January 2024
Volume 65, Issue 1
Open Access
Glaucoma  |   January 2024
Comparison of the Circumpapillary Structure-Function and Vasculature-Function Relationships at Different Glaucoma Stages Using Longitudinal Data
Author Affiliations & Notes
  • Woo Keun Song
    Department of Ophthalmology, University of Ulsan, College of Medicine, Asan Medical Center, Seoul, Korea
  • Anna Lee
    Department of Ophthalmology, University of Ulsan, College of Medicine, Asan Medical Center, Seoul, Korea
  • Jooyoung Yoon
    Department of Ophthalmology, University of Ulsan, College of Medicine, Asan Medical Center, Seoul, Korea
  • Ko Eun Kim
    Department of Ophthalmology, University of Ulsan, College of Medicine, Asan Medical Center, Seoul, Korea
  • Michael S. Kook
    Department of Ophthalmology, University of Ulsan, College of Medicine, Asan Medical Center, Seoul, Korea
  • Correspondence: Michael S. Kook, Department of Ophthalmology, University of Ulsan, College of Medicine, Asan Medical Center, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Korea; [email protected]
Investigative Ophthalmology & Visual Science January 2024, Vol.65, 30. doi:https://doi.org/10.1167/iovs.65.1.30
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Woo Keun Song, Anna Lee, Jooyoung Yoon, Ko Eun Kim, Michael S. Kook; Comparison of the Circumpapillary Structure-Function and Vasculature-Function Relationships at Different Glaucoma Stages Using Longitudinal Data. Invest. Ophthalmol. Vis. Sci. 2024;65(1):30. https://doi.org/10.1167/iovs.65.1.30.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: This study investigated the global and regional correlations between longitudinal structure-function (S-F) and vasculature-function (V-F) data using circumpapillary retinal nerve fiber layer thickness (cpRNFLT) measurements from optical coherence tomography (OCT), circumpapillary vessel density (cpVD) from OCT angiography (OCTA), and the corresponding visual field mean sensitivities at different glaucoma stages.

Methods: A total of 107 eyes from 107 glaucoma patients with progressive visual field (VF) changes followed up for an average of 3.33 ± 1.39 years were enrolled, including early-to-moderate (51 eyes) and advanced (56 eyes) stages. The rates of longitudinal change in the VF mean deviation (MD), cpRNFLT, and cpVD were evaluated using linear mixed-effects models and compared between different glaucoma stages. Longitudinal global and regional S-F and V-F relationships were assessed by repeated measures correlation analysis by glaucoma stage.

Results: No significant differences were found in the rates of VF MD and cpVD changes (P > 0.05) between the two glaucoma stage groups. CpRNFLT decreased more rapidly in the early-to-moderate stage group (P < 0.001) in which significant longitudinal global and regional correlations were found in both S-F and V-F relationships (all P < 0.05), except for the nasal sector. Significant global and regional correlations were only found in V-F relationship in advanced stage cases (all P < 0.05).

Conclusions: Significant longitudinal V-F relationships exist globally and regionally regardless of glaucoma stage but no longitudinal S-F relationship is present in advanced glaucoma. Longitudinal follow-up of cpVD parameters may be useful for monitoring glaucomatous VF progression at all disease stages.

