Unit-Based Histopathology Tissue Segmentation via Multi-Level Feature Representation

TL;DR

This paper introduces UTS, a unit-based tissue segmentation framework using a Multi-Level Vision Transformer, which improves efficiency and accuracy in histopathology analysis by classifying tissue tiles instead of pixels.

Contribution

It presents a novel unit-based segmentation approach with a Multi-Level Vision Transformer, reducing annotation effort and enhancing performance in histopathology tissue classification.

Findings

Outperforms U-Net variants and transformer baselines on tissue segmentation tasks.

Reduces annotation effort by classifying tiles instead of pixels.

Supports clinically relevant tasks like tumor-stroma quantification.

Abstract

We propose UTS, a unit-based tissue segmentation framework for histopathology that classifies each fixed-size 32 * 32 tile, rather than each pixel, as the segmentation unit. This approach reduces annotation effort and improves computational efficiency without compromising accuracy. To implement this approach, we introduce a Multi-Level Vision Transformer (L-ViT), which benefits the multi-level feature representation to capture both fine-grained morphology and global tissue context. Trained to segment breast tissue into three categories (infiltrating tumor, non-neoplastic stroma, and fat), UTS supports clinically relevant tasks such as tumor-stroma quantification and surgical margin assessment. Evaluated on 386,371 tiles from 459 H&E-stained regions, it outperforms U-Net variants and transformer-based baselines. Code and Dataset will be available at GitHub.