MUSTANG: Multi-Stain Self-Attention Graph Multiple Instance Learning Pipeline for Histopathology Whole Slide Images

TL;DR

MUSTANG introduces a novel self-attention graph-based multiple instance learning pipeline for classifying gigapixel histopathology WSIs at the patient level, achieving state-of-the-art performance without requiring detailed annotations.

Contribution

It presents a highly modular, end-to-end multi-stain self-attention graph pipeline that effectively handles heterogeneous WSIs with only patient-level labels, outperforming existing models.

Findings

Achieves state-of-the-art F1-score of 0.89 and AUC of 0.92.

Outperforms the widely used CLAM model.

Highly adaptable to different clinical datasets.

Abstract

Whole Slide Images (WSIs) present a challenging computer vision task due to their gigapixel size and presence of numerous artefacts. Yet they are a valuable resource for patient diagnosis and stratification, often representing the gold standard for diagnostic tasks. Real-world clinical datasets tend to come as sets of heterogeneous WSIs with labels present at the patient-level, with poor to no annotations. Weakly supervised attention-based multiple instance learning approaches have been developed in recent years to address these challenges, but can fail to resolve both long and short-range dependencies. Here we propose an end-to-end multi-stain self-attention graph (MUSTANG) multiple instance learning pipeline, which is designed to solve a weakly-supervised gigapixel multi-image classification task, where the label is assigned at the patient-level, but no slide-level labels or region annotations are available. The pipeline uses a self-attention based approach by restricting the operations to a highly sparse k-Nearest Neighbour Graph of embedded WSI patches based on the Euclidean distance. We show this approach achieves a state-of-the-art F1-score/AUC of 0.89/0.92, outperforming the widely used CLAM model. Our approach is highly modular and can easily be modified to suit different clinical datasets, as it only requires a patient-level label without annotations and accepts WSI sets of different sizes, as the graphs can be of varying sizes and structures. The source code can be found at https://github.com/AmayaGS/MUSTANG.