Factors Associated With Visual Field Progression in Cirrus Optical Coherence Tomography-guided Progression Analysis: A Topographic Approach

Transcription

1 ORIGINAL STUDY Factors Associated With Visual Field Progression in Cirrus Optical Coherence Tomography-guided Progression Analysis: A Topographic Approach Joong Won Shin, MD, Kyung Rim Sung, MD, PhD, Jiyun Lee, MD, and Junki Kwon, MD Purpose: To identify factors associated with visual field (VF) progression in optical coherence tomography (OCT)-guided progression analysis (GPA) using a topographic approach. Methods: Topographic components of OCT-GPA maps (Cirrus HD-OCT; Carl Zeiss Meditec) were classified according to location (temporal, superotemporal, superonasal, nasal, inferornasal, and inferotemporal), size (small, medium, and large), shape (wedge and irregular types), and pattern of retinal nerve fiber layer (RNFL) progression (widening, deepening, and new development). All positive findings in OCT-GPA (RNFL thickness maps, profiles, and average RNFL thickness) were defined as strong RNFL progression. VF progression was determined by linear regression analysis of VF mean deviation over time. Univariate and multivariate logistic regression analyses were performed to evaluate the association of VF progression with OCT-GPA parameters. Results: In total, 276 primary open-angle glaucoma patients were followed up for 5.1 years. According to OCT-GPA, 89 eyes were found to have RNFL progression. Of these, VF progression was detected in 27 (30.3%) eyes. Eyes with progressed VF group showed topographically different characteristics, which were wedge-shaped (80.6%), large-sized (40.0%), and widening (69.5%) or deepening (11.1%) pattern of RNFL progression in the inferotemporal (44.4%) and superotemporal (30.6%) regions, compared with eyes with non-progressed VF group. In multivariate analysis, strong RNFL progression and widening or deepening pattern of RNFL progression were significantly associated with VF progression (P = and 0.016, respectively). Conclusions: Topographic analysis in OCT-GPA maps showed different characteristics between progressed VF and nonprogressed VF groups. Glaucoma patients with widening or deepening pattern of RNFL progression should be carefully monitored for greater risk of VF progression. Key Words: glaucoma, OCT, GPA, progression, retinal nerve fiber layer (J Glaucoma 2017;26: ) Because glaucoma is an incurable and progressive disease, the detection of progression is of paramount importance for the clinical care of patients with glaucoma. Received for publication December 11, 2016; accepted March 20, From the Department of Ophthalmology, College of Medicine, Asan Medical Center, University of Ulsan, Seoul, Korea. Disclosure: The authors declare no conflict of interest. Reprints: Kyung Rim Sung, MD, PhD, Department of Ophthalmology, College of Medicine, Asan Medical Center, University of Ulsan, Pungnap-2-dong, Songpa-gu, Seoul , Korea ( sungeye@gmail.com). DOI: /IJG Glaucomatous changes involve both structural and functional aspects; however, disease progression may not develop concurrently with these changes. Early detection of structural or functional change is essential for the management of glaucoma. Recently, guided progression analysis (GPA) with serial retinal nerve fiber layer (RNFL) thickness maps has been shown to be an effective approach for detecting structural progression via spectral-domain optical coherence tomography (OCT). 1 3 Understanding the interpretation of OCT-GPA output and its application to clinical practice is important. However, findings from a previous study indicated that more than half of RNFL progression cases by OCT-GPA did not show corresponding visual field (VF) progression. 