Demonstrating The Risk of Imbalanced Datasets in Chest X-ray Image-based Diagnostics by Prototypical Relevance Propagation

TL;DR

This paper investigates how imbalanced datasets across sources can cause models to learn spurious correlations in chest X-ray diagnostics, emphasizing the need for balanced data and transparent models to improve reliability.

Contribution

It provides a thorough analysis of label imbalance effects in multi-source chest X-ray datasets and demonstrates how balancing sources reduces spurious correlations.

Findings

Imbalanced source domains lead models to exploit source-specific cues.

Balanced datasets improve model transparency and reduce spurious learning.

Using self-explaining models enhances detection of learned biases.

Abstract

The recent trend of integrating multi-source Chest X-Ray datasets to improve automated diagnostics raises concerns that models learn to exploit source-specific correlations to improve performance by recognizing the source domain of an image rather than the medical pathology. We hypothesize that this effect is enforced by and leverages label-imbalance across the source domains, i.e, prevalence of a disease corresponding to a source. Therefore, in this work, we perform a thorough study of the effect of label-imbalance in multi-source training for the task of pneumonia detection on the widely used ChestX-ray14 and CheXpert datasets. The results highlight and stress the importance of using more faithful and transparent self-explaining models for automated diagnosis, thus enabling the inherent detection of spurious learning. They further illustrate that this undesirable effect of learning spurious correlations can be reduced considerably when ensuring label-balanced source domain datasets.