{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,15]],"date-time":"2026-07-15T05:07:14Z","timestamp":1784092034892,"version":"3.55.0"},"reference-count":38,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,7,15]],"date-time":"2026-07-15T00:00:00Z","timestamp":1784073600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Med."],"abstract":"<jats:p>Generative artificial intelligence (GenAI) is increasingly explored for medical image synthesis, medical education, and dataset augmentation; however, the detectability and reader-perceived visual authenticity of synthetic chest radiographs generated by accessible multimodal models remain insufficiently understood. This study evaluated synthetic disease-specific chest radiographs generated by gpt-image-2 and gemini-3-pro-image-preview, hereafter referred to as the GPT-image model and the Gemini-image model, respectively, using two generation strategies: text-only generation and image-conditioned generation based on age- and sex-matched normal conditioning radiographs. We included 320 real disease-positive frontal chest radiographs covering cardiomegaly, pneumothorax, pleural effusion, and pneumonia, together with 320 matched normal conditioning radiographs. Synthetic and real disease-positive radiographs were randomized and independently assessed by four radiologists in a blinded single-image reader study, followed by paired evaluation of previously undetected image-conditioned synthetic radiographs. Image-level similarity was assessed using the structural similarity index measure (SSIM), learned perceptual image patch similarity (LPIPS), and Fr\u00e9chet Inception Distance (FID). Image-conditioned generation had significantly lower AI detection rates than text-only generation (34.0% vs. 56.1%) and consistently lower FID values. Synthetic radiographs generated by the GPT-image model were less frequently detected than those generated by the Gemini-image model and showed higher image-level structural and perceptual similarity to the conditioning radiographs. Paired comparison improved detection of previously undetected image-conditioned synthetic radiographs. These findings indicate that image-conditioned GenAI can produce synthetic chest radiographs that may appear visually authentic to radiologists, highlighting the need for transparent labeling, provenance tracking, expert review, and controlled integration of synthetic medical images into healthcare, education, and research workflows.<\/jats:p>","DOI":"10.3389\/fmed.2026.1901949","type":"journal-article","created":{"date-parts":[[2026,7,15]],"date-time":"2026-07-15T04:39:24Z","timestamp":1784090364000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Detectability and healthcare implications of generative AI\u2013synthesized chest radiographs: a blinded radiologist reader study"],"prefix":"10.3389","volume":"13","author":[{"given":"Jinghang","family":"Wang","sequence":"first","affiliation":[{"name":"Department of Radiology, The Second Affiliated Hospital of Xinjiang Medical University","place":["\u00dcr\u00fcmqi, Xinjiang, China"]},{"name":"Department of Radiology, The Second Xiangya Hospital of Central South University","place":["Changsha, Hunan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ruixin","family":"Wang","sequence":"additional","affiliation":[{"name":"Department of Radiology, The Second Xiangya Hospital of Central South University","place":["Changsha, Hunan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qijia","family":"Yi","sequence":"additional","affiliation":[{"name":"Department of Radiology, The First Hospital of Lanzhou University","place":["Lanzhou, Gansu, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hongyong","family":"Tang","sequence":"additional","affiliation":[{"name":"Department of Radiology, Pengan County People\u2019s Hospital","place":["Nanchong, Sichuan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Li","family":"Fan","sequence":"additional","affiliation":[{"name":"Department of Radiology, The Second Xiangya Hospital of Central South University","place":["Changsha, Hunan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shunan","family":"Lin","sequence":"additional","affiliation":[{"name":"Department of Radiology, Pengan County People\u2019s Hospital","place":["Nanchong, Sichuan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wenjing","family":"He","sequence":"additional","affiliation":[{"name":"Department of Radiology, Pengan County People\u2019s Hospital","place":["Nanchong, Sichuan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dan","family":"Peng","sequence":"additional","affiliation":[{"name":"Department of Radiology, Pengan County People\u2019s Hospital","place":["Nanchong, Sichuan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jun-Jie","family":"Yang","sequence":"additional","affiliation":[{"name":"Department of Radiology, The Second Affiliated Hospital of Xinjiang Medical University","place":["\u00dcr\u00fcmqi, Xinjiang, China"]},{"name":"Department of Radiology, The Second Xiangya Hospital of Central South