Purpose
To assess the relative contribution of various tomographic features on SD-OCT that influence conversion to neovascular AMD
Methods
This study was designed as a retrospective, consecutive, cohort study. Two-hundred and thirty-two fellow eyes of patients initiating treatment for unilateral neovascular AMD from January 2010 to January 2012 were included in the analysis. Baseline demographics including age, gender, and snellen visual acuity were recorded. Baseline SD-OCT (Cirrus HD-OCT, Carl Zeiss Meditec, Dublin, CA) of the non-neovascular eye was reviewed and the following features were recorded: central retinal thickness (CRT), presence or absence of vitreomacular adhesion (VMA), epiretinal membranes (ERM), drusenoid pigment epithelial detachment (dPED) and retinal pigment epithelium (RPE) saw-toothing. Each patient was followed for 12 months, with the primary outcome being conversion to neovascular AMD. The proportion or means of each SD-OCT feature was compared between the convertors and non-convertors. Multivariate logistic regression was performed to determine demographic and SD-OCT features that correlated with neovascular conversion.
Results
Participants had a mean age of 78.4 (SD 8.6), 63% female, and a median Snellen acuity of 20/30. The incidence of neovascular conversion of the cohort at 12 months was 13.4% (31/232). There was no significant difference between presence of VMA (p=0.350), ERM (p=0.117), VMA+ERM (p=0.180), dPED (p=0.730), RPE saw-toothing (p=0.457) or dPED+RPE saw-toothing (p=0.477) between convertors and non-convertors. The only significant difference was the mean CRT being thicker in the convertors group, 276.39 (SD 46.9) µm, compared to the non-convertors, 257.4 (SD 32.1) µm (p<0.01). The multivariate regression model, which included all studied features, revealed the only significant association with conversion was increasing CRT (p=0.01).
Conclusions
The presence of vitreomacular interface abnormalities or outer retinal features detectable on SD-OCT do not appear to influence conversion to neovascular AMD in high-risk patients. This study validates that increasing CRT values continues to be the most important, and conveniently most robust, SD-OCT parameter useful for monitoring those with high-risk non-neovascular AMD.
Keywords: 412 age-related macular degeneration •
550 imaging/image analysis: clinical