HomeTechnologyArtificial intelligenceFamily Medicine Doctors, AI Could Be Effective in DR Screening

Family Medicine Doctors, AI Could Be Effective in DR Screening

Posters presented at AAO 2022 showed that primary care physicians and artificial intelligence (AI) could effectively screen patients for diabetic retinopathy (DR), although false negatives in AI need to be improved.

In 2 posters presented at AAO 2022, research found that alternatives to ophthalmologist screening for diabetic retinopathy (DR) were effective. GPs and artificial intelligence (AI) both accurately identified DR in images of patients’ eyes, both for diagnostic reasons and for inclusion in clinical trials.

The first poster1 focused on GPs’ ability to diagnose DR in a telemedicine program. Telemedicine programs have previously been used to effectively screen DR to identify patients in need of eye care. GPs have reviewed the images in telemedicine programs and ophthalmologists are only needed if no diagnosis is made. The researchers wanted to evaluate the effectiveness of primary care physicians in diagnosing DR in patients they have seen.

2260 patients with type 2 diabetes were included in this study. All were photographed within 1 year with a non-mydriatic fundus camera. There were 5 masked GPs and a retinal doctor who viewed the images and the agreement between the GP and the retinal doctor was compared.

In patients with no apparent retinopathy and no additional risk factors, retinography would be repeated within 2 years. If they had no apparent retinopathy with risk factors or mild retinopathy without risk factors, they would have a retinography repeat within the year. Patients with mild retinopathy with risk factors would have their retinography repeated within 6 months. Moderate non-proliferative (NP) DR or proliferative DR or suspected diabetic macular edema or other retinal or optic nerve disorders were referred to the ophthalmology department.

Must Read
5 Jobs Which Could Be Taken By Artificial Intelligence In The Next 10 Years

There were 14 patients with mild non-proliferative (NPDR) and 27 patients (1.19%) with moderate or severe NPDR in primary care, all of which were confirmed in the ophthalmic setting. There were 180 patients referred for further assessment, 42 of which were correctly classified as “without DR”, 51 as “questionable case”, 83 as “illegible” and 4 were classified as false positive with the ophthalmologist determining that no one had DR . .

The researchers concluded that GPs were able to detect 100% of patients with moderate or severe NPDR and correctly classify 93.6% of retinographs obtained in the telemedicine program, resulting in excellent GP performance.

The second poster2 focused on using AI to determine patients who had DR for clinical trials. The FDA approves the use of AI to screen patients with a single image centered on the macula to determine who may be referable retinopathy. The screening for clinical trials involving patients with DR requires patients to be level 47/53 on the Early Treatment Diabetic Retinopathy Study scale using the 7-field stereoscopic imaging protocol to narrow down the potential participants.

There was an estimated 50% screen error in images submitted to the Wisconsin Reading Center. The study aimed to determine whether AI could help screen patients for inclusion in clinical trials.

The AI ​​used field 2 to screen patients. Patients who were ineligible by the AI ​​were not further assessed. Those eligible by the AI ​​went to the human evaluator, who would review the 7 imaging fields to confirm or decline eligibility.

The AI ​​was found to be 86.4% accurate with a sensitivity of 0.77 and a specificity of 0.89. The AI ​​had a false positive rate of 10.8% and a false negative rate of 22.6%, which was a concern in this screening model. The F1 score for the AI ​​was 0.72 and the precision was found to be 66.7%.

Must Read
Norton says the economy will shape threat landscape in 2023

The researchers found that false negatives in the AI ​​occurred because of an imbalance in pathology between the central and peripheral fields with more pathology in peripheral fields. False positives occurred due to an imbalance of pathology between central and peripheral fields with more pathology in the central field.

The researchers concluded that the AI ​​algorithm can identify patients with DR levels less than 47, but more work would need to be done to reduce the number of false negatives. AI prescreening of patients eligible for DR clinical trials before a grader confirmation could reduce screen failure rates, create cost-effectiveness, and reduce the burden on participants and clinical staff.

Automated assessment of eligible patients could also improve overall enrollment in DR clinical trials. Further research would be needed to determine this from an external source. The researchers propose to use a prospective clinical trial that would compare a traditional grader-only approach to using a clinical trial where patients were pre-screened by an AI.

These 2 posters show that DR can be screened through multiple avenues in the future, as GPs and AI can help screen patients with DR, which can help treat patients and participate in clinical trials of DR treatments.


1. Ferreras A, Pinilla I, Abecia E, Figus M, Fogagnolo P, Iester M. Ability of GPs to detect DR in a telemedicine program. Presented at: AAO 2022; September 30 – October 3, 2022; Chicago, IL. Summary PO110.

2. Domalpally A, Slater R, Barrett N, Channa R, Blodi B. AI-activated prescreening for DR clinical trials. Presented at: AAO 2022; September 30 – October 3, 2022; Chicago, IL. Summary PO341.

Must Read
Revolutionizing planetary exploration: AI-assisted mapping



Please enter your comment!
Please enter your name here

Most Popular

Recent Comments