Retinal Nerve Fiber Layer Defect Volume Deviation Analysis Using Spectral-Domain Optical Coherence Tomography

Abstract

Purpose.: To report the retinal nerve fiber layer (RNFL) defect volume deviation according to structural RNFL loss in RNFL thickness maps.

Methods.: Retinal nerve fiber layer defect is defined in RNFL thickness maps by the degree of RNFL loss. A 20% to 70% degree of RNFL loss was set with a 1% interval as the reference level for determining the boundary of RNFL defects. Each individual RNFL thickness map was compared with a normative database map and the region below the reference level was identified as an RNFL defect. The RNFL defect volume was calculated by summing the volumes of each pixel inside RNFL defect. The RNFL defect volume deviation was calculated by summing the differences between the normative database and the subject's RNFL measurements. To evaluate the glaucoma diagnostic ability, the areas under the receiver operating characteristics curves (AUCs) were calculated.

Results.: Retinal nerve fiber layer defect volume and RNFL defect volume deviation (0.984 and 0.986, respectively) had significantly greater AUCs than all circumpapillary RNFL thickness parameters (all P < 0.001). In the early stage of RNFL loss (under 31% loss of RNFL), RNFL defect volume deviation showed better diagnostic performance than the RNFL defect volume. In multivariate analysis, RNFL defect volume and RNFL defect volume deviation were significantly associated with the mean deviation in visual field tests.

Conclusions.: Retinal nerve fiber layer defect volume deviation is a useful tool for diagnosing glaucoma and monitoring RNFL change. In early stage of RNFL loss, RNFL defect volume deviation is more sensitive for detecting glaucoma than the RNFL defect volume measurements.

Introduction

Glaucoma is a progressive optic neuropathy that damages retinal ganglion cells and their axons. The progressive loss of retinal ganglion cell axons results in retinal nerve fiber layer (RNFL) loss. Retinal nerve fiber layer defect identification and assessment of progression are important for diagnosing and managing glaucoma. Optical coherence tomography (OCT) allows objective and quantitative in vivo RNFL thickness measurements.¹ RNFL thickness represents the remaining portion of retinal ganglion cell axons, retinal blood vessels, and glial cells after glaucomatous damage.^2–4 Most studies have focused on remaining RNFL thickness to diagnose and monitor glaucoma.^5–8 In a recent study, the volumetric analysis showed better diagnostic ability than RNFL thickness⁹; however, the remaining volumes of RNFL defects are not necessarily proportional to the amount of glaucomatous damage.⁹ If an RNFL defect progresses only by enlargement without further thinning of a preexisting RNFL defect, the remaining volume of the RNFL defect will increase. If an RNFL defect progresses only by thinning without enlargement, the remaining volume of the RNFL defect will decrease. The lost RNFL portion more directly reflects glaucomatous damage. 

To the best of our knowledge, there have been no reports that simultaneously evaluated the remaining and lost parts of the RNFL. This study measured RNFL defect volume (remaining part of RNFL) and RNFL defect volume deviation (estimated lost part of RNFL) using an RNFL thickness map of spectral-domain OCT (Cirrus HD-OCT; Carl Zeiss Meditec, Dublin, CA, USA). These results were compared with conventional circumpapillary RNFL thickness. The advantages and limitations of RNFL defect volume and RNFL defect volume deviation were considered. 

Methods

Participants

A total of 160 glaucoma patients and 160 healthy control subjects were enrolled. They had visited either the general health care clinic or the glaucoma clinic at Hanyang University Medical Center from September 2012 to January 2013. The institutional review board of Hanyang University Medical Center approved the study protocol and the tenets of the Declaration of Helsinki were followed. Informed consent was obtained from all participants before beginning the study. 

All subjects underwent a comprehensive ophthalmic examination, including a visual acuity test, applanation tonometry, anterior segment examination, refraction, fundus examination, pachymetry (SP-3000; Tomey, Nagoya, Japan), standard automated perimetry (Humphrey Field analyzer with SITA standard 30-2 test; Carl Zeiss Meditec), and RNFL imaging with a spectral-domain OCT (Cirrus HD-OCT; Carl Zeiss Meditec). 

