Collect data

Based on the UPGRADE-TNBC study, a high-quality TNBC sample repository was established. By integrating multi-source data-including clinical information, radiomics, pathological images, and molecular sequencing-and innovatively incorporating a meta-learning strategy, a treatment response prediction model based on multimodal small-sample learning was developed. This approach aims to optimize drug combinations and precisely identify patient subgroups likely to benefit from treatment, thereby providing a new paradigm for personalized therapy in early-stage TNBC.