CGRL: Concept-Guided Pruning and Representation Learning for Whole-Slide Image Classification

2026-07-14

AI Summary

This work addresses the limitations of existing weakly supervised whole-slide image classification methods, which are often hindered by non-informative regions and insufficient alignment between instance-level features and disease semantics. To overcome these challenges, the study introduces class-level disease prompts as prior knowledge into a multiple instance learning (MIL) framework for the first time. By constructing disease-driven concept prototypes, the authors propose concept-aware pruning to select diagnostically relevant patches and devise concept-guided contrastive learning to refine the representation space. This joint optimization of semantic-driven patch selection and feature learning significantly improves classification accuracy and macro-F1 scores on the TCGA-BRCA and TCGA-NSCLC datasets while reducing computational overhead.

Abstract

Weakly supervised whole-slide image (WSI) classification is widely used in computational pathology because slide-level labels are easier to obtain than dense region annotations. Existing multiple instance learning (MIL) methods often aggregate large bags of patch embeddings using mainly visual cues, which can retain many non-informative patches and provide weak alignment between instance features and class-level disease semantics. We propose Concept-Guided Pruning and Representation Learning (CGRL), a simple framework that introduces class-level concept prototypes derived from disease prompts into the MIL pipeline. First, concept-relevance pruning ranks patch instances by their similarity to class concepts and retains the top-K concept-relevant patches for downstream MIL aggregation. Second, concept-guided contrastive representation learning constructs class-wise positive and negative patch sets from the same similarity matrix and optimizes target-class, symmetric auxiliary, and cross-class separation objectives, thereby regularizing the projected concept space. We evaluate CGRL on TCGA-BRCA and TCGA-NSCLC using multiple representative MIL methods. Experimental results show that CGRL improves several model-dataset combinations, with gains depending on the downstream MIL model and dataset. It achieves particularly clear improvements in accuracy and macro-F1 while reducing computational cost through concept-relevance pruning. These findings demonstrate that class-level semantic concepts provide an effective and practical prior for patch selection and representation learning in weakly supervised computational pathology.

Problem

• weakly supervised learning
• whole-slide image classification
• multiple instance learning
• semantic alignment
• computational pathology

Innovation

• concept-guided pruning
• contrastive representation learning
• multiple instance learning
• weakly supervised learning
• computational pathology