Skip to content

Deep Learning to Validate EyeRoboFC Fundus Photos with Open Source/Public Fundus Datasets for Diabetic Retinopathy

Deep Learning to Validate EyeRoboFC Fundus Photos with Open Source/Public Fundus Datasets for Diabetic Retinopathy

Status
Active, not recruiting
Phases
Unknown
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2600115934
Enrollment
Unknown
Registered
2026-01-04
Start date
2026-01-05
Completion date
Unknown
Last updated
2026-01-05

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Diabetic retinopathy

Interventions

Patients without fundus disease:No
Patients with diabetic retinopathy:No

Sponsors

Tianjin Medical University Eye Hospital
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to 100 Years

Inclusion criteria

Inclusion criteria: 1. Patients with diabetic retinopathy and patients without fundus disease were evaluated by fundus photography and OCT combined with expert physicians; 2. Able to provide informed consent; 3. Can cooperate with the inspection.

Exclusion criteria

Exclusion criteria: 1. Due to turbidity of the refractive medium, imaging cannot be achieved; 2. Previous retinal surgery has resulted in an inability to grade the image.

Design outcomes

Primary

MeasureTime frame
Detection of area under the receiver operating characteristic curve (AUROC) for referrable diabetic retinopathy on EyeRoboFC images compared to ophthalmologist grading;

Secondary

MeasureTime frame
The proportion and poor quality of non-gradable images;mean absolute error (MAE) and correlation between the predicted image quality score and the grader score;Comparison of performance before and after domain adaptation.;Sensitivity, specificity, accuracy, F1 score and Cohenkappa coefficient of diabetic retinopathy severity and other lesion classifications.;

Countries

China

Contacts

Public ContactRen Xinjun

Tianjin Medical University Eye Hospital

zlrxjrsy@126.com+86 22 86428838

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Feb 4, 2026