1 Hence, it is needed to evaluate the risk of VF progression when structural progression is detected in OCT-GPA. Previous studies have identified a number of risk factors for VF progression, including a relatively higher baseline intraocular pressure (IOP), higher peak IOP, wider IOP fluctuation, disc hemorrhage, b-zone peripapillary atrophy, thinner central corneal thickness (CCT), and worse baseline VF severity However, little is known about the risk factors for VF progression when structural progression is already observed in OCT-GPA. Because structural tests can detect significant glaucomatous damage before functional tests, 12 understanding the topographic characteristics of OCT-GPA maps may be helpful to evaluate the potential risk of VF progression. Hence, in the present study, we investigated the factors associated with VF progression in terms of topographic characteristics of OCT-GPA. METHODS Subjects We performed a retrospective review of the medical records of all subjects who had visited the glaucoma clinic of the Asan Medical Center, Seoul, Korea, from May 2008 to October All the subjects underwent a comprehensive ophthalmological examination, including a review of medical history, measurement of best-corrected visual acuity, slitlamp biomicroscopy, Goldmann applanation tonometry, gonioscopy, CCT assessment (DGH-550 instrument; DGH Technology Inc., Exton, PA), fundoscopic examination, stereoscopic optic disc photography, red-free photography, standard automated perimetry [Humphrey Field analyzer (HFA) with Swedish Interactive Threshold Algorithm (SITA) standard 24-2 test; Carl Zeiss Meditec, Dublin, CA], and RNFL imaging with a spectral-domain OCT (Cirrus HD-OCT; Carl Zeiss Meditec). For inclusion, all the participants had to meet the following criteria at the initial J Glaucoma Volume 26, Number 6, June

2 Shin et al J Glaucoma Volume 26, Number 6, June 2017 assessment: best-corrected visual acuity of 20/40 or better, spherical refractive error between 6.0 and D, cylinder correction within + 3 D, and a normal anterior chamber and open-angle on slit-lamp and gonioscopic examinations. Patients with glaucoma were identified on the basis of the presence of RNFL defects on red-free photography or the presence of glaucomatous optic disc changes on optic disc photography (such as enlarged cup-to-disc ratio >0.7, diffuse or focal neural rim thinning, or disc hemorrhage) and the presence of VF defects that corresponded with RNFL defects and optic disc changes. Patients with any ophthalmic or neurological disease known to affect the optic nerve head or VF were excluded. All subjects were followed up at 6-month intervals with stereoscopic optic disc photography, red-free photography, VF testing, and RNFL imaging with Cirrus HD-OCT. If surgical or laser treatment was performed during the study follow-up period, only data obtained in the period before treatment were analyzed. If both eyes were eligible, one eye was randomly selected for analysis. The institutional review board (IRB) of the Asan Medical Center approved the present study, informed consent was waived owing to the retrospective nature of the study, and the study design was executed in accordance with the principles of the Declaration of Helsinki. VF Assessment Only reliable VF test results (ie, false-positive errors <15%, false-negative errors <15%, and fixation loss <20%) were included in the present analysis. The first set of VF test data was excluded to obviate any learning effect. For inclusion in analysis, at least 6 reliable VF tests taken at separate visits were required. Trend-based VF progression was analyzed by performing linear regression between mean deviation (MD) and age (years). Statistically significant (P < 0.05) negative slope was defined as progression. SD-OCT Assessment An optic disc cube scan on the 66mm 2 optic disc area was obtained using Cirrus HD-OCT. The RNFL thicknesses were measured at pixels and displayed as RNFL thickness maps. A built-in algorithm automatically determined the center of the optic disc and then calculated a circumpapillary RNFL thickness at a 3.46 mm diameter circle. RNFL segmentation was checked for every OCT image by 2 observers (J.W.S. and K.R.S). To be included, at least 6 OCT examinations performed at different visits were