Self-Supervised Multiple Instance Learning for Acute Myeloid Leukemia Classification

TL;DR

This paper demonstrates that self-supervised learning can effectively pre-train encoders for multiple instance learning in AML classification, reducing the need for labeled data and maintaining high performance.

Contribution

It introduces the use of SSL methods (SimCLR, SwAV, DINO) for pre-training MIL models in AML diagnosis, eliminating reliance on labeled datasets during encoder training.

Findings

SSL-pretrained encoders achieve comparable accuracy to supervised methods

SSL methods improve data efficiency and reduce annotation costs

The approach advances AI-based disease diagnosis with less labeled data

Abstract

Automated disease diagnosis using medical image analysis relies on deep learning, often requiring large labeled datasets for supervised model training. Diseases like Acute Myeloid Leukemia (AML) pose challenges due to scarce and costly annotations on a single-cell level. Multiple Instance Learning (MIL) addresses weakly labeled scenarios but necessitates powerful encoders typically trained with labeled data. In this study, we explore Self-Supervised Learning (SSL) as a pre-training approach for MIL-based AML subtype classification from blood smears, removing the need for labeled data during encoder training. We investigate the three state-of-the-art SSL methods SimCLR, SwAV, and DINO, and compare their performance against supervised pre-training. Our findings show that SSL-pretrained encoders achieve comparable performance, showcasing the potential of SSL in MIL. This breakthrough offers a cost-effective and data-efficient solution, propelling the field of AI-based disease diagnosis.