GAFR-Net: A Graph Attention and Fuzzy-Rule Network for Interpretable Breast Cancer Image Classification

Accurate classification of breast cancer histopathology images is pivotal for early oncological diagnosis and therapeutic intervention.However, conventional deep learning architectures often encounter performance degradation under limited annotations and suffer from a “blackbox” nature, hindering their clinical integration. To mitigate these limitations, we propose GAFRNet, a robust and interpretable Graph Attention and FuzzyRule Network specifically engineered for histopathology image classification with scarce supervision. GAFRNet constructs a similarity-driven graph representation to model intersample relationships and employs a multihead graph attention mechanism to capture complex relational features across heterogeneous tissue structures.Concurrently, a differentiable fuzzy-rule module encodes intrinsic topological descriptorsincluding node degree, clustering coefficient, and label consistencyinto explicit, human-understandable diagnostic logic. This design establishes transparent “IF-THEN” mappings that mimic the heuristic deduction process of medical experts, providing clear reasoning behind each prediction without relying on post-hoc attribution methods. Extensive evaluations on three benchmark datasets (BreakHis, Mini-DDSM, and ICIAR2018) demonstrate that GAFR-Net consistently outperforms various state-of-the-art methods across multiple magnifications and classification tasks. These results validate the superior generalization and practical utility of GAFR-Net as a reliable decision-support tool for weakly supervised medical image analysis.

💡 Research Summary

The paper introduces GAFR‑Net, a novel architecture that jointly leverages graph attention and a trainable fuzzy‑rule reasoning module to achieve both high classification performance and intrinsic interpretability for breast cancer histopathology image analysis under weak supervision.
Problem motivation – Conventional CNNs excel at extracting local visual features but struggle to capture global, non‑Euclidean relationships among tissue patches, especially when training data are scarce or heavily imbalanced. Recent Graph Neural Networks (GNNs) address relational modeling but remain largely black‑box, limiting clinical trust.
Core contributions –

1. Similarity‑driven graph construction – Each whole‑slide image is a node; edges are created when a cosine similarity (computed from pre‑trained CNN embeddings) exceeds a threshold τ, yielding a sparse graph that explicitly encodes inter‑sample semantic proximity.
2. Topology‑aware descriptors – For every node u three graph‑theoretic measures are extracted: clustering coefficient C(u) (local density), node degree d(u) (centrality), and two‑hop label agreement L(u) (consistency of neighboring class predictions). These form a compact vector f_u =