Machine learning applications for Radiomics: towards robust non-invasive predictors in clinical oncology

In this big-data era, like every other field, healthcare is also turning towards artificial intelligence (AI) and machine-learning (ML). In this thesis, state-of-the-art machine-learning methods were investigated for radiomic analyses. An unbiased evaluation of these advanced computational methods in terms of their accuracy and reliability is presented. Identification of optimal machine-learning methods for radiomic applications is a crucial step towards stable and clinically relevant radiomic biomarkers, providing a non-invasive way of quantifying and monitoring tumor-phenotypic characteristics in clinical practice. With ever increasing patient specific data, this work could stimulate further research towards brining AI and precision medicine in routine clinical oncology.