Glaucoma is defined by a chronic, progressive, and degenerative disorder of the optic nerve that produces characteristic visual field (VF) deterioration.1 Progressive retinal ganglion cell (RGC) and axonal loss results in progressive functional sensitivity decay with VF deficits in glaucoma.2 Glaucoma is not a cross-sectional disease but is a progressive disorder.1 Hence, detecting glaucomatous progression in a longitudinal manner is essential for preventing irreversible vision loss. 
Optical coherence tomography (OCT) has been widely used to detect and monitor glaucomatous structural changes, such as a thinning of structural parameters (e.g., the circumpapillary retinal nerve fiber layer [cpRNFL]). In addition, the recent introduction of OCT angiography (OCTA) has enabled clinicians to noninvasively assess the microvasculature perfusion status of the optic nerve head, retina, and choroid.3 Vessel density (VD) parameters measured by OCTA, including the circumpapillary VD (cpVD), are useful noninvasive and repeatable modalities for diagnosing glaucoma and monitoring glaucomatous progression.4,5 Several prior studies, including our own reports, have demonstrated the presence of significant correlations between the reduction in OCT/OCTA parameters and the loss of VF sensitivity (VFS) in glaucoma based on a cross-sectional study design.69 
The structure-function (S-F) relationships between OCT and VF parameters vary in glaucoma with its severity.10,11 In the early or pre-perimetric glaucoma stages, a thinning of OCT parameters such as the cpRNFL may precede a detectable VFS loss.12 In contrast, the OCT modality has difficulty in detecting glaucomatous progression in advanced glaucoma due to the measurement floor of the cpRNFL thickness (cpRNFLT).13,14 Notably, a significant vasculature-function (V-F) relationship, determined using VD parameters derived from OCTA measurements, has been well established in advanced glaucoma across several studies.7,8,15 Nonetheless, all of these prior studies were based on a cross-sectional design, indicating the need for longitudinal reports. As the measurement floor of VD parameters was reported to be lower than that of structural parameters across the glaucoma continuum,13,14 we hypothesized that the longitudinal S-F and V-F relationships may differ according to glaucoma stage. The purpose of this study therefore was to investigate the longitudinal global and regional S-F and V-F relationships at different stages of progressive glaucoma. 
Methods
Subjects
We performed a consecutive review of the medical records for open-angle glaucoma (OAG) patients who visited the glaucoma clinic of Asan Medical Center between January 2017 and July 2021. This retrospective, longitudinal cohort study was approved by the institutional review board of Asan Medical Center and was conducted in accordance with the tenets of the Declaration of Helsinki. Due to the retrospective study design, the institutional review board waived the requirement for informed consent from the study patients. 
OAG was defined as the presence of an open iridocorneal angle and glaucomatous optic nerve head (ONH) damage (e.g., generalized or focal neuroretinal rim loss, a vertical cup-to-disc ratio > 0.7, disc hemorrhage or an RNFL defect) with corresponding glaucomatous VF defects.16 Progressive VF loss was determined by either event- or trend-based analysis in each disease severity group. Based on the Early Manifest Glaucoma Trial criteria, VF progression was defined as a significant deterioration from the baseline pattern deviation at more than three of the same test locations evaluated on three consecutive examinations.17 Only “likely progression” was considered VF progression. Trend-based progressive VF loss was confirmed when the linear regression of the MD slope was negative with a P value < 0.05.18 Only eyes that satisfied progressive VF loss criterion according to either event- or trend-based analysis were included in the study. 
Our inclusion criteria for progressive glaucoma patients were as follows: (1) age > 18 years at baseline examination; (2) a spherical equivalent within −6 to 3 diopters (D), and cylinder refraction within −3 to 3 D; (3) OAG patients with a minimum follow-up of two years and at least five reliable serial Humphrey field analyzer Swedish Interactive Threshold Algorithm-Standard 24-2 VF tests (Carl Zeiss Meditec, Dublin, CA, USA); (4) OAG with manifested VF loss at baseline and progressive VF loss during follow-up according to event- or trend-based progression criteria; and (5) ≥ 3 spectral-domain (SD)-OCT and OCTA tests obtained at the same visit during follow-up (i.e., same number of SD-OCT and OCTA examinations per subject for analysis). To obviate any learning effects, the first perimetric result in eyes with glaucomatous VF defects was discarded from the analysis with repetition of VF testing conducted within two weeks, which was considered the first VF test in the series. Only reliable VF-test results (false-positive results: < 15%, false-negative results: < 15%, fixation loss: < 20%) for the baseline and follow-up visits were included in the analysis.16 
Patients with a history of ocular trauma, laser treatment, or intraocular surgery, including cataract extraction and glaucoma operation, as well as ocular/systemic disorders that could affect the ONH, macula, or VF, were excluded from the analysis. If both eyes of the glaucoma subjects met the inclusion criteria, one eye was randomly selected for the study. 
We divided the study patients based on the severity of baseline VF defects, early-to-moderate and advanced stage subgroups. A VF mean deviation (MD) of −15 dB at the initial presentation was used to classify the stage of eligible glaucoma subjects as follows: early-to-moderate stage, 24-2 VF MD > −15 dB; advanced stage, 24-2 VF MD ≤ −15 dB. The rationale for selecting an MD ≤ −15 dB to define advanced glaucoma was as follows. First, when the VF MD was worse than −14 dB in advanced stage glaucoma, the SD-OCT cpRNFLT experiences the measurement floor, after which no further structural changes can be detected using this modality.13,19 In addition, since there are no universally accepted criteria for the classification of advanced glaucoma, an MD ≤ −15 dB criterion was selected for this study based on previously published studies.9,20,21 
OCT and OCTA Imaging for Structure and Vasculature Parameters
SD-OCT images were acquired using Cirrus HD SD-OCT (Carl Zeiss Meditec; version 10.0) software. The average cpRNFLT was measured within a 3.46 mm diameter circle centered on the ONH. The regional cpRNFLT was also measured at six sectors in accordance with the Garway-Heath map i.e. temporal (T, 316°–45°), superotemporal (ST, 46°–90°), superonasal (SN, 91°–135°), nasal (N, 136°–225°), inferonasal (IN, 226°–270°), and inferotemporal (IT, 271°–315°) sectors (Figs. 1A1, 1B1).22 CpRNFLT measurements for each sector were calculated by integrating the clock-hour cpRNFLT from the Cirrus OCT (Figs. 1A1, 1B1).22 
Figure 1.
 
Circumpapillary regional structure and vasculature function correspondence map in the right eye. (A1) Corresponding SD-OCT cpRNFL thickness sectors with clock-hour division. (A2) Corresponding OCTA cpVD sectors. (A3) VF sectors on the VF pattern deviation map. (B1) OCT and (B2) OCTA en-face images with the cpRNFL and cpVD measurement sectors centered on the optic nerve head. The sectors on OCT and OCTA images that correspond to the regions in the VF map are identically colored.
Figure 1.
 
