Representation Disentanglement for Fair Chest X-Ray Diagnosis

2026-10-05

AI Summary

This study addresses the performance disparities across intersectional demographic groups in deep learning-based chest X-ray diagnosis caused by demographic biases. To mitigate this, we propose a single-encoder framework that integrates dual-level decorrelation with prototype-guided cross-group contrastive learning. This approach effectively decouples disease representations from sensitive attributes while preserving intra-class variations. Furthermore, we introduce the DRAR metric to quantify the extent of demographic structure elimination. The proposed mechanism successfully balances fairness and discriminative capability. Experimental results on the CheXpert dataset demonstrate that our method reduces the equalized odds gap to 10.86% and the AUC gap to 5.01%, while improving DRAR by 59.04%, indicating substantial enhancements in model fairness without compromising diagnostic performance.

Abstract

Deep learning has advanced chest X-ray (CXR) diagnosis, yet demographic biases in learned representations may contribute to performance disparities across intersectional groups. We propose a single-encoder framework combining dual-level decorrelation with prototype-guided cross-group contrastive learning to reduce demographic dependence while accounting for within-class variation. We further propose Demographic Representation Alignment Reduction (DRAR), a new metric that quantifies the reduction in demographic structure within disease representations. The framework is evaluated on four classification tasks using 34,809 CheXpert test images across eight intersectional groups, defined by age, sex and ethnicity. Compared with empirical risk minimization (ERM), our method reduces the mean equalized-odds gap from 15.41\% to 10.86\% and the AUC gap from 5.95\% to 5.01\%. Our method achieves a DRAR of 59.04\% relative to ERM, with only a slight decrease in mean AUC. These results demonstrate that representation disentanglement can reduce demographic bias and improve intersectional fairness. Code is available at \url{https://github.com/06Yujie/Fair-Medical-Imaging}.

Problem

• Representation Disentanglement
• Fairness
• Chest X-Ray Diagnosis
• Demographic Bias
• Intersectional Groups

Innovation

• Representation Disentanglement
• Contrastive Learning
• Intersectional Fairness
• Chest X-Ray Diagnosis
• DRAR