Keywords: Quantitative imaging, radiomics, NSCLC, prognostic modelling

“Longitudinal radiomics for prognosis in non-small cell lung cancer”

The 5-year survival rate of non-small cell lung cancer (NSCLC) patients is still poor. Personalized medicine has the potential to improve individual treatment decisions which lead to better patient outcome. Radiomics, the extraction of quantitative features from medical images, has been proposed to describe the phenotype of the tumor in order to predict outcome (e.g. survival). Whereas radiomic features extracted from images acquired prior to radiotherapy have shown to carry prognostic information, the change of radiomic features over time (i.e. longitudinal radiomics) was not yet investigated. This thesis shows the potential of longitudinal radiomics for prognosis in NSCLC, but improved prognostic value opposed to baseline radiomics could not yet be proven. This work leads to new insights in the longitudinal radiomics approach, which hopefully improves personalized treatments in the future.