CoChat
Sign Up
Cited by 23Retinal Imaging and AnalysisArtificial Intelligence in Healthcare and EducationOphthalmology and Visual Health Research

Autonomous Artificial Intelligence in Diabetic Retinopathy Testing—Lessons Learned on Successful Health System Adoption

Open access
Teng, Clare W., Patel, Saawan D., Barkmeier, Andrew J., Liu, T.Y. Alvin, Myung, David, Henderer, Jeffrey, Liu, James, Hansen, Eric +1 more·2026·Ophthalmology Science
DOI 10.1016/j.xops.2025.100935PMC PMC12553049
Open this paper in CoChat — free account

Abstract

Purpose: Artificial intelligence (AI)-aided diabetic retinopathy (DR) testing systems have been commercialized for 5 years, but adoption is still relatively limited. This article aims to summarize the evidence in clinical settings, describe the current state of adoption, and share themes of successful implementation. Design: Evaluation of diagnostic test or technology. Participants: Ophthalmologists. Methods: We performed literature review and conducted interviews with ophthalmologists leading implementation of AI-aided DR testing programs at several academic health systems. The study focused on the 3 currently US Food and Drug Administration-cleared AI systems: LumineticsCore, EyeArt, and AEYE Diagnostic Screening (AEYE-DS), assessing their performance and strategies utilized by health systems to effectively implement this technology in clinics. Main Outcome Measures: Diagnostic accuracy data, ophthalmologist feedback. Results: The literature review found 6 publications reporting diagnostic accuracy data of autonomous AI DR testing in primary care office settings, including 5 for LumineticsCore and 1 for EyeArt. Additional articles, of which 18 were selected for detailed review, addressed impact on patient adherence, health equity, and carbon footprint, as well as cost-effectiveness and workflow efficiency analyses. There were no studies comparing the systems on the same patients. In aggregate, adopters of the AI systems reported average nonmydriatic gradability of 49% to 75% (n = 5), sensitivity 87% to 100% (n = 3), and specificity 60% to 91% (n = 4). Based on public records at the time of writing, both LumineticsCore and EyeArt have >5 academic adopters in the United States. Limited information is available on AEYE-DS given recency of regulatory clearance. Elements of successful implementation include proper site selection, aligning AI tools with primary care clinic workflows, streamlining patient engagement and referrals, and ongoing training of staff. Health systems utilizing AI reported improved Healthcare Effectiveness Data and Information Set measures, health equity, productivity, and patient adherence to follow-up with ophthalmology. Conclusions: Artificial intelligence-aided diabetic eye examinations present a promising solution to facilitate early detection of DR, promote equitable access, and drive down system-level cost of care. Its successful implementation requires addressing technological, operational, and stakeholder engagement challenges. Our study underscores the potential of AI to revolutionize care delivery provided its adoption is strategically managed. Financial Disclosures: The author has no/the authors have no proprietary or commercial interest in any materials discussed in this article.

AI snapshot

Written from the paper’s full text.

TL;DR

This review and expert-interview study summarizes evidence and adoption lessons for the three US FDA-cleared autonomous AI diabetic retinopathy testing systems: LumineticsCore, EyeArt, and AEYE-DS. It reports variable real-world gradability and accuracy, limited comparative data, and identifies top-down stakeholder alignment, workflow redesign, site selection, and performance tracking as keys to successful health system adoption.

Key findings

  • The literature review found 6 publications reporting diagnostic accuracy of autonomous AI DR testing in primary care office settings, including 5 for LumineticsCore and 1 for EyeArt; there were no studies comparing the systems on the same patients.
  • In aggregate, adopters reported average nonmydriatic gradability of 49% to 75% (n = 5), sensitivity 87% to 100% (n = 3), and specificity 60% to 91% (n = 4).
  • Based on public records at the time of writing, both LumineticsCore and EyeArt had >5 academic adopters in the United States, while limited information was available on AEYE-DS given the recency of its regulatory clearance.
  • Elements of successful implementation included proper site selection, aligning AI tools with primary care clinic workflows, streamlining patient engagement and referrals, and ongoing staff training; health systems using AI reported improved HEDIS measures, health equity, productivity, and patient adherence to follow-up with ophthalmology.

Methods

  • Systematic literature search across PubMed, Google Scholars, Scopus, and Web of Science from January 2010 to October 2023 using keywords such as “artificial intelligence diabetic retinopathy evidence,” “autonomous eye exam,” “idx-dr,” “eyeart,” and “aeye-ds”; initial search yielded 106 articles, duplicates were removed, two independent reviewers screened titles/abstracts and full texts, and 18 articles were selected for detailed analysis.
  • Semistructured interviews with 6 ophthalmologists leading implementation of AI DR testing programs at academic health systems; interviews lasted approximately 60 minutes, covered experience, implementation learnings/challenges, practice impact, and patient experience, were recorded and transcribed, and were analyzed thematically.
  • The study focused on the 3 currently FDA-cleared AI systems—LumineticsCore, EyeArt, and AEYE-DS—assessing diagnostic accuracy data and ophthalmologist feedback.

Limitations

  • The paper states that there were no studies comparing the systems on the same patients, that limited information is available on AEYE-DS given recency of regulatory clearance, that the predictive dilation model has not been proven prospectively, and that there has been limited data to prove the observation that referrals are more likely to be disease-positive without change in overall volume after AI rollout.
  • Evident from the text: adoption data rely on publicly available information such as press releases and academic publications, and the adopter table is described as nonexhaustive.
  • Evident from the text: the expert interviews involved only 6 ophthalmologists at academic health systems, and the literature search ended in October 2023, so findings may not generalize to all practice settings or capture newer evidence.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.xops.2025.100935

Cites

38 of 38 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

32 of 69 cited sources checked · 3 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    Initial adoption of AI-aided testing systems is often limited by challenges including lack of trust, resistance to change, training gaps, and uncertainty of the cost-benefit.

