May 31, 2019 Conference


Comparison of Diabetic Retinopathy Severity Identified on Ultrawide Field Retinal Images and Ultrawide Field Fluorescein Angiograms

Return to Session

Omar Abdelal, Joslin Diabetes Center (Presenter)
Dr. Mohamed Elmasry, Joslin Diabetes Center
Siamak Shokrollahi, Joslin Diabetes Center
Dr. Jennifer K. Sun, Joslin Diabetes Center
Dr. Lloyd Paul Aiello, Joslin Diabetes Center
Dr. Paolo Silva, Joslin Diabetes Center


Early identification of diabetic retinopathy (DR) lesions and accurate assessment of DR severity is critical for the care and treatment of patients with diabetes. Standard ETDRS fundus photography evaluates only about 30% of the retinal surface area but is well associated with risk of retinopathy progression.  Fluorescein angiography of this ETDRS area was historically studied in detail and provided information associated with retinopathy progression1. Multicenter clinical trials such as the Diabetes Control and Complications Trial (DCCT) and Early Treatment Diabetic Retinopathy Study (ETDRS) has compared the prognostic value of FA using the modified 30 degree fields 1 and 2 (Figure 1) compared to 7 field stereoscopic photography and have shown shown similar sensitivities between FA and CI in detecting DR.1,2 However, the fluorescein angiographic risk factors studied did not provide a substantial clinical benefit compared with evaluating color photography alone.3

Ultrawide field imaging allows analysis of a greater retinal area than previously possible (~80%), including regions which may experience changes early in the disease and that are highly correlated with future retinopathy progression.4,5 Potentially, evaluating this area with ultrawide field angiography might provide substantial clinical benefit.6,7 Currently, it is unknown given this new imaging modality how UWF-FA and CI compare in the early detection of DR as well as grading of more advanced DR.


  • This study was a multicenter cross-sectional study that included patients over 18, with type 1 or type 2 DM imaged by both 200? UWF-CI and UWF-FA within 1 month of each other. Eyes included for analysis had adequate image quality for grading diabetic retinopathy (DR) and no evidence of panretinal photocoagulation or other retinal vascular diseases. 
  • Stereographically projected images were graded by two independent graders (O.A, M.A) who graded only one series of images (either UWF-CI or UWF-FA) while being masked  to the DR grade provided by the other imaging modality.
  • In addition to the overall ETDRS DR severity level, the grade of individual DR lesions (HMAs, venous beading, IRMAs & neovascularization) was evaluated.
    • A retina specialist experienced in grading UWF images performed side-by-side adjudication between UWF-FA and UWF-CI images with any discrepancies in DR severity level and the cause of this difference was recorded.
  • Sensitivity (SN) and specificity (SP), and simple (K) and linear weighted (KW) kappa statistics were evaluated.


  • Distribution of DR severity using UWF-CI vs UWF-FA was: no DR, 9.39 % vs 1.38%; mild nonproliferative DR (NPDR), 14.92% vs 18.23%; moderate, 34.53% vs 21.55%; severe, 28.73% vs 40.33%; proliferative DR (PDR), 10.5% vs 13.54%; and high-risk PDR, 1.93% vs 4.97%.  
  • Overall agreement between UWF-CI and UWF-FA for DR severity was K=0.510 and Kw=0.708 . DR severity agreement between UWF-CI and UWF-FA images was within 1-step in 98% of eyes and exact in 62.43%.
  • No eye was graded as having more severe DR on UWF images than on UWF-FA. Compared to UWF-CI, UWF-FA identified 1-step more severe DR in 35.6% and 2-step more severe DR in 1.93 % of cases with no discrepancy greater than 2 steps. 
  • Among eyes with PDR by UWF-FA (N =67), 35 (52.2%) were identified as having PDR on UWF-CI [SN=0.522 (CI 0.39 - 0.64), SP=1.00 (0.983 - 1.0)] Among eyes with severe NPDR or PDR by UWF-FA (N =213), 94 (44.1%) were identified as less than severe on UWF-CI [SN=0.558 (0.48-0.62), SP=1.00 (0.968-1.0)].
  • Conversely among eyes with no to mild NPDR on UWF-CI (N=88), 43 (48.8%) had more severe DR on UWF-FA [SN=0.633 (0.51-0.74), SP=1.00 (0.983 - 1.0)


  • In both research and clinical settings, the current gold standard for the diagnosis and management of DR severity remain 7-field stereoscopic ETDRS 30 degree fundus photography. Both the ETDRS and DCCT, the evaluation of ETDRS 2 posterior pole FA fields did not provide substantial benefit in increasing the ability to predict DR progression as compared to color photography alone
  • In this study, the evaluation of stereoscopically projected images UWF-FA identified substantially more severe DR than observed using UWF-CI. Furthermore, the sensitivity of color images is 63.3% and  52.2% when compared to FA in detecting mild NPDR and PDR, potentially identifying eyes with very early disease or subtle new vessels that would be not been seen on UWF-CI.
  • The progression rates to determine the frequency of follow-up and need for treatment have been largely based on findings based ETDRS fundus photography and it is unknown if the more the severe DR severity levels detected on UWF-FA have similar and increased rates of progression.
  • Given the cross sectional design of this study, further evaluation will be needed to determine if those identified as having more severe DR by UWF-FA are at higher risk of progression than UWF-CI grade and whether these findings add significant clinical utility.