required. All accepted images exhibited no motion artifacts and segmentation errors and had a signal strength of at least 7. The Cirrus OCT-GPA software determines RNFL progression by event or trend analysis in 3 strategies: RNFL thickness maps, RNFL thickness profiles, and average RNFL thickness. Event analysis is performed on RNFL thickness maps and RNFL thickness profiles. RNFL progression is classified as possible loss when the differences between the 2 baseline and follow-up measurements exceed the test-retest variability and as likely loss when the possible loss criterion is met on 2 successive visits. At least 20 adjacent pixels in RNFL thickness maps and 14 adjacent A-scans in RNFL thickness profiles are required to determine significant RNFL progression. Trend analysis is performed on an average RNFL thickness using linear regression between average RNFL thickness and age. In the present study, we defined RNFL progression as a significant negative slope detected in trend analysis and/or likely progression detected in event analysis. We defined the strength of RNFL progression based on whether all 3 strategies of OCT-GPA were satisfied. Positive findings from all 3 OCT-GPA strategies were defined as strong RNFL progression. Evaluation of Topographic Characteristics of OCT-GPA Maps RNFL thickness change maps display the RNFL progression of event-based analysis with serial RNFL thickness maps. The locations where differences between the 2 baseline and follow-up measurements exceed the testretest variability are coded in yellow. If the same changes are sustained over consecutive visits, the locations are coded in red. In the present study, topographic evaluation was performed on red-coded areas. Topographic characteristics of RNFL progression were classified by location, size, shape, and pattern of RNFL progression. The location of RNFL progression was categorized into temporal ( 30 to 30 degrees), superotemporal (30 to 90 degrees), superonasal (90 to 150 degrees), nasal (150 to 210 degrees), inferornasal (210 to 270 degrees), and inferotemporal (270 to 330 degrees) regions. If RNFL progression existed between 2 regions across the boundary, the location was determined by the larger portion. Angles were measured in a clockwise direction in right eyes and in a counterclockwise direction in left eyes, with the temporal equator set at 0 degree. The size of RNFL progression was automatically evaluated by a computer program written in MATLAB R2012a (The MathWorks Inc., Natick, MA) and measured in superpixel units composed of 44 pixels. The small, medium, and large size designations of RNFL progression were defined as having <50, 50 to 99, and Z100 red-coded superpixels on RNFL thickness change maps, respectively. The shape of RNFL progression was categorized into wedge and irregular types by 2 observers (J.W.S. and J.L.). If the opinions of the 2 observers differed, a third examiner (K.R.S.) made the final determination. The progression pattern of RNFL defects was classified as widening, deepening, or new development. 1,13 To determine the progression pattern, we compared the first baseline RNFL thickness deviation map and the final RNFL thickness change map. If multiple progressions were detected, the location, shape, and pattern of RNFL progressions were evaluated for each component. The size of RNFL progression was calculated as the sum of components in multiple progressions. Figure 1 shows 4 examples of classification by strength and topographic characteristics of RNFL progression. Statistical Analyses Statistical analyses were performed using commercial software (MedCalc, version , Mariakerke, Belgium; SPSS, version 18.0, SPSS Inc., Chicago, IL). Independent t tests were used to compare continuous variables between subjects with and without progression in VF or OCT-GPA. Categorical variables were compared using the w 2 test or the Fisher exact test. Univariate and multivariate logistic regression analyses were performed to evaluate the association of VF progression outcome with OCT-GPA parameters (strength and topographic components)

