Retinal imaging in an era of open science and privacy protection.

Retinal Imaging in an Era of Open Science and Privacy Protection - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Exp Eye Res . Author manuscript; available in PMC: 2026 Jun 1. Published in final edited form as: Exp Eye Res. 2025 Mar 14;255:110341. doi: 10.1016/j.exer.2025.110341 Retinal Imaging in an Era of Open Science and Privacy Protection Nayoon Gim 1, 2, 3 , Marian Blazes 1, 2 , Clara I Sánchez Gutiérrez 4, 5 , Luca Zalunardo 6 , Giulia Corradetti 7, 8 , Tobias Elze 9 , Naoto Honda 10 , Nadia Waheed 11 , Anne Marie Cairns 12 , MValeria Canto-Soler 13 , Amitha Domalpally 14 , Mary Durbin 15 , Daniela Ferrara 16 , Jewel Hu 7 , Prashant Nair 17 , Srinivas R Sadda 7, 8 , Tiarnan D L Keenan 18 , Cecilia S Lee 1, 2 , on behalf of the RIMR Consortium Author information Article notes Copyright and License information 1 Department of Ophthalmology, University of Washington, Seattle, Washington; 2 Roger and Angie Karalis Johnson Retina Center, Seattle, Washington; 3 Department of Bioengineering, University of Washington, Seattle, Washington; 4 Department of Biomedical Engineering and Physics, Amsterdam University Medical Center, Amsterdam; 5 Quantitative Healthcare Analysis (QurAI) Group, Informatics Institute, University of Amsterdam, Amsterdam, the Netherlands; 6 iCare, Raleigh, North Carolina; 7 Doheny Eye Institute, Pasadena, California, USA; 8 Department of Ophthalmology, David Geffen School of Medicine at UCLA, Los Angeles, California, USA; 9 Mass. Eye and Ear, Department of Ophthalmology, Harvard Medical School, Boston, MA 10 NIDEK CO., LTD., Gamagori, Japan; 11 New England Eye Center, Boston, Massachusetts, USA; 12 Optos, Dunfermline, Scotland, United Kingdom; 13 CellSight Ocular Stem Cell and Regeneration Research Program, Department of Ophthalmology, Sue Anschutz-Rodgers Eye Center, University of Colorado, Aurora, CO USA; 14 Wisconsin Reading Center, Department of Ophthalmology and Visual Sciences, University of Wisconsin, Madison, Wisconsin; 15 Topcon Healthcare, Oakland, NJ, USA; 16 Genentech, South San Francisco, CA, USA; 17 Proceedings of the National Academy of Sciences, Washington, DC; 18 Division of Epidemiology and Clinical Applications, National Eye Institute, National Institutes of Health, Bethesda, Maryland ✉ Corresponding Author : Cecilia S. Lee, MD MS, Professor, Klorfine Family Endowed Chair in Ophthalmology, University of Washington, 750 Republican St., Building F, Seattle, WA 98109, leecs2@uw.edu Issue date 2025 Jun. PMC Copyright notice PMCID: PMC12059805  NIHMSID: NIHMS2067928  PMID: 40090567 The publisher's version of this article is available at Exp Eye Res Abstract Artificial intelligence (AI) holds great promise for analyzing complex data to advance patient care and disease research. For example, AI interpretation of retinal imaging may enable the development of noninvasive retinal biomarkers of systemic disease. One potential limitation, however, is government regulation regarding retinal imaging as biometric data, which has been recently under debate in the United States. Although careful regard for patient privacy is key to maintaining trust in the widespread use of AI in healthcare, the designation of retinal imaging as biometric data would greatly restrict retinal biomarker research. There are several reasons why retinal imaging should not be considered biometric data. Unlike images of the iris, high quality images of the retina are more difficult to obtain, requiring specialized training and equipment, and often requiring pupil dilation for optimal quality. In addition, retinal imaging features can vary over time with changes in health status, and retinal images are not currently linked to any large identification databases. While the protection of patient privacy is imperative, there is also a need for large retinal imaging datasets to advance AI research. Given the limitations of retinal imaging as a source of biometric data, the research community should work to advocate for the continued use of retinal imaging in AI research. Keywords: biometric data, retina, retinal imaging, artificial intelligence, big data In an era of rapid technological innovation, artificial intelligence (AI) presents an unprecedented opportunity to revolutionize healthcare science, research, and clinical care. The integration of AI into healthcare holds promise for improved diagnostics, uncovering novel risk factors and drug targets for disease, predicting disease progression, and enabling personalized patient care. AI models have proven capability in analyzing complex medical data to generate insights beyond human reach.