{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,16]],"date-time":"2026-04-16T23:11:25Z","timestamp":1776381085993,"version":"3.51.2"},"reference-count":65,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,3,6]],"date-time":"2026-03-06T00:00:00Z","timestamp":1772755200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T00:00:00Z","timestamp":1776384000000},"content-version":"vor","delay-in-days":42,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["npj Digit. Med."],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Chest radiographs (CXRs) are among the most common tests in medicine; automated interpretation may reduce radiologists\u2019 workload and expand access. Deep learning multi-task and foundation models have shown strong CXR interpretation performance but are vulnerable to shortcut learning, where spurious correlations drive decision-making. We introduce RoentMod, a counterfactual image editing framework that generates realistic CXRs with user-specified and synthetic pathology while maintaining the original anatomical features. RoentMod combines an open-source medical image generator (RoentGen) with an image-to-image modification model without retraining. In reader studies of RoentMod-produced images, 93% appeared realistic, 89\u201399% correctly incorporated the specified finding, and all preserved native anatomy comparable to real follow-up CXRs. Using RoentMod, we demonstrate that state-of-the-art multi-task and foundation models frequently exploit off-target pathology as shortcuts, limiting their specificity. Incorporating RoentMod-generated counterfactual images during training mitigated this vulnerability, improving model discrimination across multiple pathologies by 3\u201319% AUC in internal validation and by 1\u201311% for 5 out of 6 tested pathologies in external testing. These findings establish RoentMod as a tool to probe and correct shortcut learning in medical AI. By enabling controlled counterfactual interventions, RoentMod enhances the robustness and interpretability of CXR interpretation models and provides a strategy to improve medical imaging models.<\/jats:p>","DOI":"10.1038\/s41746-026-02497-6","type":"journal-article","created":{"date-parts":[[2026,3,6]],"date-time":"2026-03-06T04:05:41Z","timestamp":1772769941000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["RoentMod: a synthetic chest X-ray modification model to identify and correct image interpretation model shortcuts"],"prefix":"10.1038","volume":"9","author":[{"given":"Lauren H.","family":"Cooke","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Matthias","family":"Jung","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jan M.","family":"Brendel","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Nora M.","family":"Kerkovits","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Borek","family":"Foldyna","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Michael T.","family":"Lu","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Vineet K.","family":"Raghu","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"297","published-online":{"date-parts":[[2026,3,6]]},"reference":[{"key":"2497_CR1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pmed.1002686","volume":"15","author":"P Rajpurkar","year":"2018","unstructured":"Rajpurkar, P. et al. Deep learning for chest radiograph diagnosis: a retrospective comparison of the CheXNeXt algorithm to practicing radiologists. PLOS Med. 15, e1002686 (2018).","journal-title":"PLOS Med."},{"key":"2497_CR2","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-024-45599-z","volume":"15","author":"W Fan","year":"2024","unstructured":"Fan, W. et al. A deep-learning-based framework for identifying and localizing multiple abnormalities and assessing cardiomegaly in chest X-ray. Nat. Commun. 15, 1347 (2024).","journal-title":"Nat. Commun."},{"key":"2497_CR3","doi-asserted-by":"publisher","first-page":"1732","DOI":"10.1200\/JCO.21.01337","volume":"40","author":"A Yala","year":"2022","unstructured":"Yala, A. et al. Multi-Institutional Validation of a Mammography-Based Breast Cancer Risk Model. J. Clin. Oncol. 40, 1732\u20131740 (2022).","journal-title":"J. Clin. Oncol."},{"key":"2497_CR4","doi-asserted-by":"publisher","first-page":"89","DOI":"10.1038\/s41586-019-1799-6","volume":"577","author":"SM McKinney","year":"2020","unstructured":"McKinney, S. M. et al. International evaluation of an AI system for breast cancer screening. Nature 577, 89\u201394 (2020).","journal-title":"Nature"},{"key":"2497_CR5","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pmed.1002683","volume":"15","author":"JR Zech","year":"2018","unstructured":"Zech, J. R. et al. Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: a cross-sectional study. PLOS Med. 15, e1002683 (2018).","journal-title":"PLOS Med."},{"key":"2497_CR6","doi-asserted-by":"publisher","DOI":"10.1186\/s12916-019-1426-2","volume":"17","author":"CJ Kelly","year":"2019","unstructured":"Kelly, C. J., Karthikesalingam, A., Suleyman, M., Corrado, G. & King, D. Key challenges for delivering clinical impact with artificial intelligence. BMC Med. 17, 195 (2019).","journal-title":"BMC Med."},{"key":"2497_CR7","doi-asserted-by":"publisher","first-page":"1981","DOI":"10.1056\/NEJMra2301725","volume":"388","author":"P Rajpurkar","year":"2023","unstructured":"Rajpurkar, P. & Lungren, M. P. The current and future state of AI interpretation of medical images. N. Engl. J. Med. 388, 1981\u20131990 (2023).","journal-title":"N. Engl. J. Med."},{"key":"2497_CR8","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-023-28633-w","volume":"13","author":"S Gaube","year":"2023","unstructured":"Gaube, S. et al. Non-task expert physicians benefit from correct explainable AI advice when reviewing X-rays. Sci. Rep. 13, 1383 (2023).","journal-title":"Sci. Rep."},{"key":"2497_CR9","doi-asserted-by":"publisher","first-page":"837","DOI":"10.1038\/s41591-024-02850-w","volume":"30","author":"F Yu","year":"2024","unstructured":"Yu, F. et al. Heterogeneity and predictors of the effects of AI assistance on radiologists. Nat. Med. 30, 837\u2013849 (2024).","journal-title":"Nat. Med."},{"key":"2497_CR10","doi-asserted-by":"publisher","first-page":"890","DOI":"10.1093\/jamia\/ocaa268","volume":"28","author":"TP Quinn","year":"2021","unstructured":"Quinn, T. P., Senadeera, M., Jacobs, S., Coghlan, S. & Le, V. Trust and medical AI: the challenges we face and the expertise needed to overcome them. J. Am. Med. Inform. Assoc. 28, 890\u2013894 (2021).","journal-title":"J. Am. Med. Inform. Assoc."},{"key":"2497_CR11","doi-asserted-by":"publisher","DOI":"10.1148\/radiol.232778","volume":"310","author":"WB Gefter","year":"2024","unstructured":"Gefter, W. B. et al. Human-AI symbiosis: a path forward to improve chest radiography and the role of radiologists in patient care. Radiology 310, e232778 (2024).","journal-title":"Radiology"},{"key":"2497_CR12","doi-asserted-by":"crossref","unstructured":"Ong Ly, C. et al. Shortcut learning in medical AI hinders generalization: method for estimating AI model generalization without external data. Npj Digit. Med. 7, 124 (2024).","DOI":"10.1038\/s41746-024-01118-4"},{"key":"2497_CR13","doi-asserted-by":"publisher","first-page":"610","DOI":"10.1038\/s42256-021-00338-7","volume":"3","author":"AJ DeGrave","year":"2021","unstructured":"DeGrave, A. J., Janizek, J. D. & Lee, S.-I. AI for radiographic COVID-19 detection selects shortcuts over signal. Nat. Mach. Intell. 3, 610\u2013619 (2021).","journal-title":"Nat. Mach. Intell."},{"key":"2497_CR14","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-39902-7","volume":"14","author":"A Brown","year":"2023","unstructured":"Brown, A. et al. Detecting shortcut learning for fair medical AI using shortcut testing. Nat. Commun. 14, 4314 (2023).","journal-title":"Nat. Commun."},{"key":"2497_CR15","doi-asserted-by":"publisher","first-page":"e406","DOI":"10.1016\/S2589-7500(22)00063-2","volume":"4","author":"JW Gichoya","year":"2022","unstructured":"Gichoya, J. W. et al. AI recognition of patient race in medical imaging: a modelling study. Lancet Digit. Health 4, e406\u2013e414 (2022).","journal-title":"Lancet Digit. Health"},{"key":"2497_CR16","doi-asserted-by":"publisher","first-page":"2226","DOI":"10.1016\/j.jcmg.2021.01.008","volume":"14","author":"VK Raghu","year":"2021","unstructured":"Raghu, V. K., Weiss, J., Hoffmann, U., Aerts, H. J. W. L. & Lu, M. T. Deep learning to estimate biological age from chest radiographs. JACC Cardiovasc. Imaging 14, 2226\u20132236 (2021).","journal-title":"JACC Cardiovasc. Imaging"},{"key":"2497_CR17","doi-asserted-by":"crossref","unstructured":"Weng, N., Pegios, P., Petersen, E., Feragen, A. & Bigdeli, S. Fast Diffusion-Based Counterfactuals for Shortcut Removal and Generation. In Computer Vision \u2013 ECCV 2024 (eds Leonardis, A. et al.) vol. 15144, 338\u2013357 (Springer Nature Switzerland, Cham, 2025).","DOI":"10.1007\/978-3-031-73016-0_20"},{"key":"2497_CR18","first-page":"11450","volume":"35","author":"Z Hu","year":"2022","unstructured":"Hu, Z. et al. Improving multi-task generalization via regularizing spurious correlation. Adv. Neural Inf. Process. Syst. 35, 11450\u201311466 (2022).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"2497_CR19","doi-asserted-by":"publisher","first-page":"488","DOI":"10.1038\/s41586-025-09079-8","volume":"643","author":"D Ma","year":"2025","unstructured":"Ma, D., Pang, J., Gotway, M. B. & Liang, J. A fully open AI foundation model applied to chest radiography. Nature 643, 488\u2013498 (2025).","journal-title":"Nature"},{"key":"2497_CR20","doi-asserted-by":"publisher","first-page":"454","DOI":"10.1148\/radiol.212482","volume":"305","author":"AB Sellergren","year":"2022","unstructured":"Sellergren, A. B. et al. Simplified transfer learning for chest radiography models using less data. Radiology 305, 454\u2013465 (2022).","journal-title":"Radiology"},{"key":"2497_CR21","doi-asserted-by":"publisher","unstructured":"Bommasani, R. et al. On the opportunities and risks of foundation models. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2108.07258 (2022).","DOI":"10.48550\/arXiv.2108.07258"},{"key":"2497_CR22","doi-asserted-by":"publisher","unstructured":"Lu, Y. et al. Fully-Adaptive Feature Sharing in Multi-Task Networks with Applications in Person Attribute Classification. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Honolulu, HI, USA, 1131\u20131140. https:\/\/doi.org\/10.1109\/CVPR.2017.126 (IEEE, 2017).","DOI":"10.1109\/CVPR.2017.126"},{"key":"2497_CR23","doi-asserted-by":"publisher","unstructured":"Chai, J., Lu, S. & Wan, X. Identifying and Mitigating Spurious Correlation in Multi-Task Learning. In 2025 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Nashville, TN, USA, 25698\u201325707 https:\/\/doi.org\/10.1109\/CVPR52734.2025.02393 (IEEE, 2025).","DOI":"10.1109\/CVPR52734.2025.02393"},{"key":"2497_CR24","doi-asserted-by":"crossref","unstructured":"Arun, N. et al. Assessing the trustworthiness of saliency maps for localizing abnormalities in medical imaging. Radiol. Artif. Intell. 3, e200267 (2021).","DOI":"10.1148\/ryai.2021200267"},{"key":"2497_CR25","unstructured":"Pan, Y. & Bareinboim, E. Counterfactual image editing. In Proc. 41st International Conference on Machine Learning (JMLR.org, 2024)."},{"key":"2497_CR26","unstructured":"Kumar, A., Kriz, A., Havaei, M. & Arbel, T. PRISM: High-Resolution & Precise Counterfactual Medical Image Generation using Language-guided Stable Diffusion. In Medical Imaging with Deep Learning (2025)."},{"key":"2497_CR27","doi-asserted-by":"crossref","unstructured":"P\u00e9rez-Garc\u00eda, F. et al. RadEdit: stress-testing biomedical vision models via diffusion image editing. In Computer Vision \u2013 ECCV 2024, (eds Leonardis, A. et al.) 358\u2013376 (Springer Nature, 2025).","DOI":"10.1007\/978-3-031-73254-6_21"},{"key":"2497_CR28","doi-asserted-by":"publisher","unstructured":"Gu, Y. et al. BiomedJourney: counterfactual biomedical image generation by instruction-learning from multimodal patient journeys. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2310.10765 (2023).","DOI":"10.48550\/arXiv.2310.10765"},{"key":"2497_CR29","doi-asserted-by":"crossref","unstructured":"Rombach, R., Blattmann, A., Lorenz, D., Esser, P. & Ommer, B. High-resolution image synthesis with latent diffusion models. In Proc. 2022 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 10674\u201310685 (IEEE, 2022).","DOI":"10.1109\/CVPR52688.2022.01042"},{"key":"2497_CR30","doi-asserted-by":"publisher","first-page":"43","DOI":"10.1016\/j.ejim.2005.07.008","volume":"17","author":"V Novack","year":"2006","unstructured":"Novack, V. et al. Disagreement in the interpretation of chest radiographs among specialists and clinical outcomes of patients hospitalized with suspected pneumonia. Eur. J. Intern. Med. 17, 43\u201347 (2006).","journal-title":"Eur. J. Intern. Med."},{"key":"2497_CR31","doi-asserted-by":"publisher","first-page":"94","DOI":"10.1186\/1471-2253-14-94","volume":"14","author":"M Hammon","year":"2014","unstructured":"Hammon, M. et al. Improving diagnostic accuracy in assessing pulmonary edema on bedside chest radiographs using a standardized scoring approach. BMC Anesthesiol. 14, 94 (2014).","journal-title":"BMC Anesthesiol."},{"key":"2497_CR32","unstructured":"Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B. & Hochreiter, S. GANs trained by a two time-scale update rule converge to a local Nash equilibrium. In Proc. 31st International Conference on Neural Information Processing Systems, 6629\u20136640 (Curran Associates Inc., 2017)."},{"key":"2497_CR33","doi-asserted-by":"crossref","unstructured":"Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J. & Wojna, Z. Rethinking the inception architecture for computer vision. In Proc. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2818\u20132826 (IEEE, 2016).","DOI":"10.1109\/CVPR.2016.308"},{"key":"2497_CR34","doi-asserted-by":"crossref","unstructured":"He, K., Zhang, X., Ren, S. & Sun, J. Deep Residual Learning for Image Recognition. In Proc. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 770\u2013778 (IEEE, 2016).","DOI":"10.1109\/CVPR.2016.90"},{"key":"2497_CR35","unstructured":"Radford, A. et al. Learning transferable visual models from natural language supervision. In Proc. 38th International Conference on Machine Learning, Vol. 139, 8748\u20138763 (PMLR, 2021)."