Individualized Compensation of Anterior Segment Birefringence during Scanning Laser Polarimetry

Abstract

purpose. To describe a method for assessment and individualized compensation of anterior segment birefringence with scanning laser polarimetry.

methods. A scanning laser polarimeter (GDx Nerve Fiber Analyzer; Laser Diagnostic Technologies, Inc., San Diego, CA) was modified to accommodate a variable compensator. The magnitude and axis of anterior segment birefringence of normal eyes were determined from a polarimetry image of the Henle fiber layer. The variable compensator was then adjusted to minimize anterior segment birefringence. Retinal nerve fiber layer (RNFL) and macular measurements were then obtained. Macular images with individualized compensation served to verify the effectiveness of the compensation. To demonstrate individualized compensation, two sets of three images each were obtained from four eyes of four normal subjects. One set was obtained with individualized compensation and another with fixed compensation, as used in the commercial polarimetry system.

results. In the tested eyes, the magnitude of anterior segment birefringence ranged from 21.7 to 86.3 nm, and the slow axis ranged from 5.7° nasally upward to 54.3° nasally downward. The maximum residual retardation resulting from compensation was 70 nm for fixed compensation and 11.5 nm for individualized compensation. The compensation residual directly affected the assessment of the RNFL by scanning laser polarimetry. RNFL images obtained with individualized compensation were more consistent with the expected anatomy of the eye. In the eyes measured, the range of RNFL thicknesses appeared to be narrower with the variable corneal and lens compensator (VCC) compared with the fixed corneal compensator (FCC).

conclusions. In eyes with a normal macula, the magnitude and axis of anterior segment birefringence can be determined from a polarimetry image of the Henle fiber layer. Individualized anterior segment compensation can be achieved with the described method so that the measured birefringence largely reflects the RNFL birefringence. Whether and how macular diseases affect this method remain to be investigated.

Retinal scanning laser polarimetry (SLP) was developed to assess the retinal nerve fiber layer (RNFL) thickness at a particular location by measuring retardation, the change of polarization of light reflected from the retina. ^{1 2 3} This technique is based on the birefringence properties of the RNFL. Each retinal nerve fiber bundle contains microtubules, cylindrical intracellular organelles with diameters much smaller than the wavelength of the illuminating light, which are form birefringent. ^{4 5} Because the RNFL is form birefringent, its optic axis is aligned with the retinal nerve fiber bundle orientation, and its retardance is proportional to its thickness. ^{1 2 4} In the case of the RNFL, the optic axis is a slow axis. ^{1 4 6} In the area near the optic disc, the axis distribution is approximately radial. In normal eyes, the superior and inferior regions exhibit higher retardance, and the temporal and nasal regions exhibit lower retardance. This corresponds to the thicker RNFL in the superior and inferior regions. SLP is a noninvasive, rapid, reproducible, and objective test. ³  

Although the RNFL exhibits substantial birefringence, it is not the only birefringent structure in the eye. The Henle fiber layer in the macula also is birefringent. This layer consists of elongated photoreceptor axons extending radially from the fovea. The Henle fiber layer is structurally similar to the RNFL and exhibits significant birefringence. ^{7 8} Further, it is more uniform than the peripapillary RNFL. The cornea and, to a lesser extent, the lens also exhibit birefringence. ^{9 10} Because all birefringent structures cause a change in the polarization of an illuminating beam, the accuracy of RNFL measurement with SLP depends on the ability to extract the RNFL retardance from the measured total retardation. 

To minimize corneal and lens birefringence, a commercial SLP system (GDx Nerve Fiber Analyzer; Laser Diagnostic Technologies, Inc., San Diego, CA) uses a fixed compensator with magnitude of 60 nm (single-pass retardance) and a fast axis oriented at 15° nasally downward (ND). The measured total retardation is thought largely to reflect the retardance of the RNFL. However, in a recent study, Knighton et al. ¹¹ measured central anterior segment birefringence in 73 subjects (146 eyes) and showed large variation in individual eyes. The slow axis of most anterior segments was oriented between 80° ND and 20° nasally upward (NU), and the magnitude ranged from 0 to 125 nm (single-pass retardance). This wide range of anterior segment birefringence causes variability in RNFL assessment with the current GDx system (Laser Diagnostic Technologies, Inc.), particularly in the eyes with an axis and/or magnitude outside the range of magnitudes used in the fixed compensator. 

