Participants

Control participants (n = 16) and those with early AMD (n = 16) were recruited for this study from among staff, students, and volunteers attending the eye clinic at the School of Optometry and Vision Sciences and the eye unit at the University Hospital of Wales. All participants had a corrected visual acuity (VA) of 0.3 logarithm of the minimal angle of resolution units (approximately 20/40) or better assessed using an Early Treatment Diabetic Retinopathy Study sight chart and a refractive error of less than ± 6 diopters. Participants were excluded if they had secondary retinal disease, significant cataract (Lens Opacities Classification System III grade 4 or more for any criteria ), or narrow iridocorneal angles (grade 1 or less assessed by Van Herick). Each participant was given a full explanation of the procedures involved, and their written informed consent was obtained before participation in the study.

Participants were categorized into either a control or early AMD group, depending on the assessment of 37-degree digital fundus images (CR-DGi nonmydriatic retinal camera; Canon Inc, Lake Success, New York, USA) and 20-degree 1060-nm OCT images. Images were assessed for AMD-related features located within a 6000μm diameter centered on the fovea. Definitions were based on the International and the Age-Related Eye Disease Study AMD classification systems. Control participants exhibited no features associated with AMD, with any drusen present being less than 125 μm in diameter (hard drusen). Early AMD was defined as the presence of soft drusen (> 125 μm diameter), pigment changes, or drusenoid pigment epithelial detachment in the absence of any feature of advanced AMD (wet or dry) as defined by the Age-Related Eye Disease Study grading system. Classification was carried out by on of the authors (A.W.) and was confirmed independently by 2 other authors (A.B. and T.M.). One drop of tropicamide 1.0% was instilled into both eyes of each participant, ensuring pupil dilation of at least 7 mm before obtaining fundus photographs and OCT images. Images were obtained from both eyes of all individuals to determine their AMD status. One eye was selected for analysis from each participant; this was the eye with a diagnosis of early AMD or, in the case of bilateral early AMD or controls, the eye with the better VA, with the left eye used as default.

Imaging

Three-dimensional OCT imaging at 1060 nm was performed with less than 2.5 mW at the cornea, below the maximum power limit for a 10-second exposure. OCT volumes were acquired across a 20 × 20-degree (5.76 × 5.76-mm) field consisting of 512 × 512 A-scans obtained at a rate of 47 000 A-scans/second (approximately 8-7 μm axial resolution). OCT volumes were centered on the fovea, aligned by participant fixation. Axial length (cornea to RPE) measurements were acquired using optical biometry (IOL Master; Zeiss, Jena, Germany) for each eye by averaging 5 measurements.

Image processing and analysis was undertaken using ImageJ software (ImageJ, Bethesda, Maryland, USA). Raw OCT images were enhanced digitally to improve the visibility of the retina and choroidal boundaries. Postprocessing procedures were carried out subjectively and included adjustments to brightness and contrast, B-scan registration, despeckling or noise removal, and application of Gaussian or convolving blur (ie, controlled blurring of the image between adjacent b-scans). Measurements were made using the caliper function in the axial plane of the images. The thickness represented by each pixel was calculated assuming a refractive index of 1.4.

Retinal thickness was measured from the most anterior hyperreflective line, which corresponds to the inner limiting membrane, to the center of the most posterior hyperreflective line, which corresponds to the RPE. Choroidal thickness was measured from the RPE to the choroid–sclera boundary. Thickness measurements were obtained for both the retina and the choroid at the fovea and then at 0.5-mm intervals to 2 mm nasally (N), temporally (T), superiorly (S), and inferiorly (I; see Figure 1 ). This produced thickness measurements at 17 individual retinal locations for each eye. At each location, the distribution of thickness measurements was checked for normality, and an independent t test was conducted between the control and early AMD groups.

Optical coherence tomography (OCT) layer boundaries and measurement locations used in this study. (Left) Cross-sectional 1060-nm OCT B-scan, the superior row of arrows indicating the location of the inner limiting membrane (ILM), the middle row indicating the retinal pigmented epithelium (RPE), and the inferior row indicating the choroid–sclera boundary. These features were used to delineate the boundaries of the retina and choroid for thickness measurements. (Right) Fundus photograph with retinal locations overlaid (solid dots) at which each retinal and choroid thickness measurement was obtained. Retinal direction is indicated by S (superior), I (inferior), N (nasal), and T (temporal) as labeled.

Additionally, the fundus and OCT images were assessed subjectively; specifically, the participants regarded at the highest risk of progression to advanced AMD (ie, those with drusenoid pigment epithelial detachment, confluent drusen, or a fellow eye with advanced AMD ) were compared with those at lowest risk (ie, those with normal fellow eyes and none of the high-risk features listed above).