3 J Glaucoma Volume 26, Number 6, June 2017 Topographic Analysis of RNFL Progression Strength Strong Location IT Size 110 Shape Wedge Pattern Widening Strength Weak Location ST Size 28 Shape Wedge Pattern Deepening A B Strength Weak Location T Size 39 Shape Irregular Pattern New Strength Strong Location IT/ST Size 143 Shape Irregular/Wedge Pattern Widening/New C D FIGURE 1. Examples of classification according to strength and topographic characteristics of RNFL progression in OCT-GPA. In summary report of GPA, if RNFL progression was identified by all of RNFL thickness map, RNFL thickness profiles, and average RNFL thickness, the strength classified as strong (A and D), or if not, weak (B and C). A, Wedge-shaped RNFL progression with 110 red-coded superpixels was identified at inferotemporal region (270 to 330 degrees) in RNFL thickness change map. Compared with the blue line of RNFL defect from baseline deviation map, RNFL defect had been widened. B, Wedge-shaped RNFL progression with 28 red-coded superpixels was identified at superotemporal region (30 to 90 degrees). RNFL progression was detected inside previous RNFL defect (blue line) and classified as deepening pattern. C, Irregular-shaped RNFL progression with 39 red-coded superpixels was identified at temporal region ( 30 to 30 degrees). It showed newly developed RNFL progression apart from previous RNFL defect (blue line). D, If multiple progressions were detected, the location (inferotemporal and superotemporal), shape (irregular and wedge), and pattern (widening and new) of RNFL progression were evaluated for each component. The size of RNFL progression was calculated as the sum (143 red-coded superpixels) of all components in multiple progressions. GPA indicates guided progression analysis; IT, inferotemporal; OCT, optical coherence tomography; ONH, optic nerve head; RNFL, retinal nerve fiber layer; ST, superotemporal; T, temporal. RESULTS In total, 276 eyes of 276 primary open-angle glaucoma patients were included in the final analysis. Subject demographics are summarized in Table 1. The mean follow-up period was 5.1 ± 0.8 years (range, 3.6 to 7.3 y). The mean follow-up duration after the first detection of RNFL progression was 2.3 ± 0.7 years (range, 1.6 to 5.8 y). The mean baseline VF MD was 2.87 ± 4.58 db. Of all eyes examined, 231 (83.7%), 27 (9.8%), and 18 (6.5%) were at mild (MDZ 6 db), moderate ( 12rMD < 6 < db), and advanced (MD < 12 db) stages of glaucoma, respectively. According to the OCT-GPA RNFL progression criteria, 89 eyes (32.2%) were classified as progressed and 187 eyes (67.8%) were classified as nonprogressed. Of the 89 eyes with RNFL progression by OCT-GPA, 77 (86.5%) and 12 (13.5%) eyes were mild and moderate-to-advanced stage of glaucoma. The proportion of eyes with VF progression significantly differed between the progressed (27:62) and nonprogressed (15:172) RNFL groups according to OCT-GPA (P < 0.001). The baseline average RNFL thickness in the nonprogressed RNFL group was significantly thinner than that in the progressed RNFL group (P < 0.001, 79.5 ± 13.0 and 85.1 ± 11.4 mm, respectively). In the 15 eyes with VF progression without RNFL TABLE 1. Demographic and Clinical Characteristics of Glaucoma Patients on Long-term Follow-up With OCT-GPA Total OCT-GPA Progressed OCT-GPA Nonprogressed P N Age (y) 60.0 ± ± ± Sex (male:female) 141:135 43:46 98: Initial IOP (mm Hg) 16.0 ± ± ± F/U mean IOP 14.7 ± ± ± CCT ± ± ± Refractive error (D) 1.20 ± ± ± F/U duration 5.1 ± ± ± MD (db) 2.87 ± ± ± HFA-GPA (progressed:stable) 42:234 27:62 15:172 < Average RNFL thickness (mm) 81.3 ± ± ± 13.0 < CCT indicates central corneal thickness; F/U, follow-up; GPA, guided progression analysis; HFA, Humphrey field analyzer; IOP, intraocular pressure; MD, mean deviation; OCT, optical coherence tomography; RNFL, retinal nerve fiber layer