Healthy subjects were included if they had a best-corrected visual acuity of 20/30 or better, a normal visual field, a normal anterior segment on slit-lamp examination, a normal-appearing optic disc head, no RNFL defects, and no history of IOP higher than 21 mm Hg. Glaucoma subjects were included based on the presence of RNFL defects on red-free photographs or the presence of the glaucomatous appearance of the optic nerve head on color fundus photographs (neuroretinal rim notching or thinning, or optic disc hemorrhage) and the presence of visual field defects that corresponded with RNFL defects and optic nerve head abnormalities. A visual field defect was defined as (1) the presence of a cluster of three or more nonedge contiguous points on a pattern deviation plot with a P value of less than 5% (one of which had a P value less than 1%) confirmed by at least two consecutive examinations, (2) a pattern standard deviation (PSD) with a P value less than 5%, or (3) a glaucoma hemifield test result outside normal limits. The severity of glaucomatous damage was classified as mild (mean deviation [MD] ≥ −6dB) or moderate-to-advanced (MD < −6dB). Visual field tests were considered reliable based on fixation losses and false-positive and false-negative results of 15% or less. 

Eyes with high myopic or hyperopic refractive errors of less than −6.0 diopters or greater than +3.0 were excluded from this study. Eyes with any ophthalmic or neurologic disease known to affect RNFL thickness or visual function also were excluded. 

Retinal Nerve Fiber Layer Image Acquisition

To obtain the RNFL thickness map, an Optic Disc Cube scan protocol was applied using a Cirrus HD-OCT (software version 5.1; Carl Zeiss Meditec). This protocol generated 200 horizontal B-scans, each composed of 200 A-scans on the 6 × 6-mm² optic disc region. Built-in analysis software automatically segmented the RNFL boundary and calculated RNFL thickness. Retinal nerve fiber layer segmentation was checked for every OCT image by two observers (JWS, MS). A built-in algorithm automatically detected the center of the optic disc, and its coordinates were displayed in a result report as the degree of movement from the center of the RNFL thickness map (e.g., “Disc Center [−0.02, 0.04] mm”). Retinal nerve fiber layer thickness maps of all the study eyes were aligned using the respective coordinates of the optic disc centers automatically provided by the Cirrus software. To account for differences in Cirrus scan center among the study eyes, a central 5.5 × 5.5–mm² area of the aligned RNFL thickness maps was used for analysis. All images had signal strength of at least 7. Images with motion artifacts were rescanned at the same visit. 

Normative RNFL Thickness Map Database

A normative database was composed of another 261 eyes of healthy Korean subjects ranging in age from 18 to 81 years (mean 56.4 ± 10.2 years).⁹ All individuals in the normative database had complete ophthalmic examinations and were selected using the same criteria as the healthy control group. Mean deviation of the visual field test was 0.27 ± 1.03 dB. Age distribution included 44 eyes in 18- to 30-year-olds, 41 eyes in 31- to 40-year-olds, 45 eyes in 41- to 50-year-olds, 48 eyes in 51- to 60-year-olds, 41 eyes in 61- to 70-year-olds, and 42 eyes in those older than 70 years. Significant negative correlations were found between age and RNFL thickness (−0.26 μm/y; P = 0.014). To accurately differentiate age-related RNFL changes from glaucomatous RNFL loss, the normative RNFL thickness map database was adjusted by age-related rates of change. 

Retinal Nerve Fiber Layer Defect Volume and RNFL Defect Volume Deviation Measurements

Retinal nerve fiber layer thickness measurements of 200 × 200 pixels on the RNFL thickness map were extracted by a Cirrus HD-OCT Research Browser (Carl Zeiss Meditec). In the RNFL thickness map, RNFL defects were defined by the degree of RNFL loss. A 20% to 70% degree of RNFL loss with a 1% interval was set as the reference level for RNFL defect determination. To identify the boundary of RNFL defects, each individual RNFL thickness map was compared with the normative database map using MATLAB software (The MathWorks, Inc., Natick, MA, USA), and the region below the reference level (20%–70% loss of RNFL with 1% interval) was detected (Fig. 1). 

