Macular imaging with optical coherence tomography in glaucoma

Abstract

With the advent of spectral-domain optical coherence tomography, imaging of the posterior segment of the eye can be carried out rapidly at multiple anatomical locations, including the optic nerve head, circumpapillary retinal nerve fiber layer, and macula. There is now ample evidence to support the role of spectral-domain optical coherence tomography imaging of the macula for detection of early glaucoma. Macular spectral-domain optical coherence tomography measurements demonstrate high reproducibility, and evidence on its utility for detection of glaucoma progression is accumulating. We present a comprehensive review of macular spectral-domain optical coherence tomography imaging emerging as an essential diagnostic tool in glaucoma.

Keywords: OCT; detection; glaucoma; imaging; macula; optical coherence tomography; progression; structure-function; variability.

Figure 1.

A) A printout of Cirrus HD-OCT’s macular OCT image displaying full macular thickness measurements presented in the Early Treatment of Diabetic Retinopathy Study (ETDRS) grid format. Top row, the thickness maps demonstrate a false color map of the central macular retinal thickness and full retinal thickness in ETDRS regions. Left middle and bottom panels provide raw OCT images. Right column images represent 3-D images of macula (top) and the top surfaces of the inner limiting membrane (middle) and the retinal pigment epithelium (bottom). The very bottom row on the right displays the central subfield thickness, macular cube volume, and macular cube average from left to right, respectively, with their corresponding color scheme for statistical significance. B) The Ganglion Cell Analysis printout provides the ganglion cell/inner plexiform (GCIPL) layer thickness measurements in an ellipse 4.8×4.0 mm in size excluding the central 1.2×1.0 mm (the central fovea). This elliptical region is divided into 6 wedge-shaped sectors where the GCIPL thickness is averaged. Top row, the thickness maps demonstrate a false color map of the GCIPL thickness; middle row, deviation maps and global, sectoral, and minimum GCIPL thickness measurements. The former show regions where the GCIPL thickness has fallen below the 5% limit based on the normative database with the yellow color representing measurements with p value <0.05 and >0.01 and the red superpixels representing superpixels where the GCIPL thickness has decreased to below the 1 percentile cutoff point in the normative database. Bottom row, raw OCT images provide the opportunity for the reviewer to inspect the quality of the layer segmentation.

Figure 2.

The ganglion cell complex (GCC) printout of the left eye of a glaucoma patient with significant and extensive inferior macular GCC thinning acquired with the RTVue SD-OCT (OptoVue Inc., Fremont, CA, USA). The GCC thickness map is represented in the upper left; the inferior temporal macular region displayed in blue (red arrow) demonstrates a significant reduction in the GCC thickness. The upper right graph represents the significance map and is pseudocolor coded; GCC thickness in areas shown in green falls within the normal range (5%−95% prediction interval in the normative database). The yellow color indicates borderline abnormal GCC thickness, i.e., the thickness measurement falls between 1 and 5 percentile cutoff points in the normative database, and the red color displays areas where the GCC thickness measurements are outside normal limits (<1 percentile cutoff in the normative database). The central masked area shown in gray on the right upper image is explained by the lack of retinal ganglion cells in the center of the fovea. The table (middle on top) provides summary GCC parameters including average, superior, and inferior thickness, intra-eye superior-inferior asymmetry, focal loss volume (FLV) and global loss volume (GLV). High-resolution B-scans with GCC segmentation results are also provided in the lower section of the printout.

Figure 3.

Printouts of the posterior Pole Algorithm of the Spectralis OCT demonstrating a grid of 64 superpixels, 3°×3° in size, providing the full macular thickness (A) or the ganglion cell layer thickness (B) in a 24°×24° region of the macula centered on the fovea. Note that the pseudocolor scales used are different.

Figure 4.

A schematic model of superimposed retinal nerve fiber layer (RNFL) and macular thickness maps (ganglion cell/inner plexiform layer in A and RNFL in B). A) Projection path of retinal ganglion cell (RGC) axons from the macular region to corrresponding areas of the optic disc. Blue lines in the inferotemproal region of the disc define the optic disc sector to which RGC axons from the macular zone of vulnerability (MZV) project. B) A representation of the MZV (region delineated in red) and its projection onto the optic disc (defined by blue lines. The temporal and superior region with a black boundary tend to be damaged later in glaucoma. (Figure published with permission from Hood et al.)

Figure 5.

A comparison of the area under the receiver operating characteristics curves (AUC) for detection of perimetric glaucoma between the inferior quadrant retinal nerve fiber layer thickness alone and the inferior RNFL combined with the minimum ganglion cell/inner plexiform layer (GCIPL) thickness (the best macular parameter in the study). The AUC was higher for the combined variable (p value =0.041 for comparison of the partial AUCs). (Figure with permission from Nouri-Mahdavi et al.)

Figure 6.

Vertical full macular thickness asymmetry across the horizontal meridian on the Posterior Pole Algorithm of the Spectralis OCT as an early sign of glaucomatous damage in the right eye of a patient with early glaucoma. The inferior macula demonstrates a large number of superpixels flagged as various shades of gray indicating thinning compared to the superior macula.

