Validation of Smartphone Imaging for Retinopathy and Diabetic Maculopathy Staging

Abstract

Objective: To validate the diagnostic accuracy of fundus images obtained with a smartphone camera, compared to a non-mydriatic camera, for the detection of diabetic retinopathy and maculopathy.

Methods: A cross-sectional descriptive observational study was performed with 78 eyes of 50 Cuban diabetic patients. Retinal images were acquired with both techniques, classified according to Modified Scottish Diabetic Retinopathy scale by 2 retina specialists in a blinded manner. Interevaluator concordance (kappa), sensitivity, specificity and predictive values were calculated.

Results: The non-mydriatic camera showed better quality. The agreement between smartphone camera and non-mydriatic camera was moderate in retinopathy (kappa 0.40) and maculopathy (kappa 0.32). The smartphone camera had good sensitivity (91.15 %) and specificity (86.25 %) for detecting proliferative retinopathy, with predictive values of 86.45 % and 90.95 %. There were no significant differences in the image quality of the Smartphone camera by lens status.

Conclusions: The non-mydriatic camera showed better image quality. The smartphone camera showed moderate inter-technique/evaluator agreement and good effectiveness in detecting proliferative retinopathy. The image quality of the smartphone camera was not affected by lens status, unlike the non-mydriatic camera.