Figure 1

 

Identification and measurement of RNFL defect volume and RNFL defect volume deviation in the thickness map. The RNFL defects are detected by comparison (B) between normative (A) and individual (C) RNFL thickness maps. Through B-scan comparison, the region ([B] interval between vertical lines, [C] red transverse line between red empty circles) below the reference level ([B] red line) is recognized as an RNFL defect ([C] red dashed line). In this case, the reference level is set as 36% loss of RNFL, which is equal to 64% of normative data ([B] 64% level of green graph equal to red graph). The RNFL defect volume is the sum of the area under a subject's RNFL thickness measurement ([B] green zone under blue graph) in every 200 B-scans. The RNFL defect volume deviation is the sum of the area between the normative database and the subject's RNFL measurement ([B] yellow zone between green and blue graph) in every 200 B-scans. The RNFL defect volume and RNFL defect volume deviation are automatically calculated within the identified RNFL defect boundary ([C] text). The RNFL thickness maps were analyzed within a 5.5-mm square ([A, C] inside black square; [B] inside diagonal pattern zone) and outside the 2.5-mm-diameter circular area ([A, C] outside black circle).

The RNFL defect volume (remaining part of RNFL) was calculated by summing of pixel volumes inside the RNFL defect region, RNFL defect volume =  

 

where n is number of pixels inside the RNFL defect.  

The RNFL defect volume deviation (estimated lost part of RNFL) was calculated by summing the differences between the normative database and the subject's RNFL measurement, RNFL defect volume deviation =  

 

where n is the number of pixels inside the RNFL defect.  

Eyes that moved more than 0.25 mm from the center of the RNFL thickness map were excluded. Considering individual variation of eye movements, RNFL thickness maps were analyzed within a 5.5-mm square. The optic disc area (range, 1.04–4.09 mm²) varied among subjects, so the 2.5-mm-diameter circular area in the RNFL thickness map was removed. 

Statistical Analyses

Statistical analyses were performed using SPSS software (version 18.0; SPSS, Inc., Chicago, IL, USA) and MedCalc software (version 12.2.1; MedCalc, Mariakerke, Belgium). Clinical characteristics and OCT measurements were compared between the healthy and glaucoma groups using an independent t-test and χ² test. The mean RNFL defect volume and RNFL defect volume deviation were compared at each reference level using a paired t-test. To evaluate the glaucoma diagnostic ability of RNFL defect volume and RNFL defect volume deviation, the areas under the receiver operating characteristic curves (AUCs) were calculated. To determine appropriate reference levels for RNFL defects, the AUCs of several reference levels were compared using a method described by DeLong et al.¹⁰ Univariate and multivariate regression analyses were performed to evaluate the association of RNFL defect volume and RNFL defect volume deviation with glaucomatous visual field damage. P values of 0.05 or less were considered statistically significant. 

Results

Table 1 describes the clinical characteristics of subjects. The mean age, sex, IOP, refractive error, and optic disc area did not differ between the healthy and glaucoma groups. The average MD of visual field tests was −0.29 ± 1.13 dB in the healthy group and −4.91 ± 4.15 dB in the glaucoma group. The proportion of mild (MD ≥ −6 dB) and moderate-to-advanced (MD < −6 dB) visual field defects in glaucoma subjects was 107:53 (66.9%:33.1%). 

Table 1

 

Comparison of Clinical Characteristics Between Healthy and Glaucoma Groups

	Healthy	Glaucoma	P Value
n	160	160
Age, y	55.4 ± 12.7	56.5 ± 13.4	0.440*
Sex, male:female	81:79	94:66	0.178†
IOP, mm Hg	14.9 ± 3.1	15.3 ± 3.6	0.243*
Signal strength	8.41 ± 0.81	7.74 ± 0.90	<0.001*
Refractive error, diopters	−0.09 ± 1.43	−0.15 ± 2.71	0.920*
MD, dB	−0.29 ± 1.13	−4.91 ± 4.15	<0.001*
MD≥−6:MD<−6)	160:0	107:53	<0.001†
PSD, dB	1.74 ± 0.70	5.04 ± 4.55	<0.001*
Disc area, mm2	2.13 ± 0.53	2.05 ± 0.48	0.146*
Rim area, mm2	1.31 ± 0.25	0.99 ± 0.96	<0.001*

*  Independent t-test.

†  χ² test.

The mean RNFL defect volume and RNFL defect volume deviation in glaucoma subjects and healthy subjects are described in Figure 2. These parameters increased with decreasing RNFL loss reference level. In the glaucoma group, RNFL defect volume deviation was significantly smaller than RNFL defect volume in the 20% to 30% RNFL loss reference level and larger in the 37% to 70% level. In the healthy group, RNFL defect volume deviation was significantly smaller than RNFL defect volume in a reference range of 20% to 41% RNFL loss (all P < 0.05, after Bonferroni correction). 

