RESUMO
PURPOSE: To conduct a feasibility study of computer-aided screening for diabetic retinopathy by developing a computerized program to automatically detect retinal changes from digital retinal images. METHODS: The study was carried out in three steps. Step 1 was to collect baseline retinal image data of 600 eyes of normal subjects with normal fundi and data of 300 eyes of diabetic patients with diabetic retinopathy. All data were recorded by digital fundus camera. Step 2 was to analyse all retinal images for normal and abnormal features. By this method, the automated computerized screening program was developed. The program preprocesses colour retinal images and recognizes the main retinal components (optic disc, fovea, and blood vessels) and diabetic features such as exudates, haemorrhages, and microaneurysms. All of the accumulated information is interpreted as normal, abnormal, or unknown. Step 3 was to evaluate the sensitivity and specificity of the computerized screening program by testing the program on diabetic patients and comparing the program's results with the results of screening by retinal specialists. RESULTS: Diabetic patients (182 patients, 336 eyes) were examined by retinal specialists; 221 eyes had a normal fundus and 115 eyes had nonproliferative diabetic retinopathy. Digital retinal images were taken of these 336 eyes and interpreted by the automated screening program. The program had a sensitivity and specificity of 74.8% and 82.7%, respectively. CONCLUSIONS: The automated screening program was able to differentiate between the normal fundus and the diabetic retinopathy fundus. The program may be beneficial for use in screening for diabetic retinopathy. Further development of the program may provide higher sensitivity.