The influence of the axis of anterior segment birefringence on these measurements has been demonstrated. ^{12 13} However, these studies did not investigate the variation in its magnitude. To achieve complete compensation for the anterior segment, a compensator must not only account for the variation in the birefringence axis, but also must account for the variation in magnitude. In the current study, we developed a method for assessment and compensation of both the axis and the magnitude of anterior segment birefringence. 

Methods

A commercial GDx was modified to incorporate a variable anterior segment compensator. In brief, macular polarimetry images with the variable compensator set to 0 nm retardance were acquired, to determine the axis and magnitude of anterior segment birefringence. The compensator was then adjusted to minimize anterior segment birefringence based on macular measurements. Peripapillary RNFL measurements were then taken with complete anterior segment compensation. To compare the measurements with the commercial GDx, the same subjects also were imaged with the compensator set to the GDx fixed corneal compensation settings, and these images were compared with the ones obtained with variable corneal compensation. 

Experimental Setup

The experimental setup was the GDx system (Laser Diagnostic Technologies, Inc.) modified so that the original fixed corneal compensator (FCC) was replaced with a variable corneal and lens compensator (VCC) as shown in Figure 1A . The magnitude of the VCC was set by Dial 1 from 0 to 120 nm, and the axis was set by Dial 2 from −90° to +90°. A diagram of the experimental setup is shown in Figure 1B . There are four retarders in the measurement beam’s path: The first two linear retarders have equal retardance and form a VCC, the third linear retarder is the combination of cornea and lens—the anterior segment (Fig. 1B 1A) —and the fourth linear retarder, with radially distributed axes, is the retinal birefringent structure (Fig. 1B , RE; either peripapillary RNFL or the Henle fiber layer in the macula). Images of the ocular fundus are formed by scanning a near-infrared beam (785 nm) point by point in a raster pattern. For each location in the image, the total retardation in the optical path is determined by detecting the ellipticity induced in a linearly polarized input beam. ² Reflection from the ocular fundus exhibits a high degree of polarization preservation and is treated as a complete polarization-preserving reflector. ^{6 14} The measurement beam double passes each retarder along the same optical path—that is, the illuminating beam path overlaps the beam path reflected from the retina. 

In SLP, the cornea, lens, and retina are all treated as linear retarders (optical elements that introduce retardation to an illuminating beam). ² A linear retarder has a slow axis and a fast axis, and the two axes are orthogonal to each other. Polarized light travels at higher speed when its electric field vector is aligned with the fast axis of a retarder. In contrast, polarized light travels at lower speed when its electric field vector is aligned with the slow axis of a retarder. 

Anterior segment birefringence is dominated by corneal birefringence. In general, corneal birefringence is modeled with a biaxial crystal, with one optic axis perpendicular to the corneal surface and another optic axis parallel to the corneal surface. ¹⁰ At normal incidence, the optic axis perpendicular to the corneal surface has no effect on the polarization of the illuminating light and therefore is negligible. The corneal birefringence is not uniform. Usually, retardance increases toward the periphery of the cornea. ¹⁰ In SLP, the scan beam paths are nearly perpendicular to the corneal surface and cover only a central small area. Therefore, the cornea can be assumed to be a uniform linear retarder for the entire scan field. However, a change in the alignment of the eye may also cause the effective corneal birefringence to change. 

A linear retarder can be described by retardance (δ) and azimuth (θ) of the fast axis. Based on Mueller calculus, for a two-retarder combination (δ₁ and θ₁, and δ₂ and θ₂), the combined retardance (δ) is determined as follows ¹⁵ :  

\[\mathrm{cos}\ (2{\pi}{\delta}/{\lambda})\ {=}\ \mathrm{cos}\ (2{\pi}{\delta}_{1}/{\lambda})\ \mathrm{cos}\ (2{\pi}{\delta}_{2}/{\lambda})\ {-}\ \mathrm{sin}\ (2{\pi}{\delta}_{1}/{\lambda})\ \mathrm{sin}\ (2{\pi}{\delta}_{2}/{\lambda})\ \mathrm{cos}\ 2({\theta}_{2}{-}{\theta}_{1})\]

The equation shows that the combined retardance varies with the angle between the azimuth of the two retarders. When θ₂ − θ₁ is 0°, the axes of the two retarders are completely aligned with each other; the total retardance is then the sum of the two retarders (δ = δ₁ + δ₂), which is the maximum retardance of the combination. When θ₂ − θ₁ is 90°, the axes of the two retarders cross each other. The net retardance is then the difference of the two retarders (δ = δ₁ − δ₂), which is the minimum retardance of the combination. 

