Radiation therapy (RT) research increasingly recognises the critical limitations of historical knowledge regarding the relation between physical radiation exposures and their biologic effects, known as the dose response. Improved understanding of dose response is urgently needed to guide treatment-plan optimisation and enable continued increases in long-term survivorship. Advanced computational methods for survivorship research, like deep learning and data mining, offer opportunities to improve our understanding of dose response. Survivorship research is especially important for children, whose growing bodies are inherently vulnerable to radiation damage and for whom survival often spans decades. Much of the work to incorporate advanced computational techniques into survivorship research, however, has focused on adult patients. Two important applications of advanced computational methods to paediatric survivorship research include automatic contouring and voxel-based methods for identifying the most important parts of the anatomy driving toxicity. The latter is called image-based data mining. Both avenues are currently under investigation in our lab, and preliminary results are promising. The continued, safe integration of advanced computational methods into paediatric radiation oncology is needed to realise the full beneficial potential of radiation therapy and provide not only prolonged survival, but also superior outcomes with reduced toxicity for children with cancer.