Assessing the reliability of automated OCT retinal thickness measurements in patients with choroidal neovascularization due to age-related macular degeneration.

BACKGROUND AND OBJECTIVE: To compare automated versus manual retinal measurement using Stratus optical coherence tomography (OCT) (Carl Zeiss Meditec, Dublin, CA) in patients with choroidal neovascularization (CNV) due to age-related macular degeneration (AMD). PATIENTS AND METHODS: Patients with subfoveal CNV due to AMD underwent Stratus OCT imaging. Each scan was evaluated for automated: foveal center point thickness, central subfield thickness, signal strength, and manual: central retinal thickness, central retinal/lesion thickness, CNV thickness, pigment epithelial detachment thickness, and number of correct retinal boundaries (CRBs). RESULTS: A total of 209 OCT visits (49 patients) were assessed. On fast macular thickness map, 52% of scans had less than 4 (of possible 6) CRBs placed by the automated OCT algorithm. There was a mean difference of 34 microm between automated foveal center point thickness and manual central retinal thickness (P < .00005). Only 40% and 27% of automated foveal center point thickness and central subfield, respectively, lay within +/- 25 microm of manual central retinal thickness. As the number of CRBs or signal strength increased, the difference between manual and automated measurements decreased. CONCLUSION: Manual retinal thickness measurements differ significantly from those calculated by the current automated Stratus OCT software in exudative AMD.