Figure 2

 

Mean RNFL defect volume and RNFL defect volume deviation in glaucoma and healthy subjects according to the degree of RNFL loss. The RNFL defect volume and RNFL defect volume deviation increased with decreasing RNFL loss reference level. Significant differences existed between RNFL defect volume and RNFL defect volume deviation in a range of intervals indicated by the arrows (20%–30% and 37%–70% RNFL loss reference level in glaucoma group; 20%–41% RNFL loss reference level in healthy group). G, glaucoma subjects; N, healthy subjects.

Figure 3A presents the AUCs of RNFL defect volume and RNFL defect volume deviation according to the degree of RNFL loss. Table 2 presents the AUCs of circumpapillary RNFL thickness. The RNFL defect volume showed the highest diagnostic performance at 37% loss of RNFL level (0.984, AUC; 95.0% sensitivity; 90.0% specificity). The RNFL defect volume deviation showed the highest diagnostic performance at 36% loss of RNFL level (0.986, AUC; 96.9% sensitivity; 92.5% specificity). There was no significant difference between the highest AUCs of RNFL defect volume and RNFL defect volume deviation (P = 0.278). The AUCs showed plateau under the 36% reference point for RNFL defect volume deviation, but it had decreasing slope at same range for RNFL defect volume (Fig. 3A, interpreted from right to left). Under 31% loss of RNFL, the AUCs between RNFL defect volume and RNFL defect volume deviation showed significant differences (Fig. 3B, all P < 0.05, after Bonferroni correction). Average RNFL thickness (0.955, AUC; 86.2% sensitivity; 91.5% specificity), analyzed by a built-in OCT algorithm on the 3.46-mm circle, showed the highest performance among circumpapillary RNFL parameters, followed by inferior and 7 o'clock. The RNFL defect volume and RNFL defect volume deviation had significantly greater AUCs than all of the circumpapillary RNFL thickness parameters at their highest diagnostic performance reference level (at 37% and 36% loss of RNFL, all P < 0.001). 

Figure 3

 

Areas under the receiver operating characteristic curve (A) and differences of AUCs (B) of RNFL defect volume and RNFL defect volume deviation according to the degree of RNFL loss. (A) At 37% and 36% loss of RNFL, glaucoma diagnostic performance was highest for RNFL defect volume and RNFL defect volume deviation. The AUCs of volumetric parameters were significantly greater than the AUCs of average RNFL thickness at their highest diagnostic performance reference level. (B) With less than 31% loss of RNFL, AUCs between RNFL defect volume and RNFL defect volume deviation showed significant differences. RNFLDV, retinal nerve fiber layer defect volume.

Table 2

 

Glaucoma Diagnostic Ability of Circumpapillary RNFL Thickness, RNFL Defect Volume, and RNFL Defect Volume Deviation Assessed by the AUC and Sensitivity at ≥90% Specificity Level

	Healthy,n = 160	Glaucoma, n = 160	P Value	AUCs (95% CI)	Sensitivity/Specificity (Sp ≥ 90%)
CpRNFL parameters, μm
Average	102.9 ± 6.5	76.2 ± 11.4	<0.001	0.955 (0.929–0.972)	86.2/91.5
Quadrant
Temporal	78.9 ± 8.3	58.1 ± 11.4	<0.001	0.862 (0.824–0.895)	68.8/90.6
Superior	128.4 ± 11.7	93.8 ± 21.2	<0.001	0.902 (0.867–0.930)	74.7/90.1
Nasal	74.3 ± 9.3	63.5 ± 9.1	<0.001	0.707 (0.658–0.752)	31.2/90.1
Inferior	130.1 ± 13.0	89.3 ± 22.5	<0.001	0.935 (0.906–0.959)	83.8/90.6
Clock-hour
1	117.5 ± 19.3	89.7 ± 22.1	<0.001	0.791 (0.746–0.830)	51.8/91.0
2	90.3 ± 14.8	74.7 ± 14.6	<0.001	0.695 (0.646–0.741)	31.8/90.6
3	65.9 ± 9.9	57.2 ± 9.6	<0.001	0.606 (0.555–0.655)	19.4/90.1
4	68.7 ± 10.7	58.5 ± 10.1	<0.001	0.672 (0.622–0.718)	21.8/91.5
5	103.5 ± 15.8	77.9 ± 19.3	<0.001	0.814 (0.771–0.852)	58.8/90.1
6	138.9 ± 21.1	93.5 ± 30.2	<0.001	0.869 (0.831–0.901)	70.6/91.0
7	145.4 ± 18.2	95.6 ± 31.1	<0.001	0.928 (0.897–0.953)	81.6/90.6
8	82.1 ± 12.4	59.5 ± 15.6	<0.001	0.821 (0.779–0.858)	54.7/91.5
9	67.2 ± 7.4	50.0 ± 9.5	<0.001	0.729 (0.682–0.773)	40.6/90.6
10	88.4 ± 12.1	66.7 ± 16.4	<0.001	0.838 (0.798–0.874)	60.0/92.5
11	137.4 ± 18.9	95.5 ± 28.4	<0.001	0.881 (0.844–0.911)	68.2/91.0
12	130.0 ± 21.3	96.1 ± 28.2	<0.001	0.791 (0.747–0.831)	52.4/91.0
Volumetric parameters, mm3
RNFLD volume*	0.0139 ± 0.0283	0.2153 ± 0.1175	<0.001	0.983† (0.962–0.994)	95.0/90.0
RNFLD volume deviation*	0.0123 ± 0.0249	0.2422 ± 0.1505	<0.001	0.986† (0.966–0.996)	96.9/92.5

