{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T03:33:44Z","timestamp":1792726424392,"version":"4.5.0"},"reference-count":26,"publisher":"Springer Science and Business Media LLC","issue":"7","license":[{"start":{"date-parts":[[2025,1,29]],"date-time":"2025-01-29T00:00:00Z","timestamp":1738108800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,1,29]],"date-time":"2025-01-29T00:00:00Z","timestamp":1738108800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100005788","name":"Medical University of Vienna","doi-asserted-by":"crossref","award":["Open Access Funding"],"award-info":[{"award-number":["Open Access Funding"]}],"id":[{"id":"10.13039\/501100005788","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100005788","name":"Medical University of Vienna","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100005788","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Eur J Nucl Med Mol Imaging"],"published-print":{"date-parts":[[2025,6]]},"abstract":"<jats:title>Abstract<\/jats:title>\n          <jats:sec>\n            <jats:title>Purpose<\/jats:title>\n            <jats:p>Advancements of deep learning in medical imaging are often constrained by the limited availability of large, annotated datasets, resulting in underperforming models when deployed under real-world conditions. This study investigated a generative artificial intelligence (AI) approach to create synthetic medical images taking the example of bone scintigraphy scans, to increase the data diversity of small-scale datasets for more effective model training and improved generalization.<\/jats:p>\n          <\/jats:sec>\n          <jats:sec>\n            <jats:title>Methods<\/jats:title>\n            <jats:p>We trained a generative model on <jats:sup>99m<\/jats:sup>Tc-bone scintigraphy scans from 9,170 patients in one center to generate high-quality and fully anonymized annotated scans of patients representing two distinct disease patterns: abnormal uptake indicative of (i) bone metastases and (ii) cardiac uptake indicative of cardiac amyloidosis. A blinded reader study was performed to assess the clinical validity and quality of the generated data. We investigated the added value of the generated data by augmenting an independent small single-center dataset with synthetic data and by training a deep learning model to detect abnormal uptake in a downstream classification task. We tested this model on 7,472 scans from 6,448 patients across four external sites in a cross-tracer and cross-scanner setting and associated the resulting model predictions with clinical outcomes.<\/jats:p>\n          <\/jats:sec>\n          <jats:sec>\n            <jats:title>Results<\/jats:title>\n            <jats:p>The clinical value and high quality of the synthetic imaging data were confirmed by four readers, who were unable to distinguish synthetic scans from real scans (average accuracy: 0.48% [95% CI 0.46\u20130.51]), disagreeing in 239 (60%) of 400 cases (Fleiss\u2019 kappa: 0.18). Adding synthetic data to the training set improved model performance by a mean (\u00b1\u2009SD) of 33(\u00b1\u200910)% AUC (<jats:italic>p<\/jats:italic>\u2009&lt;\u20090.0001) for detecting abnormal uptake indicative of bone metastases and by 5(\u00b1\u20094)% AUC (<jats:italic>p<\/jats:italic>\u2009&lt;\u20090.0001) for detecting uptake indicative of cardiac amyloidosis across both internal and external testing cohorts, compared to models without synthetic training data. Patients with predicted abnormal uptake had adverse clinical outcomes (log-rank: <jats:italic>p<\/jats:italic>\u2009&lt;\u20090.0001).<\/jats:p>\n          <\/jats:sec>\n          <jats:sec>\n            <jats:title>Conclusions<\/jats:title>\n            <jats:p>Generative AI enables the targeted generation of bone scintigraphy images representing different clinical conditions. Our findings point to the potential of synthetic data to overcome challenges in data sharing and in developing reliable and prognostic deep learning models in data-limited environments.