Comparison between diabetic retinopathy grading using nonmydriatic 1-field portable fundus cameras and ultrawide-field cameras

Objective To systematically compare the use of Optos and Aurora cameras in diabetic retinopathy (DR) Grading. Patients and methods We conducted a cross-sectional study in which, after informed consent-147 eyes (77 patients) were included and photographed using the Aurora handheld device and the ultrawide-field (UWF) Optos machine. For the Aurora camera, a single 50° field image was captured, centered on the macula. The UWF camera captured on-axis images covering a 200° field. No mydriasis was used. Each image was analyzed by certified graders regarding the level of DR and diabetic macular edema and agreement levels were calculated between both devices. Results The level of agreement between the Optos UWF and Aurora 1-field was found to be substantial regarding all levels of retinopathies (κ=0.769), in cases with predominantly peripheral lesion (κ=0.614) and referable DR (κ=0.781). In cases of vision-threatening DR (VTDR) the agreement level was moderate (κ=0.603). For diabetic macular edema a fair level of agreement (κ=0.327) was estimated. Conclusion The Aurora 1-field system showed significant limitations, missing a substantial number of Proliferative DR (PDR) cases with a detection rate of just 27.59%. It also demonstrated poor performance in detecting referable DR, vision-threatening DR (VTDR), and predominately peripheral lesion cases, with low agreement rates compared with UWF imaging.