University","place":["Changsha, Hunan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jun","family":"Liu","sequence":"additional","affiliation":[{"name":"Department of Radiology, The Second Affiliated Hospital of Xinjiang Medical University","place":["\u00dcr\u00fcmqi, Xinjiang, China"]},{"name":"Department of Radiology, The Second Xiangya Hospital of Central South University","place":["Changsha, Hunan, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,7,15]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"e232471","DOI":"10.1148\/radiol.232471","article-title":"Generating synthetic data for medical imaging","volume":"312","author":"Koetzier","year":"2024","journal-title":"Radiology"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"105174","DOI":"10.1016\/j.ebiom.2024.105174","article-title":"Synthetically enhanced: unveiling synthetic data's potential in medical imaging research","volume":"104","author":"Khosravi","year":"2024","journal-title":"EBioMedicine"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"eabb7973","DOI":"10.1126\/sciadv.abb7973","article-title":"Breaking medical data sharing boundaries by using synthesized radiographs","volume":"6","author":"Han","year":"2020","journal-title":"Sci Adv"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"2356","DOI":"10.1038\/s41598-023-28175-1","article-title":"Image Turing test and its applications on synthetic chest radiographs by using the progressive growing generative adversarial network","volume":"13","author":"Jang","year":"2023","journal-title":"Sci Rep"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"141","DOI":"10.1038\/s41746-021-00507-3","article-title":"Overcoming barriers to data sharing with medical image generation: a comprehensive evaluation","volume":"4","author":"DuMont","year":"2021","journal-title":"NPJ Digit Med."},{"key":"ref6","article-title":"Generation of anonymous chest radiographs using latent diffusion models for training thoracic abnormality classification systems","author":"Packh\u00e4user","year":"2022"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"102846","DOI":"10.1016\/j.media.2023.102846","article-title":"Diffusion models in medical imaging: a comprehensive survey","volume":"88","author":"Kazerouni","year":"2023","journal-title":"Med Image Anal"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"494","DOI":"10.1038\/s41551-024-01246-y","article-title":"A vision-language foundation model for the generation of realistic chest X-ray images","volume":"9","author":"Bluethgen","year":"2025","journal-title":"Nat Biomed Eng"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"e43110","DOI":"10.2196\/43110","article-title":"What does DALL-E 2 know about radiology?","volume":"25","author":"Adams","year":"2023","journal-title":"J Med Internet Res"},{"key":"ref10","article-title":"RoentGen: Vision-Language Foundation Model for Chest X-ray Generation","author":"Chambon","year":"2022"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"e240597","DOI":"10.1148\/radiol.240597","article-title":"Foundation models in radiology: what, how, why, and why not","volume":"314","author":"Paschali","year":"2025","journal-title":"Radiology"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"80","DOI":"10.4274\/dir.2023.232417","article-title":"Large language models in radiology: fundamentals, applications, ethical considerations, risks, and future directions","volume":"30","author":"Akinci D'Antonoli","year":"2024","journal-title":"Diagn Interv Radiol"},{"key":"ref13","volume-title":"Advances in Knowledge Discovery and Data Mining","author":"Weber","year":"2023"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"e252094","DOI":"10.1148\/radiol.252094","article-title":"The rise of Deepfake medical imaging: radiologists' diagnostic accuracy in detecting ChatGPT-generated radiographs","volume":"318","author":"Tordjman","year":"2026","journal-title":"Radiology"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"015022","DOI":"10.1088\/2632-2153\/ad9a3a","article-title":"Beware of diffusion models for synthesizing medical images-a comparison with GANs in terms of memorizing brain MRI and chest x-ray images","volume":"6","author":"Akbar","year":"2025","journal-title":"Mach Learn Sci Technol"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"601","DOI":"10.1080\/19393555.2025.2547206","article-title":"Adversarial attack detection in secure medical images using deep learning","volume":"35","author":"Kishan","year":"2025","journal-title":"Inform Sec J"},{"key":"ref17","volume-title":"RadEdit: Stress-Testing Biomedical Vision Models via Diffusion Image Editing. Computer Vision \u2013 ECCV","author":"P\u00e9rez-Garc\u00eda","year":"2024"},{"key":"ref18","doi-asserted-by":"crossref","DOI":"10.1109\/ISBI56570.2024.10635417","article-title":"Chest-diffusion: a light-weight text-to-image model for report-to-CXR generation","author":"Huang","year":"2024"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"324","DOI":"10.1038\/s41746-026-02497-6","article-title":"RoentMod: a synthetic chest X-ray modification model to identify and correct image interpretation model