( Bhinder et al. 2021 ; Rajpurkar et al. 2022 ; Sharma et al. 2024 ; Ting et al. 2019 ) Retinal imaging could help unearth systemic biomarkers of eye disease and provide a window into overall health. The process of capturing retinal images is rapid, non-invasive, and cost-effective, making retinal imaging an accessible and scalable tool for medical research and diagnostics. However, whether retinal imaging should be considered biometric data has been continuously debated in recent years. In the United States, the policy on retinal imaging as protected health information under the Health Insurance Portability and Accountability Act varies depending on the interpretation by individual academic and clinical institutions. Addressing privacy and ethical concerns is key to maintaining trust in the widespread use of AI in healthcare, and retinal imaging is no exception. However, widespread confusion in the general public regarding the difference between iris scans and retinal scans underscores the need for clear communication and education on the distinct purposes and scanning technologies involved. A recent editorial from the American Academy of Ophthalmology highlighted the critical need for retinal imaging data for enabling AI-based discovery and recommended that retinal imaging should not be considered a biometric identifier.(American Academy of Ophthalmology Board of Trustees, “Special Commentary”) The concept of biometrics evolves over time and eludes simplistic binary classifications. There are several reasons why retinal imaging is unlikely to become a source of biometric data in the future. Widely used biometric data, such as iris scans, fingerprints, and face recognition, are linked to databases specifically for identifying individuals. Some examples of such linked systems are the Automated Biometric Identification System (IDENT) of the United States (US) Department of Defense developed by the US Immigration and Naturalization Service; and the Federal Bureau of Investigation’s fingerprint database.(“Department of Defense, “Automated Biometric Identification System (ABIS) Version 1.2”; Shaffer 2008 ; Yonkers 2006 ) Other notable examples that use iris recognition include India’s UIDAI/Aadhaar program and Pakistan’s National Database & Registration Authority (NADRA), used for citizen registration.(NADRA, “Home NADRA”; UIDAI, “Aadhaar Dashboard”) Unlike iris scans, however, no database that directly links retinal images (such as fundus photographs) to other personally identifiable information, such as social security numbers. Without such a database, the utility of retinal imaging for personal identification remains limited, compared with other established biometric systems. Secondly, retinal images vary significantly in an individual over time due to the changing state of the retina, making it an unreliable biometric method. Common diseases that affect and alter the retina include diabetes, glaucoma, and retinal degenerative disorders. Communicable illnesses such as syphilis, malaria, AIDS, and chickenpox, as well as hereditary diseases such as leukemia, lymphoma, and sickle cell anemia, affect the retina.( Beare et al. 2006 ; Culbertson et al.; Feroze and Gulick 2023 ; Koshy et al. 2015 ; Koundanya and Tripathy 2023 ; Nawaiseh et al. 2022 ; To et al. 2002 ) Pregnancy ( Mackensen et al. 2014 ) as well as chronic health conditions such as congestive heart failure and atherosclerosis can leave a discernible retinal imprint.( Rakusiewicz et al. 2021 ; Wong et al. 2005 , 2008 ) Even diurnal variation can affect funduscopic findings.