 CpRNFL, circumpapillary retinal nerve fiber layer; RNFLD, retinal nerve fiber layer defect; Sp, specificity.

*  The RNFL defect volume and RNFL defect volume deviation at their highest diagnostic performance reference level (at 37% and 36% loss of RNFL).

†  The AUCs of RNFL defect volume and RNFL defect volume deviation were significantly greater than those of all cpRNFL thickness parameters (all P < 0.001).

The relationship between MD of visual field tests and various clinical factors was assessed by univariate and multivariate regression analyses (Table 3). The MD of visual field tests was associated with rim area, average RNFL thickness, RNFL defect volume, and RNFL defect volume deviation by univariate regression analysis (P < 0.05). In multivariate analysis of the significant variables from univariate analysis, RNFL defect volume and RNFL defect volume deviation were significantly associated with MD of visual field tests (P = 0.042 and 0.002, respectively). 

Table 3

 

Univariate and Multivariate Regression Analysis to Determine Factors Associated With Visual Field Damage

	Univariate		Multivariate,R2 = 0.336
	Coefficient (95% CI)	P	Coefficient (95% CI)	P
Age	−0.290 (−0.695 to 0.115)	0.159
Refractive error	0.140 (−0.086 to 0.366)	0.225
Disc area	−0.014 (−0.029 to 0.001)	0.066
Rim area	0.033 (0.003 to 0.063)	0.032	0.049 (−0.772 to 0.870)	0.906
Average RNFL thickness	1.173 (0.874 to 1.472)	<0.001	0.119 (−0.031 to 0.269)	0.117
RNFLD volume	−0.010 (−0.013 to −0.007)	<0.001	20.098 (0.466 to 39.730)	0.042
RNFLD volume deviation	−0.015 (−0.019 to −0.011)	<0.001	−25.936 (−42.684 to −9.189)	0.002

 CI, confidence interval.

Discussion

The RNFL defect volume deviation (estimated lost part of RNFL) using an RNFL thickness map was comparable to RNFL defect volume (remaining part of RNFL). In the early stage of RNFL loss, RNFL defect volume deviation was more sensitive for detecting glaucoma than RNFL defect volume. Three-dimensional volumetric analysis showed better glaucoma diagnostic ability and association with visual field damage than conventional circumpapillary RNFL thickness. In clinical practice, deviation volume analysis can provide an objective interpretation of the RNFL thickness map to detect early changes in RNFL loss. 

Three-dimensional volumetric parameters showed better glaucoma diagnostic performance than circumpapillary RNFL thickness measurements. Circumpapillary RNFL thickness may dilute glaucomatous RNFL damage because the damaged RNFL is mixed with adjacent normal RNFL in clock-hour sector.⁹ In this study, selective analysis of RNFL damage contributed to better diagnostic ability. Glaucomatous damage affects the RNFL structure in all x-y-z directions, but circumpapillary RNFL thickness evaluates only the z-axis of RNFL on the circular path. If RNFL defects exist or deteriorate outside the 3.46-mm-diameter scan circle, circumpapillary RNFL thickness may miss the RNFL defect.⁹ Although many studies have proven the efficacy of circumpapillary RNFL thickness in diagnosing and managing glaucoma,^5–8 three-dimensional volumetric analysis shows that there is still room for improvement. 