},{"key":"2497_CR36","doi-asserted-by":"publisher","DOI":"10.1186\/s12885-024-12147-3","volume":"24","author":"L Song","year":"2024","unstructured":"Song, L. et al. Investigating causal associations between pneumonia and lung cancer using a bidirectional mendelian randomization framework. BMC Cancer 24, 721 (2024).","journal-title":"BMC Cancer"},{"key":"2497_CR37","doi-asserted-by":"publisher","first-page":"843","DOI":"10.1001\/jama.2019.11456","volume":"322","author":"R Smith-Bindman","year":"2019","unstructured":"Smith-Bindman, R. et al. Trends in use of medical imaging in US health care systems and in Ontario, Canada, 2000-2016. JAMA 322, 843\u2013856 (2019).","journal-title":"JAMA"},{"key":"2497_CR38","doi-asserted-by":"crossref","unstructured":"Bluethgen, C. et al. A vision\u2013language foundation model for the generation of realistic chest X-ray images. Nat. Biomed. Eng. 9, 494\u2013506 (2024).","DOI":"10.1038\/s41551-024-01246-y"},{"key":"2497_CR39","doi-asserted-by":"publisher","unstructured":"Kiraly, A. P. et al. Health AI developer foundations. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2411.15128 (2024).","DOI":"10.48550\/arXiv.2411.15128"},{"key":"2497_CR40","doi-asserted-by":"crossref","unstructured":"Wang, X. et al. ChestX-Ray8: hospital-scale chest X-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. in Proc. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 3462\u20133471 (IEEE, 2017).","DOI":"10.1109\/CVPR.2017.369"},{"key":"2497_CR41","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-019-0322-0","volume":"6","author":"AEW Johnson","year":"2019","unstructured":"Johnson, A. E. W. et al. MIMIC-CXR, a de-identified publicly available database of chest radiographs with free-text reports. Sci. Data 6, 317 (2019).","journal-title":"Sci. Data"},{"key":"2497_CR42","first-page":"590","volume":"33","author":"J Irvin","year":"2019","unstructured":"Irvin, J. et al. CheXpert: a large chest radiograph dataset with uncertainty labels and expert comparison. Proc. AAAI Conf. Artif. Intell. 33, 590\u2013597 (2019).","journal-title":"Proc. AAAI Conf. Artif. Intell."},{"key":"2497_CR43","doi-asserted-by":"publisher","first-page":"101797","DOI":"10.1016\/j.media.2020.101797","volume":"66","author":"A Bustos","year":"2020","unstructured":"Bustos, A., Pertusa, A., Salinas, J.-M. & De La Iglesia-Vay\u00e1, M. PadChest: a large chest x-ray image dataset with multi-label annotated reports. Med. Image Anal. 66, 101797 (2020).","journal-title":"Med. Image Anal."},{"key":"2497_CR44","doi-asserted-by":"crossref","unstructured":"Zeiler, M. D. & Fergus, R. Visualizing and Understanding Convolutional Networks. In Computer Vision \u2013 ECCV 2014 (eds Fleet, D. et al.), 818\u2013833 (Springer International Publishing, 2014).","DOI":"10.1007\/978-3-319-10590-1_53"},{"key":"2497_CR45","doi-asserted-by":"crossref","unstructured":"Selvaraju, R. R. et al. Grad-CAM: visual explanations from deep networks via gradient-based localization. In Proc. 2017 IEEE International Conference on Computer Vision (ICCV) 618\u2013626 (IEEE, 2017).","DOI":"10.1109\/ICCV.2017.74"},{"key":"2497_CR46","doi-asserted-by":"crossref","unstructured":"Ancona, M., Ceolini, E., \u00d6ztireli, C. & Gross, M. Towards better understanding of gradient-based attribution methods for Deep Neural Networks. In 6th International Conference on Learning Representations (2018).","DOI":"10.1007\/978-3-030-28954-6_9"},{"key":"2497_CR47","doi-asserted-by":"publisher","unstructured":"Smilkov, D., Thorat, N., Kim, B., Vi\u00e9gas, F. & Wattenberg, M. SmoothGrad: removing noise by adding noise. In Workshop on Visualization for Deep Learning, 34th International Conference on Machine Learning. https:\/\/doi.org\/10.48550\/arxiv.1706.03825 (2017).","DOI":"10.48550\/arxiv.1706.03825"},{"key":"2497_CR48","doi-asserted-by":"publisher","unstructured":"Simonyan, K., Vedaldi, A. & Zisserman, A. Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps. In Workshop, 2nd International Conference on Learning Representations. https:\/\/doi.org\/10.48550\/arxiv.1312.6034 (2014).","DOI":"10.48550\/arxiv.1312.6034"},{"key":"2497_CR49","unstructured":"Adebayo, J. et al. Sanity Checks for Saliency Maps. In Advances in Neural Information Processing Systems, Vol. 31 (Curran Associates, Inc., 2018)."},{"key":"2497_CR50","unstructured":"Yona, G. & Greenfeld, D. Revisiting Sanity Checks for Saliency Maps. In eXplainable AI approaches for debugging and diagnosis, 35th Conference on Neural Information Processing Systems (2021)."},{"key":"2497_CR51","unstructured":"Boland, C., Goatman, K. A., Tsaftaris, S. A. & Dahdouh, S. There are no shortcuts to anywhere worth going: identifying shortcuts in deep learning models for medical image analysis. In Proc. 7nd International Conference on Medical Imaging with Deep Learning, 131\u2013150 (PMLR, 2024)."},{"key":"2497_CR52","doi-asserted-by":"crossref","unstructured":"Stanley, E. A. M., Souza, R., Wilms, M. & Forkert, N. D. Where, why, and how is bias learned in medical image analysis models? A study of bias encoding within convolutional networks using synthetic data. eBioMedicine 111 (2025).","DOI":"10.1016\/j.ebiom.2024.105501"},{"key":"2497_CR53","doi-asserted-by":"crossref","unstructured":"Wang, R. et al. Drop the shortcuts: image augmentation improves fairness and decreases AI detection of race and other demographics from medical images. eBioMedicine 102, 105501 (2024).","DOI":"10.1016\/j.ebiom.2024.105047"},{"key":"2497_CR54","doi-asserted-by":"publisher","first-page":"1166","DOI":"10.1038\/s41591-024-02838-6","volume":"30","author":"I Ktena","year":"2024","unstructured":"Ktena, I. et al. Generative models improve fairness of medical classifiers under distribution shifts. Nat. Med. 30, 1166\u20131173 (2024).","journal-title":"Nat. Med."},{"key":"2497_CR55","unstructured":"Cohen, J. P. et al. Gifsplanation via latent shift: a simple autoencoder approach to counterfactual generation for chest X-rays. In Proc. Machine Learning Research, Vol. 143, 74\u2013104 (PMLR, 2021)."},{"key":"2497_CR56","doi-asserted-by":"publisher","unstructured":"Seah, J., Tang, J., Kitchen, A. & Seah, J. Generative Visual Rationales. Preprint at https:\/\/doi.org\/10.48550\/arXiv.1804.04539 (2018).","DOI":"10.48550\/arXiv.1804.04539"},{"key":"2497_CR57","doi-asserted-by":"publisher","first-page":"139","DOI":"10.1145\/3422622","volume":"63","author":"I Goodfellow","year":"2020","unstructured":"Goodfellow, I. et al. Generative adversarial networks. Commun. ACM 63, 139\u2013144 (2020).","journal-title":"Commun. ACM"},{"key":"2497_CR58","doi-asserted-by":"publisher","unstructured":"Kingma, D. P. & Welling, M. Auto-Encoding Variational Bayes. In Proc. 2nd International Conference on Learning Representations. https:\/\/doi.org\/10.48550\/arxiv.1312.6114 (2014).","DOI":"10.48550\/arxiv.1312.6114"},{"key":"2497_CR59","doi-asserted-by":"publisher","unstructured":"ScarGAN: chained generative adversarial networks to simulate pathological tissue on cardiovascular MR scans. In Lecture Notes in Computer Science, 343\u2013350. https:\/\/doi.org\/10.1007\/978-3-030-00889-5_39. (Springer International Publishing, 2018).","DOI":"10.1007\/978-3-030-00889-5_39"},{"key":"2497_CR60","doi-asserted-by":"publisher","first-page":"514","DOI":"10.1148\/radiol.2018180887","volume":"290","author":"JCY Seah","year":"2019","unstructured":"Seah, J. C. Y., Tang, J. S. N., Kitchen, A., Gaillard, F. & Dixon, A. F. Chest radiographs in congestive heart failure: visualizing neural network learning. Radiology 290, 514\u2013522 (2019).","journal-title":"Radiology"},{"key":"2497_CR61","doi-asserted-by":"publisher","unstructured":"Wiatrak, M., Albrecht, S. V. & Nystrom, A. Stabilizing generative adversarial networks: a survey. Preprint at https:\/\/doi.org\/10.48550\/arXiv.1910.00927 (2020).","DOI":"10.48550\/arXiv.1910.00927"},{"key":"2497_CR62","unstructured":"Pawlowski, N., Coelho de Castro, D. & Glocker, B. Deep structural causal models for tractable counterfactual inference. In Advances in Neural Information Processing Systems, Vol. 33, 857\u2013869 (Curran Associates, Inc., 2020)."