4 Shin et al J Glaucoma Volume 26, Number 6, June 2017 progression, the baseline average RNFL thickness was significantly lower than that of VF progression with RNFL progression (P < 0.001, 62.5 ± 8.8 and 84.7 ± 7.5 mm, respectively). OCT-GPA classified 89 eyes as having RNFL progression, including 60 eyes (67.4%, 60/89) according to RNFL thickness maps, 36 (40.4%, 36/89) according to RNFL thickness profiles, and 69 (77.5%, 69/89) according to average RNFL thickness. Of the eyes found to have RNFL progression in OCT-GPA, VF progression was detected in 27 (30.3%, 27/89) eyes; 23 (38.3%, 23/60) were within RNFL thickness map progression, 18 (50.0%, 18/36) were within RNFL thickness profile progression, and 22 (31.9%, 22/69) were within average RNFL thickness progression (Fig. 2A). All positive findings (or strong RNFL progression) from the 3 strategies of OCT-GPA were observed in 16 (59.3%, 16/27) eyes in the progressed VF (PV) group (Fig. 2A). In contrast, 12 (19.4%, 12/62) eyes were found to have strong RNFL progression in the nonprogressed VF (NPV) group (Fig. 2B). Topographic characteristics of RNFL thickness change maps were compared between PV and NPV groups (Table 2). In brief, topographic characteristics of PV and NPV on RNFL thickness change maps could be classified as follows: (1) inferotemporal or superotemporal versus other locations, (2) large versus small or medium size, (3) wedge versus irregular shape, and (4) widening or deepening versus newly developed pattern of RNFL progression. Locational distribution of RNFL progression was concentrated in the inferotemporal (44.4%) and superotemporal (30.6%) regions in the PV group. In contrast, the temporal region (43.3%) was the most common location of RNFL progression in the NPV group. The mean size of RNFL progression was 71 ± 64 superpixels (range, 20 to 311). Small-sized or medium-sized RNFL progression was observed in 91.5% of cases in the NPV group and 60.0% of cases in the PV group. Large-sized RNFL progression was observed in 8.5% of cases in the NPV group and 40.0% of cases in the PV group. Wedge-shaped progression was the predominant type (80.6%) in the PV group, whereas irregular-shaped progression was observed in 51.4% of cases in the NPV group. The widening pattern of RNFL progression was observed in 69.5% of cases in the PV group, whereas a newly developed pattern was observed in TABLE 2. Comparisons of Topographic Characteristics in RNFL Thickness Change Maps Between PV and NPV Groups RNFL Thickness Change Map [n (%)] PV Group NPV Group P* Locationw 36 (100) 37 (100) SN 1 (2.8) 1 (2.7) 0.006z ST 11 (30.6) 7 (18.9) T 5 (13.9) 16 (43.3) IT 16 (44.4) 8 (21.6) IN 3 (8.3) 3 (8.1) N 0 2 (5.4) Size (superpixels) 25 (100) 35 (100) < (40.0) 22 (62.9) 0.009y (20.0) 10 (28.6) Z (40.0) 3 (8.5) Shapew 36 (100) 37 (100) Wedge 29 (80.6) 18 (48.6) Irregular 7 (19.4) 19 (51.4) Patternw 36 (100) 37 (100) Widening 25 (69.5) 10 (27.0) < 0.001z Deepening 4 (11.1) 2 (5.4) New developed 7 (19.4) 25 (67.6) *w 2 test. win case of multiple RNFL progression, these parameters were evaluated separately in each progression. zst or IT vs SN, T, IN, or N. yz100 vs <100. 8Wedge versus irregular. zwidening or deepening versus new developed. IN indicates inferonasal; IT, inferotemporal; N, nasal; NPV, nonprogressed VF; PV, progressed VF; RNFL, retinal nerve fiber layer; SN, superonasal; ST, superotemporal; T, temporal; VF, visual field. 67.6% of cases in the NPV group. The deepening pattern (11.1% vs. 5.4%) was more frequently observed in the PV group than in the NPV group. Overall, significant differences in topographic characteristics were observed between the PV and NPV groups (P < 0.01 for all). There were no significant differences of topographic characteristics (location, shape, size, and pattern) between mild and moderateto-advanced group (all P > 0.05, the Fisher exact test). Logistic regression analysis was performed to determine the factors associated with VF progression (Table 3). In univariate analysis, OCT-GPA parameters were FIGURE 2. RNFL progression according to OCT-GPA strategies in the progressed and nonprogressed VF groups. In progressed VF group, exclusively positive findings (strong RNFL progression) for all 3 strategies of OCT-GPA were found in 16 eyes [59.3%, 16/27 (A)], whereas only 12 eyes [19.4%, 12/62 (B)] showed strong RNFL progression in nonprogressed VF group. GPA indicates guided progression analysis; OCT, optical coherence tomography; RNFL, retinal nerve fiber layer; VF, visual field

5 J Glaucoma Volume 26, Number 6, June 2017 Topographic Analysis of RNFL Progression TABLE 3. Logistic Regression Analysis to Determine the Factors Associated With VF Progression Univariate Multivariate OR (95% CI) P OR (95% CI) P Age (y) ( ) CCT ( ) Baseline IOP ( ) Baseline VF MD ( ) Average RNFL thickness ( ) OCT-GPA parameters Strength* ( ) ( ) Locationw ( ) ( ) Sizez ( ) ( ) Shapey ( ) ( ) Pattern ( ) ( ) *Strong versus weak. wst or IT versus SN, T, IN, or N. zz100 versus <100. ywedge versus irregular. 8Widening or deepening versus new developed. CCT indicates central corneal thickness; CI, confidence interval; GPA, guided progression analysis; IN, inferonasal; IOP, intraocular pressure; IT, inferotemporal; MD, mean deviation; N, nasal; OCT, optical coherence tomography; OR, odds ratio; RNFL, retinal nerve fiber layer; SN, superonasal; ST, superotemporal; T, temporal; VF, visual field. associated with VF progression (P < 0.05 for all), whereas age, CCT, baseline IOP, baseline VF MD, and average RNFL thickness were not associated with VF progression (P > 0.05 for all). In multivariate analysis, exclusively positive findings in all 3 strategies of OCT-GPA (strong RNFL progression) and the widening or deepening pattern of RNFL progression were significantly associated with VF progression (P = and 0.016, respectively). DISCUSSION Approximately 10 years have passed since spectraldomain OCT was first introduced to clinical practice. 14 Hence, it is now possible to longitudinally analyze the change in RNFL status using OCT-GPA. Recently, Yu et al 3 reported progressive RNFL thinning in OCT-GPA as a strong predictive factor for subsequent VF progression. In the present study, characteristic differences in topographic OCT-GPA parameters were observed between the PV and NPV groups. In addition, we found that the VF progression was related to the topographic characteristics of RNFL progression as well as presence of RNFL progression in OCT-GPA. These findings may provide valuable information for determining glaucoma progression in clinical practice. A widening or deepening pattern of RNFL progression was found to be significantly associated with VF progression. Widening or deepening of a preexisting RNFL defect accumulates glaucomatous damage in a specific region. This cumulative structural damage may facilitate immediate VF progression. In contrast, new development of RNFL defects was associated with a low incidence of concurrent VF progression. Considering spatial structurefunction relationships, 15 newly developed RNFL defects are not expected to have a cumulative effect on visual function but may be independently regarded as an early stage of glaucoma. Hood and Kardon 12 reported that RNFL thickness can indicate significant changes before VF loss during early glaucomatous damage. In the present study, the mean follow-up duration was 2.3 ± 0.7 years after the first detection of RNFL progression. Therefore, new RNFL defects may require considerable intervals before developing into detectable VF deterioration. The presence of positive findings from all 3 strategies of OCT-GPA (strong RNFL progression) was found to be significantly associated with VF progression. The OCT-GPA algorithm identifies RNFL progression according to eventbased or trend-based analyses. Event-based analysis (RNFL thickness maps and profiles) determines progressive RNFL thinning based on whether the differences between baseline and follow-up measurements exceed the test-retest variability. Trend-based analysis (average RNFL thickness) detects the presence of statistically significant change over time during