CRCFound

Overview

CRCFound is a self-supervised colorectal cancer CT foundation model designed to improve preoperative risk stratification and downstream clinical prediction tasks under limited annotation conditions. The model is pretrained on large-scale unlabeled 3D CT data and then adapted to multiple diagnostic and prognostic tasks, showing strong transferability and generalization. This work provides a practical path to leverage routine CT imaging for precision diagnosis and personalized treatment planning in colorectal cancer.