Method of Measuring Anterior Segment Birefringence

To determine anterior segment birefringence, the magnitude of the VCC is set to zero by simply aligning the fast axis of the first retarder with the slow axis of the second identical retarder and then obtaining SLP images of the macula. The retardation at a locus of points along a circle centered on the fovea is calculated (Fig. 2) . The measurement circle is 80 pixels in diameter, which corresponds to 1.41 mm in an emmetropic eye. The retardation profile reflects the combined retardance of the anterior segment and that of the Henle fiber layer. 

The axis of anterior segment birefringence can be determined from the measured macular retardation profile (Fig. 2) . According to the equation, the combined retardance of two retarders varies with the angle between the axes of the two retarders. At a macular locus where the radial slow axis of the Henle fiber layer is parallel to the slow axis of the anterior segment, the combined retardance is the sum of the two. This results in a maximum in the macular retardation profile. At a macular locus where the radial slow axis of the Henle fiber layer is perpendicular to the slow axis of the anterior segment, the combined retardance is the difference of the two. This results in a minimum in the macular retardation profile. Thus, the slow axis of the anterior segment is aligned with the orientation of the retardation maxima in the macula (Fig. 2) . The location of the two maxima are determined by performing a least-squares fit to the retardation profile with the equation. The black line in Figure 2 connects the two maxima and indicates the slow axis orientation of the anterior segment. 

The magnitude of anterior segment birefringence also can be determined from the measured macular retardation profile (Fig. 2) . The shape of the macular retardation profile depends on the magnitude of anterior segment birefringence and that of the Henle fiber layer. Three examples of macular profiles were calculated according to the equation (Fig. 3A) . The three profiles represent the cases of anterior segment birefringence less than, equal to, and greater than that of the Henle fiber layer. Figure 3B illustrates the relationship between the macular retardation profile and the birefringence of the Henle fiber layer and that of the anterior segment. The average and half of the modulation (maximum –minimum) of the macular retardation profile were calculated for three cases of Henle fiber layer retardation (18, 23, and 28 nm) and a range of anterior segment retardation (0–80 nm). As seen, when the retardation of the anterior segment is higher than that of the Henle fiber layer, the average retardation of the macular profile is dominated by the anterior segment. When the retardation of the anterior segment is less than that of the Henle fiber layer, half of the modulation of the profile is exactly the retardation of the anterior segment. The magnitude of the retardation of the Henle fiber layer and that of the anterior segment can both be determined by performing a least-squares fit to the macular retardation profile with the equation. A set of rules is used to determine one of the two values generated from the least-squares fit to be the magnitude of the anterior segment and the other to be the magnitude of the Henle fiber layer. 

Method of Variable Compensation

Once the magnitude and the slow axis of anterior segment birefringence are known, the VCC can be adjusted to minimize it. The magnitude of the VCC is set to a desired value by varying the angle between the axes of the two identical retarders (Dial 1 in Fig. 1A ). The slow axis of the VCC is set to the desired orientation by rotating the two retarders together (Dial 2 in Fig. 1A ) so that the slow axis of the retarder combination is perpendicular to the slow axis of the anterior segment. A VCC consisting of two fixed retarders is a retarder plus a rotator. However, in the double-pass SLP setup, the rotation in the illuminating beam is canceled by the rotation in the reflected beam, and the combination can simply be treated as a retarder. 

The VCC method described in this study takes into account both the axis and the magnitude of anterior segment birefringence. Once the VCC is correctly adjusted, the total retardation measured with the SLP is directly the birefringence of the RNFL or that of the Henle fiber layer. As a result, an SLP image of the macula should show a radially symmetric uniform pattern, and the SLP image of the peripapillary RNFL should show a pattern consistent with the anatomy of the eye. 

