NIA AI Training Dataset Construction Project
Participated in the National Information Society Agency (NIA) AI Training Dataset Construction Project to develop large-scale multimodal medical datasets for cancer AI research.
Our work focused on building integrated datasets for lung cancer prognosis prediction, malignant lymphoma prognosis prediction, and radiotherapy data for cancer patients, combining medical imaging such as CT, PET, and CBCT with clinical EMR data and, in selected tasks, digital pathology images.
These datasets were designed to support AI applications including prognosis prediction, survival analysis, lesion and organ annotation, and clinically relevant decision-support research.