Identifying tumor phenotypes non-invasively from quantitative imaging features is a challenge faced by radiomics. This study aimed at investigating if radiomic features measured at diagnosis time from conventional structural MRI can predict histone H3 mutations and overall survival of patients with diffuse intrinsic pontine glioma. To this end, 316 features from multimodal diagnostic MRI of 38 patients were extracted. Two approaches were proposed: a conventional estimation of features inside the whole region of interest and a mean estimation inside this region of local features that are computed from fixed size patches. A feature selection pipeline was then developed. Three machine learning models for H3 mutation classification and three regression models for overall survival prediction were evaluated. Leave-one-out F1-weighted scores for SVM model combining imaging and clinical features reached 0.84, showing a good prediction of H3 mutation using structural MRI. Some encouraging results were obtained to predict overall survival but they need to be reinforced on a larger number of patients.