Diabetic Retinopathy Screening Using Smartphone-Based Fundus Photography and Deep-Learning Artificial Intelligence in the Yucatan Peninsula: A Field Study
Authors' conclusion
Medios and EyeArt AI algorithms demonstrated high levels of sensitivity and specificity for detecting DR when applied in this real-world field setting. Both programs should be considered in remote, large-scale DR screening campaigns where immediate results are desirable, and in the case of EyeArt, online access is possible.
Study context
- Population
- Patients screened for diabetic retinopathy (DR) with portable Remidio fundus-on-phone cameras in a remote outreach field setting, Yucatan Peninsula, June 2019; many with chronic visual impairment
- Participants
- 248 patients2,130 images
- Comparison
- Medios vs EyeArt artificial intelligence (AI) algorithms
- EyeArtn = 156
- Outcome
- Sensitivity and specificity for detecting DR vs reference standard (masked image analysis combined with clinical examination)
Methodological detailsfield study, retrospective analysis, masked image analysis
- Field study
- Retrospective analysis
- Masked image analysis
Performance
As reported in the abstract
Detection of DR
| Strategy | Sensitivity | Specificity |
|---|---|---|
| Medios | 94% (95% CI 88%-97%)Sensitivity on a scale from 0 to 100 percent | 94% (95% CI 88%-98%)Specificity on a scale from 0 to 100 percent |
| EyeArt | 94% (95% CI 86%-98%)Sensitivity on a scale from 0 to 100 percent | 86% (95% CI 77%-93%)Specificity on a scale from 0 to 100 percent |
Sensitivity
Medios94% (95% CI 88%-97%)
EyeArt94% (95% CI 86%-98%)
Sensitivity on a scale from 0 to 100 percentSpecificity
Medios94% (95% CI 88%-98%)
EyeArt86% (95% CI 77%-93%)
Specificity on a scale from 0 to 100 percent- Patients evaluated by EyeArt, owing primarily to photographer error
- 156
- Patients with some level of DR
- 129
- Patients with no DR
- 119
Further findings
| Outcome | Medios | EyeArt |
|---|---|---|
| Head-to-head comparison | ||
| Sensitivity | 99% (93%-100%) | 95% (87%-99%) |
| Specificity | 88% (73%-97%) | 88% (73%-97%) |
- Head-to-head comparison: 110 patients
Original publication
Diabetic Retinopathy Screening Using Smartphone-Based Fundus Photography and Deep-Learning Artificial Intelligence in the Yucatan Peninsula: A Field Study. J Diabetes Sci Technol. Published 29 August 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →