Aim: To assess a two-step automated system (RetmarkerSR) that analyzes retinal photographs to detect diabetic retinopathy for the purpose of reducing the burden of manual grading. Methods: Anonymous images from 5,386 patients screened in 2007 were obtained from a nonmydriatic diabetic retinopathy screening program in Portugal and graded by an experienced ophthalmologist. RetmarkerSR earmarked microaneurysms, generating two outputs: ‘disease’ or ‘no disease’. A second-step analysis, based on coregistration, combining two visits, was subsequently performed in 289 patients who underwent repeated examinations in 2008. The study was extended by analyzing all referrals considered urgent by the ophthalmologist from 2001 to 2007. Results were compared with those obtained by manual grading. Results: The RetmarkerSR classified in a first-step analysis 2,560 patients (47.5%) as having ‘no disease’ and 2,826 patients (52.5%) as having ‘disease’, thus requiring manual grading. RetmarkerSR detected all eyes considered urgent referrals. The two-step analysis further reduced the number of false-positive results by 26.3%, indicating an overall sensitivity of 95.8% and a specificity of 63.2%. Conclusion: Automated grading of diabetic retinopathy may safely reduce the burden of grading patients in diabetic retinopathy screening programs. The novel two-step automated analysis system offers improved sensitivity and specificity over published automated analysis systems.

Purpose: To describe the procedures of a nonmydriatic diabetic retinopathy (DR) screening program in the Central Region of Portugal and the added value of the introduction of an automated disease/no disease analysis. Methods: The images from the DR screening program are analyzed in a central reading center using first an automated disease/no disease analysis followed by human grading of the disease cases. The grading scale used is as follows: R0 - no retinopathy, RL - nonproliferative DR, M - maculopathy, RP - proliferative DR and NC - not classifiable. Results: Since the introduction of automated analysis in July 2011, a total of 89,626 eyes (45,148 patients) were screened with the following distribution: R0 - 71.5%, RL - 22.7%, M - 2.2%, RP - 0.1% and NC - 3.5%. The implemented automated system showed the potential for human grading burden reduction of 48.42%. Conclusions: Screening for DR using automated analysis allied to a simplified grading scale identifies DR vision-threatening complications well while decreasing human burden.

Using the RetmarkerDR software, the authors were able to identify patients with higher risk to develop CSME during follow-up using a threshold of 2 or more MA formation rate. Together with the high negative predictive value, the automated analysis may help to determine the individual risk of a patient to develop sight-threatening complications related to diabetic retinopathy and schedule individual screening intervals.

Diabetic retinopathy (DR) is a sight-threatening condition occurring in persons with diabetes, which causes progressive damage to the retina. The early detection and diagnosis of DR is vital for saving the vision of diabetic persons. The early signs of DR which appear on the surface of the retina are the dark lesions such as microaneurysms (MAs) and hemorrhages (HEMs), and bright lesions (BLs) such as exudates. In this paper, we propose a novel automated system for the detection and diagnosis of these retinal lesions by processing retinal fundus images. We devise appropriate binary classifiers for these three different types of lesions. Some novel contextual/numerical features are derived, for each lesion type, depending on its inherent properties. This is performed by analysing several wavelet bands (resulting from the isotropic undecimated wavelet transform decomposition of the retinal image green channel) and by using an appropriate combination of Hessian multiscale analysis, variational segmentation and cartoon+texture decomposition. The proposed methodology has been validated on several medical datasets, with a total of 45,770 images, using standard performance measures such as sensitivity and specificity. The individual performance, per frame, of the MA detector is 93% sensitivity and 89% specificity, of the HEM detector is 86% sensitivity and 90% specificity, and of the BL detector is 90% sensitivity and 97% specificity. Regarding the collective performance of these binary detectors, as an automated screening system for DR (meaning that a patient is considered to have DR if it is a positive patient for at least one of the detectors) it achieves an average 95-100% of sensitivity and 70% of specificity at a per patient basis. Furthermore, evaluation conducted on publicly available datasets, for comparison with other existing techniques, shows the promising potential of the proposed detectors.

No abstract