La radiomique suppose que des informations pertinentes non repérables visuellement peuvent être trouvées en calculant une grande quantité d’indices quantitatifs à partir des images médicales. En cancérologie, ces informations pourraient caractériser le phénotype de la tumeur et définir le pronostic du patient. Le GITC est une tumeur pédiatrique rare diagnostiquée d'après des signes cliniques et son apparence en IRM. Cette thèse présente les premières études radiomiques pour des patients atteints de GITC. Comme les intensités en IRM clinique sont exprimées en unités arbitraires, la première étape de l’étude a été la standardisation des images. Une méthode de normalisation basée sur l'estimation de l'intensité dans la matière blanche d'apparence normale s’est avérée efficace sur plus de 1500 volumes d'images. Des études méthodologiques sur le calcul des indices de texture ont abouti aux recommandations suivantes : (a) discrétiser les niveaux de gris avec une largeur constante pour tous les patients, (b) utiliser un volume d'intérêt constant ou faire attention au biais introduit par des volumes de taille et forme différentes. En s’appuyant sur ces recommandations, les indices radiomiques issus de 4 modalités d'IRM ont été systématiquement analysés en vue de prédire les principales mutations génétiques associées aux GITC et la survie globale des patients au moment du diagnostic. Un pipeline de sélection d’indices a été proposé et différentes méthodes d’apprentissage automatique avec validation croisée ont été mises en oeuvre pour les deux tâches de prédiction. La combinaison des indices cliniques avec les indices d’imagerie est plus efficace que les indices cliniques ou d’imagerie seuls pour la prédiction des deux principales mutations de l’histone H3 (H3.1 versus H3.3) associées au GITC. Comme certaines modalités d'imagerie étaient manquantes, une méthodologie adaptée à l’analyse des bases de données d’imagerie multi-modales avec données manquantes a été proposée pour pallier les limites de recueil des données d'imagerie. Cette approche permet d'intégrer de nouveaux patients. Les résultats du test externe de prédiction des deux principales mutations de l’histone H3 sont encourageants. Concernant la survie, certains indices radiomiques semblent informatifs. Toutefois, le faible nombre de patients n'a pas permis d'établir les performances des prédicteurs proposés. Enfin, ces premières études radiomiques suggèrent la pertinence des indices radiomiques pour la prise en charge des patients atteints de GITC en absence de biopsie mais l’augmentation de la base de données est nécessaire pour confirmer ces résultats. La méthodologie proposée dans cette thèse peut être appliquée à d'autres études cliniques. / Radiomics is based on the assumption that relevant, non-visually identifiable information can be found by calculating a large amount of quantitative indices from medical images. In oncology, this information could characterize the phenotype of the tumor and define the prognosis of the patient. DIPG is a rare pediatric tumor diagnosed by clinical signs and MRI appearance. This work presents the first radiomic studies for patients with DIPG. Since clinical MRI intensities are expressed in arbitrary units, the first step in the study was image standardization. A normalization method based on intensity estimation of the normal-appearing white matter has been shown to be effective on more than 1500 image volumes. Methodological studies on the calculation of texture indices have then defined the following recommendations: (a) discretize gray levels with a constant width for all patients, (b) use a constant volume of interest or pay attention to the bias introduced by volumes of different size and shape. Based on these recommendations, radiomic indices from four MRI modalities were systematically analyzed to predict the main genetic mutations associated with DIPG and the overall survival of patients at the time of diagnosis. An index selection pipeline was proposed and different cross-validated machine learning methods were implemented for both prediction tasks. The combination of clinical indices with imaging indices is more effective than the clinical or imaging indices alone for the prediction of the two main mutations in histone H3 (H3.1 versus H3.3) associated with DIPG. As some imaging modalities were missing, a methodology adapted to the analysis of multi-modal imaging databases with missing data was proposed to overcome the limitations of the collection of imaging data. This approach made it possible to integrate new patients. The results of the external prediction test for the two main mutations of H3 histone are encouraging. Regarding survival, some radiomic indices seem to be informative. However, the small number of patients did not make it possible to establish the performance of the proposed predictors. Finally, these first radiomic studies suggest the relevance of the radiomic indices for the management of patients with DIPG in the absence of biopsy but the database need to be increased in order to confirm these results. The proposed methodology can be applied to other studies.

No description available.

No description available.

No description available.

One of the most pressing global health concerns is the impact of cancer, which remains a leading cause of death worldwide. The timeliness of detection and diagnosis is critical to maximizing the chances of successful treatment. Radiomics is an emerging medical imaging analysis proposed, which refers to the high-throughput extraction of a large number of image features. Radiomics generally refers to the use of CT, PET, MRI or Ultrasound imaging as input data, extracting expressive features from massive image-based data, and then using machine learning or statistical models for quantitative analysis and prediction of disease. Feature reduction is very critical in Radiomics as a large number of quantitative features can have redundant characteristics not necessarily important in the analysis process. Due to the immense features obtained from radiological images, the main objective of our research is the application of machine learning techniques to reduce the number of dimensions, thereby rendering the data more manageable. Radiomics involves several steps including: Imaging, segmentation, feature extraction, and analysis. Extracted features can be categorized in the description of tumor gray histograms, shape, texture features, and the tumor location and surrounding tissue. For this research, a large-scale CT dataset for Lung cancer diagnosis (Lung- PET-CT-Dx) which was collected by scholars from Medical University in Harbin in China is used to illustrate the dimension reduction techniques, which is a main part of radiomics process, via R, SAS and Python. The proposed reduction and analysis techniques in our research will entail; Principal Component Analysis, Clustering analysis (Hierarchical Clustering and K-means), and Manifold-based algorithms (Isometric Feature Mapping (ISOMAP).

