TVT-PAPD: Pathology-Aware Prototype Distillation for Self-Supervised Whole Slide Image Classification

2026-07-11

AI Summary

Existing self-supervised learning approaches struggle to explicitly model pathology-specific morphological patterns that are critical for disease representation. To address this limitation, this work proposes a novel framework integrating a Tiny Vision Transformer with pathology-aware prototype distillation. By leveraging a learnable library of pathological prototypes, the method explicitly captures and preserves representative tissue morphologies during self-supervised contrastive learning, thereby enforcing semantic consistency in pathological representations. The proposed approach achieves an optimal balance between computational efficiency and discriminative power, attaining weighted F1 scores of 93.02% and 90.23% on the TCGA and IPD-Brain datasets, respectively. Furthermore, it demonstrates strong cross-cohort generalization capability, highlighting its robustness and clinical applicability.

Abstract

Self-supervised learning (SSL) has emerged as an effective paradigm for learning transferable representations from large-scale unlabeled whole slide images (WSIs). However, existing SSL methods primarily learn generic visual features and often fail to explicitly capture pathology-specific morphological patterns that are critical for disease characterization. To address this limitation, we propose Tiny Vision Transformer with Pathology-Aware Prototype Distillation (TVT-PAPD). This self-supervised pathology representation learning framework integrates a Tiny Vision Transformer (TVT) with a novel Pathology-Aware Prototype Distillation (PAPD) module. PAPD employs a learnable pathology prototype bank to discover and preserve representative tissue morphology patterns, encouraging semantically similar pathological regions to learn consistent and discriminative representations. The proposed framework enhances pathology-aware feature learning while maintaining computational efficiency with 90M parameters. Experiments on the Cancer Genome Atlas (TCGA) low-grade glioma (LGG)/glioblastoma (GBM) dataset and the Indian Pathology Brain (IPD-Brain) dataset demonstrate that TVT-PAPD achieves weighted F1-scores of 93.02% and 90.23%, respectively, for LGG-GBM classification, while exhibiting strong cross-cohort generalization across independent glioma datasets.

Problem

• self-supervised learning
• whole slide image
• pathology-specific morphology
• representation learning
• morphological patterns

Innovation

• Pathology-Aware Prototype Distillation
• Self-Supervised Learning
• Whole Slide Image Classification
• Tiny Vision Transformer
• Morphological Pattern Discovery