Functional Testing

Best-corrected visual acuity for each eye was assessed via the Electronic Visual Acuity tester (EVA; JAEB Center, Tampa, Florida, USA) under photopic conditions (100 cd/m ² ) and expressed as the logarithm of the minimal angle of resolution (logMAR).

Cone-mediated light sensitivity was measured at 8 retinal loci by the Humphrey Field Analyzer (Carl Zeiss Meditec, Inc) in the 24-degree radius field using the 24-2 SITA protocol and a white stimulus on white background. The diameter of the target stimulus was 0.43 degree of visual angle. Light sensitivity thresholds were expressed in decibel units (dB) of attenuation. In Figure 1 (Bottom left), the white dots represent the 8 test target locations where cone-mediated light sensitivity was measured within the region of SDOCT imaging, as denoted by the green rectangle.

Macular locations used for structure-function analysis. The green rectangle delineates the area of spectral-domain optical coherence tomography Spectralis imaging on a color photograph of the fundus (Top left). The large green spot represents the location and size of the test target where rod-mediated dark adaptation was evaluated. The small white spots represent the locations and sizes of 8 test targets where cone-mediated light sensitivity was measured. Total retinal pigment epithelium (RPE)-drusen complex volume, abnormal volumes, and retinal volumes (outer, inner, and total) were calculated centered at the fovea within a 12 degree diameter signified by the black circle (Top right), within a 6 degree diameter at each of 8 perimetry points outlined in white circles (Bottom left), and within a 6 degree diameter at the dark adaptation point outlined by the green circle (Bottom right). A consistent truncation of the volumes was applied at the margin of the area of OCT imaging for all eyes.

Rod-mediated dark adaptation was measured psychophysically using AdaptDx (MacuLogix, Hummelstown, Pennsulvania, USA), a computer-automated dark adaptometer described previously. Before testing, the eye was dilated with 1% tropicamide and 2.5% phenylephrine hydrochloride so that a pupil diameter of ≥6 mm was achieved. Trial lenses were added for the 30 cm viewing distance, if needed, to correct for optical blur. The fellow eye was occluded with an opaque patch. The participant then placed his or her head in the forehead-chinrest of the adaptometer. An infrared camera behind the fixation light displayed the eye on a monitor viewed by the examiner, who positioned the participant’s test eye to the red fixation light using a reticule. The procedure began with a photo-bleach exposure to a flash (0.25 ms duration, 58 000 scotopic cd/m ² s intensity; equivalent ∼83% bleach) while the participant was focused on the fixation light. This bleach has been shown to be sufficiently intense to generate impaired dark adaptation parameters in early AMD patients using a 20-minute-duration test protocol. The flash subtended 4 degrees and was centered at 5 degrees on the inferior vertical meridian (ie, superior to the fovea on the retina). This position was also used for the test target for measuring light sensitivity. During threshold measurement, the participant was instructed to maintain fixation on the red fixation light and to press a response button when a flashing target first became visible within the bleached area. Threshold was estimated using a 3-down/1-up modified staircase procedure described previously, and continued at 30-second intervals for 20 minutes. Log thresholds were expressed as sensitivity in dB as a function of time from bleach offset. The speed of dark adaptation was characterized by the “rod intercept time,” defined as the duration required for sensitivity to recover to a criterion sensitivity value of 5.0 × 10 ⁻³ scotopic cd/m ² (3.0 log units of attenuation of the stimulus). The criterion sensitivity level is located in the latter half of the second component of rod recovery. An increase in the rod intercept time is caused by a slowing of the second component of rod-mediated dark adaptation and thus a rightward shift of the dark adaptation function. In Figure 1 , the large green dot represents the location and size of the test target where rod-mediated dark adaptation was measured. Note that this is located within the region of SDOCT imaging denoted by the green rectangle.

Optical Coherence Tomography

Retinal vascular landmarks on the fundus photographs were used to register the fundus image onto the scanning laser ophthalmoscope retinal image of the 20 × 15 degree SDOCT volume (Spectralis; Heidelberg Engineering, Heidelberg, Germany), which had 60 μm between B-scans. A retinal map was developed, specifying the locations and sizes of test targets for both dark adaptometry and cone-mediated light sensitivity measurements ( Figure 1 ).

