumor size measurements, collected from brain slices stained with hematoxylin and eosin, and survival curve. Additionally, IL-12 localization will confirm the release of the output component. We anticipate that the DNA treatment will result in a decrease in tumor size, leading to smaller tumor size in the RG2/DNA group versus the RG2 group. The control group is expected to survive the longest, followed by the DNA group, then the RG2/DNA group, and finally the RG2 group. In the DNA group, IL-12 is expected to stay localized to the implantation site, remaining in its unreleased stem-loop form. On the other hand, it is expected to be released and active in the RG2/DNA group. This study provides a proof of concept to demonstrate the viability and effectiveness of a DNA system using VEGF, CAV, and B2 receptors as biomarkers and IL-12 as a therapeutic output component in the RG2 model. Further research may include varying several of the parameters used in this study, including amount of RG2 administered, choice of biomarkers, quantity and choice of output component, and choice of animal model. This system provides a promising and innovative new approach that is less invasive than surgery yet is still effective in diagnosing, targeting, and treating GBM.

Η εισαγωγή της Μαγνητικής Τομογραφίας (ΜΤ) στην κλινική πρακτική και η συμπληρωματική πληροφορία που δίνει η Φασματοσκοπία Μαγνητικού Συντονισμού (ΦΜΣ) συνιστά μια από τις πιο σημαντικές εξελίξεις στη διάγνωση ασθενών με καρκίνο εγκεφάλου [1]. Παρ’ όλα αυτά, οι εικόνες ΜΤ είναι συχνά δύσκολο να ερμηνευθούν από τους ειδικούς λόγω [2] α/ της υποκειμενικότητας και περιορισμένης εμπειρίας του παρατηρητή στην εκτίμηση εικόνων που παράγει η σχετικά νέα αυτή τεχνολογία, β/ των ποικίλων κλινικών χαρακτηριστικών των όγκων (π.χ. τύπος, διαβάθμιση κακοήθειας κλπ.) και γ/ της ιδιαιτερότητας των όγκων στην αντίθεση που παρουσιάζουν με τον περιβάλλοντα ιστό. Μόνο λιγοστές μελέτες έχουν διεξαχθεί για να χαρακτηρίσουν ιστούς εγκεφάλου μέσω της ανάλυσης ποσοτικών χαρακτηριστικών από εικόνες εγκεφάλου ΜΤ [3, 4]. Ενώ έχει ήδη τονιστεί η αναγκαιότητα συσχετισμού της διαγνωστικής και προγνωστικής πληροφορίας που προέρχεται από εικόνες ΜΤ και σήματα ΦΜΣ στη διεθνή βιβλιογραφία [5], υπάρχουν λιγοστές ανάλογες αναφορές για τον σχεδιασμό και υλοποίηση συστήματος Η/Υ αυτόματης διάγνωσης όγκων εγκεφάλου κάνοντας συνδυασμό ποσοτικής πληροφορίας προερχόμενης από εικόνες ΜΤ και σήματα ΦΜΣ [6, 7]. Οι στόχοι της παρούσας διατριβής εστιάζονται στα παρακάτω: - στη μελέτη, ανάπτυξη και η υλοποίηση υπολογιστικού συστήματος αυτόματης ταξινόμησης όγκων του εγκεφάλου μέσω της ποσοτικής ανάλυσης εικόνων ΜΤ το οποίο θα βελτιώνει την ακρίβεια ταξινόμησης σε σχέση με ήδη υπάρχοντα συστήματα [4, 8, 9], όπως αυτά περιγράφονται στην διεθνή βιβλιογραφία μεταξύ πρωτογενών και δευτερογενών όγκων εγκεφάλου καθώς και μεταξύ γλοιωμάτων και μηνιγγιωμάτων με την χρήση δέντρου ιεραρχικής απόφασης δύο επιπέδων. Επιπλέον, στην ανάδειξη πως η χρήση ενός μη-γραμμικού πολυωνυμικού μετασχηματισμού ελάχιστων τετραγώνων των χαρακτηριστικών υφής έχει ως αποτέλεσμα την βελτίωση της ακρίβειας ταξινόμησης του ταξινομητή πιθανοκρατικού νευρωνικού δικτύου. - στην επέκταση και την βελτίωση του συστήματος αυτόματης ταξινόμησης όγκων του εγκεφάλου χρησιμοποιώντας α/ ογκομετρικές ποσοτικές παραμέτρους εικόνων ΜΤ, β/ ταξινομητή μηχανών διανυσμάτων στήριξης μαζί με τη μεθοδολογία συνάθροισης αποτελεσμάτων ταξινόμησης από τυχαιοποιημένα δείγματα κατηγοριών δημιουργημένων με επαναδειγματοληψία για κάθε κόμβο δέντρου ιεραρχικής απόφασης δύο επιπέδων όπου στο πρώτο επίπεδο πραγματοποιήθηκε διαχωρισμός μεταξύ πρωτογενών και δευτερογενών όγκων εγκεφάλου και στο δεύτερο και μεταξύ γλοιωμάτων και μηνιγγιωμάτων και γ/ έναν τροποποιημένο πυρήνα ακτινικής συνάρτησης βάσης για τον ταξινομητή μηχανών διανυσμάτων στήριξης ο οποίος περιλαμβάνει την τεχνική μη-γραμμικού πολυωνυμικού μετασχηματισμού ελάχιστων τετραγώνων με στόχο την βελτίωση της ακρίβειας ταξινόμησης. - στην περαιτέρω επέκταση και την βελτίωση του συστήματος αυτόματης ταξινόμησης με την εισαγωγή χαρακτηριστικών προερχόμενων από σήματα ΦΜΣ ώστε να διερευνηθεί εάν η χρήση του μπορεί να βελτιώσει τα αποτελέσματα ταξινόμησης