Abstract
Purpose :
To analyze the long-term risk of geographic atrophy (GA) development in correlation to choroidal thickness in a consecutive series of patients with treatment-naïve neovascular age-related macular degeneration (n-AMD) on a treat-and-extend regimen (TER) of intravitreal injections of anti-vascular endothelial growth factor (anti-VEGF) agents.
Methods :
Central choroidal thickness (CCT) was retrospectively measured using spectral-domain optical coherence tomography (SD-OCT) at baseline and last follow-up in a cohort of consecutive, treatment-naïve, n-AMD patients that displayed GA formation during their TER of anti-VEGF agents. Neovascular lesion subtyping adhered to the anatomical classification system utilizing fluorescein angiography and SD-OCT, as types 1 (sub-retinal pigment epithelium), 2 (subretinal), 3 (intraretinal), or mixed neovascularization.
Results :
Eighty-six eyes of 86 patients, 24 men (28%) and 62 women (72%) were included in final analysis. They were followed for an average of 57.2 months (median 59.3, range 24.4 - 78.7) and received in average 32.8 injections (median 32.1, mean injection/month 0.62). The mean, square rooted baseline and final area of GA were respectively 0.294 mm2 and 2.443 mm2. The mean baseline and final CCT were respectively163 and 149 micron. The likelihood of GA formation was analyzed in correlation with CCT, neovascularization subtype, number of injections and other variables in univariate and multivariate analyses.
Conclusions :
Baseline CCT does not appear to be a significant independent predictor for the long-term risk of GA development in treatment-naive n-AMD eyes on a TER of anti-VEGF agents. Final CCT is a statistically significant predictor (R2 = 0.48) but the weak slope denotes marginal clinical importance. The total number of injections in multivariate analysis is not a significant predictor to the risk of GA formation. The presence of type 1 neovascular lesions remains the strongest predictor that correlates to a decrease in the risk of GA formation in uni- and multivariate analysis (p < 0.0001).
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.