the entire follow-up period using a linear regression approach. In addition, OCT-GPA evaluates changes in 66mm 2 regions of RNFL thickness maps and 3.46 mm diameter circles of RNFL thickness profiles as well as average RNFL thickness. Significant changes in both event and trend analyses may provide strong evidence of structural progression and may be associated with the rate of VF progression. Other topographic components such as location, size, and shape in OCT-GPA map were found to significantly differ between the PV and NPV groups. Large and wedgeshaped RNFL progression in superotemporal or inferotemporal regions were a characteristic of the PV group. If the pattern of RNFL progression has positive risk of VF progression and the strength of RNFL progression has negative risk of VF progression (ie, widening/deepening weak RNFL progression) or vice versa (ie, newly developed strong RNFL progression), the location, size, and shape of RNFL progression may have greater utility in determining the risk of VF progression. In the present study, corresponding VF deterioration was not observed in 66.3% of cases with RNFL progression (59/89). Although structural changes often precede functional VF loss early in glaucoma progression, 12,16 it is difficult to distinguish whether RNFL progression in OCT- GPA represents a structural change preceding VF deterioration or a false-positive result. Unfortunately, there is no available reference for differentiating a false-positive result in OCT-GPA. However, by identifying the topographic characteristics of RNFL progression in OCT-GPA, it may 559

6 Shin et al J Glaucoma Volume 26, Number 6, June 2017 be possible to obtain clues that facilitate the identification of false-positive outcomes. For example, patients with all the characteristics of the NPV group (ie, weak, small, irregular-shaped, and newly developed RNFL progression outside the superotemporal or inferotemporal region) may have a low likelihood of VF progression. In the present study, 9 subjects had all the characteristics of the NPV group (9/59, 15.3%). The false-positive rate of RNFL progression using OCT-GPA was estimated to be 15.3%. In contrast, VF progression without corresponding RNFL progression was detected in 15 eyes. These cases were moderate-to-advanced glaucoma (VF MD, ± 3.64 db; range, 8.03 to db) and showed significantly thinner baseline average RNFL thickness (62.5 ± 8.8 mm; range, 50 to 73 mm) compared with the cases of VF progression with corresponding RNFL progression (84.7 ± 7.5 mm; range, 72 to 99 mm). Mwanza et al 17 reported that progressive average RNFL thinning stops at 57.0 mm on Cirrus HD-OCT despite continuous VF loss. The limit of decrease in RNFL thickness is known to be 10 to 14 db of VF MD. 12,17 Given the floor effect of RNFL measurements, topographic analysis of OCT- GPA is appropriate during the early stage of glaucoma. The present study had several limitations. First, we assumed that structural damage preceded function progression. However, the results of previous studies indicate that functional decay develops earlier than structural progression in a proportion of cases. 18,19 Structural measures are known to be more sensitive during the early stages of glaucoma, whereas functional measures provide greater information during advanced stages. 17,20 In the present study, 6.5% of subjects had advanced-stage glaucoma, and structural assessment by OCT-GPA may have been less efficient in this group. Second, disc hemorrhage has been reported as a risk factor for VF progression. 