The Ganglion Cell Complex
Spectral-domain optical coherence tomography allows measurements of macular thickness that may have a complementary role in glaucoma diagnosis.
The measurement of the perifoveal ganglion cell layer recently emerged as a new diagnostic parameter in glaucoma. Various manufacturers of spectral-domain optical coherence tomography (SD-OCT) systems have released or will soon release software upgrades capable of measuring the thickness of the ganglion cell layer. The aims of this article are to introduce the technology and speculate on its future role in glaucoma management.
MACULAR IMAGING FOR GLAUCOMA
Current imaging technologies for diagnosing and evaluating glaucoma rely on measurements of the optic nerve and peripapillary retina. SD-OCT has been shown to be a useful tool in this regard.1 Zeimer and colleagues2 first suggested that, since a significant portion of retinal ganglion cells (RGCs) reside in the macula, a loss of tissue in this region might be a sign of glaucomatous damage. This proposal was based on prior animal studies in which primate models of glaucoma demonstrated a substantial loss of RGCs in the perifoveal region.3
To further elucidate the relationship between macular thickness and glaucomatous damage, Lederer and colleagues4 performed a case-controlled study evaluating macular volume, as measured by time-domain optical coherence tomography (TD-OCT), in normal and glaucomatous eyes. The study group included 70 control eyes, 70 eyes classified as glaucoma suspect, 70 eyes with early glaucoma, and 62 eyes with advanced glaucoma. Macular volume in normal (2.37 +0.11 mm3), glaucoma-suspect (2.33 +0.16 mm3), and early-glaucoma (2.27 +0.13 mm3) eyes was significantly greater than in eyes with advanced glaucoma (2.12 +0.23 mm3; P = .0001, P = .0001, and P = .0008, respectively). In a separate study, Greenfield and colleagues5 demonstrated a correlation between macular thickness measured with TD-OCT and visual field mean deviation (R2 = 0.47; P < .001) in 30 glaucomatous eyes. Mean macular thickness in the hemifield associated with the field defect was found to be significantly lower compared with the unaffected hemifield, further supporting a structure-function relationship. Results of these studies confirm the potential role of macular imaging in glaucoma’s diagnosis and evaluation.
Macular thickness measurements were also found to correlate with peripapillary retinal nerve fiber layer (RNFL) measurements by TD-OCT in a study by Wollstein and colleagues. 6 Peripapillary RNFL thickness measurements, however, had a higher sensitivity and specificity (area under receiver operating curve [AROC] = 0.79) for the detection of glaucomatous visual field abnormalities than the measurements of macular thickness (AROC = 0.63). In light of these results, the investigators recommended against the routine use of TD-OCT macular-thickness measurements alone in the evaluation of glaucoma.
WHAT IS THE GANGLION CELL COMPLEX?
The ganglion cell complex (GCC) is defined as the three innermost retinal layers: the nerve fiber layer, the ganglion cell layer, and the inner plexiform layer.7 Tan and colleagues7 suggested that glaucoma likely preferentially affects these layers, rather than all macular layers, because they contain the axons, cell bodies, and dendrites of ganglion cells. This idea is supported by research findings that photoreceptors do not seem to be lost in glaucoma.8
The advent of SD-OCT has allowed for increased axialimage resolution compared to TD-OCT.9 Greater resolution now allows for the discrete segmentation and thickness measurement of the perifoveal GCC.
ANALYZING THE GCC
At the time of this writing, only the RTVue Fourier- Domain OCT (RTVue FD-OCT system; Optovue, Inc., Fremont CA) system has automatic segmentation software available for GCC analysis. It measures GCC thickness after centering a 7-mm2 area scan over the fovea. The GCC is then automatically segmented by the software and displayed in three color-coded maps for analysis.7 The GCC thickness map represents absolute GCC thickness in the perifoveal area. The deviation map displays acquired signal deviation from the normative database as a fraction of the mean normative value for each acquired pixel. The significance map displays the corresponding probabilities of deviation from the normal range for each acquired pixel in the GCC map.