Circumpapillary regional structure and vasculature function correspondence map in the right eye. (A1) Corresponding SD-OCT cpRNFL thickness sectors with clock-hour division. (A2) Corresponding OCTA cpVD sectors. (A3) VF sectors on the VF pattern deviation map. (B1) OCT and (B2) OCTA en-face images with the cpRNFL and cpVD measurement sectors centered on the optic nerve head. The sectors on OCT and OCTA images that correspond to the regions in the VF map are identically colored.
All OCTA imaging of the circumpapillary regions was performed with an AngioVue OCTA device (Optovue Inc., Fremont, CA, USA) using version 8.0 to ensure data consistency among all study subjects. On these OCTA images, the VD was defined as the percentage area occupied by blood vessels displaying flow. After the automated removal of large retinal vessels, cpVD measurements were conducted using images of 4.5 × 4.5 mm² scans centered on the optic disc within the radial peripapillary capillary slab from the internal limiting membrane to the nerve fiber layer. The OCTA software automatically calculates the cpVD of the peripapillary regions, which are subdivided into eight 45° sectors (i.e. ST, SN, IT, IN, temporal upper [TU], temporal lower [TL], nasal upper [NU], and nasal lower [NL]). To calculate the average VD of six sectors according to the Garway-Heath map, the VDs of the temporal (T) and nasal (N) sectors were estimated by averaging the TU and TL for the T sector and NU and NL for the N sector (Figs. 1A2 and 1B2).15,22 Any initial or follow-up SD-OCT and OCTA images were excluded from the analysis if they met any of the following criteria: (1) poor image quality with a signal strength < 7; (2) motion artifacts; (3) localized weak signal intensity due to vitreous floaters or media opacity; (4) poor centration; or (5) segmentation failure. 
Mapping the Structure (cpRNFL) and the Vasculature (cpVD) to Function (VFS) Correlations
V-F and S-F relationships were defined as correlations between OCTA cpVD and SD-OCT cpRNFLT, and the corresponding VF mean sensitivity (VFMS) values obtained at the same visit, in accordance with the regionalization described by Garway-Heath et al.22 (Fig. 1A3). The global 24-2 VFMS was calculated as the average of threshold values obtained at 24-2 VF threshold map in 52 points of the SAP, except two points just above and below the blind spot. The regional VFMS values were also determined by averaging the threshold values in each of the hemifield and six sectors (i.e., SN, N, IN, IT, T, and ST) as described elsewhere.15,23,24 
Statistical Analyses
All statistical analyses were conducted using SPSS (version 26.0; SPSS Inc., Chicago, IL, USA), R (version 4.0.3; R Foundation for Statistical Computing, Boston, MA, USA), and SAS (version 9.4, SAS Institute Inc., Cary, NC, USA). The demographic and clinical characteristics of the study subjects were compared between the early-to-moderate stage and advanced stage groups using independent Student t-tests or Mann–Whitney U tests, depending on the normality of data distribution as determined by the Kolmogorov-Smirnov test. Categorical variables were compared between the groups using a Pearson's χ2 test. The rates of change in the global cpVD, cpRNFLT, and VF MD over time were calculated using a linear mixed-effects model. For each of the cpVD, cpRNFLT, and VF MD values, the linear mixed-effects model was a-priori fitted using the fixed effects with age, axial length (AXL), scan quality, follow-up duration, baseline intraocular pressure (IOP), mean follow-up IOP, and baseline VF MD, accepting random intercepts and coefficients at the individual levels when analyzing the effects of time. For each parameter, final equation was derived from the linear mixed-effects model. Locally-weighted scatterplot smoothing (LOWESS) curves, modeling technique that combines linear least squares regression with nonlinear regression, were used to graphically fit the relationships between the rates of change in cpRNFLT and cpVD and the baseline VF MD.25 
The correlations between the longitudinal global and regional S-F and V-F relationships were evaluated using repeated measures correlation analysis (R package "Rmcorr"), a statistical method that examines the relationship between two continuous variables with multiple repeated measurements of the same individuals over time.26 The key advantage of repeated measures correlation analysis is that it considers both within-subject variation and temporal changes, providing more accurate correlations that are robust to individual differences and temporal dynamics.26 P values < 0.05 were considered statistically significant. 
Results
Baseline Demographics and Clinical Characteristics
One hundred twenty-four OAG eyes from 124 patients who met the initial inclusion criteria were assessed. Of these, 17 eyes from 17 patients were excluded from the analyses as their SD-OCT, OCTA, or VF examinations did not fulfill our inclusion or reliability criteria for serial analyses. Consequently, a total of 107 eyes from 107 glaucoma patients with progressive VF loss and a mean follow-up of 3.33 ± 1.39 years were included in this study, comprising 51 early-to-moderate stage (VF MD > −15 dB) and 56 advanced stage (VF MD ≤ −15 dB) glaucomatous eyes. The demographics and clinical characteristics of the study patients are presented in Table 1. There were no significant differences between the two glaucoma stage groups in terms of the baseline characteristics such as age, sex, AXL, spherical equivalent, and central corneal thickness (CCT) (all P > 0.05). Patients with advanced glaucoma had a substantially higher baseline IOP (P = 0.002) and poorer logMAR best-corrected visual acuity (BCVA; P < 0.001), as well as a worse baseline and final MD (P < 0.001), visual field index (VFI; P < 0.001), cpRNFLT (baseline P < 0.001, final P = 0.011), and cpVD (P < 0.001). The VF MD at baseline for the total cohort and the early-to-moderate and advanced stage groups were −13.38 ± 6.47 dB, −7.73 ± 2.85 dB, and −19.03 ± 3.37dB, respectively (P < 0.001). 
Table 1.
 
Baseline Demographics and Clinical Characteristics of the Study Patients Stratified by Glaucoma Stage
Table 1.
 
Baseline Demographics and Clinical Characteristics of the Study Patients Stratified by Glaucoma Stage
Rates of Change in the Structure, Vasculature, and Functional Parameters
The average number of paired SD-OCT/OCTA and VF tests obtained at the same visit during follow-up in the entire cohort, early-to-moderate stage, and advanced stage groups were 4.72, 4.69, and 4.75, respectively, which did not differ statistically (Table 1; P = 0.639). Table 2 demonstrates the global rates of change and comparisons for the SD-OCT, OCTA, and VF-derived global parameters, calculated by linear mixed-effects models for early-to-moderate and advanced stage groups. There were significant reduction rates in VF MD, cpRNFLT, and cpVD in the entire and two glaucoma stage groups (all P < 0.05), except for the average cpRNFLT in the advanced stage group (P = 0.531). In terms of functional and vasculature parameters, there were no statistically significant differences in the reduction rates for the VF MD (median [range]) (−1.26 (−1.16 [−3.06 to 0.04]) dB/y vs. −1.31 (−1.09 [−3.71 to 0.35]) dB/y, P = 0.752) and cpVD (−1.19 (−1.16 [−2.22 to 0.56]) %/y vs. −1.20 (−1.23 [−1.92 to 0.23]) %/y, P = 0.981) between the two glaucoma stage groups. However, the average cpRNFLT decreased more rapidly in the early-to-moderate stage group than in the advanced stage group (−0.89 (−0.84 [−3.33 to 0.02]) µm/y vs. −0.05 (0.03 [−3.48 to 1.00]) µm/y, P < 0.001). Comprehensive statistical reports, including final equations, for each linear mixed-effects model are provided in the Supplementary Table S1 and Supplementary Figure S1. Figure 2 displays scatterplots and best-fit lines with LOWESS curves to visually illustrate the distribution patterns of cpRNFLT and cpVD rates of change in relation to the baseline VF MD. While the rates of change in cpRNFLT exhibited a bimodal pattern based on the baseline MD −15 dB criterion, those of cpVD demonstrated a unimodal pattern, regardless of the baseline MD values (i.e., glaucoma stage). 
Table 2.
 
