Radiomics is a process of extracting many quantitative data obtained from medical images and analysing them. In neuroradiology it may be used to discover magnetic resonance imaging (MRI) features of glioblastomas that are impossible to identify by human vision alone. In this article, the authors describe the methodology and their first experience in creating a predictive model based on radiomic features obtained from the preoperative MRI examination of patients with glioblastoma. Early identification of malignant glioblastoma subtypes characterized by different prognoses and responses to treatment would greatly facilitate the implementation of targeted therapy, which appears to be the future of glioblastoma treatment.
Keywords: artificial intelligence; glioblastoma; machine learning; radiomics.
© Pol J Radiol 2023.