8,21 In the present study, only a small number of eyes demonstrated disc hemorrhage during the follow-up period, and disc hemorrhage was therefore excluded from the present analysis. Lastly, all the patients included in the present study were receiving medical treatment during most of the followup period; hence, the results of the present study should be interpreted in this context. In conclusion, topographic analysis in OCT-GPA maps showed different characteristics between PV and NPV groups. The widening or deepening pattern and exclusively positive findings (indicating strong RNFL progression) in OCT-GPA were significantly associated with VF progression. Other topographic characteristics such as location, size, and shape of RNFL progression may have utility in determining the risk of VF progression when the pattern and strength of RNFL progression have different direction of risk of VF progression. These findings may facilitate improved interpretation of glaucoma progression using OCT-GPA. REFERENCES 1. Leung CK, Yu M, Weinreb RN, et al. Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography: patterns of retinal nerve fiber layer progression. Ophthalmology. 2012;119: Na JH, Sung KR, Baek S, et al. Progression of retinal nerve fiber layer thinning in glaucoma assessed by cirrus optical coherence tomography-guided progression analysis. Curr Eye Res. 2013;38: Yu M, Lin C, Weinreb RN, et al. Risk of visual field progression in glaucoma patients with progressive retinal nerve fiber layer thinning: a 5-year prospective study. Ophthalmology. 2016;123: Kim YJ, Yun SC, Na JH, et al. Glaucoma progression in eyes with a history of refractive corneal surgery. Invest Ophthalmol Vis Sci. 2012;53: Lee JY, Sung KR, Han S, et al. Effect of myopia on the progression of primary open-angle glaucoma. Invest Ophthalmol Vis Sci. 2015;56: Kostanyan T, Sung KR, Schuman JS, et al. Glaucoma structural and functional progression in American and Korean cohorts. Ophthalmology. 2016;123: Jonas JB, Martus P, Horn FK, et al. Predictive factors of the optic nerve head for development or progression of glaucomatous visual field loss. Invest Ophthalmol Vis Sci. 2004; 45: Leske MC, Heijl A, Hyman L, et al. Predictors of long-term progression in the early manifest glaucoma trial. Ophthalmology. 2007;114: Gordon MO, Beiser JA, Brandt JD, et al. The ocular hypertension treatment study: baseline factors that predict the onset of primary open-angle glaucoma. Arch Ophthalmol. 2002;120: ; discussion Musch DC, Gillespie BW, Lichter PR, et al. Visual field progression in the Collaborative Initial Glaucoma Treatment Study the impact of treatment and other baseline factors. Ophthalmology. 2009;116: Medeiros FA, Sample PA, Zangwill LM, et al. Corneal thickness as a risk factor for visual field loss in patients with preperimetric glaucomatous optic neuropathy. Am J Ophthalmol. 2003;136: Hood DC, Kardon RH. A framework for comparing structural and functional measures of glaucomatous damage. Prog Retin Eye Res. 2007;26: Suh MH, Kim DM, Kim YK, et al. Patterns of progression of localized retinal nerve fibre layer defect on red-free fundus photographs in normal-tension glaucoma. Eye. 2010;24: Adhi M, Duker JS. Optical coherence tomography current and future applications. Curr Opin Ophthalmol. 2013;24: Garway-Heath DF, Poinoosawmy D, Fitzke FW, et al. Mapping the visual field to the optic disc in normal tension glaucoma eyes. Ophthalmology. 2000;107: Tuulonen A, Lehtola J, Airaksinen PJ. Nerve fiber layer defects with normal visual fields. Do normal optic disc and normal visual field indicate absence of glaucomatous abnormality? Ophthalmology. 1993;100: ; discussion Mwanza JC, Kim HY, Budenz DL, et al. Residual and dynamic range of retinal nerve fiber layer thickness in glaucoma: comparison of three OCT platforms. Invest Ophthalmol Vis Sci. 2015;56: Miglior S, Zeyen T, Pfeiffer N, et al. Results of the European Glaucoma Prevention Study. Ophthalmology. 2005;112: Hood DC, Anderson SC, Wall M, et al. Structure versus function in glaucoma: an application of a linear model. Invest Ophthalmol Vis Sci. 2007;48: Wollstein G, Schuman JS, Price LL, et al. Optical coherence tomography longitudinal evaluation of retinal nerve fiber layer thickness in glaucoma. Arch Ophthalmol. 2005;123: Drance S, Anderson DR, Schulzer M. Collaborative Normal- Tension Glaucoma Study Group. Risk factors for progression of visual field abnormalities in normal-tension glaucoma. Am J Ophthalmol. 2001;131: