The large scale REtrospective Validation of Eyeart in the REal world (REVERE) study assessed the diagnostic efficacy of the EyeArt system screening for referable diabetic retinopathy in 107,001 consecutive diabetic patient visits.
DEFINITIONS:
- DR: diabetic retinopathy
- NPDR: non-proliferative DR
- ICDR scale: International Clinical Diabetic Retinopathy severity scale with increasing levels of disease severity: No DR, mild NPDR, moderate NPDR, severe NPDR, and proliferative DR
- CSME: Clinically significant macular edema
- CSME surrogate markers: hard exudates within one optic disk diameter of the macula
- referable DR: moderate NPDR or higher on the ICDR scale, and/or presence of CSME surrogate markers
- treatable DR: severe NPDR or higher on the ICDR scale, and/or presence of CSME surrogate markers
- vision-threatning DR: same as treatable DR
METHOD HIGHLIGHTS:
- To eliminate/minimize selection bias 850,908 images from 107,001 consecutive diabetic patient visits from 404 primary care clinics in the EyePACS telescreening program were used
- The reference standard was quality controlled grading of the images by board certified optometrists/ophthalmologists
- Study type: retrospective
- Included images from both mydriatic and non-mydriatic imaging and a wide variety of makes/models of color fundus cameras
- There was a wide variation in the quality of images
- Patient cases deemed as ungradable by EyeArt were considered as positive for referable diabetic retinopathy and included in the analysis
RESULTS HIGHLIGHTS:
- EyeArt achieved sensitivity of 91.3% and specificity of 91.1 for referable diabetic retinopathy
- The area under ROC curve was 0.965
- EyeArt sensitivity was 98.5% for treatable diabetic retinopathy
STUDY LINKS:
- Malavika Bhaskaranand, Chaithanya Ramachandra, Sandeep Bhat, Jorge Cuadros, Muneeswar G. Nittala, Srinivas R. Sadda, and Kaushal Solanki. “The Value of Automated Diabetic Retinopathy Screening with the EyeArt System: A Study of More Than 100,000 Consecutive Encounters from People with Diabetes.” Diabetes Technology & Therapeutics, July 23, 2019. https://doi.org/10.1089/dia.2019.0164. Read Study | Download PDF