Hybrid Supervision Learning for Pathology Whole Slide Image Classification

TL;DR

This paper introduces a hybrid supervision learning framework for pathology whole slide image classification that combines coarse image-level labels with fine pixel-level annotations, improving specificity while maintaining sensitivity.

Contribution

The proposed framework effectively leverages both annotation types in high-resolution images, addressing challenges in weak and hybrid supervision learning in computational pathology.

Findings

Achieved 5.2% higher specificity over existing methods.

Retained 100% sensitivity in classification tasks.

Validated on a large hybrid annotated dataset.

Abstract

Weak supervision learning on classification labels has demonstrated high performance in various tasks, while a few pixel-level fine annotations are also affordable. Naturally a question comes to us that whether the combination of pixel-level (e.g., segmentation) and image level (e.g., classification) annotation can introduce further improvement. However in computational pathology this is a difficult task for this reason: High resolution of whole slide images makes it difficult to do end-to-end classification model training, which is challenging to research of weak or hybrid supervision learning in the past. To handle this problem, we propose a hybrid supervision learning framework for this kind of high resolution images with sufficient image-level coarse annotations and a few pixel-level fine labels. This framework, when applied in training patch model, can carefully make use of coarse image-level labels to refine generated pixel-level pseudo labels. Complete strategy is proposed to suppress pixel-level false positives and false negatives. A large hybrid annotated dataset is used to evaluate the effectiveness of hybrid supervision learning. By extracting pixel-level pseudo labels in initially image-level labeled samples, we achieve 5.2% higher specificity than purely training on existing labels while retaining 100% sensitivity, in the task of image-level classification to be positive or negative.