μεταξύ μηνιγγιωμάτων και μονήρων μεταστάσεων. Τέλος κάνοντας μια περίληψη, η παρούσα διατριβή διαπραγματεύεται τον σχεδιασμό, ανάπτυξη και υλοποίηση μεθόδων και αλγορίθμων για την επεξεργασία και ανάλυση ιατρικών εικόνων, επικεντρώνοντας ειδικότερα στην εφαρμογή των μεθόδων αυτών για την διάγνωση του τύπου των όγκων εγκεφάλου. Τα πιο βασικά συμπεράσματα που απορρέουν από την παρούσα διατριβή είναι τα ακόλουθα: α/ Το σύστημα ταξινόμησης των τύπων των όγκων εγκεφάλου που σχεδιάστηκε και υλοποιήθηκε αυξάνει τα ποσοστά ορθής ταξινόμησης σε σχέση με τα υπάρχοντα. β/ Η κωδικοποίηση των ιδιοτήτων της υφής που προέρχεται από τον σύνολο του όγκου παρέχει επιπρόσθετη πληροφορία στο σύστημα ταξινόμησης αυξάνοντας τα ποσοστά επιτυχούς διαχωρισμού. γ/ Τα χαρακτηριστικά φασματοσκοπίας μαγνητικού συντονισμού αποτελούν επιπρόσθετη αξία στο χαρακτηρισμό του τύπου των όγκων εγκεφάλου μιας και οδήγησαν στην αύξηση του ποσοστού επιτυχούς διαχωρισμού του συστήματος ταξινόμησης. / The process of brain tumor characterization requires a rather intricate assessment of the various Magnetic Resonance (MR) image and spectroscopic features and is typically performed by experienced radiologists. Despite the inherently subjective nature of many of the decisions associated with this process, an expert radiologist is able to perform this task with a significant degree of precision and accuracy. However, in the effort to deliver more effective treatment, clinicians are continuously seeking for greater accuracy in the pathological characterization of brain tissues. The aim of the present thesis was to design, implement, and evaluate a software classification system for discriminating between different brain tumor types on Magnetic Resonance Imaging (MRI), employing textural and spectroscopic features. The clinical material consisted of sixty seven T1-weighted post-contrast MR brain images (21 metastases, 19 meningiomas, and 27 gliomas), obtained from patients with verified and untreated intracranial tumors. Thirty-six 2-dimensional textural features (2D), from the image histogram and the co-occurrence and run-length matrices, were extracted from each one of 67 MR-images. Similarly, an equal number of 3-dimensional textural features (3D) were also calculated in the attempt to maximize classification performances. Finally, MR-spectroscopy features were also incorporated for improving classification accuracies. Classification methods employed included i/ a modified Probabilistic Neural Network (PNN) and Support Vector Machines (SVM) algorithms, incorporating a non-linear Least Squares Features Transformation (LSFT) into the classifiers and ii/ an ensemble classification scheme employing the LSFT-SVM classifier. The LSFT improved classifiers’ performances, increased class separability, and resulted in dimensionality reduction. For evaluating the performance of the designed classification schemes, evaluations were performed by means of the external cross validation process, which is considered indicative of the generalization performance of the designed classification system to ‘unseen’ cases. It was found that the LSFT features transformation enhanced the performance of the PNN and SVM algorithms, achieving classification accuracies of 73.48 % in distinguishing metastatic from primary tumors and 88.67% in discriminating gliomas from meningiomas. When volumetric 3-dimensional features were employed, these results improved to 88.18% for discriminating between metastatic and primary tumors and 97.33% for distinguishing gliomas from meningiomas. The textural features employed in the design of the optimum classification scheme were associated primarily with image texture homogeneity. Finally, when MR-spectroscopy features were also incorporated, classification accuracy was boosted up from 95% in discriminating meningiomas from metastasis to 100%. The MR-image features that participated in the optimum feature vector were related to the degree of homogeneity, the amount of randomness and the dispersion of the gray-tone intensity values within the texture of the tumor. These textural characteristics are related to textural parameters that physicians employ in diagnosis and they were