( Lal et al. 2023 ; Tan et al. 2012 ) Therefore, retinal imaging features commonly vary over time due to a person’s health status and are not static, unlike biomarkers such as iris signatures or fingerprints. Lastly, retinal imaging is suboptimal for rapid biometric identification due to the inherent difficulties in imaging, including the need for subject cooperation and high cost. Commonly used personally identifiable features, such as fingerprints, facial features, and irises, are externally visible and easy to record non-invasively, and even potentially without the knowledge of the subject. Collecting data on fingerprints and facial features does not require medically trained experts. Iris recognition typically uses video cameras with subtle near-infrared illumination to acquire images of the externally visible iris structures. This technology is also easier to implement for security purposes, as evidenced by its commercial use in consumer-orientated devices. For example, Samsung’s Galaxy Note 7 (2016) used iris recognition technology, and Apple’s Vision Pro (2023) uses Optic ID.(Apple Support, “About Optic ID Advanced Technology”; Samsung Newsroom, “[In-Depth Look] Keeping an Eye on Security”) In contrast, retinal imaging involves close contact with specialized medical equipment and can involve cumbersome procedures such as pupil dilation for obtaining images of suitable quality. Cell phone based retinal imaging poses even a greater challenge to obtain quality images. Retinal imaging is highly suited for research applications and has the potential to unlock novel discoveries but is not practical as a real world, rapid identification method. The development and refinement of AI models depend on the availability of extensive, high-quality datasets. The tension between the need for comprehensive datasets to advance AI research and the imperative to safeguard patient privacy represents a significant challenge. This tension underscores the need for ongoing dialog among various stakeholders, including researchers and policymakers, and the general public, as the technology and data protection measures continue to evolve. However, given the current state of retinal imaging technology and its limitations as a potential biometric, we propose that the community should focus on continuing to allow and encourage open science with retinal imaging. Halting this work by treating the retina as a biometric may impose more harm than good by preventing AI-empowered research progress in healthcare. Highlights. Advances in artificial intelligence (AI)-based health research depend on large data. AI analysis of retinal imaging may enable retinal biomarkers of systemic disease. Designating retinal imaging as biometric data may limit available data for AI models. Retinal imaging is less well-suited for biometric identification than the iris. Researchers should advocate for the use of retinal imaging data for AI research. Declaration of interests Cecilia Lee reports financial support was provided by National Institutes of Health OT2OD032644 and P30 DK035816. Srinivas R. Sadda reports a relationship with 4D Molecular Therapeutics Inc that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Alexion that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Allergan that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Alnylam Pharmaceuticals Inc that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Amgen Inc that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Apellis Pharmaceuticals, Inc that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Astellas Pharma US Inc that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Bayer that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Boerhinger Ingelheim that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with ONL Therapeutics that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Biogen Inc that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with iCare Inc. that includes: consulting or advisory and non-financial support. Srinivas R. Sadda reports a relationship with Genentech that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Neurotech that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with IVERIC bio that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Janssen that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Nanoscope that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Samsung Bioepis.Co., Ltd. that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Topcon that includes: consulting or advisory and non-financial support. Srinivas R. Sadda reports a