The compensated macular image serves to confirm the effectiveness of the VCC. Similar to the determination of anterior segment birefringence from the macular retardation profile, the residual of the anterior segment compensation also can be determined from the macular polarimetry image obtained with the VCC. If the compensation is complete, the combined retardance of the VCC and the anterior segment is zero. In this case, the macular image should be uniform and dark around the fovea and yield a flat macular retardation profile. If the compensation is incomplete, the residual birefringence of the VCC and the anterior segment combination causes a modulation in the macular retardation profile. Residual birefringence can be determined using the same method as that for the anterior segment determination. 

Testing the Individualized Compensation in Normal Subjects

Measurements were obtained on normal subjects to demonstrate the VCC method and its effects on the RNFL measurements. For each eye, three macular images with VCC set to zero were obtained. Anterior segment birefringence was determined from the average of the three macular measurements. With VCC properly set to minimize the determined anterior segment birefringence, one additional macular image and three peripapillary RNFL images were obtained. The macular image served to confirm the effectiveness of the VCC. 

To compare the VCC method with the fixed compensation method of the GDx system, the same eyes were imaged with the VCC adjusted to simulate the FCC. For each eye, one macular image and three peripapillary RNFL image were acquired. The macular image was taken to measure the residual birefringence resulting from FCC. 

Four right eyes of four normal subjects were included. Each had a full ophthalmic examination, including measurement of visual acuity, intraocular pressure, Goldmann applanation tonometry, slit lamp biomicroscopy, and dilated ophthalmoscopy. None of the subjects had a history of any ocular diseases. Each had intraocular pressure less than 21 mm Hg, normal optic disc appearance (intact rim, no hemorrhage, notch, excavation, nerve fiber layer defect, or asymmetry of the vertical cup-to-disc ratio >0.2), and normal visual fields (program 24-2; Humphrey Field Analyzer; Humphrey Instruments, San Leandro, CA). Informed consent was obtained with the approval of the University of California, San Diego, Human Subjects Committee. The study protocol conformed with the provisions of the Declaration of Helsinki. 

Results

Magnitude and Axis of Anterior Segment Birefringence

Figure 4 illustrates anterior segment birefringence measurements in three eyes. The magnitude ranged from 22 to 86 nm, and the slow axis ranged from 15° ND to 54° ND. At the lower magnitude, the macular birefringence image was darker, and at the higher magnitude, the macular birefringence image was brighter. The slow axis (Fig. 4 , S) of the anterior segment was determined from the retardation maxima in the macula along the measurement circle (Fig. 4 , thick green band) centered on the fovea. The diameter of the circle was 80 pixels in diameter. The magnitude of anterior segment birefringence was approximated by half of the modulation of the retardation profile in case 1 (Fig. 4 , left) and by the average of the retardation profile in cases 2 and 3 (Fig. 4 , middle and right, respectively). 

Comparison of Variable and Fixed Compensation

In the eyes in which anterior segment birefringence deviated significantly from the NFA-assumed values, significant differences were observed between the RNFL images acquired with the VCC and those with the FCC. Measurements of the eyes are shown in Figures 5 6 7 and 8 with comparison of VCC images with FCC images. These examples were selected to demonstrate different types of deviation of anterior segment birefringence from the FCC settings. Anterior segment birefringence, residual birefringence from the FCC, and residual birefringence from the VCC are summarized in Table 1 for these examples. Anterior segment birefringence varied widely, with the magnitude ranging from 21.7 to 86.3 nm, and the slow axis ranging from 5.7° NU to 54.3° ND. The standard deviations of corneal birefringence calculated from three repeated measurements are provided in Table 1 . The variability may not be typical because of the small sample size. The slow axis of residual birefringence was determined from the two maxima of macular retardation. Reflected on the RNFL image, the residual is added to the total retardation in the retinal locations where the nerve fiber bundle is parallel to the slow axis of the residual and subtracted from the total retardation in the locations where the nerve fiber bundles are perpendicular to the residual slow axis. For the four eyes illustrated in Figures 5 6 7 8 , average peripapillary RNFL thicknesses measured with the FCC and VCC are summarized in Table 2 , which are the average thicknesses at a locus of points along an ellipse 1.75 times the disc size. 