    “While AI-aided testing systems hold promise for helping drive down the social and economic burden of DR, initial adoption of new technology is often limited by many challenges including lack of trust, resistance to change, training gaps, and uncertainty of the cost-benefit.9”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [9] Lluch M. Healthcare professionals' organisational barriers to health information technologies-a literature review Int J Med Inform 80 2011 849 862 22000677 source

      The Results section on adaptive optics imaging describes barriers to that modality’s widespread use—time-intensive analysis, limited device availability, and lack of standardization—not barriers to adopting AI-aided testing. Searches for AI adoption, trust, resistance to change, training, and cost-benefit found no passage addressing the claimed challenges; page numbers were unavailable in our copy. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
  • Interpretationcitation

    The 2016 Abramoff et al study was limited by its retrospective nature and therefore could not demonstrate the system's clinical potential.

    “However, this study remained limited given its retrospective nature and therefore could not demonstrate the system’s clinical potential.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [14] Abràmoff M.D. Lou Y. Erginay A. Improved automated detection of diabetic retinopathy on a publicly available dataset through integration of deep learning Invest Ophthalmol Vis Sci 57 2016 5200 5206 27701631 source

      The Introduction discusses an autonomous diabetic retinal-exam product and its CPT code, but does not discuss a 2016 Abramoff study. Searches for the study’s retrospective design and its ability to demonstrate clinical potential found no passage addressing either point. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
  • Premisecitation

    Two years after IDx-DR, EyeArt received 510(k) clearance on the LumineticsCore predicate for detecting more-than-mild DR and vision-threatening DR.

    “Two years later, another AI system, EyeArt (Eyenuk Inc), received 510(k) clearance on the LumineticsCore predicate for detecting more-than-mild DR (mtmDR) and vision-threatening DR.3”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [3] Anon Eyenuk announces FDA clearance for EyeArt autonomous AI system for diabetic retinopathy screening https://www.businesswire.com/news/home/20200805005495/en/Eyenuk-Announces-FDA-Clearance-EyeArt-Autonomous-AI 2020 source

      The Retinopathy > Clinical Features passages discuss diabetic retinopathy and its treatment, not AI screening systems or regulatory clearance (page unavailable). Searches for EyeArt, LumineticsCore, 510(k), and automated screening found no passage in our copy addressing the claimed clearance or its timing. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
  • Premisecitation

    In a 2016 study by Abramoff et al, an earlier version of LumineticsCore reported a referable DR sensitivity of 96.8% and specificity of 87.0% against the Messidor-2 database.

    “In a study by Abramoff et al in 2016, an earlier version of LumineticsCore reported a referable DR sensitivity of 96.8% and specificity of 87.0% against the Messidor-2 database.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [14] Abràmoff M.D. Lou Y. Erginay A. Improved automated detection of diabetic retinopathy on a publicly available dataset through integration of deep learning Invest Ophthalmol Vis Sci 57 2016 5200 5206 27701631 source

      The Introduction discusses an autonomous diabetic retinal exam product, and Current context mentions the IDx-DR device, but neither reports a 2016 study or diagnostic performance. Searches for the stated sensitivity, specificity, Messidor-2, and the 2016 Abramoff study found no passage addressing those results. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
  • Premisecitation

    Prior to autonomous AI testing, all patients were referred to ophthalmology for an annual eye examination.

    “Prior to autonomous AI testing, all patients were referred to ophthalmology for an annual eye examination.59”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [59] American Diabetes Association Professional Practice Committee 12. Retinopathy, neuropathy, and foot care: standards of medical care in Diabetes-2022 Diabetes Care 45 2022 S185 S194 34964887 source

      Section 12.3 recommends eye examinations that are generally repeated annually and permits examination by an ophthalmologist or optometrist. It does not establish that, before autonomous AI testing, all patients were referred to ophthalmology for annual examinations; it also allows 1–2-year screening intervals for some patients. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
  • Classified as probably not a substantial claim. 32
  • May combine two claims. 37

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    LumineticsCore diagnoses according to the ETDRS and Diabetic Retinopathy Clinical Research network grading system to detect ETDRS level 35 or higher or central-involved diabetic macular edema or clinically significant diabetic macular edema.

    “LumineticsCore diagnoses according to the ETDRS and Diabetic Retinopathy Clinical Research network grading system to detect ETDRS level 35 or higher (mild or moderate DR depending on the classification system) or central-involved diabetic macular edema or clinically significant diabetic macular edema.12”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [12] Channa R. Wolf R.M. Simo R. A new approach to staging diabetic eye disease: staging of diabetic retinal neurodegeneration and diabetic macular edema Ophthalmol Sci 4 2024 100420 PMC10818256 38284099 source

      The Validation of the Artificial Intelligence System and Nonmaleficence section identifies the ETDRS and Diabetic Retinopathy Clinical Research Network scales as prognostic standards and discusses center-involved versus clinically significant macular edema. It does not say that LumineticsCore uses those scales or detects ETDRS level 35 or higher. Searches for LumineticsCore, the level-35 threshold, and related diagnostic-system terms found no passage establishing those device-specific claims. (LLM, conf=0.74)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The 2018 Abramoff et al trial was developed under a strict ethical framework.

    “developed under a strict ethical framework.16,17”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [16] Abràmoff M.D. Cunningham B. Patel B. Foundational considerations for artificial intelligence using ophthalmic images Ophthalmology 129 2022 e14 e32 34478784 PMC9175066 source

      The Bioethical Foundation section says an ethical framework is essential for identifying concerns in AI development and describes principles for designing and validating AI systems. The References section identifies the 2018 Abràmoff et al. pivotal trial, but the paper does not say that this specific trial was developed under the framework or characterize its framework as strict. Page numbers were not available in our copy. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
    • cites [17] Char D.S. Abràmoff F.C. Identifying ethical considerations for machine learning healthcare applications Am J Bioeth 20 2020 7 17 PMC7737650 33103967 source

      The Abstract cites Abràmoff et al. (2018) as an example of an autonomous AI diagnostic system, and the References identify its pivotal trial. The pipeline-framework section proposes ethical considerations for machine-learning healthcare applications generally; it does not say the 2018 trial was developed under a strict ethical framework. Searches for the trial and for an ethical framework yielded no passage establishing that claim. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    LumineticsCore is directly connected to the camera and interfaces directly with all available electronic medical record for ordering, billing, and resulting directly.

    “The system, which is directly connected to the camera, interfaces directly with all available electronic medical record (EMR) for ordering, billing, and resulting directly.13”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [13] Anon TRC-NW400 non-mydriatic retinal camera from topcon healthcare https://www.ophthalmologyweb.com/5740-Digital-Retinal-Camera/6016878-TRC-NW400-Non-Mydriatic-Retinal-Camera/ source

      The Implementation, Evaluation, and Oversight section discusses electronic medical record platforms and a hypothetical diabetic-retinopathy diagnostic system, but does not describe LumineticsCore’s connections or interfaces. Searches for camera and EMR integration, ordering, billing, and resulting found no passage addressing those capabilities; page numbers were unavailable in our copy. (LLM, conf=0.74)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Digital Diagnostics filed a 510(k) application for the current version of LumineticsCore based on the previous IDx-DR as predicate device, and it was cleared in May 2021 with the same De Novo special controls in place.

    “Two years later, Digital Diagnostics filed a 510(k) application for the current version of LumineticsCore based on the previous IDx-DR as predicate device, which was subsequently cleared in May 2021, with the same De Novo special controls in place, which now apply to all 510(k)s.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [18] Anon [no title] https://www.accessdata.fda.gov/cdrh_docs/pdf20/K203629.pdf source

      The Methods > Interventions passage identifies IDx-DR as a Digital Diagnostics device and says it received FDA De Novo authorization in 2018. It does not address LumineticsCore, a 510(k) filing or predicate designation, clearance in May 2021, or whether the same special controls applied. (LLM, conf=0.80)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    The 2018 Abramoff et al trial reported a sensitivity of 87.2%, specificity of 90.7%, and gradability of 96.1% for referrable DR.

    “reporting a sensitivity of 87.2%, specificity of 90.7%, and gradability of 96.1% for referrable DR”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [15] Abràmoff M.D. Lavin P.T. Birch M. Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices NPJ Digit Med 1 2018 39 31304320 PMC6550188 source

  • May combine two claims.

    Premisecitation

    In 2018, Abramoff et al published a preregistered prospective clinical trial, a first of its kind in any field of medicine, with n = 900 fully screened patients in primary care by operators without previous experience.

    “However, in 2018, Abramoff et al15 published a preregistered prospective clinical trial, a first of its kind in any field of medicine, with n = 900 fully screened patients in primary care by operators without previous experience”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [15] Abràmoff M.D. Lavin P.T. Birch M. Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices NPJ Digit Med 1 2018 39 31304320 PMC6550188 source

      Methods (Study design and Site initiation) describes a prospective study at 10 primary care sites with 900 enrolled participants, a prespecified protocol, and operators who had no prior ocular-imaging experience before receiving four hours of training. Results (Study population) says only 892 completed all procedures and 819 could be fully analyzed, so 900 were not fully screened. The Abstract calls the system the first FDA-authorized autonomous AI diagnostic system in any field of medicine; it does not make that first-of-its-kind claim about the trial. (LLM, conf=0.91)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Eyenuk's commercial launch of EyeArt began in the European Union in 2016 after receiving a Conformité Européenne Mark, when the system was only approved for investigational use in the United States.

    “Eyenuk’s commercial launch of EyeArt began in the European Union in 2016 after receiving a Conformité Européenne Mark, when the system was only approved for investigational use in the United States.”
    Quote in the paperNot a substantial claim3 of 3 sources checked
    • cites [29] Anon Eyenuk, inc. Expands into Germany with launch of EyeArt® artificial intelligence eye screening for diabetic retinopathy - Eyenuk, Inc. ∼ Artificial Intelligence Eye Screening. Eyenuk, Inc ∼ Artificial Intelligence Eye Screening https://www.eyenuk.com/us-en/articles/blog/eyenuk-inc-expands-into-germany-with-launch-of-eyeart-artificial-intelligence-eye-screening-for-diabetic-retinopathy 2018 source

      The section “Example of an autonomous AI: IDx-DR” discusses FDA clearance in 2018 for IDx-DR, a different system. Searches for EyeArt, Eyenuk, the 2016 EU launch, CE marking, and investigational use in the United States found no passage addressing the claim in our copy. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
    • cites [30] Anon EyeArt used to screen for diabetic retinopathy in over 25 centers across Italy - Eyenuk, Inc. ∼ Artificial Intelligence Eye Screening. Eyenuk, Inc ∼ Artificial Intelligence Eye Screening https://www.eyenuk.com/us-en/success-stories/europe/eyeart-used-to-screen-for-diabetic-retinopathy-in-over-25-centers-across-italy 2020 source

      The Image Acquisition, Imageability, and Diagnosability section discusses EyeArt’s pivotal trial, but not its commercial launch. Searches for the 2016 European Union launch, CE marking, and investigational-use status in the United States found no passage addressing those claims. (LLM, conf=0.75)

      Experimental — accuracy not yet measured
    • cites [31] Anon Top German diabetes centers adopt eyenuk AI technology - eyenuk, inc. ∼ artificial intelligence eye screening. Eyenuk, inc ∼ artificial intelligence eye screening 2021 https://www.eyenuk.com/us-en/articles/eyeart/german-diabetes-centers-adopt-eyenuk-ai source

      The Introduction describes IDx-DR by Digital Diagnostics and its FDA approval, not EyeArt by Eyenuk. Searches of our copy for EyeArt, Eyenuk, the 2016 EU launch, CE marking, and investigational use in the United States found no passage addressing the claim. No page number was provided for the Introduction passage. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Artificial intelligence algorithms have demonstrated a high sensitivity of 95% to 100% in clinical implementation, whereas specificity is relatively lower and with greater variation at 60% to 90%.

    “Artificial intelligence algorithms have demonstrated a high sensitivity of 95% to 100% in clinical implementation, whereas specificity is relatively lower and with greater variation at 60% to 90%.38,40,41,46”
    Quote in the paperSubstantial claim1 of 4 sources checked
    • cites [38] Verbraak F.D. Abramoff M.D. Bausch G.C.F. Diagnostic accuracy of a device for the automated detection of diabetic retinopathy in a primary care setting Diabetes Care 42 2019 651 656 30765436 source

      The Results report 100% sensitivity for vision-threatening diabetic retinopathy, supporting one endpoint of the claimed sensitivity range. But sensitivity for more-than-mild disease was 79.4%, and specificity was 97.8% and 93.8% for the two outcomes—neither within the claimed 60%–90% range. The paper evaluates one device, not algorithms generally. (LLM, conf=0.90)

      Experimental — accuracy not yet measured
    • cites [40] Mehra A.A. Softing A. Guner M.K. Diabetic retinopathy telemedicine outcomes with artificial intelligence-based image analysis, reflex dilation, and image overread Am J Ophthalmol 244 2022 125 132 35970206 source
    • cites [41] Mokhashi N. Grachevskaya J. Cheng L. A comparison of artificial intelligence and human diabetic retinal image interpretation in an urban health system J Diabetes Sci Technol 16 2022 1003 1007 33719599 PMC9264425 source
    • cites [46] Dow E.R. Khan N.C. Chen K.M. AI-Human hybrid workflow enhances teleophthalmology for the detection of diabetic retinopathy Ophthalmol Sci 3 2023 100330 PMC10336195 37449051 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In 2018, Health Canada approved EyeArt, allowing its sale in Canada.

    “In 2018, Health Canada approved EyeArt, allowing its sale in Canada.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [32] Anon. Eyenuk, inc. Receives health Canada approval for EyeArtTM, an AI-enabled cloud-based automated diabetic retinopathy screening software - eyenuk, inc. ∼ artificial intelligence eye screening. Eyenuk, Inc ∼ Artificial Intelligence Eye Screening https://www.eyenuk.com/us-en/articles/blog/eyenuk-inc-receives-health-canada-approval-for-eyeart-an-ai-enabled-cloud-based-automated-diabetic-retinopathy-screening-software 2018 source

      Our copy discusses U.S. FDA authorization of IDx-DR in the introductory passage following Financial Disclosure(s), but does not address EyeArt’s Canadian regulatory status. Searches for EyeArt, Health Canada, and Canadian approval found no passage stating that Health Canada approved EyeArt in 2018 or permitted its sale. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Interpretationcitation

    Amidst rising DR prevalence, autonomous AI testing presents a promising solution to facilitate early detection, promote equitable access, and drive down system-level cost of care.

    “Amidst rising DR prevalence, autonomous AI testing presents a promising solution to facilitate early detection, promote equitable access, and drive down system-level cost of care.42,48,67”
    Quote in the paperSubstantial claim3 of 3 sources checked
    • cites [42] Fuller S.D. Hu J. Liu J.C. Five-year cost-effectiveness modeling of primary care-based, nonmydriatic automated retinal image analysis screening among low-income patients with diabetes J Diabetes Sci Technol 16 2022 415 427 33124449 PMC8861785 source

      Section 12.3 identifies FDA-authorized AI systems as an alternative for diabetic-retinopathy screening and explains why prompt detection matters. It attributes improved access and possible cost savings to retinal photography with remote expert reading, not specifically to autonomous AI, and says the benefits and optimal use of AI screening remain undetermined. The cited passages do not establish rising retinopathy prevalence. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites [48] Huang J. Wang J. Channa R. Autonomous artificial intelligence exams are associated with higher adherence to diabetic retinopathy testing in an integrated healthcare system Invest Ophthalmol Vis Sci 64 2023 212 source
    • cites [67] Huang X. Wang H. She C. Artificial intelligence promotes the diagnosis and screening of diabetic retinopathy Front Endocrinol 13 2022 946915 PMC9559815 36246896 source
    • Rests on an assumption: Vulnerable populations served by Federally Qualified Health Centers lack access to specialist resources.

  • May combine two claims.

    Premisecitation

    In the pivotal trial that led to the clearance of LumineticsCore, 76.4% of participants received results without pharmacologic dilation, and an additional 23.6% became gradable with dilation.

    “In the pivotal trial that led to the clearance of LumineticsCore, 76.4% of participants received results without pharmacologic dilation, and an additional 23.6% became gradable with dilation.15”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [15] Abràmoff M.D. Lavin P.T. Birch M. Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices NPJ Digit Med 1 2018 39 31304320 PMC6550188 source

      The Results (AI system characteristics; page unavailable) report that 76.4% did not require pharmacologic dilation and 23.6% required it to obtain an AI-system disease-level output, supporting the numerical claim. The Discussion says the trial helped lead to FDA authorization, but the paper identifies the system as IDx-DR, not LumineticsCore; our copy does not establish that product-name connection. (LLM, conf=0.84)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Findingnone given

    Based on public records at the time of writing, both LumineticsCore and EyeArt have more than 5 academic adopters in the United States.

    “Based on public records at the time of writing, both LumineticsCore and EyeArt have >5 academic adopters in the United States.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Additional articles, of which 18 were selected for detailed review, addressed impact on patient adherence, health equity, and carbon footprint, as well as cost-effectiveness and workflow efficiency analyses.

    “Additional articles, of which 18 were selected for detailed review, addressed impact on patient adherence, health equity, and carbon footprint, as well as cost-effectiveness and workflow efficiency analyses.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The literature review found 6 publications reporting diagnostic accuracy data of autonomous AI DR testing in primary care office settings, including 5 for LumineticsCore and 1 for EyeArt.

    “The literature review found 6 publications reporting diagnostic accuracy data of autonomous AI DR testing in primary care office settings, including 5 for LumineticsCore and 1 for EyeArt.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    The final selection was 18 articles for detailed analysis.

    “resulting in a final selection of 18 articles for detailed analysis.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    Limited information is available on AEYE-DS given the recency of its regulatory clearance.

    “Limited information is available on AEYE-DS given recency of regulatory clearance.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Findingnone given

    AI-aided diabetic retinopathy testing systems have been commercialized for 5 years, but adoption is still relatively limited.

    “Artificial intelligence (AI)–aided diabetic retinopathy (DR) testing systems have been commercialized for 5 years, but adoption is still relatively limited.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingnone given

    Anecdotally, ophthalmologists have observed that patients referred to clinics are more likely to be disease-positive without change in overall volume after AI program roll-out.

    “Anecdotally, ophthalmologists have observed that patients referred to clinics are more likely to be disease-positive without change in overall volume after AI program roll-out.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingnone given

    Universally, leaders in ophthalmology departments have leaned into the effort to implement these systems.