(1) Background: Astrocytic gliomas present overlapping appearances in conventional MRI. Supplementary techniques are necessary to improve preoperative diagnostics. Quantitative DWI via the computation of apparent diffusion coefficient (ADC) histograms has proven valuable for tumor characterization and prognosis in this regard. Thus, this study aimed to investigate (I) the potential of ADC histogram analysis (HA) for distinguishing low-grade gliomas (LGG) and high-grade gliomas (HGG) and (II) whether those parameters are associated with Ki-67 immunolabelling, the isocitratedehydrogenase-1 (IDH1) mutation profile and the methylguanine-DNA-methyl-transferase (MGMT) promoter methylation profile; (2) Methods: The ADC-histograms of 82 gliomas were computed. Statistical analysis was performed to elucidate associations between histogram features and WHO grade, Ki-67 immunolabelling, IDH1 and MGMT profile; (3) Results: Minimum, lower percentiles (10th and 25th), median, modus and entropy of the ADC histogram were significantly lower in HGG. Significant differences between IDH1-mutated and IDH1-wildtype gliomas were revealed for maximum, lower percentiles, modus, standard deviation (SD), entropy and skewness. No differences were found concerning the MGMT status. Significant correlations with Ki-67 immunolabelling were demonstrated for minimum, maximum, lower percentiles, median, modus, SD and skewness; (4) Conclusions: ADC HA facilitates non-invasive prediction of the WHO grade, tumor-proliferation rate and clinically significant mutations in case of astrocytic gliomas.

No description available.

Early assessment of tumour response has lately acquired big interest in the medical field, given the possibility to modify treatments during their delivery. Radiomics aims to quantitatively describe images in radiology by automatically extracting a large number of image features. In this context, PET/CT (Positron Emission Tomography/Computed Tomography) images are of great interest since they encode functional and anatomical information, respectively. In order to assess the patients' responses from many image features appropriate methods should be applied. Machine learning offers different procedures that can deal with this, possibly high dimensional, problem. The main objective of this work was to develop a method to classify lung cancer patients as responding or not to chemoradiation treatment, relying on repeated PET/CT images. Patients were divided in two groups, based on the type of chemoradiation treatment they underwent (sequential or concurrent radiation therapy with respect to chemotherapy), but image features were extracted using the same procedure. Support vector machines performed classification using features from the Radiomics field, mostly describing tumour texture, or from handcrafted features, which described image intensity changes as a function of tumour depth. Classification performance was described by the area under the curve (AUC) of ROC (Receiving Operator Characteristic) curves after leave-one-out cross-validation. For sequential patients, 0.98 was the best AUC obtained, while for concurrent patients 0.93 was the best one. Handcrafted features were comparable to those from Radiomics and from previous studies, as for classification results. Also, features from PET alone and CT alone were found to be suitable for the task, entailing a performance better than random.

A CT-scan of lungs has become ubiquitous as a thoracic diagnostic tool. Thus, using CT-scan images in developing predictive models for tumor types and survival time of patients afflicted with Non-Small Cell Lung Cancer (NSCLC) would provide a novel approach to non-invasive tumor analysis. It can provide an alternative to histopathological techniques such as needle biopsy. Two major tumor analysis problems were addressed in course of this study, tumor type classification and survival time prediction. CT-scan images of 109 patients with NSCLC were used in this study. The first involved classifying tumor types into two major classes of non-small cell lung tumors, Adenocarcinoma and Squamous-cell Carcinoma, each constituting 30% of all lung tumors. In a first of its kind investigation, a large group of 2D and 3D image features, which were hypothesized to be useful, are evaluated for effectiveness in classifying the tumors. Classifiers including decision trees and support vector machines (SVM) were used along with feature selection techniques (wrappers and relief-F) to build models for tumor classification. Results show that over the large feature space for both 2D and 3D features it is possible to predict tumor classes with over 63% accuracy, showing new features may be of help. The accuracy achieved using 2D and 3D features is similar, with 3D easier to use. The tumor classification study was then extended by introducing the Bronchioalveolar Carcinoma (BAC) tumor type. Following up on the hypothesis that Bronchioalveolar Carcinoma is substantially different from other NSCLC tumor types, a two-class problem was created, where an attempt was made to differentiate BAC from the other two tumor types. To make a three-class problem a two-class problem, misclassification amongst Adenocarcinoma and Squamous-cell Carcinoma were ignored. Using the same prediction models as the previous study and just 3D image features, tumor classes were predicted with around 77% accuracy. The final study involved predicting two year survival time in patients suffering from NSCLC. Using a subset of the image features and a handful of clinical features, predictive models were developed to predict two year survival time in 95 NSCLC patients. A support vector machine classifier, naive Bayes classifier and decision tree classifier were used to develop the predictive models. Using the Area Under the Curve (AUC) as a performance metric, different models were developed and analyzed for their effectiveness in predicting survival time. A novel feature selection method to group features based on a correlation measure has been proposed in this work along with feature space reduction using principal component analysis. The parameters for the support vector machine were tuned using grid search. A model based on a combination of image and clinical features, achieved the best performance with an AUC of 0.69, using dimensionality reduction by means of principal component analysis along with grid search to tune the parameters of the SVM classifier. The study showed the effectiveness of a predominantly image feature space in predicting survival time. A comparison of the performance of the models from different classifiers also indicate SVMs consistently outperformed or matched the other two classifiers for this data.

京都大学 / 新制・課程博士 / 博士(人間健康科学) / 甲第23121号 / 人健博第83号 / 新制||人健||6(附属図書館) / 京都大学大学院医学研究科人間健康科学系専攻 / (主査)教授 椎名 毅, 教授 藤井 康友, 教授 平井 豊博 / 学位規則第4条第1項該当 / Doctor of Human Health Sciences / Kyoto University / DFAM