Peripheral retinal haemorrhage density on ultra-widefield imaging as a novel biomarker for predicting diabetic retinopathy progression: a 2-year longitudinal study in Asians
AI-identified peripheral retinal bleeding patterns predict vision-threatening diabetic eye disease progression independent of blood sugar and blood pressure.
In a 2-year prospective study of 528 diabetic eyes at a Singapore eye center, automated AI quantification of peripheral retinal haemorrhage density and predominantly peripheral lesions on ultra-widefield photography predicted diabetic retinopathy progression independently of systemic risk factors. These peripheral biomarkers identify a high-risk subgroup amendable to closer surveillance and support wider integration of AI-assisted peripheral retinal assessment in diabetic eye care.
What the study was
- Study design
- Prospective longitudinal cohort study (2-year)
- Population
- Adults with diabetes and no DR or NPDR (n=282 participants, 528 eyes), multiethnic Asian cohort
- Sample size
- 528
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- British Journal of Ophthalmology
Why it surfaced
Prospective longitudinal n=528 study validating AI-quantified peripheral UWF biomarkers for DR risk stratification. Directly implementable with existing UWF equipment and automated lesion detection tools.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.