shortcuts","volume":"9","author":"Cooke","year":"2026","journal-title":"NPJ Digit Med"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"e0279349","DOI":"10.1371\/journal.pone.0279349","article-title":"Evaluating diagnostic content of AI-generated chest radiography: a multi-center visual Turing test","volume":"18","author":"Myong","year":"2023","journal-title":"PLoS One"},{"key":"ref21","article-title":"Mask-guided attention regulation for anatomically consistent counterfactual CXR synthesis","author":"Zhang","year":"2026"},{"key":"ref22","article-title":"XReal: realistic anatomy and pathology-aware X-ray generation via controllable diffusion model","author":"Hashmi","year":"2024"},{"key":"ref23","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1007\/s42979-021-00720-7","article-title":"Evaluating the clinical realism of synthetic chest X-rays generated using progressively growing GANs","volume":"2","author":"Segal","year":"2021","journal-title":"SN Comput Sci"},{"key":"ref24","first-page":"280","volume-title":"Proceedings of the Fifth Conference on Health, Inference, and Learning; Proceedings of Machine Learning Research: PMLR","author":"Lee","year":"2024"},{"key":"ref25","doi-asserted-by":"publisher","first-page":"15453","DOI":"10.1109\/ACCESS.2025.3531366","article-title":"Disentangled contrastive learning from synthetic matching pairs for targeted chest X-ray generation","volume":"13","author":"Kim","year":"2025","journal-title":"IEEE Access"},{"key":"ref26","doi-asserted-by":"publisher","first-page":"100890","DOI":"10.1016\/j.landig.2025.100890","article-title":"Exploring the potential of generative artificial intelligence in medical image synthesis: opportunities, challenges, and future directions. Lancet digit","volume":"7","author":"Khosravi","year":"2025","journal-title":"Health"},{"key":"ref27","doi-asserted-by":"publisher","first-page":"82","DOI":"10.1109\/MCG.2024.3380168","article-title":"To authenticity, and beyond! Building safe and fair generative AI upon the three pillars of provenance","volume":"44","author":"Collomosse","year":"2024","journal-title":"IEEE Comput Graph Appl"},{"key":"ref28","article-title":"CheXGenBench: a unified benchmark for Fidelity, privacy and utility of synthetic chest radiographs","author":"Dutt","year":"2025"},{"key":"ref29","article-title":"Assessing the distributional Fidelity of synthetic chest X-rays using the embedded characteristic score","author":"Tam","year":"2025"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"18573","DOI":"10.1038\/s41598-022-23081-4","article-title":"DeepFake knee osteoarthritis X-rays from generative adversarial neural networks deceive medical experts and offer augmentation potential to automatic classification","volume":"12","author":"Prezja","year":"2022","journal-title":"Sci Rep"},{"key":"ref31","article-title":"How to fool radiologists with generative adversarial networks? A visual turing test for lung cancer diagnosis","author":"Chuquicusma","year":"2017"},{"key":"ref32","first-page":"19","author":"Prakash","year":"2025"},{"key":"ref33","doi-asserted-by":"publisher","first-page":"e0001277","DOI":"10.1371\/journal.pdig.0001277","article-title":"Diffusion-synthesized chest X-rays improve fairness and diagnostic performance","volume":"5","author":"Rehman","year":"2026","journal-title":"PLoS Digit Health"},{"key":"ref34","doi-asserted-by":"publisher","first-page":"493","DOI":"10.1038\/s41551-021-00751-8","article-title":"Synthetic data in machine learning for medicine and healthcare","volume":"5","author":"Chen","year":"2021","journal-title":"Nat Biomed Eng."},{"key":"ref35","doi-asserted-by":"publisher","first-page":"458","DOI":"10.1038\/s41551-025-01468-8","article-title":"Unconditional latent diffusion models memorize patient imaging data","volume":"10","author":"Dar","year":"2026","journal-title":"Nat Biomed Eng."},{"key":"ref36","doi-asserted-by":"publisher","first-page":"1527","DOI":"10.1007\/s10278-020-00393-3","article-title":"Cybersecurity in PACS and medical imaging: an overview","volume":"33","author":"Eichelberg","year":"2020","journal-title":"J Digit Imaging"},{"key":"ref37","doi-asserted-by":"crossref","DOI":"10.1109\/ISBI56570.2024.10635746","article-title":"Assessing the efficacy of invisible watermarks in AI-generated medical images","author":"Xing","year":"2024"},{"key":"ref38","doi-asserted-by":"publisher","first-page":"771","DOI":"10.1148\/radiology.194.3.7862977","article-title":"Authentication of digital medical images with digital signature technology","volume":"194","author":"Smith","year":"1995","journal-title":"Radiology"}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1901949\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,15]],"date-time":"2026-07-15T04:40:04Z","timestamp":1784090404000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2026.1901949\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,15]]},"references-count":38,"alternative-id":["10.3389\/fmed.2026.1901949"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2026.1901949","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,15]]},"article-number":"1901949"}}