Scanning laser polarimetry in glaucoma.

Glaucoma is an acquired progressive optic neuropathy which is characterized by changes in the optic nerve head and retinal nerve fiber layer (RNFL). White-on-white perimetry is the gold standard for the diagnosis of glaucoma. However, it can detect defects in the visual field only after the loss of as many as 40% of the ganglion cells. Hence, the measurement of RNFL thickness has come up. Optical coherence tomography and scanning laser polarimetry (SLP) are the techniques that utilize the evaluation of RNFL for the evaluation of glaucoma. SLP provides RNFL thickness measurements based upon the birefringence of the retinal ganglion cell axons. We have reviewed the published literature on the use of SLP in glaucoma. This review elucidates the technological principles, recent developments and the role of SLP in the diagnosis and monitoring of glaucomatous optic neuropathy, in the light of scientific evidence so far.

Glaucoma is an acquired progressive optic neuropathy characterized by changes in the optic nerve head and retinal nerve fiber layer (RNFL). White-on-white perimetry is the gold standard for diagnosis of glaucoma. Various new technologies are coming up for the early diagnosis of glaucoma. We performed a systemic search of the PubMed using the terms-glaucoma, scanning laser polarimetry (SLP), GDx and RNFL to prepare this review, which elucidates the technological principles, recent developments and the role of SLP in the diagnosis and monitoring of glaucomatous optic neuropathy.

Glaucoma and its Diagnosis: A Background

Imaging of RNFL is of vital importance in glaucoma as structural RNFL changes often precede functional visual field changes.[123] As many as 40% of all ganglion cells can be lost before a well-defined scotoma is detected on the visual field.[4] RNFL evaluation has been found to be more sensitive for predicting future visual field loss when compared to ONH evaluation and a better predictor of damage.[5678910] Recent technology boom in ophthalmology has given us two technologies that are capable of providing highly reproducible RNFL assessments. Optical coherence tomography (OCT) based on low-coherence interferometry, generates posterior segment thickness measurements with an axial resolution of eight to ten microns while SLP provides RNFL measurements based upon the birefringence of the retinal ganglion cell axons.

Technologic Principle

Retinal nerve fiber layer is made of highly ordered parallel axon bundles which contain microtubules, cylindrical intracellular organelles with diameters smaller than the wavelength of light. The paralleled structure of the microtubules is the source of RNFL birefringence that is the splitting of a light wave by a polar material into two components. These components travel at different velocities that create a relative phase shift termed “retardation”. This retardation is proportional to the thickness of the RNFL.[1112] Total retardation of a subject's eye is the sum of cornea, lens, and RNFL birefringence. Compensation of anterior segment birefringence is thus necessary to isolate RNFL birefringence.

Scanning laser polarimetry is basically a confocal scanning laser ophthalmoscope with an integrated ellipsometer to measure retardation. It determines the RNFL thickness, point by point in the peripapillary region, by measuring the total retardation in the light reflected from the retina. Polarized light passes through the eye and is reflected off the retina.[13141516] Because the RNFL is birefringent, the two components of the polarized light are phase shifted relative to each other and this is captured by a detector, and converted into thickness (in microns).[11]

Generations of Scanning Laser Polarimetry and Calculation of Birefringence

Fixed corneal compensation

The first-generation device (nerve fiber analyzer [NFA] I) became commercially available in 1992 and was equipped with a single detector, which was later replaced by a double detector (NFA II). These earlier versions of SLP (e.g. the GDx NFA, GDx Access) were compensated for anterior segment birefringence based on fixed values for the axis and magnitude of the anterior segment birefringence. However, later it was found that the axis and magnitude are variable for each individual and using a fixed corneal compensation may not adequately account for the aforementioned variables.[17]

Variable corneal compensation

In 2002, variable corneal compensation (VCC) was introduced, allowing eye-specific compensation of anterior segment birefringence, commercially available in the 5th iteration of the SLP instrument named GDx-VCC (Carl Zeiss Meditec, Inc., Dublin, California, USA).

