Optical Coherence Tomography for Glaucoma Diagnosis

OCT employs the principles of low-coherence interferometry and is analogous to ultrasound B-mode imaging, but it uses light instead of sound to acquire high-resolution cross-sectional images of ocular structures. The original time-domain OCT (TD-OCT) has been superseded by Fourier- or spectral-domain OCT (SD-OCT), which has improved spatial resolution and image acquisition speed, resulting in enhanced image quality and better reproducibility. OCT is able to provide quantitative measurements of the peripapillary RNFL thickness, ONH topography, and macular thickness, which can discriminate glaucomatous eyes from healthy eyes.

OCT RNFL thickness measurements are usually acquired in the peripapillary area, that is, at a fixed radius around the ONH. Most commercially available OCT devices acquire RNFL thickness measurements in a peripapillary circle located at a certain distance from the ONH, usually 3.45 mm. RNFL thickness measurements are generally lower in glaucomatous eyes compared with those in nonglaucomatous eyes, although considerable interindividual variability exists. Measurement parameters presented in OCT reports include the global average peripapillary RNFL thickness, which corresponds to the average of all thickness measurements in the peripapillary circle, as well as average RNFL thickness in quadrants (superior, inferior, temporal, nasal) or in small clock-hour sectors. Figure 5-13 shows an example of RNFL analysis provided by SD-OCT in a patient with glaucomatous RNFL loss in the right eye and normal RNFL thickness in the left eye. Sensitivities and specificities for detection of glaucomatous damage vary depending on the specific parameter evaluated and the characteristics of the studied population. In general, the parameters with best diagnostic accuracy are the average peripapillary RNFL thickness and thicknesses in the inferior and superior quadrants. This is supported by studies that demonstrated that the superior and inferior areas of the optic nerve are most commonly affected in glaucoma.

Note that, although sectorial RNFL parameters may increase the chance of detecting localized RNFL damage in glaucoma, these parameters may suffer from lower reproducibility, because measurements are averaged over relatively small areas. By contrast, the global average RNFL thickness has generally been shown to be the most reproducible parameter, which is not surprising considering that its calculation involves averaging measurements over a relatively large area. The improved reproducibility may offer gains in the ability to detect progression over time. The gain in reproducibility, however, may come at the expense of discovering localized RNFL defects, which are averaged out in the calculation. However, some commercially available OCT devices are able to acquire and visualize a 3-dimensional map of the RNFL thickness in the peripapillary region; such maps may facilitate identification of localized arcuate RNFL defects that may be missed with summary parameters.

SD-OCT devices are also able to provide topographical measurements of the ONH, including measurements of the optic disc area, neuroretinal rim area, and cup–disc ratio. These parameters and the methods for calculating them differ across platforms. Although previous versions of the OCT technology were also able to provide such measurements, considerable data interpolation was required, resulting in poor reproducibility of the measurements. The improved resolution of SD-OCT has greatly reduced the need for interpolation, resulting in much better delineation of the ONH structures.

Figure 5-13 Example of RNFL analysis with spectral-domain optical coherence tomography (SD-OCT). Top: The right eye shows diffuse RNFL thinning on SD-OCT (A), which is consistent with the neuroretinal rim thinning and enlarged cup shown in the photographs of the ONH (B), and with the visual field loss that is evident on standard automated perimetry (SAP) (C).Bottom: The left eye shows normal RNFL thickness on SD-OCT (A), normal appearance of the ONH (B), and normal visual field (C).

(Courtesy of Felipe A. Medeiros, MD, PhD.)

In general, with OCT, the border of the optic disc is defined by the termination of Bruch membrane, or its “opening.” The Bruch membrane opening–minimum rim width (BMO-MRW) is a parameter for measuring the neuroretinal rim thickness at the optic disc border with SD-OCT. BMO-MRW is defined as the minimum distance from the termination of Bruch membrane (BMO) to the inner aspect of the neuroretinal rim averaged around the disc. Studies suggest that BMO-MRW may be more sensitive than other rim-based parameters for diagnosing glaucoma. Figure 5-14 shows an example of an SD-OCT printout with calculations of BMO-MRW measurements.

Increased attention has been directed toward the macular region for evaluation of glaucomatous damage. Because much of the total macular thickness consists of RNFL and ganglion cell bodies, this region is an attractive area for identifying structural glaucoma damage. The macular RGC layer contains more than 50% of the RGCs of the entire retina. Investigations have also suggested that, contrary to previous belief, glaucomatous damage frequently affects the macular region even early in the disease process, leading to central visual field loss that can go undetected with conventional perimetry. SD-OCT enables quantitative assessment of either the entire macular thickness or the thickness of specific layers that may be important in glaucoma. The retinal layers that are included in the macular thickness measurements for glaucoma evaluation vary with the different OCT platforms. The thickness of the ganglion cell layer combined with the inner plexiform layer (GCIPL) is a commonly used parameter, as is the ganglion cell complex (GCC), which is composed of the RNFL, the ganglion cell layer, and the inner plexiform layer. The diagnostic ability of macular parameters is similar to that of the peripapillary RNFL. In eyes with myopic discs or large areas of PPA, which can present with artifacts on peripapillary RNFL assessment, macular evaluation may provide an enhanced ability to diagnose and monitor glaucomatous damage. Figure 5-15 shows an example of macular damage detected in a glaucomatous eye with SD-OCT, along with corresponding RNFL scans and optic disc photographs.

Figure 5-14 SD-OCT scan showing results for the Bruch membrane opening–minimum rim width (BMO-MRW) parameter. The B scan on the right shows how the BMO-MRW measurement is obtained. It corresponds to the minimum distance extending from the border of Bruch membrane to the internal limiting membrane (arrows). The global and sectoral BMO-MRW measurements are then calculated and compared with a normative database.