<\/jats:p>\n          <\/jats:sec>\n          <jats:sec>\n            <jats:title>Graphical abstract<\/jats:title>\n          <\/jats:sec>","DOI":"10.1007\/s00259-025-07091-8","type":"journal-article","created":{"date-parts":[[2025,1,29]],"date-time":"2025-01-29T05:50:04Z","timestamp":1738129804000},"page":"2355-2368","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":11,"title":["Generative artificial intelligence enables the generation of bone scintigraphy images and improves generalization of deep learning models in data-constrained environments"],"prefix":"10.1007","volume":"52","author":[{"given":"David","family":"Haberl","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jing","family":"Ning","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kilian","family":"Kluge","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Katarina","family":"Kumpf","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Josef","family":"Yu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zewen","family":"Jiang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Claudia","family":"Constantino","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alice","family":"Monaci","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maria","family":"Starace","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alexander R.","family":"Haug","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Raffaella","family":"Calabretta","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Luca","family":"Camoni","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Francesco","family":"Bertagna","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Katharina","family":"Mascherbauer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Felix","family":"Hofer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Domenico","family":"Albano","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Roberto","family":"Sciagra","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Francisco","family":"Oliveira","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Durval","family":"Costa","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Christian","family":"Nitsche","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Marcus","family":"Hacker","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4409-8324","authenticated-orcid":false,"given":"Clemens P.","family":"Spielvogel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,1,29]]},"reference":[{"key":"7091_CR1","doi-asserted-by":"publisher","first-page":"e114","DOI":"10.1016\/S2589-7500(23)00203-0","volume":"6","author":"I H\u00e4ggstr\u00f6m","year":"2024","unstructured":"H\u00e4ggstr\u00f6m I, Leithner D, Alv\u00e9n J, et al. Deep learning for [18F]fluorodeoxyglucose-PET-CT classification in patients with lymphoma: a dual-centre retrospective analysis. Lancet Digit Health. 2024;6:e114\u201325.","journal-title":"Lancet Digit Health"},{"key":"7091_CR2","doi-asserted-by":"publisher","first-page":"e251","DOI":"10.1016\/S2589-7500(23)00265-0","volume":"6","author":"CP Spielvogel","year":"2024","unstructured":"Spielvogel CP, Haberl D, Mascherbauer K, et al. Diagnosis and prognosis of abnormal cardiac scintigraphy uptake suggestive of cardiac amyloidosis using artificial intelligence: a retrospective, international, multicentre, cross-tracer development and validation study. Lancet Digit Health. 2024;6:e251\u201360.","journal-title":"Lancet Digit Health"},{"key":"7091_CR3","doi-asserted-by":"publisher","first-page":"e230024","DOI":"10.1148\/ryai.230024","volume":"5","author":"J Wasserthal","year":"2023","unstructured":"Wasserthal J, Breit H-C, Meyer MT, et al. TotalSegmentator: robust segmentation of 104 anatomic structures in CT images. Radiol Artif Intell. 2023;5:e230024.","journal-title":"Radiol Artif Intell"},{"key":"7091_CR4","doi-asserted-by":"publisher","first-page":"1925","DOI":"10.2967\/jnumed.121.263501","volume":"63","author":"KB Girum","year":"2022","unstructured":"Girum KB, Rebaud L, Cottereau A-S, et al. 18F-FDG PET maximum-intensity projections and Artificial Intelligence: a Win-Win Combination to easily measure prognostic biomarkers in DLBCL patients. J Nucl Med. 2022;63:1925\u201332.","journal-title":"J Nucl Med"},{"key":"7091_CR5","doi-asserted-by":"publisher","first-page":"2747","DOI":"10.1038\/s41467-024-46977-3","volume":"15","author":"RJH Miller","year":"2024","unstructured":"Miller RJH, Killekar A, Shanbhag A, et al. Predicting mortality from AI cardiac volumes mass and coronary calcium on chest computed tomography. Nat Commun. 