Comparisons of the Rates of Change in Global VF MD, SD-OCT, and OCTA-Derived Parameters According to the Glaucoma Stage
Table 2.
 
Comparisons of the Rates of Change in Global VF MD, SD-OCT, and OCTA-Derived Parameters According to the Glaucoma Stage
Figure 2.
 
Scatterplots and best-fit lines with LOWESS curves showing the relationships between rates of global cpRNFLT thinning/cpVD loss and baseline VF MD. (A) Relationship between rate of global cpRNFLT change and baseline VF MD. (B) Relationship between rate of global cpVD change and baseline VF MD LOWESS curves were appended to each scatterplot in blue, and vertical red lines were represented to highlight the baseline VF MD −15 dB criteria.
Figure 2.
 
Scatterplots and best-fit lines with LOWESS curves showing the relationships between rates of global cpRNFLT thinning/cpVD loss and baseline VF MD. (A) Relationship between rate of global cpRNFLT change and baseline VF MD. (B) Relationship between rate of global cpVD change and baseline VF MD LOWESS curves were appended to each scatterplot in blue, and vertical red lines were represented to highlight the baseline VF MD −15 dB criteria.
Longitudinal Correlations Between the cpRNFLT, cpVD, and Corresponding VFMS Values
Correlations for the longitudinal global, hemifield, and regional S-F (cpRNFLT–VFMS) and V-F (cpVD–VFMS) relationships according to the glaucoma stage are presented in Table 3. Global, hemifield, and regional analyses revealed significant longitudinal correlations for both S-F and V-F relationships in the early-to-moderate stage group, except for the nasal sector (all P < 0.05, S-F nasal sector P = 0.335, V-F nasal sector P = 0.266, Table 3). Figure 3 illustrates the repeated measures correlation plots used to provide a graphic understanding of the longitudinal S-F and V-F correlations of each patient in the early-to-moderate stage group. 
Table 3.
 
Longitudinal Global and Regional Correlations Between cpRNFL Thinning/cpVD Loss and the Corresponding VFMS Deterioration (S-F and V-F Relationships) at Different Glaucoma Stages
Table 3.
 
Longitudinal Global and Regional Correlations Between cpRNFL Thinning/cpVD Loss and the Corresponding VFMS Deterioration (S-F and V-F Relationships) at Different Glaucoma Stages
Figure 3.
 
Repeated measures correlation plots representing the longitudinal global and regional correlations between cpRNFL thinning/cpVD loss and the corresponding VFMS loss of each patient in the early-to-moderate glaucoma stage group. (Left figure) Structure-function relationships (cpRNFLT–VFMS). (Right figure) Vasculature-function relationships (cpVD–VFMS). The data from each patient and corresponding regression line are visually distinguished by different colors. Correlation coefficient (r) values are based on repeated measures correlation analyses.
Figure 3.
 
Repeated measures correlation plots representing the longitudinal global and regional correlations between cpRNFL thinning/cpVD loss and the corresponding VFMS loss of each patient in the early-to-moderate glaucoma stage group. (Left figure) Structure-function relationships (cpRNFLT–VFMS). (Right figure) Vasculature-function relationships (cpVD–VFMS). The data from each patient and corresponding regression line are visually distinguished by different colors. Correlation coefficient (r) values are based on repeated measures correlation analyses.
In the advanced stage subjects, significant positive correlations were found in the global, hemifield, and every regional V-F relationship over time (all P < 0.05). In contrast, longitudinal S-F relationships revealed no significant associations in any of the global or regional analyses (all P > 0.05, Table 3). Figure 4 also depicts the longitudinal global and regional S-F and V-F relationships of each patient for the advanced stage group. 
Figure 4.
 
Repeated-measures correlation plots representing the longitudinal global and regional correlations between cpRNFL thinning/cpVD loss and the corresponding VFMS loss of each patient in the advanced glaucoma stage group. (Left figure) Structure-function relationships (cpRNFLT–VFMS). (Right figure) Vasculature-function relationships (cpVD–VFMS). The data from each patient and corresponding regression line are visually distinguished by different colors. Correlation coefficient (r) values are based on repeated-measures correlation analyses.
Figure 4.
 