GDx-VCC measures and individually compensates for anterior segment birefringence, using a 780-790 nm laser diode.[141516] The axis of the anterior segment birefringence is determined by the orientation of the ‘bow-tie’ birefringent pattern [Fig. 1] in the macula and the magnitude of the anterior segment birefringence is calculated by analyzing the circular profile of the birefringence in the macula according to standard equations.[11] In cases of macular pathology, an alternative method is available that accurately compensates for the anterior segment birefringence.[18]

Figure 1

“Bow-tie pattern” (arrow) seen in the macula in an uncompensated scan

GDx-Enhanced corneal compensation

The enhanced corneal compensation (ECC) algorithm is implemented in the GDx-VCC by a software modification. A known birefringence bias is introduced into the measurement beam path to shift the measurement of total retardation into a more sensitive region of the curve of detection of polarization of the instrument. The bias retarder is formed by the combination of the variable corneal compensator and cornea. However, instead of completely canceling corneal birefringence, the retarder is adjusted and hence that the combination has retardance close to 55 nm and slow axis of polarization close to vertical. After image acquisition, the birefringent bias is removed mathematically, point-by-point, to yield the RNFL retardation values that are converted to thickness (in micrometers) using a fixed conversion factor.

The computerized export of the temporal-superior-nasal-inferior-temporal (TSNIT) plots on the GDx-ECC printout includes the mean RNFL thickness from 64 polar sectors (5.625°/arc). The mean for each sector is computed along the 3.2 mm diameter measurement circle surrounding the optic nerve head. The mean RNFL thickness for the superior (0–180°) and inferior (181–360°) retinal region is computed separately. ECC scans show a stronger structure-function relationship with perimetry compared with VCC.[1920]

Modified diameter scan

GDx-VCC uses a fixed scan circle of 3.2 mm diameter centered on the optic disc (conventional diameter scan). This may be affected by peripapillary atrophy (PPA). So it has measurement options with two additional scan diameters – medium and large. Dada et al.[21] found that medium and large diameter scans can also be used to discriminate between normal and glaucomatous eyes. Nerve fiber indicator (NFI) has been seen to be the best discriminating parameter across all diameters.

Retinal nerve fiber layer measurements

The GDx-VCC measurements are taken by scanning the beam of a near-infrared laser (780 nm) in a raster pattern[12] which captures an image with a field 40° horizontally by 20° vertically, and including both the peripapillary and the macular region.[22] It generates two images: A reflectance image and a retardation image [Fig. 2]. Each image is made up of 256 (horizontal) ×128 (vertical) pixels, or 32,768 total pixels. For an emmetropic eye, 1 pixel is 0.0465 mm in size, and the total scan field is 11.9 mm (horizontal) × 5.9 mm (vertical).[23]

Figure 2

Images generated by the GDx-variable corneal compensation: Left image is the reflectance image displayed as a colored intensity map. The right image is the retardation map converted to retinal nerve fiber layer thickness, color coded based on the color spectrum

Clinical Interpretation of the GDx Printout

For each GDx-VCC scan, an age-matched comparison is made to the normative database, and any significant deviations from normal limits are flagged as abnormal with a ‘P’ value. Quantitative RNFL evaluation is provided through four key elements of the printout [Fig. 3]:

Figure 3

Print out showing various parameters for quantitative retinal nerve fiber layer evaluation

Thickness map

Deviation map

TSNIT graph [Fig. 4a and b]

Figure 4

(a and b) Calculation circle, characteristic double hump pattern

Parameter table [Fig. 5].

Figure 5

Temporal-superior-nasal-inferior-temporal parameters

Nerve fiber indicator

The NFI is a global measure based on the entire RNFL thickness map. NFI ranges from 1 to 100, with lower values (<25) indicating a normal RNFL.[24252627]

Clinical applications

Scanning laser polarimetry technology can help in early diagnosis of glaucoma by picking up RNFL changes that may precede visual field damage by 5–6 years [Table 1].

Table 1

Summary of studies using GDx in glaucoma management

Healthy subjects

Age has been found to significantly affect all RNFL measurements with the ECC protocol of SLP, whereas typical scan score (TSS) and residual anterior segment retardance affect the overall average and the superior average RNFL measurements, respectively.[28]

Glaucoma suspect

Medeiros et al.[29] found GDx-VCC to be superior versus Heidelberg retina tomograph 3 (HRT3) for detecting early damage in glaucoma suspects. They studied 82 glaucoma suspects. The area under the receiver operator characteristic (AUROC) curve for the best parameter from GDx-VCC (NFI) was significantly larger than that of the best parameter from the HRT (rim volume) (0.83 vs. 0.70).

