An investigation into the impact of deep learning model choice on sex and race bias in cardiac MR segmentation

TL;DR

This study examines how different deep learning models influence sex and race biases in cardiac MRI segmentation, revealing that model choice significantly affects bias severity and nature, impacting fairness in medical AI applications.

Contribution

It provides a comparative analysis of multiple models showing how model selection impacts bias in medical image segmentation tasks.

Findings

Significant sex bias found in three models

Racial bias present in all models

Bias severity varies with model choice

Abstract

In medical imaging, artificial intelligence (AI) is increasingly being used to automate routine tasks. However, these algorithms can exhibit and exacerbate biases which lead to disparate performances between protected groups. We investigate the impact of model choice on how imbalances in subject sex and race in training datasets affect AI-based cine cardiac magnetic resonance image segmentation. We evaluate three convolutional neural network-based models and one vision transformer model. We find significant sex bias in three of the four models and racial bias in all of the models. However, the severity and nature of the bias varies between the models, highlighting the importance of model choice when attempting to train fair AI-based segmentation models for medical imaging tasks.