The RNFL defect volume converged to 0 mm³ as the reference level became close to 70% RNFL loss (Fig. 2). This indicates that RNFL is difficult to diminish below the remaining 30% of RNFL. Hood and Kardon⁴ reported that the residual RNFL after complete axonal loss contains retinal blood vessels and glial cells, which are estimated to be 33% of healthy RNFL thickness. These nonaxonal components could dilute glaucomatous RNFL changes in RNFL defect volume analysis, and the dilution effect is maximized in early stages of RNFL loss (Fig. 4). This may explain the abrupt drop in the diagnostic performance of the RNFL defect volume under 30% RNFL loss (Fig. 3A). In addition, nonaxonal components vary considerably among individuals,³ and there is no available information regarding nonaxonal components outside the arcuate region on the 3.46-mm-diameter scan circle. This indicates that RNFL defect volume deviation may be efficient for evaluating glaucomatous RNFL changes in early stage of RNFL loss. 

Figure 4

 

Diagram of glaucomatous RNFL changes in remaining and lost RNFL. If a healthy RNFL was 120 μm thick before glaucoma development, nonaxonal components may occupy 40 μm (33% of healthy RNFL thickness), according to Hood and Kardon's study.¹⁵ (A) Retinal nerve fiber layer loss in this stage may be either a natural aging process or due to glaucomatous change. (B) If RNFL loss continues, glaucomatous damage should be suspected. In this stage, although lost RNFL doubles, remaining RNFL decreases by 9%. (C) This stage includes 20% to 30% loss of RNFL in which RNFL defect volume deviation showed better diagnostic performance than RNFL defect volume. Although lost RNFL doubles, remaining RNFL decreases by 20%. (D) The highest diagnostic performance (36% and 37% loss of RNFL) is measured at the early part of this stage. Although lost RNFL doubles, remaining RNFL decreases by 50%. The decreasing rate of remaining RNFL is balanced with an increasing rate of lost RNFL. If nonaxonal components are removed from the remaining RNFL, the pure axonal decreasing rates are 14.3% (B), 33.3% (C), and 100% (D) in each stage. Nonaxonal components dilute the rate of glaucomatous RNFL changes in the RNFL defect volume. The dilution effect is maximized in early stages of RNFL loss.

In this study, we found that RNFL defects were mature enough to diagnose glaucomatous damage at approximately 36% to 37% RNFL loss. The RNFL defect volume deviation analysis showed better diagnostic performance than RNFL defect volume and circumpapillary RNFL thickness measurements in shallow immature defects under 30% RNFL loss. The border of an RNFL defect is the most active site for loss of retinal ganglion cells and their axons.¹¹ Shallow immature RNFL defects are highly suspicious areas in progression because they usually exist near mature RNFL defects. Sometimes, they may appear as islands apart from mature RNFL defects. Recent studies of RNFL defect progression announced that widening is the most frequent pattern of progression, followed by new development or deepening.^12,13 However, it is difficult to predict the pattern of RNFL defect progression in a particular individual. Figure 5 shows the RNFL loss mapping analysis, which displays the distribution pattern of RNFL loss. Mapping of RNFL loss allows individualization of information regarding shallow immature defects that are vulnerable to additional glaucomatous damage. Clinicians can infer widening (Fig. 5A), deepening (Figs. 5A, 5B), or new development (Fig. 5A) of RNFL defects by mapping RNFL loss. Efforts can then be focused on the vulnerable zones. Visualization of immature areas around RNFL defects can assist clinical prediction and management of RNFL defect progression. Further longitudinal research is needed to interpret the progression in RNFL loss mapping. 

Figure 5

 

The RNFL defect volume deviation analysis is superimposed on the OCT en face fundus image. The distribution pattern of RNFL loss is displayed by color map with a 5% interval. The RNFL defects are defined as 36% loss of RNFL based on the results of diagnostic ability. (A) Localized RNFL defect (red line) is observed at the superotemporal region and surrounded by shallow RNFL loss (bluish zone around RNFL defect). The border of the RNFL defect is the most active site of retinal ganglion cell and axon loss. If the RNFL defect deteriorates, this shallow RNFL loss is a highly suspicious area for widening progression. Another shallow loss (blue islands) apart from the RNFL defect has a chance to develop new RNFL defects. Deepening is also possible inside the RNFL defect, because it still has more than 55% remaining RNFL. (B) Diffuse RNFL defects (red line) are observed at the superotemporal and inferotemporal regions. In contrast to the above case, there is not enough shallow RNFL loss to extend the RNFL defect or develop a new RNFL defect. The most likely pattern of progression is deepening of the RNFL defect at the superotemporal region in which depth of RNFL defect is relatively shallow (diagonal pattern zone). Mapping of RNFL loss allows visualization of zones vulnerable to additional glaucomatous damage. Clinicians should focus their effort on these vulnerable zones.

