Offline AI Screening

How it works

Risks & contraindications

Comparisons

Evidence quality

Highlighted claims

• The Medios offline AI software was embedded in the smartphone fundus camera application and could run without internet connectivity. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study
• The offline AI output was binary and treated signs of DR as referable diabetic retinopathy. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study
• Offline AI did not grade diabetic retinopathy severity or diabetic macular oedema. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study
• Community AI-assisted screening had declining sensitivity after misclassified images were excluded. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study
• Offline AI screening met the cited US FDA diagnostic thresholds for referable diabetic retinopathy sensitivity and specificity. — Evaluation of non-ophthalmologist-led and offline AI-assisted models for diabetic retinopathy screening in India: a pragmatic diagnostic accuracy study