Mohammadi et al.[30] studied one eye from each of 160 glaucoma suspects with normal standard automated perimetry (SAP) visual fields. They found thinner baseline SLP RNFL measurements to be independent predictors of future visual field damage.

Badalà et al.[31] in their study compared OCT, GDx-VCC, HRT and stereophotographs in 46 healthy subjects and 46 glaucoma patients with early defect. The largest AUROC for each technique was 0.96 for OCT average thickness, 0.92 for GDx-VCC NFI, 0.91 for HRT3 FSM discriminant function and 0.97 for stereophotographs.

Shaikh et al.[32] studied 39 “glaucoma suspects.” Twenty-two of 39 (56.4%) were considered to be at risk of developing progressive glaucoma. In 19 of these patients, abnormal GDx-VCC results were found, particularly inter-eye asymmetry in the RNFL thickness. However, in two of 39 (5.1%) patients the GDx-VCC was normal, despite the presence of a neuroretinal rim defect in the optic disc with corresponding visual field loss, and in one patient with POAG. Hence, they concluded that GDx-VCC should not be used in isolation.

Li et al. found NFI to be the best performing parameter using a cutoff of 35 with a sensitivity of 75%, specificity of 95%, a positive predictive value of 25 and a negative predictive value of 99 for diagnosing glaucoma.[33]

In a study to assess the relationship between high-definition OCT (HD-OCT) and SLP with VCC and ECC in measuring RNFL in healthy eyes and those with early-to-moderate glaucomatous VF loss, HD-OCT parameters of RNFL thickness were found to be significantly higher than VCC and ECC parameters, and therefore those thickness values are not interchangeable in normal eyes and in early glaucoma patients.[34]

Early glaucoma

According to Zheng et al.,[35] NFI and inferior average are the most effective indicators for the early diagnosis of glaucoma. AUROCs were 0.81 for NFI and inferior average with better differentiation capability.

Da Pozzo et al.[36] evaluated the diagnostic accuracy of GDx-VCC comparing 62 healthy and 48 glaucomatous age-matched patients with early glaucomatous field defect. The three parameters with largest AUROCs were NFI (0.870), superior average (0.817) and normalized superior area (0.816).

Parikh et al.[37] evaluated 74 eyes with early glaucoma and 104 eyes of normal subjects. TSNIT SD had the best combination of sensitivity and specificity-61.3 and 95.2, respectively-followed by NFI > 50 (sensitivity, 52.7%; specificity, 99%).

Fortune et al. in their study found that the onset of progressive loss of RNFL retardance occurs earlier than the onset of RNFL thinning. Endpoints of progressive loss from baseline also occurred more frequently and earlier for RNFL retardance as compared with a thickness.[38]

Moderate-to-severe glaucoma

Reus and Lemij[39] compared 77 healthy eyes and 162 glaucomatous patients. AUROC for main GDx-VCC parameters ranged from 0.90 to 0.98.

Ocular Hypertension

Henderson et al.[40] studied 44 ocular hypertension (OHT) patients and 48 healthy subjects, all of similar age. Higher NFI scores were correlated significantly with thinner CCT measurements in OHT patients. In multivariate analysis, only age and CCT measurement were associated significantly with RNFL measurements in OHT eyes.

Normal tension glaucoma

Choi et al.[41] evaluated RNFL in the retinal segments without visual field loss in eyes with 56 Asian normal tension glaucoma (NTG) patients who had localized visual field defects confined to 1 hemifield and 43 normal controls. They found that perimetrically normal hemifields of NTG eyes had significantly lower RNFL thickness parameters than the corresponding regions of healthy eyes.

Normal tension glaucoma versus high tension glaucoma

Jung et al.[42] found GDx-VCC to be more sensitive in detecting RNFL damage in high tension glaucoma (HTG) patients compared to NTG patients. Compared with eyes with NTG, eyes with HTG showed reduced RNFL thickness.

Juvenile Open-Angle Glaucoma

Zareii[43] compared the ability of GDx-VCC and OCT to discriminate eyes with Juvenile Open-Angle Glaucoma from normal eyes. They observed statistically significant correlations between the two. The greatest AUROC parameter on OCT (inferior average: 0.92) had a lower area than that in GDx-VCC (NFI: 0.99).

Exfoliation syndrome

Dimopoulos et al.[44] compared RNFL thickness of normotensive eyes with exfoliation syndrome (XFS) and healthy eyes using GDx-VCC. Polarimetry-determined RNFLT was lower in XFS eyes with normal IOP. Therefore, close monitoring of RNFLT may facilitate early identification of those XFS eyes that convert to exfoliative glaucoma.

Detecting Progression

The new software used in GDx that helps in investigating progression is GPA™ (progression analysis for GDx) [Fig. 6]. The GDx-GPA algorithms are designed to have 95% specificity for “Likely progression.”

Figure 6

Overview of progression analysis using GPA software

GDx-GPA uses two different algorithms:

Change from baseline (CFB): Based on changes from two baseline exams compared to measurement variability

Statistical image mapping: Based on trend analysis. All visits contribute to change detection here in contrast to CFB where data from the first two and last two visits are used.

Atypical Retardation Patterns

Atypical retardation patterns (ARPs) have been characterized as irregular patches of elevated retardation values that are against the expected retardation distribution based on the RNFL anatomy.[45] ARPs may provide fallacious RNFL measurements[46] and occur in approximately 10–25% of healthy eyes and 15–51% of glaucomatous eyes,[454748] in elderly subjects,[49] in lightly pigmented fundi and in high myopes. SLP images are generally excluded from the analysis when the software flags them as incompatible with its normative database. The software flags an image as such when the TSS is below 25.[50]

Typical scan score

TSS is a proprietary measure provided by the GDx-VCC software that indicates whether the observed retardation pattern is typical of the human healthy or glaucomatous RNFL (range: 0–100). In scans that display an ARP, the TSS is lower. It is highly predictive of the presence of ARPs.[47] In clinical studies, a cut-off value for TSS of 80 is often used to discriminate between normal and atypical retardation patterns.[455152] The ability of SLP to discriminate between healthy and glaucomatous eyes has been shown to decrease when ARPs are present.[464748] Diagnostic accuracy has been found to be comparable for SD-OCT and GDx-VCC if typical scans (TSS = 100) are investigated.[53] Decreasing TSS is associated with a decrease in diagnostic accuracy for discriminating healthy and glaucomatous eyes by SLP. NFI has been found to be less influenced than the global or sector RNFL thickness. GDx ECC helps in the neutralization of ARPs.

Wang et al. evaluated the changes of visual field and RNFL during 24 months follow-up in primary open-angle glaucoma (POAG) patients. Visual field and RNFL were detected by using GDx-VCC system and Octopus perimeter in 60 patients with POAG in early stage (60 eyes), 32 in moderate stage (32 eyes) and 30 in advanced stage (30 eyes). They found GDx-VCC system to be useful to evaluate the progression of POAG in early and moderate stage by detecting the change of the RNFL thickness with long-term follow-up.[54]

In a study by Medeiros et al.[55] 213 eyes of 213 patients with an average follow-up time of 4.5 years were studied annually with the ECC and VCC, along with optic disc stereophotographs and SAP. Thirty-three patients (15%) showed progression over time on visual fields and/or stereophotographs. Mean rates of average RNFL thickness change measured by the GDx-ECC were significantly different in progressors versus nonprogressors (−1.24 μ/year vs. −0.34 μ/year; P < 0.001). AUROC curve for discriminating progressors from nonprogressors was significantly higher for rates of change measured by ECC compared to VCC (0.89 vs. 0.65; P <.001). Rates of RNFL change detected by the GDx ECC were significantly greater in eyes with progressive glaucoma compared to eyes with stable disease. Also, the ECC performed significantly better than the VCC for detection of change.

Makabe et al.[56] also found that longitudinal progression in NFI obtained with GDxVCC was significantly correlated with that in Humphrey field analyzer humphrey field analyzer (HFA) parameters, such as mean deviation and pattern standard deviation and recommended that GDx-VCC is a useful tool for longitudinal follow-up assessment of glaucoma.

Special Scenarios

Effect of cataract surgery

Scanning laser polarimetry measures the shift in polarization of light that is associated with the natural birefringence of RNFL. Other birefringent structures within the eye include cornea and lens. The GDx compensates for the birefringence of the cornea in part but that caused by lens is not compensated and hence there can be change in RNFL thickness after cataract surgery.

