Using deep learning to detect diabetic retinopathy on handheld non-mydriatic retinal images acquired by field workers in community settings
Source check: Source checked. Not independent scientific confirmation.
Finding: Two-field handheld-image screening achieved AUROC 0.99, sensitivity 93.86% and specificity 96.00% for referable DR in the field-worker study. 20-site prospective study with manual grading and cross-validation; two-field performance evaluated in the assembled India dataset. Generalization beyond the mono-ethnic cohort and curated gradable images remains uncertain.
Technology: A deep-learning image classifier evaluated macula- and optic-disc-centered fundus photographs for referable diabetic retinopathy and diabetic macular edema.
Publication: Peer-reviewed paper · Test context: prospective or independent check. Catalog categories, not an evidence grade.
- AI contribution
- A deep-learning image classifier evaluated macula- and optic-disc-centered fundus photographs for referable diabetic retinopathy and diabetic macular edema.
- Test and result
- Two-field handheld-image screening achieved AUROC 0.99, sensitivity 93.86% and specificity 96.00% for referable DR in the field-worker study.
- Limitation / still unknown
- 20-site prospective study with manual grading and cross-validation; two-field performance evaluated in the assembled India dataset. Generalization beyond the mono-ethnic cohort and curated gradable images remains uncertain.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.nature.com/articles/s41598-023-28347-z.pdf
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary source full text Methods and Results; version/date checked against bibliographic record..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 4/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.