    “Universally, leaders in ophthalmology departments have leaned into the effort to implement these systems.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Today, LumineticsCore is utilized in >1000 US clinical sites spanning both academic health systems and community-based health organizations including Stanford Medicine, Johns Hopkins Health System, Cahaba Medical Care, and LabCorp, with expansion to sites outside the United States as well.
    • Rests on an assumption: To date, EyeArt has been utilized in the United States by institutions such as Temple Health, University of Utah Health, and Cedars Sinai Health System.
    • Rests on an assumption: To our best knowledge, AEYE-DS has been implemented in select community-based health organizations in the United States, such as the Bethesda Health Clinic.

  • May combine two claims.

    Findingown data

    When a health system pursued implementation in a research context without top-down support, research staff tested patients without disrupting the regular visit workflow, and after the study the camera was returned and the system was not implemented.

    “In contrast, when a health system pursued implementation in a research context without top-down support, research staff visited primary care offices and tested patients without disrupting the regular visit workflow. Upon conclusion of the research study, the camera was returned to the manufacturer and the system was not implemented.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Based on publicly available information, both LumineticsCore and EyeArt each have around 5 academic adopters in the United States today.

    “Based on publicly available information (from press release, academic publications), both LumineticsCore and EyeArt each have around 5 academic adopters in the United States today (Table 1, Table 2).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Today, LumineticsCore is utilized in >1000 US clinical sites spanning both academic health systems and community-based health organizations including Stanford Medicine, Johns Hopkins Health System, Cahaba Medical Care, and LabCorp, with expansion to sites outside the United States as well.
    • Rests on an assumption: To date, EyeArt has been utilized in the United States by institutions such as Temple Health, University of Utah Health, and Cedars Sinai Health System.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Ophthalmology departments contribute numerous peer-reviewed articles and abstracts to describe their practical experience since 2018.

    “These departments also contribute numerous peer-reviewed articles and abstracts to describe their practical experience since 2018 (Table 3).”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Today, LumineticsCore is utilized in >1000 US clinical sites spanning both academic health systems and community-based health organizations including Stanford Medicine, Johns Hopkins Health System, Cahaba Medical Care, and LabCorp, with expansion to sites outside the United States as well.
    • Rests on an assumption: To date, EyeArt has been utilized in the United States by institutions such as Temple Health, University of Utah Health, and Cedars Sinai Health System.
    • Rests on an assumption: To our best knowledge, AEYE-DS has been implemented in select community-based health organizations in the United States, such as the Bethesda Health Clinic.

  • May combine two claims.

    Findingown data

    The department collecting payment for autonomous AI eye examinations varies across health systems; among adopters interviewed, about half attribute payments to the primary care department and the other half to ophthalmology.

    “Interestingly, the department collecting the payment varies across health systems. Among the adopters we interviewed with, about half attribute payments to the primary care department, and the other half to ophthalmology.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Two hospitals became interested in autonomous AI testing of DR to improve primary care HEDIS scores; with this objective articulated and priorities set, implementation followed as a coordinated effort between ophthalmology and primary care.

    “For instance, 2 hospitals became interested in autonomous AI testing of DR as an effort to improve primary care HEDIS scores. With this objective articulated and priorities set, implementation followed as a coordinated effort between ophthalmology and primary care.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In terms of staffing, 80% of the health systems did not add dedicated full-time equivalent to launch the AI program, although more than half of adopters interviewed expressed "wish we could."

    “In terms of staffing, 80% of the health systems did not add dedicated full-time equivalent to launch the AI program, although more than half of adopters we interviewed expressed “wish we could.””
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    The study adhered to the Declaration of Helsinki and informed consent was waived and was institutional review board exempt.

    “Our study adhered to the Declaration of Helsinki and informed consent was waived and was institutional review board exempt.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisenone given

    Currently, there are 3 commercially available US FDA-cleared systems for DR testing: LumineticsCore, EyeArt, and AEYE-DS.

    “Currently, there are 3 commercially available US FDA–cleared systems for DR testing: LumineticsCore (previously IDx-DR) by Digital Diagnostics, EyeArt by Eyenuk, Inc, and AEYE-DS by AEYE Health, Inc.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Successful implementation of AI-aided diabetic eye examinations requires addressing technological, operational, and stakeholder engagement challenges.

    “Its successful implementation requires addressing technological, operational, and stakeholder engagement challenges.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The experiences and perspectives of the six interviewed ophthalmologists are representative of successful AI DR testing implementation across academic health systems.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisenone given

    In late 2022, AEYE-DS was cleared through 510(k) using the same predicate as LumineticsCore.

    “In late 2022, a third system, AEYE Diagnostic Screening (AEYE-DS), was also cleared through 510(k) using the same predicate.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    Today, DR remains the only indication in medicine where an AI system can provide an image analysis decision without clinician interpretation.

    “Today, DR remains the only indication in medicine where an AI system can provide an image analysis decision without clinician interpretation.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Imaging-based diagnostics is one of the earliest frontiers of AI development and has been cited by healthcare professionals as the most common application of AI solutions in health care today.

    “Imaging-based diagnostics is one of the earliest frontiers of AI development and has been cited by healthcare professionals as the most common application of AI solutions in health care today.1”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [1] Spatharou A. Hieronimus S. Jenkins J. Transforming healthcare with AI: The impact on the workforce and organizations 2020 McKinsey & Company https://www.mckinsey.com/industries/healthcare/our-insights/transforming-healthcare-with-ai

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    Digital Diagnostics is the first autonomous AI company to receive the US FDA De Novo pathway clearance for DR testing.

    “Digital Diagnostics is the first autonomous AI company to receive the US FDA De Novo pathway clearance for DR testing.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisenone given

    LumineticsCore is an AI diagnostic system designed to autonomously diagnose patients for mtmDR, central-involved diabetic macular edema traditionally diagnosed with OCT, and clinically significant diabetic macular edema diagnosed with fundus photographs, using images acquired with a Topcon TRC-NW400 camera.

    “Their product, LumineticsCore (originally known as IDx-DR), is an AI diagnostic system designed to autonomously diagnose patients for mtmDR, central-involved diabetic macular edema traditionally diagnosed with OCT, and clinically significant diabetic macular edema diagnosed with fundus photographs, using images acquired with a Topcon TRC-NW400 camera (Topcon).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    Two independent reviewers screened titles and abstracts for relevance, followed by full-text assessment.