Figures 5 and 6 illustrate the effect of the VCC on the eyes in which the anterior segment magnitude is close to the NFA-assumed value, but the slow axis of the anterior segment is oriented farther ND (Fig. 5 , subject 1) and farther NU (Fig. 6 , subject 2) than the FCC’s assumed orientation. The three macular polarimetry images in the top row from left to right were obtained with the VCC adjusted to zero, the VCC adjusted the same as the FCC, and the VCC adjusted to cancel out anterior segment birefringence. The left macular image was used to measure anterior segment birefringence, the middle one to measure FCC residual birefringence, and the right one to measure VCC residual birefringence. The three images in the bottom row, from left to right, are reflectance image of the peripapillary retina, the RNFL image obtained with the FCC, and the RNFL image obtained with the VCC, adjusted to cancel out anterior segment birefringence, respectively. Figure 5 demonstrates that a further ND anterior segment axis caused very significant residual birefringence with slow axis approximately vertical in the FCC images. Reflected on the RNFL image, the superior and inferior nerve fiber bundles are significantly elevated. The temporal and nasal magnitudes in this subject were also somewhat elevated in the FCC image, because anterior segment residual birefringence is much higher than temporal and nasal RNFL birefringence, and net birefringence (difference between the residual and that of the RNFL) is still higher than the temporal and nasal RNFL alone. Figure 6 demonstrates that a further NU axis caused a rotated RNFL pattern in the FCC image. The rotation of the RNFL pattern correlates with the slow axis orientation of the residual of the compensation. 

Figure 7 (subject 3) illustrates the effect of the VCC for an anterior segment with axis close to the GDx’s FCC assumed value, but with the magnitude significantly higher. Arranged in the same format as that of Figures 5 and 6 , the three macular polarimetry images in the top row, from left to right, were obtained with the VCC adjusted to zero, to the same setting as the FCC, and to cancel out anterior segment birefringence, respectively. The three images in the bottom row, from left to right, are reflectance image of the peripapillary retina, the RNFL image obtained with the FCC, and the RNFL image obtained with the VCC adjusted to cancel out anterior segment birefringence, respectively. In this case, the FCC was undercompensating the anterior segment, and resulted in significant residual with slow axis approximately horizontal. In this case, less total retardation was measured in the superior and inferior retina, where the slow axes of the nerve fiber bundles were almost perpendicular to the residual birefringence. In contrast, higher total retardation was measured in the temporal and nasal retina where the slow axes of the nerve fiber bundles were almost parallel to the residual birefringence. As seen, the FCC RNFL image showed a very rotated RNFL pattern, and the VCC image showed a RNFL pattern that is more consistent with the expected anatomy of a normal eye. 

Finally, Figure 8 (subject 4) illustrates the effect of the VCC at a much lower anterior segment magnitude and slightly farther ND anterior segment axis. The images were arranged in the same format as that of Figures 5 6 and 7 . As seen, the slow axis of the residual anterior segment birefringence is nearly vertically imaged with the FCC, because of overcompensation. The magnitude of the residual was not as high as that in subject 1 (Fig. 5) , imaged with the FCC. As a result, the RNFL image acquired with the FCC shows less-elevated retardation compared with values in subject 1. Compared with the VCC’s RNFL image, the FCC’s RNFL image still showed significantly elevated RNFL thicknesses. 

When the combined retardance of the compensator and the anterior segment is zero, the compensation is considered complete. In this case, the compensated macular image is uniform and dark around the fovea. When the compensation is incomplete, a residual birefringence from the combination of the compensator and the anterior segment causes a nonuniform retardance profile around the fovea. In the normal subjects, the VCC method achieved residual segment birefringence of less than 12 nm, measured by half of the modulation (difference between the maximum and minimum values) of the macular retardation profile. 