2024;15:2747.","journal-title":"Nat Commun"},{"key":"7091_CR6","doi-asserted-by":"publisher","first-page":"1594","DOI":"10.2967\/jnumed.123.265725","volume":"64","author":"R Gutsche","year":"2023","unstructured":"Gutsche R, Lowis C, Ziemons K, et al. Automated brain tumor detection and segmentation for treatment response assessment using amino acid PET. J Nucl Med. 2023;64:1594\u2013602.","journal-title":"J Nucl Med"},{"key":"7091_CR7","first-page":"AIoa2300030","volume":"1","author":"W Kevin","year":"2023","unstructured":"Kevin W, Eric W, Brandon T, et al. Characterizing the Clinical Adoption of Medical AI Devices through U.S. Insurance claims. NEJM AI. 2023;1:AIoa2300030.","journal-title":"NEJM AI"},{"key":"7091_CR8","doi-asserted-by":"publisher","first-page":"5882","DOI":"10.1038\/s41467-022-33562-9","volume":"13","author":"R Guo","year":"2022","unstructured":"Guo R, Xue S, Hu J, et al. Using domain knowledge for robust and generalizable deep learning-based CT-free PET attenuation and scatter correction. Nat Commun. 2022;13:5882.","journal-title":"Nat Commun"},{"key":"7091_CR9","doi-asserted-by":"publisher","first-page":"e232471","DOI":"10.1148\/radiol.232471","volume":"312","author":"LR Koetzier","year":"2024","unstructured":"Koetzier LR, Wu J, Mastrodicasa D, et al. Generating synthetic data for medical imaging. Radiology. 2024;312:e232471.","journal-title":"Radiology"},{"key":"7091_CR10","doi-asserted-by":"publisher","first-page":"141","DOI":"10.1038\/s41746-021-00507-3","volume":"4","author":"A DuMont Sch\u00fctte","year":"2021","unstructured":"DuMont Sch\u00fctte A, Hetzel J, Gatidis S, et al. Overcoming barriers to data sharing with medical image generation: a comprehensive evaluation. NPJ Digit Med. 2021;4:141.","journal-title":"NPJ Digit Med"},{"key":"7091_CR11","doi-asserted-by":"crossref","unstructured":"Han T, Nebelung S, Haarburger C et al. Breaking medical data sharing boundaries by using synthesized radiographs. Sci Adv. 2020;6.","DOI":"10.1126\/sciadv.abb7973"},{"key":"7091_CR12","doi-asserted-by":"crossref","unstructured":"Ktena I, Wiles O, Albuquerque I et al. April. Generative models improve fairness of medical classifiers under distribution shifts. Nat Med. 2024.","DOI":"10.21203\/rs.3.rs-2976332\/v1"},{"key":"7091_CR13","doi-asserted-by":"publisher","first-page":"115783","DOI":"10.1016\/j.bone.2020.115783","volume":"158","author":"C Ryan","year":"2022","unstructured":"Ryan C, Stoltzfus KC, Horn S, et al. Epidemiology of bone metastases. Bone. 2022;158:115783.","journal-title":"Bone"},{"key":"7091_CR14","doi-asserted-by":"publisher","first-page":"1091","DOI":"10.1056\/NEJMoa2107322","volume":"385","author":"O Sartor","year":"2021","unstructured":"Sartor O, de Bono J, Chi KN, et al. Lutetium-177-PSMA-617 for metastatic castration-resistant prostate cancer. N Engl J Med. 2021;385:1091\u2013103.","journal-title":"N Engl J Med"},{"key":"7091_CR15","doi-asserted-by":"publisher","first-page":"213","DOI":"10.1056\/NEJMoa1213755","volume":"369","author":"C Parker","year":"2013","unstructured":"Parker C, Nilsson S, Heinrich D, et al. Alpha emitter radium-223 and survival in metastatic prostate cancer. N Engl J Med. 2013;369:213\u201323.","journal-title":"N Engl J Med"},{"key":"7091_CR16","doi-asserted-by":"publisher","first-page":"2404","DOI":"10.1161\/CIRCULATIONAHA.116.021612","volume":"133","author":"JD Gillmore","year":"2016","unstructured":"Gillmore JD, Maurer MS, Falk RH, et al. Nonbiopsy diagnosis of Cardiac Transthyretin Amyloidosis. Circulation. 2016;133:2404\u201312.","journal-title":"Circulation"},{"key":"7091_CR17","doi-asserted-by":"publisher","first-page":"1007","DOI":"10.1056\/NEJMoa1805689","volume":"379","author":"MS Maurer","year":"2018","unstructured":"Maurer MS, Schwartz JH, Gundapaneni B, et al. Tafamidis Treatment for patients with transthyretin amyloid cardiomyopathy. N Engl J Med. 2018;379:1007\u201316.","journal-title":"N Engl J Med"},{"key":"7091_CR18","doi-asserted-by":"publisher","first-page":"2060","DOI":"10.1002\/ejhf.2974","volume":"25","author":"P Elliott","year":"2023","unstructured":"Elliott P, Gundapaneni B, Sultan MB, Ines M, Garcia-Pavia P. Improved long-term survival with tafamidis treatment in patients with transthyretin amyloid cardiomyopathy and severe heart failure symptoms. Eur J Heart Fail. 2023;25:2060\u20134.","journal-title":"Eur J Heart Fail"},{"key":"7091_CR19","doi-asserted-by":"publisher","first-page":"106221","DOI":"10.1016\/j.compbiomed.2022.106221","volume":"151","author":"Z Huang","year":"2022","unstructured":"Huang Z, Pu X, Tang G, et al. BS-80K: the first large open-access dataset of bone scan images. Comput Biol Med. 2022;151:106221.","journal-title":"Comput Biol Med"},{"key":"7091_CR20","doi-asserted-by":"publisher","first-page":"1723","DOI":"10.1007\/s00259-016-3415-4","volume":"43","author":"T Van den Wyngaert","year":"2016","unstructured":"Van den Wyngaert T, Strobel K, Kampen WU, et al. The EANM practice guidelines for bone scintigraphy. Eur J Nucl Med Mol Imaging. 