    “Two independent reviewers (C.T. and S.P.) screened titles and abstracts for relevance, followed by full-text assessment.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    A systematic search was conducted across electronic databases including PubMed, Google Scholars, Scopus, and Web of Science.

    “A systematic search was conducted across electronic databases, including PubMed, Google Scholars, Scopus, and Web of Science.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Without a screening program, all patients diagnosed with type 1 diabetic mellitus and type 2 diabetic mellitus are referred for an annual eye examination.

    “Without a screening program, all patients diagnosed with type 1 diabetic mellitus and type 2 diabetic mellitus are referred for an annual eye examination;46”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [46] Dow E.R. Khan N.C. Chen K.M. AI-Human hybrid workflow enhances teleophthalmology for the detection of diabetic retinopathy Ophthalmol Sci 3 2023 100330 PMC10336195 37449051 source

  • May combine two claims.

    Interpretationcitation

    The accuracy and low dilation requirement distinguish the AEYE-DS system from the other 2 commercially available systems.

    “Though the enhanced efficiency and potential pitfalls in clinical workflows of taking 1 image versus 2 images per eye have yet to be studied, the accuracy and low dilation requirement distinguish the AEYE-DS system from the other 2 commercially available systems.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [37] Anon. CTG labs - NCBI https://clinicaltrials.gov/study/NCT04612868 source

  • May combine two claims.

    Interpretationreasoning

    When autonomous AI performs the first round of testing and refers only patients with positive or ungradable examinations, ophthalmologists can focus attention on those most likely to have disease and patient access to appointments and treatment improves.

    “When autonomous AI performs the first round of testing and refers only the patients with positive or ungradable examinations, this allows ophthalmologists to focus attention on those who are most likely to have disease and improve patient access to appointments and treatment.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    In 2021, Centers for Medicare and Medicaid Services established relative value units for autonomous AI eye examination and a national payment rate.

    “In 2021, Centers for Medicare and Medicaid Services established relative value units for this service and national payment rate.57”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [57] Larson P.M. Medicare: what's new for 2022 https://www.reviewofophthalmology.com/article/medicare-whats-new-for-2022 2022

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisecitation

    AEYE-DS requires only 1 image per eye (macula-centered color fundus images with ≥45° field of view) for diagnosis and requires dilation in <1% of patients.

    “The system requires only 1 image per eye (ie, macula-centered color fundus images with ≥45° field of view) for diagnosis and requires dilation in <1% of patients.37”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [37] Anon. CTG labs - NCBI https://clinicaltrials.gov/study/NCT04612868 source

  • May combine two claims.

    Premisecitation

    A 2023 study by Najac et al found that overall image quality in telemedical general practitioner-based diabetic eye disease testing with EyeArt was sufficient and would be accepted by both medical assistants and patients in most cases.

    “Their study found that the overall image quality achieved in this telemedical general practicioner-based diabetic eye disease testing was sufficient and would be accepted by both medical assistants and patients in most cases.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [35] Wintergerst M.W.M. Bejan V. Hartmann V. Telemedical diabetic retinopathy screening in a primary care setting: quality of retinal photographs and accuracy of automated image analysis Ophthalmic Epidemiol 29 2022 286 295 34151725 source
    • cites [36] Hutton D. Eyenuk receives FDA clearance for additional camera for autonomous detection of diabetic retinopathy. Ophthalmology Times https://www.ophthalmologytimes.com/view/eyenuk-receives-fda-clearance-for-additional-camera-for-autonomous-detection-of-diabetic-retinopathy 2023

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    If the scheduling system is fully integrated into the EMR, primary care can schedule the ophthalmology appointment before the patient leaves.

    “If the scheduling system is fully integrated into the EMR, primary care can schedule the ophthalmology appointment before the patient leaves.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Compared to a human-interpretation model, autonomous AI testing frees ophthalmologists from the time commitment of image batch reads and allows focus on in-person patient care.

    “Compared to a human-interpretation model (Fig 1, Workflow 2), this also frees ophthalmologists from the time commitment of image batch reads and focuses on in-person patient care.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    A pilot affords the opportunity to gather feedback, refine processes, and build confidence across the network to prepare for scaling.

    “A pilot also affords the opportunity to gather feedback, refine processes, and build confidence across the network to prepare for scaling.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    When executed well, an AI-aided DR testing program can create value and efficiency for the health system.

    “When executed well, an AI-aided DR testing program can create value and efficiency for the health system in the following ways:”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The 6 academic ophthalmology programs interviewed are representative of health systems adopting autonomous AI testing, so their experiences can be generalized.

  • May combine two claims.

    Interpretationreasoning

    From the patients' perspective, integrating scheduling into the EMR improves access by eliminating the need to call to schedule and allowing the option to find earlier appointments for more urgent cases.

    “From the patients’ perspective, there is also improved access as this eliminates the need to call to schedule and allows the option to find earlier appointments for more urgent cases.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    The 2023 study by Najac et al highlighted challenges in ensuring good image quality and integrating the system into the existing workflow.

    “However, the study also highlighted challenges in ensuring good image quality and integrating the system into the existing workflow.35,36”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [35] Wintergerst M.W.M. Bejan V. Hartmann V. Telemedical diabetic retinopathy screening in a primary care setting: quality of retinal photographs and accuracy of automated image analysis Ophthalmic Epidemiol 29 2022 286 295 34151725 source
    • cites [36] Hutton D. Eyenuk receives FDA clearance for additional camera for autonomous detection of diabetic retinopathy. Ophthalmology Times https://www.ophthalmologytimes.com/view/eyenuk-receives-fda-clearance-for-additional-camera-for-autonomous-detection-of-diabetic-retinopathy 2023

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Eyenuk offers a "spellcheck"-like, assistive-AI function for ophthalmologists who read retinal images for other diseases such as glaucoma and age-related macular degeneration, which are not currently FDA-cleared for autonomous diagnosis.

    “Of note, they also offer a “spellcheck”-like, assistive-AI function for ophthalmologists who read these images for other diseases such as glaucoma and age-related macular degeneration, which are not currently FDA-cleared for autonomous diagnosis.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [27] Lim G. Bellemo V. Xie Y. Different fundus imaging modalities and technical factors in AI screening for diabetic retinopathy: a review Eye Vis (Lond) 7 2020 21 32313813 PMC7155252 source

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    Integrating scheduling into the EMR will drive internal referral and patient retention in the health system and reduce the workload to call patients for scheduling on the back end.