},{"key":"2497_CR63","doi-asserted-by":"publisher","unstructured":"Reinhold, J. C. et al. A structural causal model for MR images of multiple sclerosis. 782\u2013792. https:\/\/doi.org\/10.1007\/978-3-030-87240-3_75 (Springer International Publishing, 2021).","DOI":"10.1007\/978-3-030-87240-3_75"},{"key":"2497_CR64","unstructured":"Cohen, J. P. et al. TorchXRayVision: a library of chest X-ray datasets and models. In Proc. 5th International Conference on Medical Imaging with Deep Learning, Vol. 172, 231\u2013249 (2022)."},{"key":"2497_CR65","doi-asserted-by":"publisher","unstructured":"Huang, G., Liu, Z., Van Der Maaten, L. & Weinberger, K. Q. Densely connected convolutional networks. In Proc. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). https:\/\/doi.org\/10.1109\/cvpr.2017.243 (IEEE, 2017).","DOI":"10.1109\/cvpr.2017.243"}],"container-title":["npj Digital Medicine"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41746-026-02497-6","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41746-026-02497-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41746-026-02497-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,16]],"date-time":"2026-04-16T22:34:01Z","timestamp":1776378841000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41746-026-02497-6"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,6]]},"references-count":65,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["2497"],"URL":"https:\/\/doi.org\/10.1038\/s41746-026-02497-6","relation":{},"ISSN":["2398-6352"],"issn-type":[{"value":"2398-6352","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,6]]},"assertion":[{"value":"3 October 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"17 February 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"6 March 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"The authors declare the following competing interests: Matthias Jung is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) #518480401. Jan M. Brendel is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) #540505270. Michael T. Lu reports funding to his institution from the American Heart Association, Amgen, AstraZeneca, Ionis, Johnson & Johnson Innovation, Kowa Pharmaceuticals America, MedImmune, National Academy of Medicine, National Heart, Lung, and Blood Institute, and Risk Management Foundation of the Harvard Medical Institutions outside the submitted work. Michael also holds common stock in Intel, NVIDIA, and AMD. Borek Foldyna reports institutional research support from NIH\/NHLBI, AstraZeneca, MedImmune, Cleerly, and MedTrace, all outside the submitted work. Author Borek Foldyna is a member of the editorial board of the International Journal of Cardiovascular Imaging and a member of the editorial board for Radiology: Cardiothoracic Imaging. Borek Foldyna was not involved in the journal\u2019s review of, or decisions related to, this manuscript. Vineet K. Raghu is funded by AHA Career Development Award 935176 and NHLBI K01HL168231. Vineet also holds common stock in NVIDIA, Alphabet, Apple, and Amazon. The other authors do not have a competing interest.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"324"}}