RESUMEN
PURPOSE: To evaluate the trustworthiness of saliency maps for abnormality localization in medical imaging. MATERIALS AND METHODS: Using two large publicly available radiology datasets (Society for Imaging Informatics in Medicine-American College of Radiology Pneumothorax Segmentation dataset and Radiological Society of North America Pneumonia Detection Challenge dataset), the performance of eight commonly used saliency map techniques were quantified in regard to (a) localization utility (segmentation and detection), (b) sensitivity to model weight randomization, (c) repeatability, and (d) reproducibility. Their performances versus baseline methods and localization network architectures were compared, using area under the precision-recall curve (AUPRC) and structural similarity index measure (SSIM) as metrics. RESULTS: All eight saliency map techniques failed at least one of the criteria and were inferior in performance compared with localization networks. For pneumothorax segmentation, the AUPRC ranged from 0.024 to 0.224, while a U-Net achieved a significantly superior AUPRC of 0.404 (P < .005). For pneumonia detection, the AUPRC ranged from 0.160 to 0.519, while a RetinaNet achieved a significantly superior AUPRC of 0.596 (P <.005). Five and two saliency methods (of eight) failed the model randomization test on the segmentation and detection datasets, respectively, suggesting that these methods are not sensitive to changes in model parameters. The repeatability and reproducibility of the majority of the saliency methods were worse than localization networks for both the segmentation and detection datasets. CONCLUSION: The use of saliency maps 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.Keywords: Technology Assessment, Technical Aspects, Feature Detection, Convolutional Neural Network (CNN) Supplemental material is available for this article. © RSNA, 2021.
RESUMEN
BACKGROUND: Cytokine-hormone network deregulations underpin pathologies ranging from autoimmune disorders to cancer, but our understanding of these networks in physiological/pathophysiological states remains patchy. We employed Bayesian networks to analyze cytokine-hormone interactions in vivo using murine lactation as a dynamic, physiological model system. RESULTS: Circulatory levels of estrogen, progesterone, prolactin and twenty-three cytokines were profiled in post partum mice with/without pups. The resultant networks were very robust and assembled about structural hubs, with evidence that interleukin (IL)-12 (p40), IL-13 and monocyte chemoattractant protein (MCP)-1 were the primary drivers of network behavior. Network structural conservation across physiological scenarios coupled with the successful empirical validation of our approach suggested that in silico network perturbations can predict in vivo qualitative responses. In silico perturbation of network components also captured biological features of cytokine interactions (antagonism, synergy, redundancy). CONCLUSION: These findings highlight the potential of network-based approaches in identifying novel cytokine pharmacological targets and in predicting the effects of their exogenous manipulation in inflammatory/immune disorders.