Neovascular and Non-Neovascular Subretinal Fluid Associated with Angioid Streaks: Comparative Imaging Analysis with Long-Term Follow-up.
TL;DR
The findings emphasize the importance of multimodal imaging confirmation of CNV to distinguish neovascular from non-neovascular fluid and avoid interpreting SRF alone as evidence of CNV activity and avoid interpreting SRF alone as evidence of CNV activity.
Abstract
Purpose
To characterize the multimodal imaging features and long-term visual and structural outcomes of subretinal fluid (SRF) associated with angioid streaks (AS), and to compare neovascular and non-neovascular phenotypes.
Design
Multicenter retrospective cohort study.
Participants
Fifty-three eyes of 43 patients with AS-associated SRF.
Methods
Medical records of patients with AS and SRF were retrospectively reviewed across multiple centers. Clinical and multimodal imaging data, including color fundus photography, fundus autofluorescence, optical coherence tomography (OCT), and OCT angiography, were analyzed at baseline and final follow-up. Eyes were classified as neovascular or non-neovascular based on evidence of choroidal neovascularization (CNV). OCT parameters included central macular thickness (CMT), subfoveal choroidal thickness (SFCT), SRF dimensions, and outer retinal integrity. Longitudinal visual and anatomical outcomes were compared between groups. MAIN OUTCOME MEASURES Visual acuity (VA), SRF evolution, OCT structural changes, development of complete retinal pigment epithelium and outer retinal atrophy (cRORA), and comparison of outcomes between neovascular and non-neovascular SRF phenotypes.
Results
Mean baseline VA was 0.36 ± 0.34 LogMAR (∼20/45 Snellen). CNV was identified in 37 eyes (69.8%), whereas 16 eyes (30.2%) demonstrated non-neovascular SRF. Mean follow-up duration was 4.1 ± 4.0 years. In the overall cohort, both CMT and SFCT significantly decreased over time (p<0.001 for both), whereas the prevalence of cRORA increased from 54.7% at baseline to 64.2% at final follow-up. SRF prevalence decreased in both groups, from 100% to 40.5% in neovascular eyes and from 100% to 72.7% in non-neovascular eyes. Final VA was similar between neovascular and non-neovascular groups (0.40 ± 0.30 LogMAR [∼20/50 Snellen] vs. 0.32 ± 0.20 LogMAR [∼20/40 Snellen]; p=0.62). Rates of cRORA at final follow-up were likewise comparable between groups (64.3% vs. 63.6%).
Conclusions
SRF associated with AS may result from exudative CNV or non-neovascular mechanisms likely related to retinal pigment epithelium dysfunction. Despite distinct imaging phenotypes, long-term visual and structural outcomes were comparable between groups. These findings emphasize the importance of multimodal imaging confirmation of CNV to distinguish neovascular from non-neovascular fluid and avoid interpreting SRF alone as evidence of CNV activity.