Although it is frequently observed (Figs. 5 8) , “fixed” compensation does not always produce a vertical double-hump pattern in the macula (Figs. 6 7) . The vertical pattern occurs when the corneal slow axis is farther ND than 15° (Fig. 5) , the magnitude is less than 60 nm (Fig. 8) , or both. These types of corneal patterns are typical in most of the population. ¹¹  

Discussion

The magnitude and axis of anterior segment birefringence can be determined from a polarimetry image of the Henle fiber layer with this method. Individualized anterior segment compensation can be achieved so that the measured retardation largely reflects the RNFL retardance. The variation of the anterior segment birefringence clearly indicates the need for individualized compensation in the assessment of the RNFL by scanning laser polarimetry. For a variable corneal compensation method to work correctly, both the corneal axis and its retardation magnitude must be addressed. Therefore, it may be important to reassess previously published reports that were based on fixed corneal compensation. 

The described method resulted in residual birefringence less than 12 nm. The variable retarder consisting of two identical retarders worked reliably and was easy to control. The magnitude range of the VCC from 0 through 120 nm may be sufficient to cover the population variation. 

In general, when the FCC failed to compensate for the corneal birefringence, the RNFL images appeared to have either elevated retardation (Figs. 5 8) or a rotated RNFL pattern (Figs. 6 7) . The former was usually accompanied by bright retinal vessels. The residual birefringence could be measured from the macular polarimetry image. RNFL thickness images were generally darker (lower values) with the VCC and the RNFL pattern appeared to be a better match of the expected anatomy of the eyes. 

Comparison of the current estimates of macula-derived anterior segment birefringence with birefringence measurements obtained by Knighton et al. ¹¹ using the fourth Purkinje image does not validate this technique. Although there is good correlation between these two techniques, corneal birefringence varies with the location and incidence angle of the illuminating beam. ¹⁶ The SLP beam path and the fourth Purkinje beam path are not necessarily the same, and therefore their effective corneal birefringence may be different. Because it is important to compensate for the effective corneal birefringence of the specific scanning beam path, one technique or the other cannot be considered the gold standard. Rather, validation of this method, including test–retest variability and sensitivity and specificity determinations, should be performed in a larger group of normal subjects and patients with glaucoma before it is recommended for widespread clinical use. 

The long-term stability of anterior segment birefringence and the effect of corneal refractive surgery on the birefringence measurement remain to be investigated. Regardless of the outcome, however, an SLP with a VCC would allow the user to reassess anterior segment birefringence of the eye whenever needed, and new compensation could be used if significant change occurred. This particular VCC method is based on the assumption that the Henle fiber layer is uniformly distributed around the fovea. Curve-fitting to the macular retardation profile with the equation helps to reduce the impact of macular irregularity when determining the axis and the magnitude of anterior segment birefringence. For eyes without macular disease, residual birefringence was very low, indicating that both the method of measuring anterior segment birefringence and the method of variable compensation worked well. Whether and how macular disease affects the variable corneal compensation method remain to be investigated. In the absence of a healthy Henle fiber layer, alternative methods to estimate anterior segment birefringence may have to be considered. 

 Submitted for publication October 3, 2001; revised February 25, 2002; accepted March 6, 2002.

 Commercial relationships policy: E (QZ); C, F (RNW).

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A) The experimental setup is based on a commercial GDx system (Laser Diagnostic Technologies, Inc., San Diego, CA), but with the addition of a VCC. Dial 1 sets the VCC’s magnitude and Dial 2 sets the VCC’s axis. (B) The VCC experimental setup includes the following components: the SLP, the VCC consisting of two identical retarders, the anterior segment of the eye (A), the retinal birefringent structures (RE), either the RNFL or the Henle fiber layer, and the fundus as the polarization-preserving reflector.

Figure 2.

 

Macular polarimetry with the VCC adjusted to zero was used to determine anterior segment birefringence. Left: reflectance image. Right: color-coded retardation map with brighter colors representing higher retardation and darker colors representing lower retardation. The retardation at a locus of points along a circle (thick green band) centered on the fovea was calculated and is shown on the right side above the retardation map. The measurement circle was 80 pixels in diameter. The slow axis of anterior segment birefringence was determined from the orientation of the retardation maxima in the macula (black line), and the magnitude was determined from the retardation profile (yellow line). T, temporal, S, superior, N, nasal, and I, inferior.

Figure 3.

