{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T20:49:40Z","timestamp":1792097380689,"version":"4.4.0"},"reference-count":50,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["JMIR AI"],"abstract":"Abstract<\/jats:title>\n \n Background<\/jats:title>\n AI has increasingly been used to generate synthetic images to overcome the scarcity of large, regulatory-compliant medical image datasets. However, their quality and actual utility for algorithm development across deep learning architectures, medical settings, and image types have been poorly assessed.<\/jats:p>\n <\/jats:sec>\n \n Objective<\/jats:title>\n This study aimed to review the use of synthetic images for algorithm development and to assess their quality and effects on downstream algorithms trained on these images, according to AI architecture, medical setting, and image type.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n The authors conducted a systematic review of studies published between January 2020 and September 2025. PubMed, Scopus, IEEE Xplore, and Embase were searched for peer-reviewed studies evaluating synthetic images generated from human medical imaging data. Data were extracted for imaging modalities, generative architectures, image-quality assessment methods, downstream model performance, explainability approaches, reproducibility practices, and risk of bias. Given substantial heterogeneity and incomplete statistical reporting, findings were synthesized descriptively.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n \n A total of 160 studies were included. Generative adversarial network\u2013based architectures dominated image generation (119 studies, 74.4%), although hybrid and diffusion-based approaches became increasingly represented in recent years. Image-quality evaluation was highly heterogeneous: 50 studies included human-reader evidence, and 101 reported an automated image-quality assessment. Synthetic-only training generally underperformed real-data training, including a median Dice similarity coefficient change of \u201310.2% for segmentation. Adding synthetic data to real training sets was associated with median improvements in classification precision (+7.8%), recall (+5.3%),\n F<\/jats:italic>\n 1<\/jats:sub>\n -score (+5.5%), accuracy (+4.3%), and area under the curve (+4.0%), while the median Dice coefficient improvement was +1.2%. Explainability evidence was reported in 68 (42.5%) studies and was rarely validated against expert annotations. Reproducibility reporting was inconsistent, and only 38% (60\/160) of studies were judged to have low overall risk of bias.\n <\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n Unlike previous reviews restricted to particular diseases, imaging modalities, or generative architectures, this review provides a cross-domain quantitative synthesis of both synthetic-image quality assessment and downstream model performance. Synthetic images currently appear more valuable as a complement to real training data than as a replacement, with modest improvements in classification performance, smaller gains in segmentation, and generally poorer performance with synthetic-only training. For researchers, developers, and regulators, these findings highlight the need for standardized multimethod image-quality evaluation, complete reporting, and rigorous external validation before synthetic-image pipelines are translated into clinical AI applications.<\/jats:p>\n <\/jats:sec>","DOI":"10.2196\/81472","type":"journal-article","created":{"date-parts":[[2026,9,8]],"date-time":"2026-09-08T16:40:06Z","timestamp":1788885606000},"page":"e81472-e81472","source":"Crossref","is-referenced-by-count":0,"title":["Quality Assessment and Model Training Impact of Synthetic Images in Medical Diagnostic Imaging: Systematic Review"],"prefix":"10.2196","volume":"5","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-2519-0053","authenticated-orcid":false,"given":"J\u00falia","family":"Folguera-Profit\u00f3s","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9017-4981","authenticated-orcid":false,"given":"Marina","family":"Ramiro-Pareta","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5982-7975","authenticated-orcid":false,"given":"Gerard","family":"Carot-Sans","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-2286-8384","authenticated-orcid":false,"given":"Alba","family":"Jim\u00e9nez-Rueda","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-1160-4808","authenticated-orcid":false,"given":"J\u00falia","family":"Aznar-Roig","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0258-6194","authenticated-orcid":false,"given":"Didier","family":"Dom\u00ednguez-Herrera","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5925-1929","authenticated-orcid":false,"given":"Maria