A foundation model for generalizable disease diagnosis in chest X-ray images

TL;DR

CXRBase is a large self-supervised foundation model trained on over a million unlabelled chest X-ray images, designed to improve disease diagnosis accuracy and generalizability across diverse clinical settings.

Contribution

This work introduces CXRBase, a novel foundation model trained with self-supervised learning on unlabelled data, enabling versatile and efficient adaptation to various chest X-ray diagnostic tasks.

Findings

Achieved robust disease classification performance.

Reduced reliance on task-specific labeled data.

Enhanced generalization across clinical environments.

Abstract

Medical artificial intelligence (AI) is revolutionizing the interpretation of chest X-ray (CXR) images by providing robust tools for disease diagnosis. However, the effectiveness of these AI models is often limited by their reliance on large amounts of task-specific labeled data and their inability to generalize across diverse clinical settings. To address these challenges, we introduce CXRBase, a foundational model designed to learn versatile representations from unlabelled CXR images, facilitating efficient adaptation to various clinical tasks. CXRBase is initially trained on a substantial dataset of 1.04 million unlabelled CXR images using self-supervised learning methods. This approach allows the model to discern meaningful patterns without the need for explicit labels. After this initial phase, CXRBase is fine-tuned with labeled data to enhance its performance in disease detection, enabling accurate classification of chest diseases. CXRBase provides a generalizable solution to improve model performance and alleviate the annotation workload of experts to enable broad clinical AI applications from chest imaging.