Assessing the Trustworthiness of Saliency Maps for Localizing Abnormalities in Medical Imaging
Document Type
Article
Publication Date
11-2021
Abstract
Saliency maps have become a widely used method to make deep learning models more interpretable by providing post-hoc explanations of classifiers through identification of the most pertinent areas of the input medical image. They are increasingly being used in medical imaging to provide clinically plausible explanations for the decisions the neural network makes. However, the utility and robustness of these visualization maps has not yet been rigorously examined in the context of medical imaging. We posit that trustworthiness in this context requires 1) localization utility, 2) sensitivity to model weight randomization, 3) repeatability, and 4) reproducibility. Using the localization information available in two large public radiology datasets, we quantify the performance of eight commonly used saliency map approaches for the above criteria using area under the precision-recall curves (AUPRC) and structural similarity index (SSIM), comparing their performance to various baseline measures. Using our framework to quantify the trustworthiness of saliency maps, we show that all eight saliency map techniques fail at least one of the criteria and are, in most cases, less trustworthy when compared to the baselines. We suggest that their usage in the high-risk domain of medical imaging warrants additional scrutiny and recommend that detection or segmentation models be used if localization is the desired output of the network. Additionally, to promote reproducibility of our findings, we provide the code we used for all tests performed in this work at this link: this https URL.
Source Publication
Radiology: Artificial Intelligence
Recommended Citation
Arun, N., Gaw, N., Singh, P., Chang, K., Aggarwal, M., Chen, B., Hoebel, K., Gupta, S., Patel, J., Gidwani, M., Adebayo, J., Li, M. D., & Kalpathy-Cramer, J. (2021). Assessing the Trustworthiness of Saliency Maps for Localizing Abnormalities in Medical Imaging. Radiology: Artificial Intelligence, 3(6), e200267. https://doi.org/10.1148/ryai.2021200267
Comments
© RSNA, 2021
The "Link to Full Text" on this page opens the article's version of record, hosted at the publisher website.
Downloads of this article are for personal use only. All further rights reserved by the publisher, the Radiological Society of North America, Inc.