Mercedes","family":"Nogueras","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2613-2696","authenticated-orcid":false,"given":"Esmaeil","family":"Nadimi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7355-7913","authenticated-orcid":false,"given":"Ferran","family":"Torres","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6471-9979","authenticated-orcid":false,"given":"Jordi","family":"Piera-Jim\u00e9nez","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,10,15]]},"reference":[{"issue":"10","key":"R1","doi-asserted-by":"publisher","DOI":"10.2196\/40238","article-title":"Artificial intelligence applications in health care practice: scoping review","volume":"24","author":"Sharma","journal-title":"J Med Internet Res"},{"issue":"1","key":"R2","doi-asserted-by":"publisher","DOI":"10.1186\/s12909-023-04698-z","article-title":"Revolutionizing healthcare: the role of artificial intelligence in clinical practice","volume":"23","author":"Alowais","journal-title":"BMC Med Educ"},{"issue":"11","key":"R3","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s10462-023-10453-z","article-title":"Medical image data augmentation: techniques, comparisons and interpretations","volume":"56","author":"Goceri","journal-title":"Artif Intell Rev"},{"issue":"1","key":"R4","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-023-00811-0","article-title":"Self-supervised learning for medical image classification: a systematic review and implementation guidelines","volume":"6","author":"Huang","journal-title":"NPJ Digit Med"},{"key":"R5","unstructured":"Koutsoubis N Yilmaz Y Ramachandran RP Schabath M Rasool G . Privacy preserving federated learning in medical imaging with uncertainty estimation. arXiv. Preprint posted online on Jun 18, 2024. [doi: 10.48550\/arXiv.2406.12815]"},{"issue":"3","key":"R6","doi-asserted-by":"publisher","DOI":"10.1148\/radiol.232471","article-title":"Generating synthetic data for medical imaging","volume":"312","author":"Koetzier","journal-title":"Radiology"},{"key":"R7","doi-asserted-by":"publisher","DOI":"10.3389\/fdgth.2025.1563991","article-title":"Synthetic data generation: a privacy-preserving approach to accelerate rare disease research","volume":"7","author":"Mendes","journal-title":"Front Digit Health"},{"issue":"15","key":"R8","doi-asserted-by":"publisher","first-page":"6831","DOI":"10.3390\/app14156831","article-title":"Synthetic medical imaging generation with generative adversarial networks for plain radiographs","volume":"14","author":"McNulty","journal-title":"Appl Sci"},{"issue":"1","key":"R9","doi-asserted-by":"publisher","first-page":"60","DOI":"10.1186\/s40537-019-0197-0","article-title":"A survey on image data augmentation for deep learning","volume":"6","author":"Shorten","journal-title":"J Big Data"},{"key":"R10","doi-asserted-by":"publisher","DOI":"10.1016\/j.cmpb.2020.105611","article-title":"Knowledge-guided synthetic medical image adversarial augmentation for ultrasonography thyroid nodule classification","volume":"196","author":"Shi","journal-title":"Comput Methods Programs Biomed"},{"key":"R11","doi-asserted-by":"publisher","DOI":"10.1051\/itmconf\/20213701005","article-title":"Review of medical image synthesis using GAN techniques","volume":"37","author":"Krithika alias Anbu Devi","journal-title":"ITM Web Conf"},{"issue":"7","key":"R12","doi-asserted-by":"publisher","DOI":"10.3390\/s23073440","article-title":"Evaluating synthetic medical images using artificial intelligence with the GAN algorithm","volume":"23","author":"Abdusalomov","journal-title":"Sensors (Basel)"},{"key":"R13","doi-asserted-by":"publisher","DOI":"10.1016\/j.health.2024.100340","article-title":"A systematic review of deep learning data augmentation in medical imaging: recent advances and future research directions","volume":"5","author":"Islam","journal-title":"Healthc Anal"},{"issue":"8","key":"R14","doi-asserted-by":"publisher","first-page":"1175","DOI":"10.1016\/j.acra.2019.12.024","article-title":"Creating artificial images for radiology applications using generative adversarial networks (GANs)\u2013a systematic review","volume":"27","author":"Sorin","journal-title":"Acad Radiol"},{"issue":"1","key":"R15","doi-asserted-by":"publisher","DOI":"10.4103\/jpi.jpi_103_20","article-title":"Generative adversarial networks in digital pathology and histopathological image processing: a review","volume":"12","author":"Jose","journal-title":"J Pathol Inform"},{"issue":"9","key":"R16","doi-asserted-by":"publisher","DOI":"10.2196\/27414","article-title":"Accuracy of using generative adversarial networks for glaucoma detection: systematic review and bibliometric analysis","volume":"23","author":"Saeed","journal-title":"J