Artificial intelligence for end-to-end radiotherapy optimization: Radiomics, synthetic data, deep learning, and large language models

This thesis explored how artificial intelligence can support different steps of cancer care, especially in radiotherapy. It investigated ways to make medical images more reliable for analysis, used synthetic data to improve model training when patient data are limited, and developed deep learning tools for tasks such as tumor detection and organ or tumor segmentation. In the later stage, the research also examined how large language models and AI agents can help organize medical information and support clinical decision-making. The overall aim of this work was to improve the accuracy, efficiency, and personalization of cancer treatment.