June 2020
Volume 61, Issue 7
ARVO Annual Meeting Abstract  |   June 2020
Predictive Value of ERG, OCT-A, and UWF-FA in Patients with Diabetic Retinopathy
Author Affiliations & Notes
  • Mitchell G Brigell
    Aerpio Therapeutics, Belmont, Massachusetts, United States
  • Quentin Davis
    LKC Technologies, Gaithersburg, Maryland, United States
  • Nadia K Waheed
    Ophthalmology, New England Eye Center/Tufts University, Boston, Massachusetts, United States
  • Footnotes
    Commercial Relationships   Mitchell Brigell, Aerpio Pharmaceuticals (E); Quentin Davis, LKC Technologies (E); Nadia Waheed, Apellis (C), Apellis (S), Astellas (C), Astellas (S), Boehringer Ingelheim (C), Boehringer Ingelheim (S), Boston Image Reading Center (I), Carl Zeiss Meditec (S), Gyroscope (S), Nidek Medical Products (S), Novartis (C), Novartis (S), Ocudyne (I), Regeneron (C), Roche/Genentech (C), Roche/Genentech (S), Topcon (S), Topcon (C)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2020, Vol.61, 4038. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Mitchell G Brigell, Quentin Davis, Nadia K Waheed; Predictive Value of ERG, OCT-A, and UWF-FA in Patients with Diabetic Retinopathy. Invest. Ophthalmol. Vis. Sci. 2020;61(7):4038.

      Download citation file:

      © ARVO (1962-2015); The Authors (2016-present)

  • Supplements

Purpose : To assess the value of electroretinography (ERG), OCT-angiography (OCT-A), and ultra-wide field fluorescein angiography (UWF-FA) parameters to predict progression of diabetic retinopathy in eyes with moderate to severe non-proliferative diabetic retinopathy (NPDR).

Methods : In the TIME-2b trial 167 patients with NPDR were randomized to receive 48 weeks of treatment with AKB-9778 (a Tie2 activator) 15 mg once or twice daily, or placebo. Eligible eyes had ETDRS DR severity (DRSS) of 43, 47 or 53, BCVA of 20/30 or better, and no center-involved DME. Measurements at baseline included non-mydriatic flicker ERGs (71 patients; RETeval device, DR Assessment protocol, Maa et al. 2016), OCT-A (74 patients), and OPTOS UWF-FA (101 patients). DRSS was graded on 7-field fundus photographs. Progression to DME was defined as an increase of central subfield thickness of ≥ 20% or an adverse event of DME confirmed by the image reading center. PDR was defined as a DRSS of >53, an AE involving neovascularization, or treatment for PDR. ROC analysis was used to assess predictive value of baseline measurements and pairs thereof.

Results : AKB-9778 did not have a statistically significant effect on DRSS or progression to DME/PDR, so data from all study arms were combined. 42 eyes progressed to DME/PDR; 21 of these eyes had ERGs, 17 had OCT-A, and 13 had UWF-FA. UWF-FA ischemia index, leakage index, and micro-aneurysm count were not predictive of progression. Statistically significant predictors (from worst to best, AUROC, 95% CI) include increased DRSS level (.60, .51–.69), decreased pupil response during ERG testing (.67, .55–.77), increased OCT-A FAZ area (.68, .52–.82), increased OCT-A total vessel length (.68, .56–.80), increased flicker implicit time (.73, .61–.83), and increased DR Assessment score (.79, .70–.87). The best pair was a logistic regression of the ERG-based DR Assessment score and OCT-A non-perfusion area (.86, .75–.94). When both tests were abnormal 73% (8/11) of eyes progressed. When both tests were normal 3.6% (1/28) of eyes progressed.

Conclusions : Prediction of progression of NPDR is important in the management of the disease. Pre-treatment ERG and OCT-A parameters were better predictors of progression to DME/PDR than DRSS or UWF-FA over 48-weeks of follow-up. The combination of functional and structural measures was better than either alone.

This is a 2020 ARVO Annual Meeting abstract.


This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.