Advancing diagnostic performance and clinical usability of neural networks via adversarial training and dual batch normalization

TL;DR

This study shows that adversarial training combined with dual batch normalization improves the interpretability and diagnostic performance of neural networks in medical imaging, matching standard accuracy while enhancing clinical usability.

Contribution

It introduces a novel training approach using adversarial methods and dual batch normalization to improve neural network interpretability and diagnostic accuracy in medical imaging.

Findings

Adversarially trained models have better interpretability according to radiologist ratings.

Dual batch normalization further enhances model performance and interpretability.

Models trained with this method achieve comparable accuracy to standard models on large datasets.

Abstract

Unmasking the decision-making process of machine learning models is essential for implementing diagnostic support systems in clinical practice. Here, we demonstrate that adversarially trained models can significantly enhance the usability of pathology detection as compared to their standard counterparts. We let six experienced radiologists rate the interpretability of saliency maps in datasets of X-rays, computed tomography, and magnetic resonance imaging scans. Significant improvements were found for our adversarial models, which could be further improved by the application of dual batch normalization. Contrary to previous research on adversarially trained models, we found that the accuracy of such models was equal to standard models when sufficiently large datasets and dual batch norm training were used. To ensure transferability, we additionally validated our results on an external test set of 22,433 X-rays. These findings elucidate that different paths for adversarial and real images are needed during training to achieve state of the art results with superior clinical interpretability.