    “This will not only drive internal referral and patient retention in the health system but also reduce the workload to call patients for scheduling on the back end.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    Leadership by ophthalmology has the additional benefit of breeding future expansion of this technology into other ocular diseases, such as age-related macular degeneration and glaucoma.

    “On this front, leadership by ophthalmology has the additional benefit of breeding future expansion of this technology into other ocular diseases, such as age-related macular degeneration and glaucoma.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Dow et al found that the AI algorithm is more sensitive than remote image interpretation by a retinal specialist (95.5% vs 69.5%) but had lower specificity (60.3% vs 96.9%).

    “Stanford’s experience with this hybrid system was reported by Dow et al, who found that the AI algorithm is more sensitive than remote image interpretation by a retinal specialist (95.5% vs 69.5%) but had lower specificity (60.3% vs 96.9%).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Dow et al

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    In September 2022, the FDA completed the Software Precertification Pilot Program, which was started in 2017, aimed to explore innovative approaches to regulatory oversight of medical device software.

    “In September 2022, the FDA completed the Software Precertification Pilot Program, which was started in 2017, aimed to explore innovative approaches to regulatory oversight of medical device software.66”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [66] U.S. Food and Drug Administration, Others, Digital health software precertification (pre-cert) pilot program https://www.fda.gov/medical-devices/digital-health-center-excellence/digital-health-software-precertification-pre-cert-pilot-program 2022

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisecitation

    In Singapore, autonomous testing costs $11-15 less per patient per year, leading to an annual national savings of $15 million by 2050.

    “In Singapore, it is estimated that autonomous testing costs $11-15 less per patient per year, leading to an annual national savings of $15 million by 2050.63”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [63] Xie Y. Nguyen Q.D. Hamzah H. Artificial intelligence for teleophthalmology-based diabetic retinopathy screening in a national programme: an economic analysis modelling study Lancet Digit Health 2 2020 e240 e249 33328056 source

  • May combine two claims.

    Premisecitation

    Multiple studies have shown that diagnostic rates can be increased by a reflexive dilation protocol, but utilization has been found to be challenging.

    “While multiple studies have shown that diagnostic rates can be increased by a reflexive dilation protocol,60,61 utilization has been found to be challenging.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [60] Deb-Joardar N. Germain N. Thuret G. Screening for diabetic retinopathy by ophthalmologists and endocrinologists with pupillary dilation and a nonmydriatic digital camera Am J Ophthalmol 140 2005 814 821 16310458 source
    • cites [61] Banaee T. Ansari-Astaneh M.-R. Pourreza H. Utility of 1% tropicamide in improving the quality of images for tele-screening of diabetic retinopathy in patients with dark irides Ophthalmic Epidemiol 24 2017 217 221 28658588 source

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    Among the 3 commercially available US FDA-cleared systems, both LumineticsCore and EyeArt have been implemented outside of trial conditions for 3 to 5 years with adequate levels of imageability and accuracy reported.

    “Among the 3 commercially available US FDA–cleared systems, both LumineticsCore and EyeArt have been implemented outside of trial conditions for 3 to 5 years with adequate levels of imageability and accuracy reported.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationcitation

    Reflexive dilation protocol introduces another step in the primary care workflow in addition to a small patient risk, requiring the operator to be comfortable performing the procedure and understand relevant contraindications, as well as accommodation for added visit time and ensuring patients have transportation means.

    “This introduces another step in the primary care workflow in addition to a small patient risk,62 requiring the operator to be comfortable performing the procedure and understand relevant contraindications, as well as accommodation for added visit time and ensuring patients have transportation means.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [62] Pandit R.J. Taylor R. Mydriasis and glaucoma: exploding the myth. A systematic review Diabet Med 17 2000 693 699 11110501 source
    • Rests on an assumption: Image quality in fundus photography is affected by factors such as pupil size, patient positioning and compliance, and media opacities such as cataracts.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Future algorithm improvements may trigger the requirement for a 510(k) premarket notification submission under guidelines for software changes published in 2017.

    “future algorithm improvements may trigger the requirement for a 510(k) premarket notification submission under guidelines for software changes published in 2017.65”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [65] Center for Devices, Radiological Health Deciding when to submit a 510(k) for a software change to an existing device. US Food and Drug Administration https://www.fda.gov/regulatory-information/search-fda-guidance-documents/deciding-when-submit-510k-software-change-existing-device 2020

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

57 of 57

  • The paper relies on this without support.

    The experiences and perspectives of the six interviewed ophthalmologists are representative of successful AI DR testing implementation across academic health systems.

    “interviews were conducted with 6 ophthalmologists who lead implementation of AI DR testing programs at academic health systems.”

  • The paper relies on this without support.

    In 2022, Centers for Medicare and Medicaid Services announced that they would be reimbursing LumineticsCore procedures under the physician fee schedule, and private payers soon followed.

    “In 2022, Centers for Medicare and Medicaid Services announced that they would be reimbursing LumineticsCore procedures under the physician fee schedule, and private payers soon followed.”

  • The paper relies on this without support.

    Today, LumineticsCore is utilized in >1000 US clinical sites spanning both academic health systems and community-based health organizations including Stanford Medicine, Johns Hopkins Health System, Cahaba Medical Care, and LabCorp, with expansion to sites outside the United States as well.

    “Today, LumineticsCore is utilized in >1000 US clinical sites spanning both academic health systems and community-based health organizations including Stanford Medicine, Johns Hopkins Health System, Cahaba Medical Care, and LabCorp, with expansion to sites outside the United States as well.”

  • The paper relies on this without support.

    To date, EyeArt has been utilized in the United States by institutions such as Temple Health, University of Utah Health, and Cedars Sinai Health System.

    “To date, EyeArt has been utilized in the United States by institutions such as Temple Health, University of Utah Health, and Cedars Sinai Health System.”

  • The paper relies on this without support.

    In 2020, EyeArt received FDA clearance under the 510(k) pathway using IDx-DR as the predicate device, and began its expansion into the United States.