Med Internet Res"},{"issue":"1","key":"R17","doi-asserted-by":"publisher","DOI":"10.1186\/s13244-022-01315-3","article-title":"Enhancing cancer differentiation with synthetic MRI examinations via generative models: a systematic review","volume":"13","author":"Dimitriadis","journal-title":"Insights Imaging"},{"issue":"2","key":"R18","doi-asserted-by":"publisher","first-page":"137","DOI":"10.1007\/s10278-021-00556-w","article-title":"Systematic review of generative adversarial networks (GANs) for medical image classification and segmentation","volume":"35","author":"Jeong","journal-title":"J Digit Imaging"},{"key":"R19","doi-asserted-by":"publisher","DOI":"10.2174\/0115734056258198230920042358","article-title":"Medical image processing based on generative adversarial networks: a systematic review","volume":"20","author":"Liu","journal-title":"Curr Med Imaging"},{"issue":"8","key":"R20","doi-asserted-by":"publisher","DOI":"10.3390\/children10081372","article-title":"Generative adversarial network (generative artificial intelligence) in pediatric radiology: a systematic review","volume":"10","author":"Ng","journal-title":"Children (Basel)"},{"key":"R21","doi-asserted-by":"publisher","DOI":"10.3389\/fradi.2024.1385742","article-title":"A systematic literature review: deep learning techniques for synthetic medical image generation and their applications in radiotherapy","volume":"4","author":"Sherwani","journal-title":"Front Radiol"},{"key":"R22","doi-asserted-by":"publisher","DOI":"10.1016\/j.actbio.2022.12.009","article-title":"A systematic review and comparison of automated tools for quantification of fibrous networks","volume":"157","author":"de Vries","journal-title":"Acta Biomater"},{"key":"R23","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2025.109834","article-title":"Generative AI for synthetic data across multiple medical modalities: a systematic review of recent developments and challenges","volume":"189","author":"Ibrahim","journal-title":"Comput Biol Med"},{"key":"R24","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2022.106391","article-title":"Data augmentation for medical imaging: a systematic literature review","volume":"152","author":"Garcea","journal-title":"Comput Biol Med"},{"issue":"9","key":"R25","doi-asserted-by":"publisher","DOI":"10.3390\/bioengineering10091078","article-title":"Machine learning for medical image translation: a systematic review","volume":"10","author":"McNaughton","journal-title":"Bioengineering (Basel)"},{"key":"R26","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2022.105382","article-title":"Generative adversarial networks in medical image augmentation: a review","volume":"144","author":"Chen","journal-title":"Comput Biol Med"},{"key":"R27","doi-asserted-by":"publisher","DOI":"10.1016\/j.cmpb.2026.109331","article-title":"A systematic review of generative artificial intelligence techniques for synthetic medical image datasets: quality, models, public availability and applications","volume":"280","author":"Fu","journal-title":"Comput Methods Programs Biomed"},{"issue":"2","key":"R28","doi-asserted-by":"publisher","DOI":"10.1002\/cl2.1230","article-title":"PRISMA2020: an R package and Shiny app for producing PRISMA 2020-compliant flow diagrams, with interactivity for optimised digital transparency and Open Synthesis","volume":"18","author":"Haddaway","journal-title":"Campbell Syst Rev"},{"issue":"1","key":"R29","doi-asserted-by":"publisher","DOI":"10.1186\/s13643-020-01542-z","article-title":"PRISMA-S: an extension to the PRISMA statement for reporting literature searches in systematic reviews","volume":"10","author":"Rethlefsen","journal-title":"Syst Rev"},{"key":"R30","doi-asserted-by":"publisher","DOI":"10.1136\/bmj.l6890","article-title":"Synthesis without meta-analysis (SWiM) in systematic reviews: reporting guideline","volume":"368","author":"Campbell","journal-title":"BMJ"},{"key":"R31","unstructured":"Carot-Sans G Folguera J Torres F et al. Synthetic images for diagnosis and follow-up decision-support algorithms: a systematic review of image generation techniques and validation strategies. PROSPERO (International Prospective Register of Systematic Reviews). 2023. URL: https:\/\/www.crd.york.ac.uk\/prospero\/display_record.php?RecordID=369125 [Accessed 10-02-2025]"},{"key":"R32","unstructured":"STROBE - strengthening the reporting of observational studies in epidemiology. 