Survival Modeling from Whole Slide Images via Patch-Level Graph Clustering and Mixture Density Experts

arXiv — cs.CVThursday, November 13, 2025 at 5:00:00 AM
The recent development of a modular framework for predicting cancer-specific survival from whole slide pathology images represents a significant advancement in medical imaging and oncology. This framework integrates several innovative techniques, including quantile-based patch filtering to identify prognostically informative tissue regions, and graph regularized patch clustering to model phenotype-level variations. By learning both intra and inter-cluster dependencies through hierarchical feature aggregation, the model captures the complex organization of tumors. The final component, an expert-guided mixture density model, estimates survival distributions with greater precision. Evaluated on cohorts from TCGA LUAD, TCGA KIRC, and TCGA BRCA, the model achieved concordance indices of 0.653, 0.719, and 0.733, respectively, surpassing existing state-of-the-art approaches. This advancement not only enhances the accuracy of survival predictions but also has the potential to improve patient o…
— via World Pulse Now AI Editorial System

Was this article worth reading? Share it