AI-Powered Eye Exam for Diabetes

(ACCESS2 Trial)

Overseen By Alvin Liu, MD

Age: < 65

Sex: Any

Trial Phase: Academic

Sponsor: Johns Hopkins University

Disqualifiers: Recent diabetic eye exam

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Approved in 5 Jurisdictions This treatment is already approved in other countries

Compensation: Varies

Number of Visits: 1

500 Participants Needed

Trial Summary

What is the purpose of this trial?

The purpose of this study is to determine if use of a nonmydriatic fundus camera using autonomous artificial intelligence software at the point of care increases the proportion of underserved youth with diabetes screened for diabetic retinopathy.

Do I need to stop my current medications for the trial?

The trial information does not specify whether you need to stop taking your current medications.

What data supports the effectiveness of the AI-Powered Eye Exam for Diabetes treatment?

The AI system for diabetic retinopathy screening has shown high sensitivity (87.2%) and specificity (90.7%) in detecting more than mild diabetic retinopathy and diabetic macular edema, making it effective in primary care settings and authorized by the FDA to help prevent vision loss in people with diabetes. ^{1 2 3 4 5}

Is the AI-powered eye exam for diabetes safe for humans?

The AI-powered eye exams, such as IDx-DR and EyeArt, have been studied extensively and are considered safe for use in humans. These systems have been evaluated in real-world settings and have shown to be effective in screening for diabetic retinopathy without any reported safety concerns. ^{1 5 6 7 8}

How does the AI-Powered Eye Exam for Diabetes differ from other treatments for diabetic retinopathy?

The AI-Powered Eye Exam for Diabetes is unique because it uses artificial intelligence to autonomously detect diabetic retinopathy from eye images, offering a high sensitivity rate. Unlike traditional methods that rely solely on human interpretation, this AI system can quickly analyze large volumes of images, potentially predicting disease progression and reducing the burden on eye care professionals. ^{1 2 3 5 9}

Research Team

Risa M Wolf, MD

Principal Investigator

Johns Hopkins University

Eligibility Criteria

This trial is for young people with Type 1 diabetes for at least 3 years, aged 11 or older and in puberty, or those diagnosed with Type 2 diabetes. It's aimed at helping underserved youth who haven't had a diabetic eye exam in the past year.

Inclusion Criteria

I have had Type 1 diabetes for 3+ years and am at least 11 years old or in puberty.

I have been diagnosed with Type 2 diabetes.

Exclusion Criteria

You had an eye exam for diabetes in the past year.

Timeline

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Diabetic Retinopathy Exam

Participants undergo a point-of-care diabetic retinopathy eye exam using autonomous AI. Immediate results are provided, and those with abnormal results are referred for a dilated eye exam.

1 day

1 visit (in-person)

Follow-up

Participants are monitored for safety and effectiveness after the initial exam, with a focus on agreement in interpretation of retinal images over time.

2 years

Treatment Details

Interventions

• Point of Care Autonomous AI diabetic retinopathy exam

Trial Overview The study tests if an AI-powered camera can help screen more kids with diabetes for eye problems caused by their condition (diabetic retinopathy) when used during regular care visits.

Participant Groups

Group I: Diabetic Retinopathy Exam at the point of care Experimental Treatment 1 Intervention

Participants will undergo a point of care diabetic retinopathy eye exam using autonomous AI. Those that test positive will be referred to Eye Care Provider for dilated eye exam.

Who Is Running the Clinical Trial?

Johns Hopkins University

Lead Sponsor

Trials

2,366

Recruited

15,160,000+

Juvenile Diabetes Research Foundation

Collaborator

Trials

237

Recruited

142,000+

National Eye Institute (NEI)

Collaborator

Trials

572

Recruited

1,320,000+

Findings from Research

The IDx autonomous diabetic retinopathy screening program demonstrated a perfect sensitivity of 100% in detecting referable diabetic retinopathy, but had a lower specificity of 82%, leading to a high rate of unnecessary referrals.

With a positive predictive value of only 19%, the program may overwhelm ophthalmologists and primary care clinics, suggesting a need for improved AI systems that can provide better specificity and detailed lesion annotations to enhance patient management and treatment adherence.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

The Real-World Impact of Artificial Intelligence on Diabetic Retinopathy Screening in Primary Care. Cuadros, J.[2021]

The global prevalence of diabetes is projected to reach 700 million individuals in the next 25 years, increasing the risk of vision loss from diabetic eye disease and highlighting the need for innovative detection tools.

Artificial intelligence (AI) has shown great promise in detecting diabetic retinopathy (DR) by analyzing fundus photographs, using machine learning to grade DR and potentially predict its progression, which could help alleviate the burden on eye care professionals.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial Intelligence in the assessment of diabetic retinopathy from fundus photographs. Gilbert, MJ., Sun, JK.[2021]

The autonomous diagnostic AI system demonstrated 100% sensitivity for detecting derivable diabetic retinopathy (RDR) and diabetic retinopathy with decreased vision (VTDR), indicating it can accurately identify all cases of these conditions.