Repeated-measures correlation plots representing the longitudinal global and regional correlations between cpRNFL thinning/cpVD loss and the corresponding VFMS loss of each patient in the advanced glaucoma stage group. (Left figure) Structure-function relationships (cpRNFLT–VFMS). (Right figure) Vasculature-function relationships (cpVD–VFMS). The data from each patient and corresponding regression line are visually distinguished by different colors. Correlation coefficient (r) values are based on repeated-measures correlation analyses.
Discussion
Our present study findings have demonstrated that significant longitudinal S-F and V-F relationships are present globally and regionally (except for the nasal sector) in early-to-moderate stage glaucoma. In advanced glaucoma, however, only the V-F relationship was found to be globally and regionally significant, based on our longitudinal data. These findings thus suggest that there is a potential use for cpVD parameters in detecting glaucomatous VF progression throughout the glaucoma continuum and highlight the importance of monitoring the cpVD rather than cpRNFLT in advanced stage OAG eyes. To our knowledge, this study is the first to evaluate the longitudinal circumpapillary S-F and V-F relationships in glaucoma patients with progressive VF loss according to glaucoma stage. 
No significant differences were found in the present analyses between the early-to-moderate and advanced stage groups in the rate of cpVD and VF MD change; however, there was a significant difference in the cpRNFLT rate of change between the groups (Table 2Fig. 2). Of note in this regard, the mean VF MD value was −7.73 dB for the early-to-moderate stage and −19.03 dB for the advanced stage group, which was below the cpRNFLT measurement floor. Thus, despite our present study patients having progressive VF loss, the rate of longitudinal cpRNFLT loss was minimal in the advanced stage cases (−0.05 µm/y) due to the floor effect. In contrast, the rates of cpVD change were comparable between the two groups: −1.19 %/y in the early-to-moderate group and −1.20 %/y in the advanced group. 
Shin et al.27 reported previously that VF progressors had a faster rate of change in the cpVD than non-progressors regardless of the glaucoma stage, whereas the rate of cpRNFLT thinning in progressors did not differ from non-progressors at a moderate-to-advanced stage of glaucoma. In our advanced patients, the cpVD showed considerable rate of longitudinal loss (−1.20%/y, baseline VF MD −19.03 dB), which is in agreement with those of previous studies by Lee et al.27 (−1.76 %/y, baseline VF MD −16.62 dB), and Shin et al.18 (−1.61 %/y, baseline VF MD −13.95 dB), despite differences in the baseline VF severity. While the underlying cause for the cpVD reaching a measurement threshold at a later stage of disease than cpRNFLT remains uncertain, this property underscores the importance of longitudinal evaluations of VD parameters for monitoring advanced glaucoma, as only the rate of change in cpVD showed a significant association with VF progression in the advanced glaucoma eyes with progressive VF loss in our current cohort. 
While most of the previous studies on S-F and V-F relationships were cross-sectional in design, longitudinal assessments may be critical for a better understanding of S-F and V-F relationships as glaucoma is a progressive disease. Regarding the longitudinal S-F relationship, Mohammadzadeh et al.28 reported that correlations between the macular OCT structure and central VF rates of change were weak to fair with a correlation coefficient (r) range of 0.24 to 0.41. For the longitudinal V-F relationship, another prior study has reported that the longitudinal rate of macular VD loss and ganglion cell complex thinning has significant correlations with central VF damage, with the V-F correlation (r = 0.42) being superior to the S-F correlation (r = 0.27).29 Nonetheless, these studies only evaluated the relationship between central VF sensitivities and macular structure/vasculature parameters in the relatively early stages of glaucoma. 
In the present study, the overall magnitude of global longitudinal S-F and V-F correlations were stronger than previously reported, with correlation coefficients of 0.690 in the S-F and 0.720 in the V-F in the early-to-moderate stage group (Table 3). Significant longitudinal S-F and V-F correlations were found not only for the global area, but also for the hemifield and regional analyses, except for the nasal sector, according to a Garway-Heath map. The reason for the higher correlation coefficient in our current study compared to previous analyses may be attributable to the following explanations. First, in contrast to prior studies,2830 our present study cohort comprised OAG patients with VF progression. Patients in our study population with progressive VF loss exhibited more rapid changes in structure, vasculature, and function parameters compared to the general glaucoma population, thereby making the correlations more pronounced. Second, our study used a more robust statistical method known as repeated measures correlation analysis, as opposed to the conventional approach of using simple correlation analysis. This method allowed us to precisely analyze the longitudinal associations between two variables over time. Repeated measures correlation analysis also provides more accurate results than simple correlation analysis in terms of the strength and direction of the relationship between variables by taking into account within-subject variability, temporal changes, and non-independence of the data.26,31 The absence of violation of the assumption of independence between observations and lack of the necessity for averaging or aggregation in addressing intra-individual research concerns contribute to the superior statistical power of this approach.26 This statistical method is now being used not only in ophthalmology but also in a variety of academic fields.32,33 Further studies are required to validate other longitudinal data using this technique. 