relationship with Optos Inc. that includes: consulting or advisory, non-financial support, and speaking and lecture fees. Srinivas R. Sadda reports a relationship with Oxurion that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Regeneron that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Vertex that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Novartis that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Heidelberg Engineering Inc that includes: consulting or advisory, non-financial support, and speaking and lecture fees. Srinivas R. Sadda reports a relationship with Carl Zeiss Meditec Inc that includes: consulting or advisory and non-financial support. Srinivas R. Sadda reports a relationship with CharacterBio that includes: consulting or advisory. Srinivas R. Sadda reports a relationship with Nidek that includes: non-financial support and speaking and lecture fees. Srinivas R. Sadda reports a relationship with Roche that includes: speaking and lecture fees. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Abbreviations: AI artificial intelligence IDENT Automated Biometric Identification System NADRA Pakistan’s National Database & Registration Authority US United States Footnotes Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Declaration of Interests: SRS reports receiving honoraria as a consultant/advisor for 4DMT, Alexion, Allergan, Alnylam, Amgen Inc., Apellis, Astellas, Bayer, Boerhinger Ingelheim, ONL Therapeutics, Biogen, iCare Inc., Genentech/Roche, Neurotech, Iveric Bio, Janssen, Nanoscope, Samsung Bioepis, Topcon, Optos Inc., Oxurion/ ThromboGenics, Regeneron, Vertex, Novartis, Heidelberg Engineering, Carl Zeiss Meditec, and CharacterBio; lecture fees from Optos, Heidelberg Engineering, Nidek, and Roche; research instruments from Carl Zeiss Meditec, Heidelberg Engineering, Nidek, Optos, iCare, and Topcon. Data Availability Statement: No new data was generated or analyzed in this study. The article is based on a review of existing literature and sources. References “Aadhaar Dashboard.” n.d. Accessed June 14, 2024. https://uidai.gov.in/aadhaar_dashboard/auth_trend.php?auth_id=bioiris . “About Optic ID Advanced Technology.” n.d. Apple Support. Accessed June 14, 2024. https://support.apple.com/en-us/118483 . American Academy of Ophthalmology Board of Trustees. Electronic address: flum@aao.org. 2024. “Special Commentary: Balancing Benefits and Risks: The Case for Retinal Images to Be Considered as Nonprotected Health Information for Research Purposes.” Ophthalmology, August. 10.1016/j.ophtha.2024.07.031. [ DOI ] [ PubMed ] [ Google Scholar ] Beare Nicholas A. V., Taylor Terrie E., Harding Simon P. et al. 2006. “Malarial Retinopathy: A Newly Established Diagnostic Sign in Severe Malaria.” Am J Trop Med Hyg 75 (5): 790–97. [ PMC free article ] [ PubMed ] [ Google Scholar ] Bhinder Bhavneet, Gilvary Coryandar, Madhukar Neel S. et al. 2021. “Artificial Intelligence in Cancer Research and Precision Medicine.” Cancer Discov 11 (4): 900–915. 10.1158/2159-8290.CD-21-0090 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Culbertson WW, Brod RD, Flynn HW Jr. et al. 1991. “Chickenpox-Associated Acute Retinal Necrosis Syndrome.” Ophthalmology 98 (11): 1641–45; discussion 145–46. 10.1016/S0161-6420(91)32061-X [ DOI ] [ PubMed ] [ Google Scholar ] “Department of Defense (DOD) Automated Biometric Identification System (ABIS) Version 1.2.” 2015. May 2015. 10.21236/ada489271. [ DOI ] [ Google Scholar ] Feroze KB, Gulick PG. HIV Retinopathy. [Updated 2023 Jul 17]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2025. Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK470530/ [ PubMed ] [ Google Scholar ] “Home - NADRA.” 2024. NADRA - National Database & Registration Authority. NADRA. January 14, 2024. https://www.nadra.gov.pk/ . [ Google Scholar ] “[In-Depth Look] Keeping an Eye on Security: The Iris Scanner of the Galaxy Note7.” n.d. Accessed June 14, 2024. https://news.samsung.com/global/in-depth-look-keeping-an-eye-on-security-the-iris-scanner-of-the-galaxy-note7 . Koshy Jacob, John M. Joseph, Thomas Satish et al. 2015. “Ophthalmic Manifestations of Acute and Chronic Leukemias Presenting to a Tertiary Care Center in India.” Indian J Ophthalmol 63 (8): 659–64. 