proportional to the textural imprint of brain tumors, i.e. gliomas have heterogeneous texture while meningiomas appear to be homogeneous in MR imaging. The MR-spectroscopy feature that participated in the optimum feature vector was the Choline (Cho) / N-Acetyl Aspartate (NAA) metabolite integral ratio. It was found that both meningiomas and metastases are characterized from low concentrations of NAA while meningiomas exhibit higher concentrations of Cho than metastases, which could be attributed to increased synthesis of tumor cell membranes. Finally, the proposed system might be of value as an assisting tool for brain tumor characterization on volumetric MRI series.

Meduloblastoma é o tumor intracranial mais comum em crianças, provavelmente derivado de células precursoras da camada granular externa do cerebelo durante seu desenvolvimento. O tratamento padrão consiste em cirurgia, radioterapia e quimioterapia, que produzem graves sequelas nos pacientes e garantem uma sobrevida baixa, o que demonstra a necessidade de novas alternativas terapêuticas para a doença. Evidências demonstram que o receptor do peptídeo liberador de gastrina (GRPR) está superexpresso em diversos tumores humanos, assim como seu agonista (GRP) pode atuar como um fator de crescimento autócrino em tumores cerebrais. No presente estudo, avaliamos a expressão de GRPR e o efeito de seus agonistas, bombesina (BB) e GRP, além do antagonista RC-3095, sobre a viabilidade celular de linhagens de meduloblastoma humano DAOY, D283 e ONS76. Mostramos que meduloblastomas, apesar de expressarem GRPR, não têm sua viabilidade celular afetada por agonistas e antagonista desse receptor. Uma vez que há evidências de que BDNF (fator neurotrófico derivado de cérebro) esteja relacionado à diferenciação celular em meduloblastomas, também avaliamos o efeito de BDNF sobre a viabilidade celular das linhagens de meduloblastoma humano. As linhagens DAOY e D283 tiveram sua viabilidade celular reduzida pela presença de BDNF. Uma vez que a via da PKA tem sido implicada na iniciação e progressão de vários tumores, também avaliamos o efeito de rolipram, um inibidor de fosfodiesterase tipo IV, sobre a viabilidade celular das linhagens de meduloblastoma humano, sendo que rolipram reduziu a viabilidade celular de todas as linhagens estudadas. Os receptores de BDNF e a via da PKA podem, portanto, ser alvos moleculares promissores para o desenvolvimento de novas terapias para meduloblastomas. / Medulloblastoma is the most common intracranial tumor in children and is believed to arise from the precursor cells of the external granule layer of the developing cerebellum. The standard treatment, consisting of surgery, craniospinal radiotherapy and chemotherapy, produces severe sequelae in patients and provides a poor overall survival, indicating the need for new therapeutic alternatives for treating this disease. Evidences show that the gastrin releasing peptide receptor (GRPR) is overexpressed in various human tumors and its agonist (GRP) can act as an autocrine growth factor in brain tumors. In the present study, we evaluated GRPR expression, as well as the effect of its agonists, bombesin (BB) and GRP, and its antagonist RC-3095, over cell viability of the human medulloblastoma cell lines DAOY, D283 and ONS76. We found that medulloblastomas, in spite of expressing GRPR, do not have its viability affected by the presence of agonists and antagonist of this receptor. Since there are evidences that BDNF (brain-derived neurotrophic factor) is related to cell differentiation in medulloblastomas, we also evaluated the effect of BDNF over the viability of medulloblastoma cell lines. The viability of the cell lines DAOY and D283 was reduced by the presence of BDNF. Since the PKA pathway has been implicated in the initiation and progression of various tumors, we also evaluated the effect of rolipram, a phosphodiesterase IV inhibitor, over the viability of the same medulloblastoma cell lines and we found that rolipram inhibited the viability of all the cell lines studied. BDNF receptors, as well as the PKA pathway, may be therefore promising molecular targets for the development of new therapies for treating medulloblastomas.

