Researchers have developed a new framework called AM$^2$ES for predicting cancer survival by integrating histopathological Whole-Slide Images (WSIs) and transcriptomic data. This framework addresses limitations in existing methods by explicitly aligning multi-modal features and adaptively prioritizing relevant tissue scales. AM$^2$ES utilizes an Anchor-driven Multi-modal Fusion module for semantic alignment and a Hierarchical Mixture-of-Experts module for discriminative region identification and informative resolution level selection. Experiments on TCGA cancer cohorts show that AM$^2$ES achieves state-of-the-art performance and provides interpretability by visualizing how molecular pathways influence expert routing decisions across different tissue scales. AI
IMPACT This framework could improve the accuracy and interpretability of cancer survival predictions by better integrating diverse biological data.
RANK_REASON The cluster contains a research paper detailing a new computational framework for a specific scientific problem. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