GCC analysis will become available in an expected software upgrade for the Cirrus HD-OCT (Carl Zeiss Meditec, Inc., Dublin, CA). This software is under development and not yet commercially available. Figure 1 illustrates the optic nerve head photograph and automated visual field study from a representative patient with severe primary openangle glaucoma in the left eye. The corresponding optic disc cube printout and ganglion cell analysis for this patient are shown in Figures 2 and 3, respectively. An exemplary structure-function relationship can be seen in this case, where the superior visual field defect corresponds to inferior thinning of the optic nerve head, inferior RNFL depression, and inferior GCC thinning.
DIAGNOSTIC ACCURACY AND REPRODUCIBILITY
Tan and colleagues7 investigated the diagnostic accuracy and reproducibility of novel GCC parameters as measured by the RTVue FD-OCT system compared with the standard peripapillary RNFL thickness measurements obtained with TD-OCT (Stratus OCT; Carl Zeiss Meditec, Inc.). In this cross-sectional study, the researchers employed existing data from patients enrolled in the Advanced Imaging for Glaucoma Study. Participants were categorized into three groups: normal (65 eyes), perimetric glaucoma (78 eyes), and preperimetric glaucoma (52 eyes). Each eye underwent scanning with the RTVue FD-OCT system for GCC analysis as well as scanning with the Stratus TD-OCT system for peripapillary RNFL analysis. AROC was then used to compare the diagnostic power of the GCC and RNFL parameters. The coefficient of variation was used to assess the reproducibility of GCC parameters.
In this study, the AROC for the three GCC parameters (overall average thickness, focal loss volume, and global loss volume) were 0.90, 0.92, and 0.92, respectively. These values did not differ significantly from the AROC for average RNFL thickness (0.92; P > .1) as measured by TD-OCT in the diagnosis of perimetric glaucoma. The AROCs for GCC and RNFL parameters also did not differ for the diagnosis of preperimetric glaucoma. With regard to reproducibility, the GCC parameters outperformed (smaller coefficients of variation) RNFL parameters in normal (P = .0002) and perimetric glaucomatous (P > .001) eyes but not preperimetric glaucomatous (P = .11) eyes.
The results of this study indicate that, although GCC and average RNFL parameters perform similarly in terms of glaucoma diagnosis, the former are more reproducible and therefore may better detect glaucomatous progression.
The limitations of SD-OCT imaging of the optic nerve head and peripapillary RNFL also apply to imaging of the GCC. Specifically, GCC analysis may be limited by signal quality,10 image artifact,11 and errors in the software algorithm. 12 In addition, because the GCC is imaged in the perifoveal region, any coexisting macular pathology (eg, agerelated macular degeneration, macular edema, epiretinal membrane) may affect GCC thickness measurements. Imaging of the perifoveal GCC limits analysis to ganglion cells associated with the central visual field. In early glaucoma, visual field defects (and corresponding damage to the ganglion cells) may occur far from fixation and therefore escape detection when imaging only the perifoveal GCC.
In their study comparing GCC parameters with standard RNFL parameters, Tan and colleagues7 also found that combining the two parameters significantly increased the detection rate of both preperimetric and perimetric glaucoma. Newer software algorithms will likely combine RNFL, optic nerve head, and GCC parameters to further increase the diagnostic yield of SD-OCT and will also lead to better reproducibility for accurate tracking of glaucomatous progression. Further investigation of the GCC is likely to lead to the standardization of current software algorithms and their availability across all manufacturers of SD-OCT technology.
Ahmad A. Aref, MD, is an instructor and clinical glaucoma fellow at the Bascom Palmer Eye Institute, Miller School of Medicine, University of Miami. He acknowledged no financial interest in the products or companies mentioned herein. Dr. Aref may be reached at (305) 326-6384; firstname.lastname@example.org.
Donald L. Budenz, MD, MPH, is a professor of ophthalmology, epidemiology, and public health at the Bascom Palmer Eye Institute, Miller School of Medicine, University of Miami. He acknowledged no financial interest in the products or companies mentioned herein. Dr. Budenz may be reached at (305) 326-6384.