In the advanced group in the present study, no significant S-F correlation was observed, because the structural parameters had already reached the measurement floor (Table 3). Notably, however, a significant V-F correlation was observed across global, hemifield, and regional sectors. In contrast to structural parameters, vascular structures may be less susceptible to the floor effect in advanced stages, resulting in ongoing VD loss even beyond the measurement floor, as experienced by the cpRNFL.18,27,34 A recent study by Moghimi et al.13 supports this notion, indicating that the cpVD can continue to decrease even after reaching the measurement floor of the cpRNFLT, as the VF progresses at advanced stages of glaucoma. A plausible explanation for this phenomenon in eyes with advanced glaucoma is the persistence of vasculature adjacent to the impaired RGCs and their axons until a functional shutdown occurs. These findings highlight the clinical significance of the cpVD in terms of monitoring glaucoma progression and V-F correlation in advanced disease. Moreover, they underscore the necessity for further research to explore the pathophysiology of RGC damage and its relationship to vascular changes. 
Mansoori et al.35 reported that cpRNFLT measurements for nasal quadrant showed higher variability than superior and inferior quadrants in normal and glaucomatous eyes based on cross-sectional study design. This may partially explain our finding that a significant S-F or V-F relationship was absent in the nasal sectors of early-to-moderate glaucoma cases. Our finding is also in agreement with that of previous cross-sectional study by Shin et al.,7 despite two studies having a difference in the study design (longitudinal vs. cross-sectional). Nonetheless, the clinical implication of the stronger longitudinal association between cpVD and VFMS than between cpRNFLT and VFMS, at all sectors of ONH, including nasal and temporal sectors in advanced glaucoma, is that cpVD may be a better biomarker than cpRNFLT for monitoring of glaucoma in advanced cases. 
This study had several limitations of note. First, although we only used good-quality images for study inclusion, the impact of poor-quality or artifacts on OCT/OCTA images in real-world clinical settings should not be overlooked.36 Moreover, the high variability in VF examinations observed in advanced glaucoma may have influenced the S-F and V-F relationships.37 However, although these limitations may be pronounced in cross-sectional studies, they can be mitigated in longitudinal studies.36 Hence, longitudinal research conducted over extended periods, such as in the present study, remains crucial for accurate interpretation. Second, the definition of advanced glaucoma (MD < −15 dB) that we here used was somewhat arbitrary. While we carefully selected this criterion by considering the measurement floor of SD-OCT thickness parameters and the same threshold value used by previous reports,9,20,21 the definition of advanced glaucoma may be considered differently in relation to the location of VF defects or the application of different VF grading system. Third, while OCTA provides the automatic calculation of average VD values in the six sectors of circumpapillary region through built-in software, Cirrus SD-OCT, on the other hand, does not offer direct automatic calculation of six sectoral cpRNFLT based on the Garway-Heath map. Therefore sectoral cpRNFLT estimated manually using Cirrus SD-OCT clock-hour map may be less accurate compared to OCTA-based sectoral cpVD value. Fourth, six OCT/OCTA sectors based on the Garway-Heath map may not exhibit a complete topographical correspondence with VFMS sectors due to individual variations in macular anatomy, including variations in the foveal and cpRNFL distribution. For instance, the displacement of fovea-to-optic disc center in myopic eyes owing to optic disc tilt/rotation or axial elongation may render structure- and vasculature-function correlation maps imprecise due to deviations from the Garway-Heath map.22,38 Nonetheless, it is important to note that these limitations are consistently applied to the evaluation of both S-F and V-F relationships within each individual in the current study. Fifth, we did not assess the potential impact of systemic hypotensive medications or topical anti-glaucomatous eye treatments on VD measurements.39 Furthermore, modifications to these medications during follow-up periods were not taken into account. Sixth, trend analysis of global indices such as VF MD to define VF progression may underestimate the rate of progression in advanced glaucoma because of the influence of test points without detectable sensitivity (blind locations) and a floor effect.40 Our current findings should therefore be interpreted with caution in light of these confounding factors. Finally, as our study subjects were enrolled from a single tertiary academic center and were solely of Korean ethnicity, our findings may not be fully applicable more generally. 
In conclusion, we have here evaluated global and regional longitudinal S-F and V-F relationships in progressive glaucoma patients at different glaucoma stages with a statistical method used for calculating the longitudinal correlations of multiple repeated measurements of the same individuals over time. Significant longitudinal V-F relationships were observed globally and regionally in different stages of glaucoma, whereas longitudinal S-F relationships were not evident in the advanced stage. Longitudinal changes in VD parameters measured by OCTA may be useful biomarkers for monitoring glaucomatous VF progression at various glaucoma stages, including advanced disease. 
Acknowledgments
Disclosure: W.K. Song, None; A. Lee, None; J. Yoon, None; K.E. Kim, None; M.S. Kook, None 
References
Weinreb RN, Garway-Heath D, Leung C, Medeiros F, Liebmann J. Diagnosis of Primary Open Angle Glaucoma: WGA consensus series-10. vol. 10. Amsterdam, Netherlands: Kugler Publications; 2017.
Quigley HA, Dunkelberger GR, Green WR. Retinal ganglion cell atrophy correlated with automated perimetry in human eyes with glaucoma. Am J Ophthalmol. 1989; 107: 453–464. [CrossRef] [PubMed]
Jia Y, Wei E, Wang X, et al. Optical coherence tomography angiography of optic disc perfusion in glaucoma. Ophthalmology. 2014; 121: 1322–1332. [CrossRef] [PubMed]
Rao HL, Pradhan ZS, Weinreb RN, et al. Relationship of optic nerve structure and function to peripapillary vessel density measurements of optical coherence tomography angiography in glaucoma. J Glaucoma. 2017; 26: 548–554. [CrossRef] [PubMed]
Chung JK, Hwang YH, Wi JM, Kim M, Jung JJ. Glaucoma diagnostic ability of the optical coherence tomography angiography vessel density parameters. Curr Eye Res. 2017; 42: 1458–1467. [CrossRef] [PubMed]
Yarmohammadi A, Zangwill LM, Diniz-Filho A, et al. Relationship between optical coherence tomography angiography vessel density and severity of visual field loss in glaucoma. Ophthalmology. 2016; 123: 2498–2508. [CrossRef] [PubMed]
