Self-Supervised Pretraining Enables High-Performance Chest X-Ray Interpretation Across Clinical Distributions

Abstract

Chest X-rays (CXRs) are a rich source of information for physicians, essential for disease diagnosis and treatment selection. Recent deep learning models aim to alleviate strain on medical resources and improve patient care by automating the detection of diseases from CXRs. However, shortages of labeled CXRs can pose a serious challenge when training models. Currently, models are generally pretrained on ImageNet, but they often need to then be finetuned on hundreds of thousands of labeled CXRs to achieve high performance. Therefore, the current approach to model development is not viable on tasks with only a small amount of labeled data. An emerging method for reducing reliance on large amounts of labeled data is self-supervised learning (SSL), which uses unlabeled CXR datasets to automatically learn features that can be leveraged for downstream interpretation tasks. In this work, we investigated whether self-supervised pretraining methods could outperform traditional ImageNet pretraining for chest X-ray interpretation. We found that SSL-pretrained models outperformed ImageNet-pretrained models on thirteen different datasets representing high diversity in geographies, clinical settings, and prediction tasks. We thus show that SSL on unlabeled CXR data is a promising pretraining approach for a wide variety of CXR interpretation tasks, enabling a shift away from costly labeled datasets.

Vision–language foundation model for echocardiogram interpretation

“…Given the broad range of data used to train these models, the performance of foundation models are often more robust than with conventional convolutional neural networks 14,15 . In biomedical applications, foundation models have been developed to organize biological [16][17][18] and medical 19 datasets, including modality-specific models for chest X-rays, retinal imaging, wearable waveforms and pathology images [20][21][22][23][24][25] . Training of foundation models on medical imaging has been bottlenecked by dataset size and is often limited to publicly available data that may not represent the range of disease severities and possible presentations.…”