Head-to-Head Comparative Evaluation of Four Commercially Available Artificial Intelligence Systems for Detecting Referable Diabetic Retinopathy in a Tanzanian Population
Authors' conclusion
Several commercially available AI systems demonstrated high sensitivity in detecting referable and proliferative DR, supporting their potential implementation.
Study context
- Population
- People in a Tanzanian diabetic retinopathy (DR) screening program (retinal images); 379 (55.0%) had referable DR, 93 (13.5%) proliferative DRtest data set
- Participants
- 689 people
- Comparison
- Four commercially available artificial intelligence (AI) systems for detecting referable DR
- Outcome
- Sensitivity and specificity in detecting referable DR
Methodological detailshead-to-head comparison, scoping review and expert consultation to identify systems
- Head-to-head comparison
- Scoping review and expert consultation to identify systems
Performance
As reported in the abstract
Referable DR detection
Sensitivity range across the four systems
83.9% to 93.7%
Specificity range across the four systems (lower)
70.3% to 79.0%
Further findings
Proliferative DR detection
- Sensitivity in all four AI systems
- >98%
Implementation features
- All systems Conformité Européenne-marked medical devices; one (Medios AI/Remidio) functions offline as standard
Original publication
Head-to-Head Comparative Evaluation of Four Commercially Available Artificial Intelligence Systems for Detecting Referable Diabetic Retinopathy in a Tanzanian Population. Diabetes Care. Published 10 June 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →