Refining Tuberculosis Detection in CXR Imaging: Addressing Bias in Deep Neural Networks via Interpretability

TL;DR

This paper enhances tuberculosis detection in chest X-ray images by improving deep neural network interpretability and bias mitigation, leading to more reliable and generalizable models aligned with medical expert reasoning.

Contribution

It introduces a pre-training strategy and mixed objective optimization to align model decision regions with medical expertise, improving reliability and generalization.

Findings

Improved alignment of model and expert decision regions.

Maintained high classification accuracy (AUROC).

Enhanced generalization on unseen datasets.

Abstract

Automatic classification of active tuberculosis from chest X-ray images has the potential to save lives, especially in low- and mid-income countries where skilled human experts can be scarce. Given the lack of available labeled data to train such systems and the unbalanced nature of publicly available datasets, we argue that the reliability of deep learning models is limited, even if they can be shown to obtain perfect classification accuracy on the test data. One way of evaluating the reliability of such systems is to ensure that models use the same regions of input images for predictions as medical experts would. In this paper, we show that pre-training a deep neural network on a large-scale proxy task, as well as using mixed objective optimization network (MOON), a technique to balance different classes during pre-training and fine-tuning, can improve the alignment of decision foundations between models and experts, as compared to a model directly trained on the target dataset. At the same time, these approaches keep perfect classification accuracy according to the area under the receiver operating characteristic curve (AUROC) on the test set, and improve generalization on an independent, unseen dataset. For the purpose of reproducibility, our source code is made available online.