2025. URL: https:\/\/www.strobe-statement.org\/ [Accessed 16-09-2026]"},{"key":"R33","doi-asserted-by":"publisher","DOI":"10.1136\/bmj-2023-078378","article-title":"TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods","volume":"385","author":"Collins","journal-title":"BMJ"},{"key":"R34","doi-asserted-by":"publisher","DOI":"10.1136\/bmj.i4919","article-title":"ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions","volume":"355","author":"Sterne","journal-title":"BMJ"},{"issue":"12","key":"R35","doi-asserted-by":"publisher","first-page":"4960","DOI":"10.1109\/TMI.2025.3585372","article-title":"3D MedDiffusion: a 3D medical latent diffusion model for controllable and high-quality medical image generation","volume":"44","author":"Wang","journal-title":"IEEE Trans Med Imaging"},{"issue":"10","key":"R36","doi-asserted-by":"publisher","first-page":"3648","DOI":"10.1109\/TMI.2024.3415032","article-title":"MedSyn: text-guided anatomy-aware synthesis of high-fidelity 3-D CT images","volume":"43","author":"Xu","journal-title":"IEEE Trans Med Imaging"},{"issue":"6","key":"R37","doi-asserted-by":"publisher","first-page":"065007","DOI":"10.1088\/1361-6560\/adb9b3","article-title":"Guided synthesis of annotated lung CT images with pathologies using a multi-conditioned denoising diffusion probabilistic model (mDDPM)","volume":"70","author":"Krishna","journal-title":"Phys Med Biol"},{"issue":"4","key":"R38","doi-asserted-by":"publisher","first-page":"045015","DOI":"10.1088\/1361-6560\/adac25","article-title":"Diffusion transformer model with compact prior for low-dose PET reconstruction","volume":"70","author":"Huang","journal-title":"Phys Med Biol"},{"issue":"14","key":"R39","doi-asserted-by":"publisher","DOI":"10.1126\/sciadv.ads4593","article-title":"FairDiffusion: enhancing equity in latent diffusion models via fair Bayesian perturbation","volume":"11","author":"Luo","journal-title":"Sci Adv"},{"issue":"1","key":"R40","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-025-14359-4","article-title":"Clinical evaluation of medical image synthesis: a case study in wireless capsule endoscopy","volume":"15","author":"Gatoula","journal-title":"Sci Rep"},{"issue":"1","key":"R41","doi-asserted-by":"publisher","DOI":"10.1038\/s44172-025-00450-1","article-title":"Scorecard for synthetic medical data evaluation","volume":"4","author":"Zamzmi","journal-title":"Commun Eng"},{"issue":"4","key":"R42","doi-asserted-by":"publisher","DOI":"10.1148\/ryai.240300","article-title":"Checklist for artificial intelligence in medical imaging (CLAIM): 2024 update","volume":"6","author":"Tejani","journal-title":"Radiol Artif Intell"},{"key":"R43","unstructured":"Budding C Eitel F Ritter K Haufe S . MedNeurips_2021. GitHub. URL: https:\/\/github.com\/cebudding\/MedNeurips_2021 [Accessed 14-07-2025]"},{"issue":"1","key":"R44","doi-asserted-by":"publisher","first-page":"133","DOI":"10.2478\/popets-2019-0008","article-title":"Logan: membership inference attacks against generative models","volume":"2019","author":"Hayes","journal-title":"Proc Priv Enhanc Technol"},{"key":"R45","unstructured":"Carlini N Hayes J Nasr M et al. Extracting training data from diffusion models. Presented at: SEC \u201923: Proceedings of the 32nd USENIX Conference on Security Symposium; Aug 9-11, 2023; Anaheim CA, USA. URL: https:\/\/dl.acm.org\/doi\/abs\/10.5555\/3620237.3620531 [Accessed 16-09-2026]"},{"key":"R46","doi-asserted-by":"crossref","unstructured":"Mitchell M Wu S Zaldivar A et al. Model cards for model reporting. Presented at: FAT* \u201919: Proceedings of the Conference on Fairness, Accountability, and Transparency; Jan 29-31, 2019; Atlanta GA, USA. [doi: 10.1145\/3287560.3287596]","DOI":"10.1145\/3287560.3287596"},{"issue":"12","key":"R47","doi-asserted-by":"publisher","first-page":"86","DOI":"10.1145\/3458723","article-title":"Datasheets for datasets","volume":"64","author":"Gebru","journal-title":"Commun ACM"},{"issue":"1","key":"R48","doi-asserted-by":"publisher","first-page":"e64","DOI":"10.1016\/S2589-7500(24)00224-3","article-title":"Tackling algorithmic bias and promoting transparency in health datasets: the STANDING Together consensus recommendations","volume":"7","author":"Alderman","journal-title":"Lancet Digit Health"},{"key":"R49","unstructured":"EQUATOR Network. URL: https:\/\/www.equator-network.org\/ [Accessed 17-06-2025]"},{"key":"R50","unstructured":"2507-synthetic-images-sysrev. GitHub. URL: https:\/\/github.com\/DS3-SisCat\/2507-synthetic-images-sysrev\/ [Accessed 16-09-2026]"}],"container-title":["JMIR AI"],"original-title":[],"language":"en","deposited":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T20:43:32Z","timestamp":1792097012000},"score":1,"resource":{"primary":{"URL":"https:\/\/ai.jmir.org\/2026\/1\/e81472"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,15]]},"references-count":50,"URL":"https:\/\/doi.org\/10.2196\/81472","relation":{},"ISSN":["2817-1705"],"issn-type":[{"value":"2817-1705","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,15]]},"article-number":"v5i10e81472"}}