    “More recently, in 2020, the system received FDA clearance under the 510(k) pathway using IDx-DR as the predicate device, and began its expansion into the United States.”

  • The paper relies on this without support.

    In October 2022, AEYE-DS received FDA clearance under the 510(k) pathway using IDx-DR as the predicate device.

    “In October 2022, AEYE-DS received FDA clearance under the 510(k) pathway using IDx-DR as the predicate device.”

  • The paper relies on this without support.

    To our best knowledge, AEYE-DS has been implemented in select community-based health organizations in the United States, such as the Bethesda Health Clinic.

    “To our best knowledge, AEYE-DS has been implemented in select community-based health organizations in the United States, such as the Bethesda Health Clinic.”

  • The paper relies on this without support.

    AI is not a standalone tool but a component in the overall design of a testing program for eye diseases.

    “It is worth noting that AI is not a standalone, but rather a component in the overall design of a testing program for eye diseases.”

  • The paper relies on this without support.

    The 6 academic ophthalmology programs interviewed are representative of health systems adopting autonomous AI testing, so their experiences can be generalized.

    “Based on our interviews, the 6 academic ophthalmology programs agreed that buy-in from the primary care team is critical for initiating utilization of AI examinations.”

  • The paper relies on this without support.

    In the traditional model, a considerable proportion of patients fail to follow up after being referred for ophthalmology assessment.

    “In the traditional model, one of the challenges is that a considerable proportion of patients fail to follow up after being referred for ophthalmology assessment.”

  • The paper relies on this without support.

    Image quality in fundus photography is affected by factors such as pupil size, patient positioning and compliance, and media opacities such as cataracts.

    “Image quality in fundus photography is affected by a number of factors, such as pupil size, patient positioning and compliance, and media opacities such as cataracts.”

  • The paper relies on this without support.

    Following Medicare and Medicaid, a growing number of commercial payers established reimbursement for autonomous AI eye examinations.

    “Following Medicare and Medicaid, a growing number of commercial payers established reimbursement.”

  • The paper relies on this without support.

    Previously, retinal eye examination required patients to schedule and attend a second appointment with ophthalmology after receiving a referral from primary care.

    “Previously, this required patients to schedule and attend a second appointment with ophthalmology after receiving a referral from primary care.”

  • The paper relies on this without support.

    LabCorp launched DR examinations with LumineticsCore at 9 locations in Alabama in April 2023.

    “Another notable adopter on the horizon is LabCorp, who launched DR examinations with LumineticsCore at 9 locations in Alabama in April 2023.”

  • The paper relies on this without support.

    Payment for the point-of-care diagnostic service is currently covered under CPT code 92229.

    “Payment for the point-of-care diagnostic service is currently covered under CPT code 92229”

  • The paper relies on this without support.

    Vulnerable populations served by Federally Qualified Health Centers lack access to specialist resources.

    “Given the lack of access to specialist resources among the vulnerable population served by FQHCs”
16 of 16

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Sensitivity-study count conflicts with Table 1

    Minor

    The abstract gives n = 3 for reported sensitivities, whereas Table 1 reports sensitivities for four institutions: Iowa, Mayo, Stanford, and Temple.

  • Conflicting creation dates for CPT code 92229

    Minor

    One passage says the CPT Editorial Panel created code 92229 in September 2020; another says it was created in 2019.

  • Pivotal-trial gradability percentages conflict

    Minor

    The stated 76.4% receiving results without dilation plus an additional 23.6% becoming gradable with dilation implies 100% overall gradability. The paper reports 96.1% gradability for the pivotal trial.

3 of 3

Cites

Flagged first, then by citations

1 of the 39 references couldn’t be matched to a known paper.

Research works printed but not linked: 6

  1. 2023, Fed Regist, 88, 7881
  2. Huang, 2023, Autonomous artificial intelligence exams are associated with higher adherence to diabetic retinopathy testing in an integrated healthcare system, Invest Ophthalmol Vis Sci, 64, 212
  3. Najac, 2023, Identifying causes of ungradable fundus photos in an artificial intelligence assisted screening program for diabetic retinopathy, Invest Ophthalmol Vis Sci, 64, 228
  4. Nelson, 2023, Evaluating Follow-Up metrics in an AI-Assisted telemedicine screening program for diabetic retinopathy in primary care clinics after hiring a patient care navigator, Invest Ophthalmol Vis Sci, 64, 2274
  5. Leong, 2023, Autonomous artificial intelligence (AI) increases health equity for patients who are more at risk for poor visual outcomes due to diabetic eye disease (DED), Invest Ophthalmol Vis Sci, 64, 243
  6. Knapp, 2023, Real world outcomes from artificial intelligence to detect diabetic retinopathy in the primary care setting: 12 month experience, Invest Ophthalmol Vis Sci, 64, 252

Cited by

Sorted by citations

Related

Authors
Teng, C. W., Patel, S. D., Barkmeier, A. J., Liu, T. Y. A., Myung, D., Henderer, J., Liu, J., Hansen, E., & Al-Aswad, L. A.
Affiliations
  • University of Pennsylvania (US)
  • Mayo Clinic (US)
  • Johns Hopkins University (US)
  • Stanford University (US)
  • Washington University in St. Louis (US)
  • University of Utah (US)
ORCID
Venue
Ophthalmology Science
Publisher
Elsevier BV
Published
2026-01
Volume
6
Issue
1
Pages
100935
ISSN
2666-9145 (ISSN-L)
DOI
10.1016/j.xops.2025.100935
PMC
PMC12553049
OpenAlex
W4414001469
Semantic Scholar
8211d352251820b3e63b1afe34e6376965ada102
Cited by
23
References
39
Citations by year
2026: 22 · 2025: 1
FWCI
12.53
Type
journal article
Language
en
Open access
Licence
Locations
  • Ophthalmology Science (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • Europe PMC (PubMed Central) (submittedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Research to Prevent Blindness
Topics
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Retinal Imaging and Analysis
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Health Sciences › Medicine › Ophthalmology › Ophthalmology and Visual Health Research
Keywords
artificial intelligence, technology implementation, primary care, diagnostic accuracy, health equity, patient adherence, cost-effectiveness, workflow efficiency

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-13.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-13.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review