Optical coherence tomography-raster scanning and manual segmentation in determining drusen volume in age-related macular degeneration

Abstract

Purpose: Drusen are the hallmark of age-related macular degeneration (AMD), and substantial evidence exists that the amount of drusen and their effect on retinal pigment epithelium is a strong predictor of progression of AMD and vision loss. Until recently, it was not possible to quantitate the volume of the drusen. However, the use of image-stabilized scanning laser ophthalmoscope or spectral domain-optical coherence tomography (OCT) has enabled determination of drusen volume of this abnormal material. The purpose of this study was to assess the correlation of drusen volume with Age-Related Eye Disease Study (AREDS) grade and drusen area in dry AMD.

Methods: Thirty-six eyes from 18 patients with nonexudative AMD with visual acuity between 20/16 and 20/160 were studied. Spectral domain-OCT or simultaneous OCT scans were taken as color fundus photographs (35 degrees ) of each eye. Early Treatment Diabetic Retinopathy Study visions were also recorded. The full AREDS score excluding late-stage AMD was determined by agreement between two trained observers. Drusen volume was determined by examination of a series of 96 spectral domain-OCT scans taken from arcade to arcade for a length of 6 mm. The volume was determined by calculating the drusen area in each scan and determining the drusen volume by calculating the effective volume of each cut using National Institutes of Health Image J. Drusen were identified and outlined manually, not using an automated algorithm.

Results: There was a strong and significant correlation between drusen volume and AREDS-determined drusen area (P < 0.0001, r = 0.78). In addition, there was a correlation between AREDS classification and drusen volume (P = 0.023, r = 0.43) as determined by pairwise correlation.

Conclusion: Drusen volume as determined by spectral domain-OCT correlates with AREDS-determined drusen area and AREDS grade in nonexudative AMD. The correlation is not perfect, however, because drusen area and volume average 40% and 82% of the variation, respectively. Drusen volume can provide additional information in grading the severity of eyes with dry AMD.