Discoveries from AAO
Discoveries from AAO
Source/Disclosures
Source:

Alauddin S. PA057. Presented at: American Academy of Ophthalmology annual meeting; Nov. 13-15, 2020 (virtual meeting).

Disclosures: Alauddin reports no relevant financial disclosures.
December 16, 2020
1 min read
Save

AI-based telemedicine platform may allow for large-scale AMD screening

Source/Disclosures
Source:

Alauddin S. PA057. Presented at: American Academy of Ophthalmology annual meeting; Nov. 13-15, 2020 (virtual meeting).

Disclosures: Alauddin reports no relevant financial disclosures.
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Artificial intelligence-based telemedicine platform may allow for large-scale screening of patients with age-related macular degeneration while lowering the cost of care, according to a poster at AAO Annual Meeting.

The positive results of this study permit for further clinical validation of AI-powered telemedicine as well as potential widespread deployment.

“We have focused on AMD because late AMD is the leading cause of blindness in the developed countries,” Sharmina Alauddin, FCPS, said in her virtual presentation. “Our study is the first of an AI fundus photo-based telemedicine platform for larger scale screening of referable AMD in outpatient settings,”

In a prospective clinical evaluation, 299 men and women aged 50 years and older underwent non-dilated retinal color imaging for AMD analysis by FDA approved camera Eidon. Researchers uploaded the captured retinal images (n = 503 gradable images) to AI algorithm iHealth screen and received an immediate grading of either referable (intermediate macula and late AMD) or non-referable (healthy macula and early AMD) cases. Ophthalmologists further analyzed the images following international AMD grading protocol (n = 161 eyes referable, 342 eyes non-referable) and compared results against the AI algorithm for sensitivity, specificity and accuracy.

Results showed iHeath AI-screening properly identified 307 images with an accuracy of 88.7% (95% CI, 85.6% to 91.2%), a sensitivity of 86.3% (95% CI, 80.1% to 81.2%) and a specificity of 89.8% (95% CI, 86.1% to 92.8%). Further analysis yielded a kappa score of 0.75 (95% CI, 0.68 to 0.81).

“Such an AI-based telemedicine tool can speed the delivery and lower the cost of care while easing access to the healthcare system; this tool will allow large-scale screening of patients from remote and rural areas, underserved population, family physicians, optometry offices and general ophthalmology clinics,” Alauddin concluded. “The results, validated by human expert graders, warrant further clinical validation and potential remote telemedical deployment.”