alt_text: A vibrant cover image depicting lung tissue with GNN elements and metrics showcasing model success.

Integrating Graph Neural Networks and Mamba for Enhanced Tissue Spatial Analysis in Whole Slide Images

A new model combining graph neural networks (GNNs) with the state space model Mamba enables capturing both local and global spatial relationships in whole slide images (WSIs). This approach significantly improves the accuracy of progression-free survival predictions in early-stage lung adenocarcinoma patients. By efficiently analyzing large tissue samples, the method benefits computational pathology workflows and precision medicine. The model achieved a c-index of 0.70, outperforming existing techniques.

Leave a Reply

Your email address will not be published. Required fields are marked *