Image processing, radiomics and model selection for prediction of treatment outcome of anal cancer using CT-, PET-, and MR-sequences

Abstract

The goal of this master thesis was to predict treatment outcome of anal cancer using features extracted from medical image sequences. The medical images had beforehand been provided from and registered at the Norwegian Radium Hospital. They consisted of CT,- PET- and MRsequences, in which MR was the modality of interest as former projects at NMBU has involved prediction of treatment outcome using PET- and CT-sequences. The treatment outcome of interest was PFS, Progression-Free Survival.