Shin JW, Lee J, Kwon J, Choi J, Kook MS. Regional vascular density–visual field sensitivity relationship in glaucoma according to disease severity. Br J Ophthalmol. 2017; 101: 1666–1672. [CrossRef] [PubMed]
Shin JW, Lee J, Kwon J, et al. Relationship between macular vessel density and central visual field sensitivity at different glaucoma stages. Br J Ophthalmol. 2019; 103: 1827–1833. [CrossRef] [PubMed]
Song WK, Kim KE, Yoon JY, Lee A, Kook MS. Association of macular structure, function, and vessel density with foveal threshold in advanced glaucoma. Sci Rep. 2022; 12(1): 19771. [CrossRef] [PubMed]
Rao HL, Zangwill LM, Weinreb RN, Leite MT, Sample PA, Medeiros FA. Structure-function relationship in glaucoma using spectral-domain optical coherence tomography. Arch Ophthalmol. 2011; 129: 864–871. [CrossRef] [PubMed]
Lee J-W, Morales E, Sharifipour F, et al. The relationship between central visual field sensitivity and macular ganglion cell/inner plexiform layer thickness in glaucoma. Br J Ophthalmol. 2017; 101: 1052–1058. [CrossRef] [PubMed]
Sommer A, Katz J, Quigley HA, et al. Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss. Arch Ophthalmol. 1991; 109: 77–83. [CrossRef] [PubMed]
Moghimi S, Bowd C, Zangwill LM, et al. Measurement floors and dynamic ranges of OCT and OCT angiography in glaucoma. Ophthalmology. 2019; 126: 980–988. [CrossRef] [PubMed]
Ghahari E, Bowd C, Zangwill LM, et al. Association of macular and circumpapillary microvasculature with visual field sensitivity in advanced glaucoma. Am J Ophthalmol. 2019; 204: 51–61. [CrossRef] [PubMed]
Lee A, Shin JW, Lee JY, Baek MS, Kook MS. Vasculature–function relationship in open-angle glaucomatous eyes with a choroidal microvasculature dropout. Sci Rep. 2022; 12(1): 19507. [CrossRef] [PubMed]
Anderson DR. Collaborative normal tension glaucoma study. Curr Opin Ophthalmol. 2003; 14(2): 86–90. [CrossRef] [PubMed]
Leske MC, Heijl A, Hyman L, Bengtsson B, Group EMGT. Early Manifest Glaucoma Trial: design and baseline data. Ophthalmology. 1999; 106: 2144–2153. [CrossRef] [PubMed]
Lee A, Sung KR, Shin JW. Progression detection capabilities of circumpapillary and macular vessel density in advanced glaucomatous eyes. Sci Rep. 2022; 12(1): 12109. [CrossRef] [PubMed]
Mwanza J-C, 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: 6344–6351. [CrossRef] [PubMed]
Jammal AA, Ferreira BG, Zangalli CS, et al. Evaluation of contrast sensitivity in patients with advanced glaucoma: comparison of two tests. Br J Ophthalmol. 2020; 104: 1418–1422. [CrossRef] [PubMed]
Araie M, Hori J, Koseki N. Comparison of visual field defects between normal-tension and primary open-angle glaucoma in the late stage of the disease. Graefes Arch Clin Exp Ophthalmol. 1995; 233: 610–616. [CrossRef] [PubMed]
Garway-Heath DF, Poinoosawmy D, Fitzke FW, Hitchings RA. Mapping the visual field to the optic disc in normal tension glaucoma eyes. Ophthalmology. 2000; 107: 1809–1815. [CrossRef] [PubMed]
Jung KI, Kang MK, Choi JA, Shin H-Y, Park CK. Structure–function relationship in glaucoma patients with parafoveal versus peripheral nasal scotoma. Invest Ophthalmol Vis Sci. 2016; 57: 420–428. [CrossRef] [PubMed]
Bozdogan H. Model selection and Akaike's information criterion (AIC): the general theory and its analytical extensions. Psychometrika. 1987; 52: 345–370. [CrossRef]
Cleveland WS, Devlin SJ. Locally weighted regression: an approach to regression analysis by local fitting. J Am Stat Assoc. 1988; 83(403): 596–610. [CrossRef]
Bakdash JZ, Marusich LR. Repeated measures correlation. Front Psychol. 2017; 8: 456. [CrossRef] [PubMed]
Shin JW, Song MK, Kook MS. Association between progressive retinal capillary density loss and visual field progression in open-angle glaucoma patients according to disease stage. Am J Ophthalmol. 2021; 226: 137–147. [CrossRef] [PubMed]
Mohammadzadeh V, Rabiolo A, Fu Q, et al. Longitudinal macular structure–function relationships in glaucoma. Ophthalmology. 2020; 127: 888–900. [CrossRef] [PubMed]
Mohammadzadeh V, Moghimi S, Nishida T, et al. Longitudinal structure–function relationship between macular vessel density and thickness and central visual field in early glaucoma. Ophthalmol Glaucoma. 2022; 5: 648–657. [CrossRef] [PubMed]
Suda K, Hangai M, Akagi T, et al. Comparison of longitudinal changes in functional and structural measures for evaluating progression of glaucomatous optic neuropathy. Invest Ophthalmol Vis Sci. 2015; 56: 5477–5484. [CrossRef] [PubMed]
Cheloni R, Dewsbery SD, Denniss J. A simple subjective evaluation of enface OCT reflectance images distinguishes glaucoma from healthy eyes. Transl Vis Sci Technol. 2021; 10(6): 31–31. [CrossRef] [PubMed]
Schartmüller D, Schwarzenbacher L, Meyer EL, Schriefl S, Leydolt C, Menapace R. Comparison of long-term rotational stability of three commonly implanted intraocular lenses. Am J Ophthalmol. 2020; 220: 72–81. [CrossRef] [PubMed]
Geropeppa M, Papadatou I, Sarantis P, et al. Receptor-binding-domain-specific B cell responses induced by mRNA immunization against SARS-CoV-2. Vaccines. 2023; 11: 1148. [CrossRef] [PubMed]
Holló G. Peripapillary capillary vessel density progression in advanced glaucoma: a case report. BMC Ophthalmol. 2019; 19: 1–5. [CrossRef] [PubMed]
Mansoori T, Viswanath K, Balakrishna N. Reproducibility of peripapillary retinal nerve fibre layer thickness measurements with spectral domain optical coherence tomography in normal and glaucomatous eyes. Br J Ophthalmol. 2011; 95: 685–688. [CrossRef] [PubMed]
Kamalipour A, Moghimi S, Hou H, et al. OCT angiography artifacts in glaucoma. Ophthalmology. 2021; 128: 1426–1437. [CrossRef] [PubMed]
Heijl A, Lindgren A, Lindgren G. Test-retest variability in glaucomatous visual fields. Am J Ophthalmol. 1989; 108: 130–135. [CrossRef] [PubMed]
Ferreras A, Pablo LSE, Garway-Heath DF, Fogagnolo P, Garcia-Feijoo J. Mapping standard automated perimetry to the peripapillary retinal nerve fiber layer in glaucoma. Invest Ophthalmol Vis Sci. 2008; 49: 3018–3025. [CrossRef] [PubMed]
Mayama C, Araie M. Effects of antiglaucoma drugs on blood flow of optic nerve heads and related structures. Jpn J Ophthalmol. 2013; 57: 133–149. [CrossRef] [PubMed]
Rao HL, Kumar AU, Babu JG, Senthil S, Garudadri CS. Relationship between severity of visual field loss at presentation and rate of visual field progression in glaucoma. Ophthalmology. 2011; 118: 249–253. [CrossRef] [PubMed]
Figure 1.
 
