Whole-slide image analysis models tumor microenvironment interactions dynamically
Modeling Whole-Slide Images as Dynamic Tumor Microenvironment Fields
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Whole-slide images of tissue samples are huge and hard to analyze all at once. The authors designed a new method, TMEvolve, that groups nearby tissue patches into regions and models how these regions interact and change over time, like a dynamic environment. This helps capture important spatial relationships that simpler methods miss. Their method improves predictions about cancer survival, gene activity, and tissue types compared to older techniques.
What this means in practice
- •For medical image analysts: Improve cancer prognosis and subtype classification models by accounting for dynamic tissue region interactions in whole-slide images.
- •For biomedical software developers: Create advanced pathology tools that model evolving tumor microenvironments from tissue images for more accurate slide-level predictions.
Authors
Lei Wu, Jiashuai Liu, Di Zhang, Zhangpeng Gong, Yingkang Zhan, Yi Niu, Jiusong Ge, Chunze Yang, Kai Yi, Mireia Crispin-Ortuzar, Chen Li, Zeyu Gao
Abstract
Due to the gigapixel-scale nature of whole-slide images (WSIs), weakly supervised WSI analysis is commonly formulated as a multiple instance learning (MIL) problem, where patch-level features are aggregated into slide-level representations. However, diagnostic and prognostic evidence often arises from spatially coherent tumor microenvironment regions and their interactions, rather than isolated patches alone. Existing patch-level or static region-based methods usually overlook how tissue regions should be adaptively formed and subsequently evolved through microenvironment interactions across heterogeneous boundaries. In this paper, we propose Concept-Guided Tumor Microenvironment Evolution (TMEvolve), a reaction-diffusion-inspired framework that models WSIs as latent tumor microenvironment fields over discrete patch graphs. TMEvolve instantiates this view as a learnable graph-discretized evolution process over patch neighborhoods. It first forms adaptive soft tissue regions as coherent microenvironment units, then performs pseudo-time evolution through two complementary local dynamics: intra-region diffusion, which stabilizes latent states within coherent tissue compartments, and concept-guided boundary flux, which propagates visual feature signals and language-derived concept signals across heterogeneous region interfaces. The evolved microenvironment regions are finally aggregated for slide-level prediction. We evaluate TMEvolve on six datasets across three weakly supervised WSI tasks: survival prediction, gene expression prediction, and histological subtype classification. TMEvolve consistently improves over representative MIL methods, pathology foundation models, and concept-guided baselines. Ablation studies and visualizations further support the effectiveness and interpretability of TMEvolve, highlighting the value of dynamic region modeling and boundary interaction.