The SDOCT structures were qualitatively and quantitatively analyzed for regions of 12 degrees (3.46 mm) diameter centered on the fovea ( Figure 1 , Top right), 6 degrees (1.73 mm) diameter centered on each of the 8 cone sensitivity test sites ( Figure 1 , Bottom left), and 6 degrees (1.73 mm) diameter centered on the dark adaptation site ( Figure 1 , Bottom right). SDOCT analyses were masked to all color fundus photograph, functional, and clinical data. Qualitative analysis included assessment of the vitreoretinal interface, hyperreflective foci within the retina, external limiting membrane, photoreceptor inner segment ellipsoid band, and interdigitation zone (where RPE apical processes and photoreceptor outer segments interface). In our dataset, we found no SDOCT findings (eg, fluid or gross atrophy) that would typically be associated with advanced AMD.

For quantitative SDOCT analysis, semi-automated segmentation of the inner retina, outer retina, and RPE was performed using proprietary software (Duke SDOCT Retinal Analysis Program, DOCTRAP V14.1.2; Duke University, Durham, North Carolina).

Automated segmentation was reviewed for errors by trained graders. The foveal center was chosen manually based on the deepest site in the foveal pit. Retinal thickness was measured from the inner aspect of the internal limiting membrane to the inner aspect of the RPE plus drusen complex ( Figure 2 , between lines 1 and 3). Inner retinal layers were measured from the inner border of the internal limiting membrane to the inner border of the outer plexiform layer ( Figure 2 , between lines 1 and 2). Thickness of the photoreceptor layer was measured from the outer plexiform layer to the inner border of the RPE or any overlying SDD material ( Figure 2 , between lines 2 and 3). RPE-drusen complex thickness extended from the inner aspect of the RPE plus drusen material to the outer aspect of the Bruch membrane ( Figure 2 , between lines 3 and 4). Thus, according to previously published definitions, the RPE-drusen complex volume contained all extracellular deposits internal to the RPE (SDD) and external to the RPE (basal laminar deposits, basal linear deposits/drusen), as well as RPE apical processes and cell bodies, whether normal, hypertrophic, or atrophied.

Semi-automated segmentation of the inner retina, outer retina, and retinal pigment epithelium (RPE)-drusen complex by Duke SDOCT Retinal Analysis Program V14.1.2. The inner retina is delineated from the internal limiting membrane to the inner aspect of the outer plexiform layer (between lines 1 and 2), while the outer retina is delineated from the inner aspect of the outer plexiform layer to the inner border of the RPE-drusen complex (between lines 2 and 3). The RPE-drusen complex extends from the inner aspect of the retinal pigment epithelium plus drusen material to the outer aspect of Bruch membrane (between lines 3 and 4), as seen in the color legend (Bottom right). Representative examples are shown for: No Apparent Aging (Top left), Early AMD (Top right), Intermediate AMD (Bottom left). Note the interdigitation zone is less apparent in the Early (Top right) and Intermediate AMD groups (Bottom left) than in the No Apparent Aging group (Top left).

We calculated the volumes of the retina, inner and outer retina, and RPE-drusen complex by first centering the layer boundary positions on the fovea, then generating thickness maps 73 × 1024 pixels in size. These maps were interpolated to 1001 × 1001 pixels to achieve equivalent resolutions in both en face (X-Y) directions. Thickness values were converted from pixels to micrometers according to the Spectralis SDOCT imaging axial resolutions. The volumes consisted of the thicknesses integrated over the 12 degree (3.46 mm) diameter region centered on the fovea ( Figure 1 , Top right, black ring), within a 6 degree (1.73 mm) diameter region centered on each of the 8 cone sensitivity test sites ( Figure 1 , Bottom left, white rings), and within a 6 degree (1.73 mm) diameter region centered on the dark adaptation site ( Figure 1 , Bottom right, green ring). Several of the regions of OCT volumes were truncated at the margin of the SDOCT scan; these were at the same location for all eyes ( Figure 3 ). Overlaid thickness maps and representative B-scans for each group are illustrated in Figure 3 .

