Polarized light undergoes a phase shift as it passes through a birefringent structure such as the RNFL. The scanning laser polarimeter (SLP) (GDx, Zeiss Meditec, Inc., Dublin, California) scans the optic nerve with a 780-nm laser beam and detects this change in phase shift or retardation. Since retardation is proportional to RNFL thickness, the detected signal can be used to create an RNFL thickness map. A circumlinear scan, 3.2 mm in diameter and centered on the optic disc, is used to generate what is commonly known as a “TSNIT” plot (Temporal-Superior-Nasal–Inferior-Temporal). The inferior and superior poles represent the thickest portions of the RNFL in normal controls; therefore, the TSNIT plot appears as a “double hump.” Some of the parameters presented are based on the RNFL thickness measurements of the calculation circle; however, the nerve fiber indicator (NFI, a summary value that is intended to represent the likelihood of glaucomatous RNFL loss) is based on the entire RNFL thickness map. A progression algorithm utilizes data based on the entire image, outside the calculation circle.

The cornea and lens also possess birefringent properties. Early versions of the GDx assumed this anterior segment birefringence was the same for all patients, which often created aberrant RNFL thickness maps. The most current version of the GDx incorporates a corneal compensation mechanism (GDx Variable Corneal Compensator (VCC)). A scan taken through the fovea, which lacks any RNFL tissue, contains the total anterior segment birefringence. This value is subtracted from the global scan to isolate the RNFL birefringence.

Optical coherence tomography (OCT) produces high-resolution cross-sectional images of ocular tissues based on optical backscattering, akin to ultrasound technology. The optical backscattering of low-coherence laser light (850 nm) as it passes through layers of differing optical density is recorded by an interferometer to construct a 2-dimensional image of the scanned area. Mirrors are used to scan the laser beam physically through the tissue in the x–y plane and along the z-axis as a function of time. This acquisition technique is referred to as time-domain OCT (TD-OCT). The third-generation Stratus OCT (Zeiss Meditec, Inc., Dublin, California) is the TD-OCT in most widespread use. Similar to the GDx the Stratus OCT generates a circumlinear scan (slightly larger at 3.4 mm diameter) centered on the optic disc, resulting in a TSNIT plot with a “double hump” configuration. The patient’s TSNIT RNFL profile is compared to a normative database in order to highlight diffuse and focal defects. An optic disc algorithm captures three axial scans through the center of the disc, determines the disc margin in these scans by locating the edge of the retinal pigment epithelium, and utilizes data interpolation to produce an image of the optic nerve head. Various optic nerve structural parameters similar to those displayed on the HRT are then determined. In addition, algorithms are available to calculate and display total macular thickness.

The recently developed technique of spectral-domain OCT (SD-OCT) takes advantage of the fact that structural information unique to the tissue of interest is encoded in the frequency spectrum of the backscattered signal. Processing this signal through a spectrometer extracts this structural information much more rapidly than the manual scanning employed in TD-OCT, allowing for faster acquisition times, better resolution (4–5 μm range), less motion artifact, and larger volumes of structural data. Three SD-OCT devices approved for clinical use are the Cirrus (Zeiss Meditec, Inc., Dublin, California), Spectralis (Heidelberg Engineering, Heidelberg, Germany), and RTVue (Optovue, Inc., Fremont, California). All three machines provide RNFL thickness maps, TSNIT plots, and ONH analysis. Using a segmentation algorithm, the RTVue and Cirrus also display the thickness of the macular ganglion cell complex (GCC) comprised of the RNFL, ganglion cell layer and inner plexiform layer.

A recent evidence-based assessment of ONH and RNFL analyzers was conducted and published by the Ophthalmic Technology Assessment Committee of the American Academy of Ophthalmology [4]. In the vast majority of publications, the glaucomatous (“diseased”) group was comprised of patients who had visual field loss. These patients, as noted previously, have more advanced glaucoma. The ability to distinguish normal from advanced glaucoma is biased toward greater detectability. In practice, the devices have more potential usefulness when they can detect glaucoma at a pre-perimetric stage when visual field defects are absent.

Published evidence will be provided for each device. The ability of a test to distinguish between pathologic and normal states is described in the literature by the area under the receiver operator curve (AUROC). It is a graphic representation of the sensitivity plotted against (1-specificity). Values for the AUROC range from 0.5 to 1, with the latter value representing the ideal test to discriminate disease from normal.

Numerous studies have demonstrated the ability of the HRT I to discriminate glaucoma associated with VF loss from normal (Fig. 8.1a–d). In a paper by Miglior et al. [5] the Moorfields Regression Analysis (MRA) was compared with the multivariate discriminant analysis. The authors found better specificity with the Moorfields algorithm (94 % vs. 75 % by multivariate discriminant analysis), but better sensitivity using the multivariate analysis (83 % vs. 74 % by MRA). Others have also used different software algorithms to separate glaucomatous from control cases with varying success [6–9].

In the Ocular Hypertension Treatment Study, the HRT I was used in a subset of the participating sites, and a retrospective analysis of parameters that predicted glaucoma was analyzed [10]. Glaucoma was defined as an endpoint when there was a change in the optic disc consistent with glaucoma damage and/or development of repeatable VF defects. The most predictive parameters were mean height contour, rim area, and mean cup depth.

Studies on HRT II and HRT III have also shown very good detection of perimetric glaucoma [11–13]. Harizman et al. compared the MRA of HRT II and the GPS of HRT III for the detection of perimetric glaucoma [13]. Sensitivity and specificity were 77 and 90 % for GPS and 71 and 92 % for MRA, respectively.

There is limited available data on the ability of these devices to detect glaucoma progression. Overall, there is significant overlap in the detection of progression by structural and functional criteria, but there are also substantial percentages of glaucoma patients who progress by only one of the measures.

Comparison of the HRT I with VF testing found that although there was considerable overlap of eyes that progressed by both tests, significant numbers of eyes were found to have progressed by only functional or structural criteria [76]. It may be that these tests represent independent indicators of progression. In a study [77] that evaluated glaucoma progression among four groups with different glaucomatous disc types (focal, myopic, senile sclerotic, and generalized), progression rates by HRT (44–82 %) were greater than those by standard automated perimetry (33–57 %), echoing the groups’ earlier reported results [78]. Moreover, there were similarly high rates of progression by either test alone (20–61 %) across the groups. A more recent study found that the HRT II also detected a higher rate of glaucomatous progression when compared to visual field [79].

With regard to TD-OCT, Wollstein et al. [80] also reported higher rates of progression when the structural test was used (25 %) as compared with standard automated perimetry (12 %), although their study population included glaucoma suspects and pre-perimetric glaucoma patients in addition to patients with baseline VF defects. It is unclear if this represents greater sensitivity for progression or hypersensitivity (false positives) of the structural test.

In an observational cohort study of 253 patients by Medeiros et al. [81] glaucoma progression was detected in 13 % by either visual field criteria or serial disc photos. Mean rates of change in average RNFL thickness were significantly higher for progressors compared with nonprogressors. Similar findings were observed in a study using four different criteria for visual field progression [82].