Fairness and Robustness of CLIP-Based Models for Chest X-rays

TL;DR

This paper evaluates the fairness and robustness of CLIP-based models in chest X-ray classification, revealing performance disparities across patient subgroups and reliance on spurious correlations, highlighting areas for improvement in medical AI fairness.

Contribution

It provides a comprehensive assessment of CLIP-based models' fairness and robustness in radiology, including analysis of embeddings and identification of biases and reliance on shortcuts.

Findings

Performance gaps across age groups

Models rely on spurious features like chest drains

Embeddings reveal sensitive attribute information

Abstract

Motivated by the strong performance of CLIP-based models in natural image-text domains, recent efforts have adapted these architectures to medical tasks, particularly in radiology, where large paired datasets of images and reports, such as chest X-rays, are available. While these models have shown encouraging results in terms of accuracy and discriminative performance, their fairness and robustness in the different clinical tasks remain largely underexplored. In this study, we extensively evaluate six widely used CLIP-based models on chest X-ray classification using three publicly available datasets: MIMIC-CXR, NIH-CXR14, and NEATX. We assess the models fairness across six conditions and patient subgroups based on age, sex, and race. Additionally, we assess the robustness to shortcut learning by evaluating performance on pneumothorax cases with and without chest drains. Our results indicate performance gaps between patients of different ages, but more equitable results for the other attributes. Moreover, all models exhibit lower performance on images without chest drains, suggesting reliance on spurious correlations. We further complement the performance analysis with a study of the embeddings generated by the models. While the sensitive attributes could be classified from the embeddings, we do not see such patterns using PCA, showing the limitations of these visualisation techniques when assessing models. Our code is available at https://github.com/TheoSourget/clip_cxr_fairness