 

(A) Three macular retardation profiles were calculated corresponding to anterior segment (AS) birefringence less than (AS < HFL), equal to (AS = HFL), and greater than (AS > HFL) that of the Henle fiber layer (HFL), respectively. (B) Relationships between the macular retardation profile to the birefringence of the Henle fiber layer and that of the anterior segment were calculated. The horizontal axis is the anterior segment retardance. The average (Avg) and the half of the modulation (Mod/2) of the macular retardation profile were plotted for HFL retardation of 18, 23, and 28 nm, respectively, over the anterior segment retardation range of 0 to 80 nm.

Figure 4.

 

Three examples of anterior segment birefringence measurements are displayed. The slow axis is marked (either a white or a black line), and the magnitude is measured from macular retardation profile along the measurement circle (thick green band) centered on the fovea. From left to right: anterior segment birefringence was measured to be 22 nm and 23° ND, 52 nm and 54° ND, and 86 nm and 15° ND, respectively. Note the change of brightness of the macular retardation images.

Figure 5.

 

Images of subject 1, OD, are displayed. Top: macular polarimetry images. (Left to right): measurement of anterior segment birefringence (51.7 nm and 54.3° ND), measurement of FCC residual (70 nm and 89° ND), and measurement of VCC residual (9.5 nm and 25° ND). Bottom (left to right): reflectance image of the peripapillary retina, the RNFL image obtained with the FCC, and the RNFL image obtained with the individualized compensation, respectively. Note that the elevated FCC RNFL image correlates with the high residual in the FCC macular image.

Figure 6.

 

Images of subject 2, OD, are displayed. Top: macular polarimetry images. (Left to right): measurement of anterior segment birefringence (56.7 nm and 5.7° NU), measurement of FCC residual (23 nm and 37° NU), and measurement of VCC residual (9.5 nm and 22° NU). Bottom (left to right): reflectance image of the peripapillary retina, the RNFL image obtained with the FCC, and the RNFL image obtained with the individualized compensation, respectively. Note that the rotated FCC RNFL pattern correlates with the residual pattern in the FCC macular image.

Figure 7.

 

Images of subject 3, OD, are displayed. Top: macular polarimetry images. (Left to right): measurement of anterior segment birefringence (86.3 nm and 15.3° ND), measurement of FCC residual (47 nm and 22° ND), and measurement of VCC residual (11.5 nm and 41° ND). Bottom (left to right): reflectance image of the peripapillary retina, the RNFL image obtained with the FCC, and the RNFL image obtained with the individualized compensation, respectively. Note that the very rotated FCC RNFL image correlates with the residual pattern in the FCC macular image.

Figure 8.

 

Images of subject 4, OD, are displayed. Top: macular polarimetry images. (Left to right): measurement of anterior segment birefringence (21.7 nm and 23.3° ND), measurement of FCC residual (43 nm and 81° NU), and measurement of VCC residual (8.5 nm and 1° ND). Bottom (left to right): reflectance image of the peripapillary retina, the RNFL image obtained with the FCC, and the RNFL image obtained with the individualized compensation, respectively. Note that the very elevated FCC RNFL image correlates with the residual pattern in the FCC macular image.

Table 1.

 

Anterior Segment Birefringence and Residual Birefringence Data of the Normal Subjects

Subject	Anterior Segment			FCC Residual			VCC Residual
		Magnitude (nm)	Axis (deg)	Magnitude (nm)	Axis (deg)	Magnitude (nm)	Axis (deg)
1	51.7 ± 2.1	54.3 ± 2.1	70	89°	9.5	25
2	56.7 ± 1.5	−5.7 ± 1.5	23	−37°	9.5	−22
3	86.3 ± 1.2	15.3 ± 0.6	47	22°	11.5	41
4	21.7 ± 0.6	23.3 ± 1.2	43	−81°	8.5	1

 Positive axis values represent nasally downward and negative axis values represent nasally upward.

Table 2.

 

Mean Peripapillary RNFL Thickness Values

Subject	Average RNFL Thickness by FCC	Average RNFL Thickness by VCC
1	122.84	56.14
2	61.83	55.53
3	61.54	63.84
4	76.69	51.68

 Measurements of a locus of points along an ellipse 1.75 times the disc size for images illustrated in Figures 5 6 7 8 . Data are expressed as micrometers.

The authors thank Robert W. Knighton, PhD, and Xiangrun Huang, PhD, for valuable discussions on the corneal compensation method.