CGAM: An end-to-end causality graph attention Mamba network for esophageal pathology grading

Yingbo Qu, Xiangli Zhou, Pan Huang, Yanan Liu, Francesco Mercaldo, Antonella Santone, Peng Feng

Abstract: Highlights•Graph Attention in CGAM for ESCC Grading.•Improves relational representation learning with graph attention.•Mamba Causality in CGAM.•Enhances model receptive field for tumor-TME relationship.•Superior Performance of CGAMOutperforms SOTA models in grading and interpretability on two pathology datasets.