Enhancing Network Initialization for Medical AI Models Using Large-Scale, Unlabeled Natural Images

TL;DR

This study demonstrates that self-supervised learning pre-training on large-scale natural images significantly improves medical image analysis performance, surpassing traditional supervised methods and enhancing diagnostic accuracy in chest radiograph AI models.

Contribution

The paper shows that SSL pre-training on unlabeled natural images can outperform supervised pre-training on both natural and medical images for chest radiograph analysis.

Findings

SSL pre-training outperforms ImageNet-based pre-training (P<0.001) across datasets.

SSL pre-training sometimes exceeds supervised pre-training on medical images.

Pre-training strategy critically impacts AI diagnostic accuracy in medical imaging.

Abstract

Pre-training datasets, like ImageNet, have become the gold standard in medical image analysis. However, the emergence of self-supervised learning (SSL), which leverages unlabeled data to learn robust features, presents an opportunity to bypass the intensive labeling process. In this study, we explored if SSL for pre-training on non-medical images can be applied to chest radiographs and how it compares to supervised pre-training on non-medical images and on medical images. We utilized a vision transformer and initialized its weights based on (i) SSL pre-training on natural images (DINOv2), (ii) SL pre-training on natural images (ImageNet dataset), and (iii) SL pre-training on chest radiographs from the MIMIC-CXR database. We tested our approach on over 800,000 chest radiographs from six large global datasets, diagnosing more than 20 different imaging findings. Our SSL pre-training on curated images not only outperformed ImageNet-based pre-training (P<0.001 for all datasets) but, in certain cases, also exceeded SL on the MIMIC-CXR dataset. Our findings suggest that selecting the right pre-training strategy, especially with SSL, can be pivotal for improving artificial intelligence (AI)'s diagnostic accuracy in medical imaging. By demonstrating the promise of SSL in chest radiograph analysis, we underline a transformative shift towards more efficient and accurate AI models in medical imaging.