(Courtesy of Felipe A. Medeiros, MD, PhD.)

Several studies have compared the diagnostic ability of different RNFL, ONH, and macular parameters. The results of these studies have not always been consistent; the differences in the results may be related to differences in the criteria used to select the glaucoma and control subjects, as well as the different characteristics of the populations included in the studies. As in any diagnostic accuracy study, a reference standard needs to be employed to select cases and controls. The ability to distinguish cases and controls, as defined by the reference standard, is then determined for the parameters under investigation. In most studies, the reference criteria for glaucoma cases are glaucomatous field losses with “compatible” optic nerve damage. If the reference criteria include clinician assessment of the optic disc, this increases the chance that patients with obvious abnormalities of optic disc features such as rim or cup, as opposed to those with RNFL abnormalities, will be those selected for inclusion as glaucoma cases in the study. Reference criteria for control cases frequently require “normal” optic discs, which clinically equate to a normal-appearing neuroretinal rim and cup. Such selection criteria can bias studies toward favoring the accuracy of topographic optic disc–based parameters. Furthermore, potential controls with anomalous optic disc characteristics such as tilted discs are often excluded from such studies and from normative databases. As a result, the diagnostic performance of OCT-based parameters for distinguishing eyes with glaucoma from nonglaucomatous eyes in clinical practice is not as good as one might expect on the basis of published findings.

Although imaging technologies are generally helpful in the detection of glaucoma, their diagnostic performance decreases for detection of early disease compared with moderate or advanced disease. Most studies evaluating the diagnostic accuracy of SD-OCT have evaluated the instruments’ ability to differentiate the eyes of patients with well-defined glaucomatous visual field defects from those of healthy subjects. Such studies are important for providing an initial exploratory evaluation of newly developed methods to detect glaucomatous damage. However, in clinical practice, a diagnostic test is used to diagnose disease in patients suspected of having the disease, not in patients with a confirmed diagnosis. Therefore, from a practical standpoint, it is of little utility to demonstrate that an imaging device is able to diagnose glaucoma in a patient with a confirmed visual field defect, because such a patient would already have a clear diagnosis established. In fact, studies with a case-control design that includes patients with well-established disease and a separate group of control subjects without glaucoma tend to substantially overestimate the performance of the tests. Therefore, if a test succeeds in initial exploratory diagnostic studies, further steps are necessary to evaluate whether it is able to provide clinically relevant information.

Figure 5-15 Macular (posterior pole) thickness maps of a patient with glaucoma. A, The grayscale plots illustrate hemi sphere asymmetries for each eye (ie, differences in thickness between the superior and inferior corresponding regions), and the asymmetry map shows differences between the right and left eyes. Note that the right eye shows thinning in the inferior arcuate region, which is easily seen in both the hemi sphere asymmetry graph and the color-coded map. The right eye also shows a superior arcuate defect, clearly suggested by the OD–OS asymmetry map. The left eye shows an extensive inferior arcuate defect that clearly involves the paracentral region. The superior region of the left eye seems preserved in the posterior pole analysis. B, RNFL scans of the same eyes. The right eye shows localized losses in both the superior and inferior temporal regions, which correspond to the damage shown in the posterior pole scan. The left eye shows damage that seems to be limited to the inferior temporal region, again in agreement with the posterior pole analysis. C, Optic disc photographs showing rim thinning and localized RNFL defects in the superior and inferior temporal regions of the right eye. The left eye demonstrates notching and complete loss of the rim in the inferior temporal location, but the superior rim appears well preserved.

(Courtesy of Felipe A. Medeiros, MD, PhD.)

In fact, the optimal design for assessing a diagnostic test’s accuracy is considered to be a prospective, masked comparison of the test and the reference test in a consecutive series of patients from a relevant clinical population—that is, those suspected of having the disease. Some studies have prospectively investigated the ability of SD-OCT to diagnose glaucoma in individuals who were suspected of having the disease by conventional examination at the time of enrollment. In 1 study, assessment of the RNFL with SD-OCT detected abnormalities in one-third of subjects up to 5 years before the first appearance of a visual field defect on standard automated perimetry. Importantly, other studies have shown that RNFL thickness measurements are predictive of development of future visual field losses in glaucoma suspect eyes. Figure 5-16 shows an eye in which the OCT measurements of RNFL thickness were abnormal before the development of visual field defects.

• Kuang TM, Zhang C, Zangwill LM, Weinreb RN, Medeiros FA. Estimating lead time gained by optical coherence tomography in detecting glaucoma before development of visual field defects. Ophthalmology. 2015;122(10):2002–2009.

  Figure 5-16 Structural damage often precedes detectable visual field damage. A, photograph of the optic nerve head shows marked neuroretinal rim thinning and an enlarged cup. B, Standard automated perimetry, by contrast, shows that the visual field is still within normal limits. C, Retinal nerve fiber layer (RNFL) analysis with spectral-domain optical coherence tomography shows diffuse loss of the RNFL.

  

  (Courtesy of Felipe A. Medeiros, MD, PhD.)

• Lisboa R, Leite MT, Zangwill LM, Tafreshi A, Weinreb RN, Medeiros FA. Diagnosing preperimetric glaucoma with spectral domain optical coherence tomography. Ophthalmology. 2012;119(11):2261–2269.

Excerpted from BCSC 2020-2021 series: Section 10 - Glaucoma. For more information and to purchase the entire series, please visit https://www.aao.org/bcsc.