Vetrugno et al.[57] examined 68 eyes of 68 patients undergoing phacoemulsification with foldable intraocular lens (IOL) implantation (silicone and acrylic) by SLP before and after surgery. They could not find any statistically significant difference between SLP parameters before and after cataract surgery, regardless of the type of IOL implanted.

In contrast to this, Park et al.,[58] Iacono et al.[59] and Dada et al.[60] stressed on new baseline SLP measurements after cataract extraction with IOL. Gazzard et al.[61] have also shown a significant change in SLP measurements after cataract extraction, especially in patients with posterior subcapsular cataract.

Brittain et al.[62] investigated changes in GDx-VCC parameters caused by posterior capsular opacification. Typical scan score significantly improved after laser from 33 to 55.1 (P = 0.001; n = 19) and TSNIT score was significantly lower, dropping from 62.3 to 58.9 (P = 0.03; n = 19).

Effect of refractive surgery (laser in situ keratomileusis (LASIK), photo refractive keratectomy)

Dada et al.[63] and Zangwill et al.[64] could not find any statistically significant change in any of the parameters of RNFL before and after LASIK. Katsanos et al.,[65] however, found transient changes in RNFL by SLP after LASIK, which became stable by the 3rd postoperative month.

Comparison with other Diagnostic Modalities

Scanning laser polarimetry versus optical coherence tomography

Leung et al.,[66] Chung[67] and Kim et al.[68] have reported that the total average RNFL thickness measured with Stratus OCT and GDx-VCC were highly correlated. Yoo et al.[69] found that overall performance of Stratus OCT and GDx-VCC about their internal normative database was not significantly different.

Aptel et al.[70] tested 120 eyes of 120 patients (40 with healthy eyes, 40 with suspected glaucoma, and 40 with glaucoma) on Cirrus-OCT, GDx-VCC, and SAP. With spectral-domain OCT, the correlations (r(2)) between RNFL thickness and visual field sensitivity ranged from 0.082 to 0.726. By comparison, with GDx-VCC, the correlations ranged from 0.062 to 0.362. The largest AUROC were seen for OCT superior thickness (0.963 ± 0.022; P < 0.001) in eyes with glaucoma and for OCT average thickness (0.888 ± 0.072; P < 0.001) in eyes with suspected glaucoma. They concluded that the structure–function relationship was significantly stronger with OCT than with GDx-VCC.

Lee et al.[71] found AUROCs of the Cirrus OCT to be significantly higher than those of GDx-VCC in glaucoma detection. The best-performing parameter was the NFI in GDx-VCC and inferior RNFL thickness in cirrus-OCT (AUROC = 0.912, 0.961, P = 0.045). There was good agreement between the two instruments with respect to abnormal classifications (kappa, 0.611–0.766).

Xu et al.[72] investigated the performance of SDOCT and SLP to detect progressive RNFL changes in glaucoma and found that at a comparable level of specificity, progressive RNFL thinning was detected more often than progressive reduction of retardance.

Bagga et al.[47] showed that images with ARPs have a weaker correlation with RNFL thickness measured with OCT compared with images with normal patterns. These results were subsequently confirmed in a study by Sehi et al.[73]

Garas[74] found that the diagnostic accuracy of the GDx-VCC/ECC NFI and RTVue-OCT average RNFLT were similar in glaucoma cases.

Schallenberg et al.[75] used GDx-VCC, GDx-ECC, and Spectralis-OCT to study RNFL measurements in 92 eyes of 92 glaucoma patients with various amount of glaucomatous damage and found that they cannot be directly compared because of differences in method of the devices.

Scanning Laser Polarimetry Versus Frequency Doubling Technology/Standard Automated Perimetry

In the Groningen Longitudinal Glaucoma Study[76] seventy glaucoma suspect patients with normal test results at baseline were followed prospectively for 4 years with SAP, frequency doubling technology and GDx. Conversion of suspects was better picked up by GDx.

Scanning laser polarimetry versus Heidelberg retina tomograph

Alencar et al.[77] found that the measurements of rates of change in GDx-VCC RNFL thickness were superior to HRT Rim area in identifying eyes with progression.

Kanamori et al.[78] also found GDx-VCC, HRT and OCT to be useful in preperimetric and early glaucoma. In their study of glaucomatous eyes with or without early visual field defects, SLP and OCT performed similarly or had better discriminating abilities compared with CSLO.