Meduloblastoma é o tumor intracranial mais comum em crianças, provavelmente derivado de células precursoras da camada granular externa do cerebelo durante seu desenvolvimento. O tratamento padrão consiste em cirurgia, radioterapia e quimioterapia, que produzem graves sequelas nos pacientes e garantem uma sobrevida baixa, o que demonstra a necessidade de novas alternativas terapêuticas para a doença. Evidências demonstram que o receptor do peptídeo liberador de gastrina (GRPR) está superexpresso em diversos tumores humanos, assim como seu agonista (GRP) pode atuar como um fator de crescimento autócrino em tumores cerebrais. No presente estudo, avaliamos a expressão de GRPR e o efeito de seus agonistas, bombesina (BB) e GRP, além do antagonista RC-3095, sobre a viabilidade celular de linhagens de meduloblastoma humano DAOY, D283 e ONS76. Mostramos que meduloblastomas, apesar de expressarem GRPR, não têm sua viabilidade celular afetada por agonistas e antagonista desse receptor. Uma vez que há evidências de que BDNF (fator neurotrófico derivado de cérebro) esteja relacionado à diferenciação celular em meduloblastomas, também avaliamos o efeito de BDNF sobre a viabilidade celular das linhagens de meduloblastoma humano. As linhagens DAOY e D283 tiveram sua viabilidade celular reduzida pela presença de BDNF. Uma vez que a via da PKA tem sido implicada na iniciação e progressão de vários tumores, também avaliamos o efeito de rolipram, um inibidor de fosfodiesterase tipo IV, sobre a viabilidade celular das linhagens de meduloblastoma humano, sendo que rolipram reduziu a viabilidade celular de todas as linhagens estudadas. Os receptores de BDNF e a via da PKA podem, portanto, ser alvos moleculares promissores para o desenvolvimento de novas terapias para meduloblastomas. / Medulloblastoma is the most common intracranial tumor in children and is believed to arise from the precursor cells of the external granule layer of the developing cerebellum. The standard treatment, consisting of surgery, craniospinal radiotherapy and chemotherapy, produces severe sequelae in patients and provides a poor overall survival, indicating the need for new therapeutic alternatives for treating this disease. Evidences show that the gastrin releasing peptide receptor (GRPR) is overexpressed in various human tumors and its agonist (GRP) can act as an autocrine growth factor in brain tumors. In the present study, we evaluated GRPR expression, as well as the effect of its agonists, bombesin (BB) and GRP, and its antagonist RC-3095, over cell viability of the human medulloblastoma cell lines DAOY, D283 and ONS76. We found that medulloblastomas, in spite of expressing GRPR, do not have its viability affected by the presence of agonists and antagonist of this receptor. Since there are evidences that BDNF (brain-derived neurotrophic factor) is related to cell differentiation in medulloblastomas, we also evaluated the effect of BDNF over the viability of medulloblastoma cell lines. The viability of the cell lines DAOY and D283 was reduced by the presence of BDNF. Since the PKA pathway has been implicated in the initiation and progression of various tumors, we also evaluated the effect of rolipram, a phosphodiesterase IV inhibitor, over the viability of the same medulloblastoma cell lines and we found that rolipram inhibited the viability of all the cell lines studied. BDNF receptors, as well as the PKA pathway, may be therefore promising molecular targets for the development of new therapies for treating medulloblastomas.

