Purchase this article with an account.
M. Prakash, J. D. Cavallerano, T. J. Murtha, J. K. Sun, A. C. Clermont, R. N. Hamam, S. E. Bursell, L. P. Aiello; Comparison of Automated Digital Retinal Images, ETDRS 35mm 7-Standard Field Color Stereoscopic Retinal Photography and Retinal Examination for Assessment of Diabetic Retinopathy Severity. Invest. Ophthalmol. Vis. Sci. 2007;48(13):1396.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
ETDRS protocol 7-standard field 35 mm color stereoscopic retinal photos (ETDRS photos) are the standard for assessing diabetic retinopathy (DR) severity. Acquisition of suitable ETDRS photos requires substantial experience, technical skill, and specific certification for clinical trials. The Automated Retinal Imaging System (ARIS, Prescott, AZ) acquires 7, pseudo-color stereoscopic digital images approximating the ETDRS photo fields by a semi-automated process that requires only a technician to operate. We compared correlation between ETDRS photos, ARIS photos and clinical exam in 141 eyes of patients with diabetes mellitus (DM).
All evaluations were fully masked. Following clinical exam with a retinal specialist to determine DR level, all patients were imaged using both the ARIS and ETDRS protocols. One reader graded all ARIS images and a different reader graded all ETDRS photos.
141 eyes of 71 pts with DM were evaluated. Mean age was 50 years (range 19-75). Clinical exam revealed no DR in 15%, mild NPDR in 30%, moderate NPDR in 23%, severe NPDR in 1%, PDR in 4%, PDR with high-risk characteristics in 6%, and quiescent PDR in 21% of eyes. To date, ARIS images from 32 patients (45%) and ETDRS photos from 19 patients (27%) have been graded. Ungradable images occurred 3 (4.7%) and 0 times for ARIS and ETDRS photos, respectively. Exact agreement of retinopathy severity between ARIS and clinical exam was 77.6% (k=0.65, p<0.001), comparing favorably with ETDRS photo agreement with clinical exam of 65.8% (k=0.28, p=0.035). ARIS and ETDRS photos were within 1 retinopathy severity level of clinical exam in 91.6% and 100%, respectively. Exact agreement between ETDRS photos and ARIS occurred in 78.9% (k=0.48, p<0.001) and agreement within 1 level occurred in 100%. ARIS was graded as more severe than ETDRS photos in 3 eyes and as less severe in 4 eyes.
ARIS automated digital images compare favorably with both clinical exam and ETDRS protocol 7-standard field photography, suggesting that automated digital imaging may prove useful in evaluation of retinopathy severity. Findings from the entire cohort analyzed by retinopathy severity level and specific lesion will also be presented.
This PDF is available to Subscribers Only