Figure 7.

Schematic representation of the vertical asymmetry algorithm developed by Yamada et al. for Spectralis OCT. The minimum asymmetry unit was calculated as the absolute value of the logarithm of the ratio of the average upper (Ux) and lower thickness (Lx) measurements (|log₁₀ (Ux/Lx)|) for each of the 10 vertical scans (in red). The Asymmetry Index for each eye was then estimated by averaging the asymmetry values for the 10 vertical scans. (Printed with permission from Yamada et al.)

Figure 8.

Definition of the temporal vertical asymmetry on Cirrus HD-OCT according to Sharifipour et al.. The 200×200 macular cube of Cirrus SD-OCT was arranged as a 20×20 grid of superpixels. The best performing asymmetry index compared average thickness measurements of the 8 superpixels temporal to the fovea on the 3 rows above and below the temporal raphe. (Figure published with permission from Sharifipour F. et al.)

Figure 9.

Microcystic macular edema (MME) manifests as hyporeflective cystic and lacunar areas within the inner nuclear layer on OCT cross-sectional images (red arrow).

Figure 10.

A) The Fovea-Bruch membrane opening (FoBMO) angle as measured on co-registered images of the optic disc and macular cubes of Cirrus HD-OCT. B) The Spectralis OCT automaticallay measures the FoBMO after delineating the BMO and the fovea. C) The ganglion cell/inner plexiform thickness asymmetry across the horizontal meridian is influenced by the FoBMO angle; a more (negatively) tilted FoBMO angle is associated with relatively thinner inferior GCIPL thickness along the horizontal raphe compared with the superior region. (Figure published with permission from Ghassabi et al.)

Figure 11.

A one-page OCT report providing an overlay of the retinal nerve fiber layer (RNFL) and ganglion cell/inner plexiform layer (GCIPL) OCT measurements and 24–2 and 10–2 visual field test locations. A) The raw image of the circumpapillary scan displayed in the NSTIN format, i.e., starting from the nasal quadrant going counterclockwise in the right eye. B) The circumpapillary-RNFL thickness plot obtained from the disc cube scan from panel C, right; it corresponds to the raw image of the OCT scan in panel A and is displayed shown in NSTIN orientation. C) The RNFL thickness maps from the OCT cube scan of the macula (left) and the optic disc (right). D) The GCIPL thickness color map from the OCT cube scan of the macula. For both panels C and D, warmer colors represent thicker RNFL or GCIPL measurements. E) The RNFL probability map is shown in visual field view with the 24–2 locations overlaid. Warmer colors represent deeper defects. F) The GCIPL probability map based upon the thickness maps in panel D with superimposed 10–2 visual field locations. The black circles in panels E and F demonstrate the boundaries (±8°) of the macula; the color bars on the right provide the probability cutoffs for RNFL (E) and GCIPL (F) thickness measurements. (Figure published with permission from Hood et al.)

Figure 12.

The scatter plots demonstrate that the structure-function relationship between ganglion cell inner plexiform layer thickness and total deviation at central 10–2 visual field locations improves after adjustment for retinal ganglion cell displacement (A, before adjustment, B, after adjustment).

Figure 13.

A broken-stick model of cross-sectional structure-function relationships for 5 different macular variables (full macular thickness, ganglion cell complex, ganglion cell/inner plexiform layer, ganglion cell layer and inner plexiform layer). The Y-axis represents thickness measurements at 3°×3° superpixels measurements and the X-axis shows total deviation values at corresponding 10–2 visual field locations. The red brackets demonstrate the dynamic range for each macular parameter. (Figure published with permission from Miraftabi A et al.)

Figure 14.

A) The 6 sectors defined on the Cirrus HD-OCT printout: S = superior, ST = superior temporal, IT = inferior temporal, I = inferior, IN = inferior temporal, SN = superior nasal, SN = superior nasal. B) 10–2 visual field locations demonstrating the highest correlations with the ganglion cell/inner plexiform layer (GCIPL) thickness in various sectors; 21 out of 68 visual field locations were significantly correlated with sectoral GCIPL thickness measurements and are displayed by colored squares. Two locations superiorly and 1 inferior location marked by black circles are locations already tested with 24–2 pattern. (Figure published with permission from Lee JW et al.)

Figure 15.

A) Spatial distribution of progressive thinning of the ganglion cell/inner plexiform layer (GCIPL) in a study by Shin et al.; 292 glaucoma eyes were followed over an average of 6 years. Warmer colors display more frequent occurrence of progressive thinning. B) Spatial distribution of ganglion cell/inner plexiform layer (GCIPL) thinning in the central macula at baseline created by overlaying the magnitude of thickness deviation from the normative database on the GCIPL thickness map. The inferotemporal region was the most frequently affected area in the central macula. The dotted region represents the area demonstrating the most prominent progressive changes. (Figure published with permission from Shin et al.)