This study aims to utilise a real-world data platform to integrate multi-omics data-including radiomics, gut microbiota, pathological quality control and liquid biopsy-to construct a multidimensional predictive model for the efficacy of rectal cancer treatment following neoadjuvant therapy. By integrating multimodal data, the study aims to accurately assess the efficacy of neoadjuvant therapy and identify patients suitable for a 'watch-and-wait' strategy, thereby achieving tumour control and preserving organ function without the need for surgery. Furthermore, it seeks to provide scientific evidence for the efficacy of the 'watch-and-wait' strategy and the selection of optimal timing for surgery, whilst validating the model's effectiveness and assessing its clinical feasibility through prospective clinical trials.