Prediction Models for Personalized Treatment Delivery in Oncology

Cancer care is advancing rapidly, yet every patient’s journey remains unique and complex. This thesis investigated innovative ways to improve cancer care by using clinical prediction models to make treatments smarter and more personalized. It addressed key challenges in treatment, such as identifying patients at risk of heart complications after lung radiotherapy, distinguishing between different types of lung inflammations, predicting the likelihood of brain metastasis and cognitive decline. By analyzing patient data and medical images, the research developed tools that can help healthcare professionals make earlier, more accurate decisions, ultimately leading to better outcomes. This work highlights the potential of combining data science with clinical practice to create a future where cancer care is more precise, proactive, and tailored to individual patients’ needs.