Spectral domain optical coherence tomography imaging of drusen in nonexudative age-related macular degeneration

Abstract

Purpose: To measure drusen area and volume in eyes with nonexudative age-related macular degeneration (AMD) using spectral domain optical coherence tomography imaging (SD-OCT).

Design: Evaluation of diagnostic technology.

Participants: One hundred three eyes from 74 patients with drusen.

Methods: Patients with drusen secondary to nonexudative AMD were enrolled in this study. Five separate SD-OCT scans, each consisting of 40 000 uniformly spaced A-scans organized as 200 A-scans in each B-scan and 200 horizontal B-scans, were performed on each eye. Each scan covered a retinal area of 6×6 mm centered on the fovea. A novel algorithm was used to quantitatively assess drusen area and volume. Measurements from the entire scans, as well as in regions contained within 3- and 5-mm circles centered on the fovea, were analyzed. Test-retest standard deviations of drusen area and volume measurements were calculated for each eye.

Main outcome measures: Drusen area and volume.

Results: The algorithm created drusen maps that permitted both qualitative and quantitative assessment of drusen area and volume. Both the qualitative appearance and the quantitative measurements of drusen area and volume were highly reproducible over the 5 different datasets. The intraclass correlation coefficient was >0.99 for both area and volume measurements on the entire dataset as well as the 3- and 5-mm circles. The correlation between lesion size and the test-retest standard deviations can be eliminated by performing a square root transformation of the area measurements and a cube root transformation of the volume measurements. These transformed data allowed for the inclusion of all drusen sizes in the calculation of an estimated single pooled test-retest standard deviation, which will be useful for longitudinal studies of drusen natural history.

Conclusions: A novel algorithm for the qualitative and quantitative assessment of drusen imaged using SD-OCT was shown to be highly reproducible. The ability to assess drusen volume reliably represents a new quantitative parameter to measure in AMD and may be useful when assessing disease progression, particularly in trials for treatments of nonexudative AMD.

Figure 1

Imaging small to medium drusen with spectral domain optical coherence tomography (SDOCT). A. A color fundus photo is shown with the superimposed location of the SDOCT dataset indicated by the blue square. This location can be determined with the help of the OCT fundus image. B. Two B-scans are shown from the SDOCT that correspond to the white lines on the fundus photo. The segmentation of the retinal pigment epithelium (RPE) is shown in red and the drusen floor is shown in yellow. C. Surface rendering of the RPE segmentation. D. Retinal thickness map. E. Surface rendering of the virtual RPE free of any deformations and referred to as the RPE floor. F. Drusen thickness map.

Figure 2

Imaging large drusen with spectral domain optical coherence tomography (SDOCT). A. A color fundus photo is shown with the superimposed location of the SDOCT dataset indicated by the blue square. This location can be determined with the help of the OCT fundus image. B. Two B-scans are shown from the SDOCT that correspond to the white lines on the fundus photo. The segmentation of the retinal pigment epithelium (RPE) is shown in red and the drusen floor is shown in yellow. C. Surface rendering of the RPE segmentation. D. Retinal thickness map E. Drusen thickness map.

Figure 3

Identification of the foveal center from within the spectral domain optical coherence tomography (SDOCT) dataset. From within the SDOCT dataset, the location of a particular B-scan (right) and the location of the A-scan corresponding to the foveal center were identified. The A-scan containing the fovea is marked by a white line. This choice determines crosshairs intersecting at the fovea on the OCT fundus image (left). The black horizontal line corresponds to the location of the B-scan on the right. The white vertical line indicates the position of the particular A-scan also marked in white on the right.

Figure 4

Registration of the spectral domain optical coherence tomography (SDOCT) scan on a color fundus image using the OCT fundus image. A. The SDOCT dataset can be registered to a color fundus photo of the same eye by overlapping the OCT fundus image on the color fundus image. The fovea location is marked by a green cross on the OCT fundus image, while the white lines show circles centered on the foveal center with 3 and 5 mm diameters, respectively. B. Using the registration in part A, the foveal location and the position of the circles can be superimposed on the fundus photo.

Figure 5

Reproducibility of drusen maps obtained from spectral domain optical coherence tomography (SDOCT) datasets. Five drusen maps are shown from five different SDOCT datasets of the same eye. The location of the 3mm circle centered at the fovea is shown in white. Drusen area and volume within the 3mm circle are shown under each drusen map.

Figure 6

Relationship between the standard deviations of the measurements and the mean drusen volume measurements: A. Full datasets. B. 5mm circles. C. 3mm circles.

Figure 7

Relationship between the standard deviation of the measurements and the mean drusen area measurements: A. Full datasets. B. 5mm circles. C. 3mm circles.

Figure 8

Relationship between the standard deviation of the measurements and the mean for the cube root of drusen volume: A. Full datasets. B. 5mm circles. C. 3mm circles.

Figure 9

Relationship between the standard deviation of the measurements and the mean for the square root of drusen area: A. Full datasets. B. 5mm circles. C. 3mm circles.

Figure 10

Relationship between drusen area and drusen volume: A. Drusen area plotted against volume measurements for the 3mm circle. B. Square root of drusen area plotted against the cube root of drusen volume for the 3mm circle.