Medeiros et al.[79] compared the GDx-VCC, the HRT II and the Stratus OCT best parameters (NFI, the linear discriminant function and Inferior Average, respectively) in their ability to discriminate healthy eyes from glaucomatous patients. The AUROC curve of the three parameters indicated excellent discriminating power, without significant differences between them. For the GDx-VCC the three best performing parameters were NFI (0.91), Normalized Inferior Area (0.86) and TSNIT average (0.85).

Scanning Laser Polarimetry Versus Matrix Perimetry, Optical Coherence Tomography, and Retinal Nerve Fiber Layer Photography using Heidelberg Retina Angiograph 1

Hong et al.[80] compared the efficacy of detecting early glaucoma using all these in 72 POAG patients with early-stage visual field defects and 48 healthy controls. The AUROCs of Matrix perimetry, GDx-VCC, OCT, and RNFL photography using Heidelberg retina angiograph 1 with the best discriminating parameter were 0.990, 0.906, 0.794, and 0.751, respectively.

Strengths of Scanning Laser Polarimetry

Scanning laser polarimetry is a sophisticated technology that provides highly reproducible, objective measurements of the peripapillary RNFL. It is easy to operate, does not require pupillary dilatation, is highly reproducible, does not employ a reference plane, does not require correction for ocular magnification and incorporates an age-matched normative database. GDx-ECC has largely reduced ARP and is capable of neutralizing any residual anterior segment birefringence not completely compensated by VCC.

Retinal nerve fiber layer Integrity™

The RNFL-I measurement is the phase shift that occurs when polarized light passes through the RNFL. SLP captures birefringence, a tissue property that depends on the integrity of retinal ganglion cell axon microtubules and neurofilaments and hence SLP has the capacity not only to corroborate findings of RNFL thinning as determined by other methods, but may also provide insight into structural damage due to changes in density and orientation of the RNFL microstructures that may precede or occur in the absence of RNFL thinning.

Limitations

Peripapillary atrophy

Image quality may be compromised if PPA falls under the measurement circle. Kunimatsu et al.[81] found reproducibility of GDx parameters to be similar among three circles in normal eyes (P > 0.05), whereas coefficients of reproducibility of TSNIT average (P = 0.006) and superior average (P = 0.035) were smaller in the smaller circles in OAG eyes. They recommended using medium circles in OAG patients as the obstructing influences of PPA on GDx measurement could be avoided.

Refractive error

Manufacturer recommendation is to avoid scans in eyes with refraction outside the + 5/−10 spherical diopter range as the probability of obtaining low-quality scans increases markedly. Patients wearing contact lenses can undergo GDx-VCC. Wang et al.[82] found that the average RNFL thickness measured with Cirrus HD-OCT decreases as the degree of myopia increases while no such correlation was detected in GDx-ECC. Wang et al.[83] in another study found RNFL thickness to be lower in all but the nasal quadrant in patients with POAG and high myopia (HM), compared to patients with only HM. GDx-VCC was found to be better than OCT in detecting POAG in HM patients. Dada et al.[84] studied RNFL thickness in myopia with SLP. Moderate myopes showed a significant thinning of the RNFL. In HM due to PPA and contribution of scleral birefringence, the RNFL values were abnormally high.

Tilted disc

Yu et al.[85] and Bozkurt et al.[86] have reported GDx to be unreliable in tilted disc.

Macular pathology

A large area centered on the fovea (6° × 6° square region) should be used in macular pathology.[87] Dada et al.[88] studied the effect of two macular birefringence protocols (bow-tie retardation and irregular macular scan) using GDx-VCC on RNFL thickness parameters and found the standard protocol to overestimate RNFL thickness in eyes with macular lesion which was normalized using the irregular macular pattern protocol.

Conclusion

GDx-VCC has emerged as a valuable, objective tool of clinical value in the management of glaucoma. GDx-ECC has overcome the presence of ARPs and improved the diagnostic accuracy and the relationship between functional and structural measurements. It is useful for early detection of structural damage to RNFL and longitudinal follow-up to pick up progression. However the clinician should be aware of its limitations, reconfirm an abnormal test, check the image quality and always use the GDx test report in conjunction with clinical evaluation of the optic nerve head and functional testing for starting or adjusting glaucoma therapy.