The AI system also showed high specificity (81.82% for RDR and 94.64% for VTDR), suggesting it can effectively distinguish between affected and non-affected individuals, which could enhance accessibility to screening in primary care settings.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Use in clinical practice of an automated screening method of diabetic retinopathy that can be derived using a diagnostic artificial intelligence system. Peris-Martínez, C., Shaha, A., Clarida, W., et al.[2021]

In a pivotal trial involving 900 participants, an AI system demonstrated high sensitivity (87.2%) and specificity (90.7%) in detecting diabetic retinopathy, surpassing pre-specified performance benchmarks.

The FDA has authorized this AI diagnostic system for use in primary care settings, marking it as the first autonomous AI diagnostic tool approved in medicine, which could significantly enhance early detection and prevention of vision loss in diabetic patients.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices. Abràmoff, MD., Lavin, PT., Birch, M., et al.[2020]

The AI-based diagnostic tool IDx-DR demonstrated high sensitivity (95.5%) for detecting more-than-mild diabetic retinopathy (MTMDR), but had lower specificity (60.3%) compared to human retina specialists, who had a sensitivity of 69.5% and a much higher specificity of 96.9%.

A hybrid workflow combining AI assessment with human expert review significantly improved diagnostic accuracy, achieving a sensitivity of 95.5% and a specificity of 98.2%, while also enhancing the gradability of fundus photographs from 62.5% to 95.6%.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

AI-Human Hybrid Workflow Enhances Teleophthalmology for the Detection of Diabetic Retinopathy. Dow, ER., Khan, NC., Chen, KM., et al.[2023]

The EyeArt system, an automated AI technology, demonstrated high diagnostic efficacy for detecting referral-warranted diabetic retinopathy (DR) with a sensitivity of 91.3% and specificity of 91.1% across 850,908 fundus images from 101,710 patient visits.

For encounters with potentially treatable DR, the EyeArt system achieved an impressive sensitivity of 98.5%, indicating its effectiveness in identifying patients who need urgent care, making it a valuable tool for healthcare providers in managing diabetic eye health.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

The Value of Automated Diabetic Retinopathy Screening with the EyeArt System: A Study of More Than 100,000 Consecutive Encounters from People with Diabetes. Bhaskaranand, M., Ramachandra, C., Bhat, S., et al.[2020]

In a study of 310 youth with diabetes, an autonomous AI system demonstrated a high diagnosability rate of 97.5% for detecting diabetic retinopathy (DR), with a sensitivity of 85.7% and specificity of 79.3% compared to retinal specialists' assessments.

The implementation of AI for diabetic eye exams significantly improved adherence to screening guidelines from 49% to 95%, highlighting its potential to enhance early detection of DR in young patients.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

The SEE Study: Safety, Efficacy, and Equity of Implementing Autonomous Artificial Intelligence for Diagnosing Diabetic Retinopathy in Youth. Wolf, RM., Liu, TYA., Thomas, C., et al.[2021]

The AI algorithm Selena+ in the handheld Optomed Aurora fundus camera demonstrated a high sensitivity and specificity of 96.8% for detecting diabetic retinopathy (DR) in a study of 256 patients, making it a reliable screening tool.

The study found a perfect agreement (100%) between the AI analysis and the expert retina specialist's evaluation, suggesting that this technology could effectively support large-scale screening efforts for DR in real-world clinical settings.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial intelligence in diabetic retinopathy screening: clinical assessment using handheld fundus camera in a real-life setting. Lupidi, M., Danieli, L., Fruttini, D., et al.[2023]

Artificial intelligence (AI) technology can significantly enhance the diagnosis and treatment of diabetic retinopathy, a leading cause of preventable blindness, by providing early detection and monitoring capabilities.

The effectiveness of AI in this field is hindered by inconsistent labeling standards in retinal exam datasets, highlighting the need for standardized classification methods to improve the reliability of AI algorithms.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Diabetic retinopathy classification for supervised machine learning algorithms. Nakayama, LF., Ribeiro, LZ., Gonçalves, MB., et al.[2022]

References

1. United States pubmed.ncbi.nlm.nih.gov

The Real-World Impact of Artificial Intelligence on Diabetic Retinopathy Screening in Primary Care. [2021]

2. England pubmed.ncbi.nlm.nih.gov

Artificial Intelligence in the assessment of diabetic retinopathy from fundus photographs. [2021]

3. Spain pubmed.ncbi.nlm.nih.gov

Use in clinical practice of an automated screening method of diabetic retinopathy that can be derived using a diagnostic artificial intelligence system. [2021]

4. England pubmed.ncbi.nlm.nih.gov

Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices. [2020]

5. Netherlands pubmed.ncbi.nlm.nih.gov

AI-Human Hybrid Workflow Enhances Teleophthalmology for the Detection of Diabetic Retinopathy. [2023]

6. United States pubmed.ncbi.nlm.nih.gov

The Value of Automated Diabetic Retinopathy Screening with the EyeArt System: A Study of More Than 100,000 Consecutive Encounters from People with Diabetes. [2020]

7. United States pubmed.ncbi.nlm.nih.gov

The SEE Study: Safety, Efficacy, and Equity of Implementing Autonomous Artificial Intelligence for Diagnosing Diabetic Retinopathy in Youth. [2021]

8. Germany pubmed.ncbi.nlm.nih.gov

Artificial intelligence in diabetic retinopathy screening: clinical assessment using handheld fundus camera in a real-life setting. [2023]

9. England pubmed.ncbi.nlm.nih.gov

Diabetic retinopathy classification for supervised machine learning algorithms. [2022]