Large-scale audit validates automated screening system for diabetic retinopathy

A previous study of 6,722 patients in Scotland showed automated grading software has the potential to reduce the manual grading workload within diabetic retinopathy screening programmes by more than 45 percent. However, before the software could be implemented into the Scottish national diabetic retinopathy screening programme, it was necessary to demonstrate that its performance was maintained when used in other Scottish screening centres. As a result, Scotland's National Diabetic Retinopathy Screening Collaborative requested this audit in a large, unselected population of people with diabetes.

The automated screening system, based on microaneurysm/dot haemorrhage detection and image quality assessment, was used on 78,601 images, obtained from 33,535 consecutive patients, which had been manually graded at one of two regional diabetic retinopathy screening programmes. It performed as well as in the previous study. All of the patients with proliferative, referable background and observable background retinopathy, and at least 97 percent of those with referable maculopathy, observable maculopathy, or ungradable images were detected.