Color fundus photographs and retinal pigment epithelium (RPE)-drusen complex volume maps in the 4 age-related macular degeneration study groups. (Top row) Color fundus photographs with schematic overlay of dark adaptation site and perimetry sites (See Figure 1 ). (Middle row) Overlay of retinal thickness maps from Spectralis SDOCT on color fundus photography. (Bottom row) Representative foveal B-scans are also shown for each: No Apparent Aging (first column), Normal Aging (second column), Early AMD (third column), Intermediate AMD (fourth column).

The RPE-drusen complex abnormal thickening and abnormal thinning volumes were also computed. The volume of RPE-drusen complex abnormal thickening is a measure of RPE-drusen complex that exceeds 3 standard deviations above the mean of each SDOCT axial measurement point from the normative dataset composed of the 15 No Apparent Aging subjects ( Figure 4 , orange tones). RPE-drusen complex abnormal thinning volume (mm ³ ) is defined as the cumulative loss of volume from sites with volume loss of 2 or more standard deviations below the mean at each location based on the mapped SDOCT thicknesses from 15 No Apparent Aging subjects ( Figure 4 , blue tones). RPE-drusen complex abnormal thinning is thus the cumulative absolute value of deviation of RPE-drusen complex thickness measurements below the macular site-specific normal range. Detailed methods have been published for these analyses in a larger AMD and control population in the Age-Related Eye Disease Study 2 (AREDS2) Ancillary SDOCT Study.

A map of abnormal retinal pigment epithelium (RPE)-drusen complex thickness (Left) within the 12-degree-diameter macular field of an eye with intermediate age-related macular degeneration (AMD), and an exaggerated schematic diagram (Right) of cross-sectional RPE-drusen complex thickness within the region of the white box. At right, the black line is the segmentation line of the inner border of the RPE-drusen complex layer in this AMD eye, and the white line is mean RPE-drusen complex thickness calculated from the No (retinal) Aging dataset. In both images, the dark gray area is the normal range of RPE-drusen complex thickness based on the non-AMD dataset, with upper limit defined as ≥3 standard deviations above the mean and lower limit defined as ≥2 standard deviations below the mean in the No Apparent Aging dataset. In both images, the orange tones are the areas included in RPE-drusen complex abnormal thickening volume, obtained from RPE-drusen complex thickness measurements greater than the normal range, and the blue tones are the areas included in RPE-drusen complex abnormal thinning volume, defined as the volume lost owing to RPE-drusen complex thickness measurements below the normal range. Note that RPE-drusen complex abnormal thickening volume can decrease without necessarily increasing RPE-drusen complex abnormal thinning volume.

A post hoc analysis was performed to compare OCT-based-drusen volume in intermediate AMD from this study (n = 46) to the baseline AREDS 2 Ancillary SDOCT Study Intermediate AMD group without geographic atrophy (n = 184). All volumes were calculated as described above over the same retinal area of 12 degrees (3.46 mm) diameter centered on the fovea. The mean and standard deviation of drusen volume per mm ² were compared between groups using the Wilcoxon rank sum test.

Biostatistical Approach

We compared the 3 different functional assessments with their corresponding structural SDOCT assessment (by region)—visual acuity (foveal centered region, black ring in Figure 1 ), cone-mediated light sensitivity (6 regions, white rings in Figure 1 ), rod-mediated dark adaptation as assessed by rod intercept time (single superior region, green ring in Figure 1 )—and to SDOCT assessments for the entire macular region. Analyses were performed for AMD subgroups based on fundus appearance using the Clinical Classification System, across the entire study population.

Statistical analysis was performed using JMP Pro V11 (SAS, Cary, North Carolina, USA) and SAS V9.4 (SAS). Demographic and visual functional characteristics were compared between groups using χ ² tests and analysis of variance (or nonparametric equivalents). Age-adjusted Spearman correlation coefficients were calculated for the association between functional and structural variables. Analysis of variance was also used to compare functional and structural variables between eyes with and without SDD, adjusted for age.