10.4103/0301-4738.169789 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Koundanya VV, Tripathy K. Syphilis Ocular Manifestations. [Updated 2023 Aug 25]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2025. Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK558957/ [ PubMed ] [ Google Scholar ] Lal Barsha, Alonso-Caneiro David, Read Scott A. et al. 2023. “Diurnal Changes in Choroidal Optical Coherence Tomography Angiography Indices over 24 Hours in Healthy Young Adults.” Sci Rep 13 (1): 3551. 10.1038/s41598-023-30433-1 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Mackensen Friederike, Paulus Wolfgang E., et al. 2014. “Ocular Changes during Pregnancy.” Dtsch Arztebl Int 111 (33–34): 567–75; quiz 576. 10.3238/arztebl.2014.0567 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Nawaiseh Mohammed, Roto Allaa, Nawaiseh Yara et al. 2022. “Risk Factors Associated with Sickle Cell Retinopathy: Findings from the Cooperative Study of Sickle Cell Disease.” Int J Retina Vitreous 8 (1): 68. 10.1186/s40942-022-00419-8 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Rajpurkar Pranav, Chen Emma, Banerjee Oishi et al. 2022. “AI in Health and Medicine.” Nat Med 28 (1): 31–38. 10.1038/s41591-021-01614-0 [ DOI ] [ PubMed ] [ Google Scholar ] Rakusiewicz Klaudia, Kanigowska Krystyna, Hautz Wojciech et al. 2021. “The Impact of Chronic Heart Failure on Retinal Vessel Density Assessed by Optical Coherence Tomography Angiography in Children with Dilated Cardiomyopathy.” J Clin Med 10 (12): 2659. 10.3390/jcm10122659 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Shaffer Diane. 2008. “Privacy Impact Assessment for the Fingerprint Identification Records System (FIRS) Integrated Automated Fingerprint Identification System (IAFIS) Outsourcing for Noncriminal Justice Purposes - Channeling.” May 5, 2008. https://www.fbi.gov/how-we-can-help-you/more-fbi-services-and-information/freedom-of-information-privacyact/department-of-justice-fbi-privacy-impact-assessments/firs-iafis . Sharma Alok, Lysenko Artem, Jia Shangru et al. 2024. “Advances in AI and Machine Learning for Predictive Medicine.” J Hum Genetics 69, 487–497 (2024). 10.1038/s10038-024-01231-y. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Tan Colin S., Ouyang Yanling, Ruiz Humberto et al. 2012. “Diurnal Variation of Choroidal Thickness in Normal, Healthy Subjects Measured by Spectral Domain Optical Coherence Tomography.” Invest Ophthalmol Vis Sci 53 (1): 261–66. 10.1167/iovs.11-8782 [ DOI ] [ PubMed ] [ Google Scholar ] Ting Daniel Shu Wei, Pasquale Louis R., Peng Lily et al. 2019. “Artificial Intelligence and Deep Learning in Ophthalmology.” Br J Ophthalmol 103 (2): 167–75. 10.1136/bjophthalmol-2018-313173 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] To King W., Thirkill Charles E., Jakobiec Frederick A. et al. 2002. “Lymphoma-Associated Retinopathy.” Ophthalmology 109 (11): 2149–53. https://doi.org/0.1016/s0161-6420(02)01261-7 [ DOI ] [ PubMed ] [ Google Scholar ] Wong Tien Y., Cheung Ning, Islam F. M. Amirul et al. 2008. “Relation of Retinopathy to Coronary Artery Calcification: The Multi-Ethnic Study of Atherosclerosis.” Am J Epidemiol 167 (1): 51–58. 10.1093/aje/kwm256 [ DOI ] [ PubMed ] [ Google Scholar ] Wong Tien Y., Rosamond Wayne, Chang Patricia P. et al. 2005. “Retinopathy and Risk of Congestive Heart Failure.” JAMA 293 (1): 63–69. 10.1001/jama.293.1.63 [ DOI ] [ PubMed ] [ Google Scholar ] Yonkers Steve. 2006. “Privacy Impact Assessment for the Automated Biometric Identification System (IDENT).” July 31, 2006. https://www.dhs.gov/xlibrary/assets/privacy/privacy_pia_usvisit_ident_final.pdf . Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Data Availability Statement No new data was generated or analyzed in this study. The article is based on a review of existing literature and sources. ACTIONS View on publisher site PDF (58.2 KB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Simultaneous functional imaging of neuronal and photoreceptor layers in living human retina#978601
• The Molecular and Cellular Basis of Retinal Diseases#486680
• Optic Nerve Head Strain in Non-glaucoma Subjects#833750
• DiffReID: Discriminative Diffusion Model for Object Re-Identification#1132095
• PEEL: Physics-Enabled Evidential Learning for Identifiable Uncertainty in CT Imaging#1077527