Can We Go Beyond Visual Features? Neural Tissue Relation Modeling for Relational Graph Analysis in Non-Melanoma Skin Histology

TL;DR

This paper introduces Neural Tissue Relation Modeling (NTRM), a novel framework that enhances histopathology image segmentation by explicitly modeling tissue relationships with graph neural networks, improving accuracy over existing CNN-based methods.

Contribution

NTRM combines CNNs with a tissue-level graph neural network to explicitly encode inter-tissue dependencies, advancing structural coherence in histological segmentation.

Findings

NTRM outperforms state-of-the-art methods on a skin cancer segmentation dataset.

NTRM achieves a 4.9% to 31.25% higher Dice coefficient than previous models.

Relational modeling improves context-awareness and interpretability in tissue segmentation.

Abstract

Histopathology image segmentation is essential for delineating tissue structures in skin cancer diagnostics, but modeling spatial context and inter-tissue relationships remains a challenge, especially in regions with overlapping or morphologically similar tissues. Current convolutional neural network (CNN)-based approaches operate primarily on visual texture, often treating tissues as independent regions and failing to encode biological context. To this end, we introduce Neural Tissue Relation Modeling (NTRM), a novel segmentation framework that augments CNNs with a tissue-level graph neural network to model spatial and functional relationships across tissue types. NTRM constructs a graph over predicted regions, propagates contextual information via message passing, and refines segmentation through spatial projection. Unlike prior methods, NTRM explicitly encodes inter-tissue dependencies, enabling structurally coherent predictions in boundary-dense zones. On the benchmark Histopathology Non-Melanoma Skin Cancer Segmentation Dataset, NTRM outperforms state-of-the-art methods, achieving a robust Dice similarity coefficient that is 4.9\% to 31.25\% higher than the best-performing models among the evaluated approaches. Our experiments indicate that relational modeling offers a principled path toward more context-aware and interpretable histological segmentation, compared to local receptive-field architectures that lack tissue-level structural awareness. Our code is available at https://github.com/shravan-18/NTRM.