Circumpapillary regional structure and vasculature function correspondence map in the right eye. (A1) Corresponding SD-OCT cpRNFL thickness sectors with clock-hour division. (A2) Corresponding OCTA cpVD sectors. (A3) VF sectors on the VF pattern deviation map. (B1) OCT and (B2) OCTA en-face images with the cpRNFL and cpVD measurement sectors centered on the optic nerve head. The sectors on OCT and OCTA images that correspond to the regions in the VF map are identically colored.
Figure 1.
 
Circumpapillary regional structure and vasculature function correspondence map in the right eye. (A1) Corresponding SD-OCT cpRNFL thickness sectors with clock-hour division. (A2) Corresponding OCTA cpVD sectors. (A3) VF sectors on the VF pattern deviation map. (B1) OCT and (B2) OCTA en-face images with the cpRNFL and cpVD measurement sectors centered on the optic nerve head. The sectors on OCT and OCTA images that correspond to the regions in the VF map are identically colored.
Figure 2.
 
Scatterplots and best-fit lines with LOWESS curves showing the relationships between rates of global cpRNFLT thinning/cpVD loss and baseline VF MD. (A) Relationship between rate of global cpRNFLT change and baseline VF MD. (B) Relationship between rate of global cpVD change and baseline VF MD LOWESS curves were appended to each scatterplot in blue, and vertical red lines were represented to highlight the baseline VF MD −15 dB criteria.
Figure 2.
 
Scatterplots and best-fit lines with LOWESS curves showing the relationships between rates of global cpRNFLT thinning/cpVD loss and baseline VF MD. (A) Relationship between rate of global cpRNFLT change and baseline VF MD. (B) Relationship between rate of global cpVD change and baseline VF MD LOWESS curves were appended to each scatterplot in blue, and vertical red lines were represented to highlight the baseline VF MD −15 dB criteria.
Figure 3.
 
Repeated measures correlation plots representing the longitudinal global and regional correlations between cpRNFL thinning/cpVD loss and the corresponding VFMS loss of each patient in the early-to-moderate glaucoma stage group. (Left figure) Structure-function relationships (cpRNFLT–VFMS). (Right figure) Vasculature-function relationships (cpVD–VFMS). The data from each patient and corresponding regression line are visually distinguished by different colors. Correlation coefficient (r) values are based on repeated measures correlation analyses.
Figure 3.
 
Repeated measures correlation plots representing the longitudinal global and regional correlations between cpRNFL thinning/cpVD loss and the corresponding VFMS loss of each patient in the early-to-moderate glaucoma stage group. (Left figure) Structure-function relationships (cpRNFLT–VFMS). (Right figure) Vasculature-function relationships (cpVD–VFMS). The data from each patient and corresponding regression line are visually distinguished by different colors. Correlation coefficient (r) values are based on repeated measures correlation analyses.
Figure 4.
 
Repeated-measures correlation plots representing the longitudinal global and regional correlations between cpRNFL thinning/cpVD loss and the corresponding VFMS loss of each patient in the advanced glaucoma stage group. (Left figure) Structure-function relationships (cpRNFLT–VFMS). (Right figure) Vasculature-function relationships (cpVD–VFMS). The data from each patient and corresponding regression line are visually distinguished by different colors. Correlation coefficient (r) values are based on repeated-measures correlation analyses.
Figure 4.
 
Repeated-measures correlation plots representing the longitudinal global and regional correlations between cpRNFL thinning/cpVD loss and the corresponding VFMS loss of each patient in the advanced glaucoma stage group. (Left figure) Structure-function relationships (cpRNFLT–VFMS). (Right figure) Vasculature-function relationships (cpVD–VFMS). The data from each patient and corresponding regression line are visually distinguished by different colors. Correlation coefficient (r) values are based on repeated-measures correlation analyses.
Table 1.
 
Baseline Demographics and Clinical Characteristics of the Study Patients Stratified by Glaucoma Stage
Table 1.
 
Baseline Demographics and Clinical Characteristics of the Study Patients Stratified by Glaucoma Stage
Table 2.
 
Comparisons of the Rates of Change in Global VF MD, SD-OCT, and OCTA-Derived Parameters According to the Glaucoma Stage
Table 2.
 
Comparisons of the Rates of Change in Global VF MD, SD-OCT, and OCTA-Derived Parameters According to the Glaucoma Stage
Table 3.
 
Longitudinal Global and Regional Correlations Between cpRNFL Thinning/cpVD Loss and the Corresponding VFMS Deterioration (S-F and V-F Relationships) at Different Glaucoma Stages
Table 3.
 
Longitudinal Global and Regional Correlations Between cpRNFL Thinning/cpVD Loss and the Corresponding VFMS Deterioration (S-F and V-F Relationships) at Different Glaucoma Stages
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×