Quantification of Nonperfusion and Neovascularization on Ultrawide-Field Fluorescein Angiography in Patients with Diabetes and Association with Vitreous Hemorrhage, Macular Edema, and Vitrectomy

Abstract

Purpose : To quantify the surface areas of nonperfusion (NP) and neovascularization (NV) in patients with diabetes using ultrawide-field (UWF) fluorescein angiography (FA) and evaluate associations with diabetic retinopathy risk factors, and to calculate a threshold NP area associated with increased risk for proliferative disease

Methods : We evaluated nine years of clinical data via retrospective chart review of adult patients diagnosed with diabetes that received a UWF FA obtained from Optos 200Tx or California machines at the University of Michigan Kellogg Eye Center (January 2009 and May 2018) after approval from the Institutional Review Board. Trained, masked graders segmented NP and NV regions for analysis. Eyes with severe media opacities, indistinguishable areas of NP or NV, or prior panretinal photocoagulation were excluded. Surface area was calculated with Optos research software and ITK-SNAP for a nominal eye diameter of 24mm. Multiple linear regression analysis was completed to determine significant associations

Results : A total of 651 eyes were included: 76 eyes with no diabetic retinopathy (DR), 92 with mild non-proliferative diabetic retinopathy (NPDR), 144 with moderate NPDR, 101 with severe NPDR, 220 with proliferative diabetic retinopathy (PDR), and 18 with DR of unknown severity. Of 363 patients, 299 had type 2 diabetes (82.3%), 61 (16.8%) had type 1, and 3 had unknown type (0.8%). Vitreous hemorrhage (VH) was found in 18 eyes (2.8%) and had a significant positive association with total NP area (difference = 30.00, 95% CI: 5.26-54.75, p=0.02). Presence of diabetic macular edema (374 eyes, 57.4%) and history of pars plana vitrectomy (35 eyes, 5.4%) had no significant associations with NP or NV. We identified a threshold total area of NP of 77.48 mm², above which patients have an increased risk of developing PDR (sensitivity of 59.5% and specificity of 73.6%).

Conclusions : Our results indicate that the surface areas of NP and NV can be quantified on UWF FA. VH is significantly associated with a greater total area of NP in patients with diabetes. Eyes with at least 77.48 mm² of total NP are at increased risk of developing PDR.

This is a 2020 ARVO Annual Meeting abstract.

 

Example segmentation of patient with PDR. Green indicates focal avascular zone, red indicates nonperfusion, and purple indicates neovascularization.