abstract: Following diagnosis of a glioblastoma (GBM) brain tumor, surgical resection, chemotherapy and radiation together yield a median patient survival of only 15 months. Importantly, standard treatments fail to address the dynamic regulation of the brain tumor microenvironment that actively supports tumor progression and treatment resistance. Moreover, specialized niches within the tumor microenvironment maintain a population of highly malignant glioblastoma stem-like cells (GSCs). GSCs are resistant to traditional chemotherapy and radiation therapy and are likely responsible for near universal rates of tumor recurrence and associated morbidity. Thus, disrupting microenvironmental support for GSCs could be critical to more effective GBM therapies. Three-dimensional (3D) culture models of the tumor microenvironment are powerful tools for identifying key biochemical and biophysical inputs that may support or inhibit malignant behaviors. Here, we developed synthetic poly(N-isopropylacrylamide-co-Jeffamine M-1000® acrylamide) or PNJ copolymers as a model 3D system for culturing GBM cell lines and low-passage patient-derived GSCs in vitro. These temperature responsive scaffolds reversibly transition from soluble to insoluble in aqueous solution by heating from room temperature to body temperature, thereby enabling easy encapsulation and release of cells in a 3D scaffold. We also designed this system with the capacity for presenting the cell-adhesion peptide sequence RGD for adherent culture conditions. Using this system, we identified conditions that promoted GBM proliferation, invasion, GSC phenotypes, and radiation resistance. In particular, using two separate patient-derived GSC models, we observed that PNJ scaffolds regulated self-renewal, provided protection from radiation induced cell death, and may promote stem cell plasticity in response to radiation. Furthermore, PNJ scaffolds produced de novo activation of the transcription factor HIF2α, which is critical to GSC tumorigenicity and stem plasticity. All together, these studies establish the robust utility of PNJ biomaterials as in vitro models for studying microenvironmental regulation of GSC behaviors and treatment resistance. / Dissertation/Thesis / Doctoral Dissertation Biomedical Engineering 2017

Class of 2012 Abstract / Specific Aims: This study was aimed to evaluate inpatient clinical treatment characteristics associated with the use of intracranial implantation of chemotherapeutic wafers for malignant brain neoplasms within United States, and assess inpatient mortality and total charges regarding treatment with wafer versus without. Methods: A retrospective cohort investigation was conducted utilizing inpatient discharge records from the Agency for Healthcare Research and Quality (AHRQ) Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample from 2005 to 2009. From this nationally-representative sample, 9,455 adults aged 18 years or older were identified with malignant neoplasms of the brain treated with implantable chemotherapeutic wafers. Outcomes of inpatient mortality and charges were assessed via multivariate regression analysis, controlling for patient characteristics, hospital structure, comorbidities, and clinical complications. Main Results: The average age of patients with brain neoplasms was 56.6 (±16.5) years, and of those patients, 42.9% were female. The odds ratio for inpatient mortality of patients treated with implantable chemotherapeutic wafers was OR=0.380 (P<0.001), and patients that received wafer treatment had increased charges exp(b)=2.147 (P<0.001). Conclusions: Multiple factors were associated with inpatient mortality and charges among the 247,829 patients that were diagnosed with malignant brain neoplasms from 2005-2009. With regards to these patients, implantable chemotherapeutic wafers were associated with increased inpatient survival and increased charges.

Os gliomas são tumores do sistema nervoso central (SNC) que evoluem a partir das células da glia. O tipo mais frequente e mais agressivo destes tumores é conhecido como glioblastoma multiforme (GBM) e entre as características biológicas de agressividade associadas a esse tumor estão o seu rápido crescimento e ausência de apoptose. O seu prognóstico desfavorável está associado à dificuldade de tratamento dessas células, pois possuem resistência à quimioterapia e a radioterapia. A expressão gênica das enzimas ciclooxigenase-1 (COX-1), ciclooxigenase-2 (COX-2), prostaglandina E sintase-1 microssomal (mPGES-1), prostaglandina E sintase-2 microssomal (mPGES-2), prostaglandina E sintase citosólica (cPGES) e os produtos da síntese destas enzimas, incluindo a prostaglandina E1 (PGE1) e a prostaglandina E2 (PGE2) estão diretamente relacionados com a malignidade dos gliomas. A PGE1 e a PGE2 podem atuar de modo autócrino e parácrino, interagindo com suas células alvos através de ligação aos receptores da superfície celular que estão ligados a proteína G. Estes receptores são conhecidos como receptores EPs e dividem-se em quatro subtipos: EP-1, EP-2, EP-3 e EP-4 sendo que cada um deles ativa vias distintas de sinalização intracelular. Desta forma, este estudo teve por objetivo analisar in vitro o papel da PGE1, PGE2 e seus receptores na proliferação e apoptose em glioma humano, e a expressão das enzimas COX-1, COX-2, mPGES-1, mPGES-2 e cPGES. / Gliomas are tumors of the central nervous system (CNS) that evolve from glial cells. The most common and most aggressive form of these tumors is known as glioblastoma multiforme (GBM). The biological aggressiveness of GBM is associated with its rapid growth and lack of apoptosis. Its poor prognosis is strongly associated with the difficulty of treating these cells as they are resistant to chemotherapy and radiotherapy. The gene expression of