The MD of visual field tests was associated with the RNFL defect volume and RNFL defect volume deviation in multivariate regression analysis. Care should be taken when interpreting the RNFL defect volume, because it does not necessarily represent the amount of glaucomatous damage. Table 4 and Figure 6 demonstrate changes in RNFL defect volume and RNFL defect volume deviation when RNFL defects progress. If an RNFL defect progresses only by enlargement without further thinning of the preexisting RNFL defect, the RNFL defect volume will increase. If an RNFL defect progresses only by thinning without area enlargement, the RNFL defect volume will decrease. The RNFL defect volume may cause confusion to interpret change of RNFL defect when RNFL defect progresses simultaneously via both enlargement and thinning. However, the RNFL defect volume deviation is advantageous because it proportionally reflects the amount and change of glaucomatous damage. 

Figure 6

 

Schematic RNFL defect progression examples. Initial RNFL thickness displayed with gray line. After focal RNFL defect progression (red arrows), final RNFL thickness displayed with blue line. (A) If an RNFL defect progresses only by enlargement without further thinning of the preexisting RNFL defect, remaining RNFL (asterisked green zone) and lost RNFL (asterisked yellow zone) within defect region increased. (B) If an RNFL defect progresses only by thinning without area enlargement, remaining RNFL within the defect region decreased as increased lost RNFL (asterisked yellow zone).

Table 4

 

Comparison of RNFL Defect Volume and RNFL Defect Volume Deviation Changes When RNFL Defects Progress

Factors Associated With Progression of RNFL Defects			Volumetric Parameters
(Changes When RNFL Defects Progress)			(Effects of RNFL Defect Progression)
Defect area	×	RNFL thickness*	=	RNFLD volume
(↑)		(↓)		(?)
(→)		(↓)		(↓)
(↑)		(→)		(↑)
Defect area	×	Estimated RNFL loss†	=	RNFLD volume deviation
(↑)		(↑)		(↑)
(→)		(↑)		(↑)
(↑)		(→)		(↑)

*  Averaged RNFL thickness within RNFL defects region.

†  Averaged subtraction between normative database and RNFL thickness within RNFL defect region.

The RNFL defect volume deviation analysis is different from RNFL deviation map in Cirrus HD-OCT software. Both RNFL deviation map and RNFL defect volume deviation analysis are results of interpretation from RNFL thickness map. Difference is what they recognize as an RNFL defect. The RNFL deviation map detects areas in which RNFL thickness is less than the lower 95% or 99% of the centile ranges. These criteria are based on statistical distribution. Because prevalence rate of glaucoma is 1.9% (2010, USA), the lower 95th or 99th percentile of RNFL measurement are high-risk groups for glaucoma. The RNFL defect volume deviation analysis identifies RNFL defects according to the structural loss. Because an RNFL defect is a result of axonal loss of retinal ganglion cells, it is more reasonable to define RNFL defect by how much RNFL loss is progressed rather than by where RNFL measurement is placed in statistical distribution. How much RNFL loss is required for clinical detection of RNFL defect in human eyes? Quigley and Addicks¹⁴ reported that clinical detection of RNFL defect was possible after a loss of 50% of the neural tissue in primate eyes. There is no available information in humans, so we designed a new method to detect RNFL defect according to structural loss.⁹ The most optimal reference level for glaucomatous RNFL defect identification was 36% to 37% loss of RNFL in this study, and 42% loss of RNFL in our previous study.⁹ 

In conclusion, RNFL defect volume deviation is a useful tool for diagnosing glaucoma and monitoring RNFL change. The RNFL defect volume deviation is more sensitive at detecting glaucoma than RNFL defect volume measurements in the early stage of RNFL loss. The lost portion of RNFL can directly reflect glaucomatous damage. 

Acknowledgments

Disclosure: J.W. Shin, None; K.B. Uhm, None; M. Seong, None