2016;43:1723\u201338.","journal-title":"Eur J Nucl Med Mol Imaging"},{"key":"7091_CR21","doi-asserted-by":"publisher","first-page":"2065","DOI":"10.1007\/s12350-019-01760-6","volume":"26","author":"S Dorbala","year":"2019","unstructured":"Dorbala S, Ando Y, Bokhari S, et al. ASNC\/AHA\/ASE\/EANM\/HFSA\/ISA\/SCMR\/SNMMI expert consensus recommendations for multimodality imaging in cardiac amyloidosis: part 1 of 2-evidence base and standardized methods of imaging. J Nucl Cardiol. 2019;26:2065\u2013123.","journal-title":"J Nucl Cardiol"},{"key":"7091_CR22","unstructured":"Karras T, Aittala M, Hellsten J, Laine S, Lehtinen J, Aila T. June. Training Generative Adversarial Networks with Limited Data. arXiv [csCV]. 2020."},{"key":"7091_CR23","doi-asserted-by":"crossref","unstructured":"McInnes L, Healy J, Melville J. February. UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction. arXiv [statML]. 2018.","DOI":"10.21105\/joss.00861"},{"key":"7091_CR24","doi-asserted-by":"crossref","unstructured":"Deng J, Dong W, Socher R, Li L-J, Li K, Fei-Fei L. ImageNet: A large-scale hierarchical image database. In: 2009 IEEE Conference on Computer Vision and Pattern Recognition. IEEE; 2009:248\u2013255.","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"7091_CR25","doi-asserted-by":"publisher","first-page":"1906","DOI":"10.2967\/jnumed.122.264041","volume":"63","author":"C Nitsche","year":"2022","unstructured":"Nitsche C, Mascherbauer K, Calabretta R, et al. Prevalence and outcomes of Cardiac Amyloidosis in All-Comer referrals for bone scintigraphy. J Nucl Med. 2022;63:1906\u201311.","journal-title":"J Nucl Med"},{"key":"7091_CR26","doi-asserted-by":"crossref","unstructured":"Hajianfar G, Sabouri M, Manesh AS et al. Stable diffusion model-based scintigraphy image synthesis: data augmentation toward enhanced multiclass thyroid diagnosis. EUVIP September 2024:1\u20136.","DOI":"10.1109\/EUVIP61797.2024.10772863"}],"container-title":["European Journal of Nuclear Medicine and Molecular Imaging"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00259-025-07091-8.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s00259-025-07091-8\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00259-025-07091-8.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,5,28]],"date-time":"2025-05-28T10:18:07Z","timestamp":1748427487000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s00259-025-07091-8"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,1,29]]},"references-count":26,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2025,6]]}},"alternative-id":["7091"],"URL":"https:\/\/doi.org\/10.1007\/s00259-025-07091-8","relation":{},"ISSN":["1619-7070","1619-7089"],"issn-type":[{"value":"1619-7070","type":"print"},{"value":"1619-7089","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,1,29]]},"assertion":[{"value":"23 October 2024","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"11 January 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"29 January 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and\/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This article does not contain any studies with animals performed by any of the authors.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval"}},{"value":"For the patient cohort from Careggi University Hospital, the patient\u2019s signature on the examination authorization form included consent for the use of anonymized data for research purposes. The patient cohort from West China Hospital is an open-access dataset. For all other cohorts, the requirement to obtain informed consent was waived.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent to participate"}},{"value":"Christian Nitsche reports speaker fees or institutional research grants from Pfizer and advisory board honoraria from Prothena.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}]}}