the enzymes cyclooxygenase-1 (COX-1), cyclooxygenase-2 (COX-2), microsomal prostaglandin E synthase-1 (mPGES-1), microsomal prostaglandin E synthase-2 (mPGES-2), cytosolic prostaglandin E synthase (cPGES) and the products of the activity of these enzymes, including prostaglandin E1 (PGE1) and prostaglandin E2 (PGE2), are directly related to the malignancy of gliomas. PGE1 and PGE2 can act in an autocrine and paracrine manner, by interacting with their target cells via binding to cell surface receptors that are linked to G-proteins. These receptors are known as EP receptors and are divided into four subtypes: EP1, EP2, EP3 and EP4; each of which activates distinct intracellular signaling pathways. Therefore, this study aimed to analyze, in vitro, the role of PGE1, PGE2 and their receptors in the proliferation and apoptosis of human glioma and the expression of COX-1, COX-2, mPGES-1, mPGES-2 and cPGES.

Brain tumors present the second highest cause of death among pediatric cancers. About 60% are located in the posterior fossa region of the brain; among the most frequent types the ones considered for this project were astrocytomas, medulloblastomas, and ependymomas. Diagnosis can be done either through invasive histopathology exams or by non-invasive magnetic resonance (MR) scans. The tumors listed can be difficult to diagnose, even for trained radiologists, so machine learning methods, in particular deep learning, can be useful in helping to assess a diagnosis. Deep learning has been investigated only in a few other studies.The dataset used included 115 different subjects, some with multiple scan sessions, for which there were 142 T2-w, 119 T1Gd-w, and 89 volumes that presented both MR modalities. 2D slices have been manually extracted from the registered and skull-stripped volumes in the transversal, sagittal, and frontal anatomical plane and have been preprocessed by normalizing them and selecting the slices containing the tumor. The scans employed are T2-w, T1Gd-w, and a combination of the two referred to as multimodal images. The images were divided session-wise into training, validation, and testing, using stratified cross-validation and have also been augmented. The convolutional neural networks (CNN) investigated were ResNet50, VGG16, and MobileNetV2. The model performances were evaluated for two-class and three-class classification tasks by computing the confusion matrix, accuracy, receiver operating characteristic curve (ROC), the area under the curve (AUROC), and F1-score. Moreover,  explanations for the behavior of networks were investigated using GradCAMs and occlusion maps. Preliminary investigations showed that the best plane and modality were the transversal one and T2-w images. Overall the best model was VGG16, for the two-class tasks the best classification was between astrocytomas and medulloblastomas which reached an F1-score of 0.86 for both classes on multimodal images, followed by astrocytomas and ependymomas with an F1-score of 0.76 for astrocytomas and 0.74 for ependymomas on T2-w, and last F1-score of 0.30 for ependymomas and 0.65 for medulloblastomas on multimodal images. The three-class classification reached F1-score values of 0.59 for astrocytomas, 0.46 for ependymomas, and 0.64 for medulloblastomas on T2-w images. GradCAMs and occlusion maps showed that VGG16 was able to focus mostly on the tumor region but that there also seemed to be other information in the background of the images that contributed to the final classification.To conclude, the classification of infratentorial pediatric brain tumors can be achieved with acceptable results by means of deep learning and using a single MR modality, though one might have to account for the dataset size, number of classes and class imbalance. GradCAMs and occlusion maps offer important insights into the decision process of the networks

The Wide Range Assessment of Memory and Learning (WRAML) is a relatively new instrument used in the assessment of memory in children. The purpose of this study was to examine the validity of the WRAML by comparing the performance of children on both the WRAML and the Woodcock-Johnson Tests of Cognitive Ability- Revised (WJTCA-R). Subjects for the study were children in treatment for a brain tumor at a regional children's medical center. Fifty children participated in the study ranging from ages 6 to 17. A multiple regression analysis was conducted to determine which of four selected clusters from the WJTCA-R would have the highest correlation with the Verbal Memory Index (VERI) from the WRAML. The Short-Term Memory (GSM) cluster had the highest correlation ( r = .82) as predicted. A Pearson's product-moment correlational analysis was conducted between the Visual Processing (GV) cluster from the WJTCA-R and the Visual Memory Index (VISI) from the WRAML. GV was found to have a high positive correlation ( r = .63) with VISI. A similar analysis was conducted between the Long-Term Retrieval (GLR) cluster from the WJTCA-R and the Learning Index (LRNI) from the WRAML. GLR was found to have a high positive correlation ( r = .81) with LRNI. Finally, a correlational analysis was conducted between the Broad Cognitive Ability (BCA) scale from the WJTCA-R and the General Memory Index (GENI) from the WRAML. A high positive correlation ( r = .87) was found between these most global measures from the two batteries. The observed correlation between BCA and GENI was much higher than anticipated. The author concluded that neurological impairment had affected subject memory and intellectual functioning in similar ways. The results do not generalize to children who have not had similar decrements in cognitive functioning. Future research should establish a baseline correlation between the two instruments with a non-impaired population.

Les tumeurs gliales et glioneuronales sont les tumeurs cérébrales les plus fréquentes chez l'enfant. Elles sont généralement d'excellent pronostic. En revanche, les astrocytomes pilocytiques (AP) hypothalamo-chiasmatiques, ont un potentiel évolutif plus agressif. Ce travail de thèse propose une nouvelle stratégie thérapeutique pour ce sous-type d'AP selon la méthode du « drug repositioning », en employant la combinaison du celecoxib et de la fluvastatine. Nos travaux ont montré in vitro que cette association de molécules était synergique, capable d'arrêter le cycle cellulaire, de diminuer la prolifération et d'induire l'apoptose des cellules tumorales. Cette combinaison a également été testée avec succès chez une patiente souffrant d'un AP multifocal et réfractaire aux traitements conventionnels dans le cadre d'une thérapie métronomique. Ce manuscrit décrit également l'étude histo-moléculaire de plusieurs séries de tumeurs gliales et glioneuronales pédiatriques menées afin d'améliorer leur caractérisation et leur diagnostic. Nos travaux ont confirmé la présence de la fusion KIAA1549:BRAF dans les AP analysés ainsi que le caractère péjoratif de la topographie hypothalamo-chiasmatique, du variant histologique pilomyxoïde et de l'âge au diagnostic inférieur à 36 mois. Ils ont également montré l'absence de différence moléculaire entre les gliomes corticaux de grade II et des DNT. Enfin, nos travaux ont montré que les DNT, les GG et les PXA partagent la mutation BRAFV600E et l'expression de CD34. Ces travaux confirment l'implication majeure de l'altération de la voie des MAPKinases dans la tumorigenèse de ces tumeurs, constituant ainsi une cible thérapeutique prometteuse. / Glial and glioneuronal tumors are the most frequent brain tumors in children. They are characterized by an excellent prognosis. However, hypothalamic-chiasmatic pilocytic astrocytomas (PA) have a more aggressive outcome. In the first part, we propose a new therapeutic strategy for hypothalamic-chiasmatic PA according to drug repositioning method, by using celecoxib, and fluvastatin. We showed that, in vitro, this combination was synergistic, stopped cell cycle, inhibited cell proliferation and increased apoptosis. In addition, this combination was tested with success, under a metronomic chemotherapy, for a girl suffering from a multifocal PA and refractory to conventional treatment. This new strategy of treatment appears promising for this type of tumor because it is less toxic than conventional chemotherapy and not too expensive. In the second part, this manuscript describes the histo-molecular study of several retrospective series of glial and glioneuronal pediatric tumors conducted to improve their characterization and their diagnosis. We confirmed the presence of the fusion gene KIAA1549: BRAF in PA as well as the pejorative nature of the hypothalamic-chiasmatic topography, pilomyxoïde histology and the age at diagnosis less than 36 months. We also showed no molecular difference between cortical grade II gliomas associated with chronic epilepsy and the DNT group. Finally, we showed that DNT, GG and PXA share BRAFV600E mutation and expression of CD34. These studies confirm the major implication of the MAPKinase altered pathway in tumorigenesis of glial and glioneuronal pediatric tumors, constituting a promising therapeutic target.