Discriminating Between Compressive Optic Neuropathy With Glaucoma-Like Cupping and Glaucomatous Optic Neuropathy Using OCT and OCTA

Abstract

Purpose: To discriminate between compressive optic neuropathy with glaucoma-like cupping (GL-CON) and glaucomatous optic neuropathy (GON) by comparing the peripapillary retinal nerve fiber layer (pRNFL) thickness and retinal microvasculature using optical coherence tomography (OCT) and optical coherence tomography angiography (OCTA).

Methods: In this retrospective cross-sectional study, OCT scans were performed on 28 eyes of GL-CON, 34 eyes of GON, and 41control eyes to determine the pRNFL thickness, ganglion cell complex thickness, and cup/disc ratio. OCTA scans were conducted for 12 eyes of GL-CON, 15 eyes of GON, and 15 control eyes to measure the vessel density of the peripapillary and macular areas. Analysis of covariance was used to perform the comparisons, and the area under the curve was calculated.

Results: The GON eyes had a significantly thinner pRNFL in the inferior quadrant and greater vertical cup/disc ratio than the GL-CON eyes. In the radial peripapillary capillary segment, the vessel density of the GON in the inferior sectors was significantly lower than in the GL-CON. The superficial macular vessel density in the whole-image, peritemporal, perinasal, and peri-inferior sectors was significantly smaller in the GON group than in the GL-CON group. The best parameter for discriminating between GL-CON and GON was the superficial macular vessel density in the peritemporal sector.

Conclusions: GL-CON eyes showed a characteristic pattern of pRNFL and retinal microvascular changes.

Translational Relevance: GL-CON can be effectively distinguished from GON by detecting the alterations in the pRNFL and retinal microvasculature using OCT and OCTA.

Introduction

Although the pathogenesis of glaucoma and compressive optic neuropathy (CON) differs, both lead to chronic optic nerve atrophy and visual field (VF) defects, resulting in misdiagnosis in clinical practice. Some studies have attempted to find useful clinical parameters to discriminate CON from glaucomatous optic neuropathy (GON).^1–5 Compared with GON patients, CON patients are younger and have worse visual acuity.¹ The VF defects in CON respect the vertical meridian, while horizontal arcuate scotomas typically present with GON.¹ In addition, the pallor of the neuroretinal rim is more common in nonglaucomatous cupping, while thinning of the neuroretinal rim is a typical sign for GON.^2,3 Moreover, the temporal and nasal peripapillary retinal nerve fiber layer (pRNFL) of CON is much thinner than that of GON, while the inferior and superior pRNFL are more vulnerable in GON.^4,5 

However, the clinical manifestation of some CON is not always as typical as pallor of the neuroretinal rim, temporal and nasal pRNFL thinning, and temporal hemianopia. In clinical practice, enlargement and excavation of the cup, which is widely recognized as a clinical hallmark of glaucoma, can also be found in CON.^6–9 Qu et al.¹⁰ previously detected 34 patients with glaucoma-like cupping from 501 patients with tumors in the sellar region of the brain. The prevalence of disc glaucoma is much higher in the CON group than in the control group (6.8% vs. 1.3%). Differentiating these CON with glaucoma-like cupping (GL-CON) from GON is challenging even for experienced ophthalmologists.⁶ Although the mechanism of cupping still needs to be clarified, this misdiagnosis would delay the treatment of brain tumors, resulting in irreversible vision loss and even life-threatening risk. 

Optical coherence tomography (OCT) provides detailed structural information about the optic nerve head, pRNFL, and ganglion cell complex (GCC) in vivo. More recently, optical coherence tomography angiography (OCTA), an advanced imaging technique, was developed to assess the peripapillary and macular microcirculation of retinal vascular diseases noninvasively.¹¹ Previous studies have shown significantly decreased retinal vessel density in the peripapillary and macular area of CON.^12,13 However, as far as we know, no study has evaluated the pRNFL and retinal microvasculature in GL-CON eyes. We hypothesized that GON and GL-CON had different patterns of pRNFL and retinal microvascular changes, which might provide clinicians with insight into the two conditions. 

The study aimed to quantitatively analyze the pRNFL, GCC, and vessel density of the peripapillary and macular areas in patients with GL-CON and to determine the effectiveness of OCT and OCTA parameters in distinguishing between GL-CON and GON. 

Methods

Participants

This retrospective study protocol was approved by the Medical Ethics Committee of Beijing Tiantan Hospital and followed the tenets of the Declaration of Helsinki. Patients with intrasellar or perisellar tumors with glaucomatous cupping, patients with GON, and healthy controls were enrolled in this study. 

Inclusion and Exclusion Criteria

All the GL-CON and GON patients underwent comprehensive ophthalmic examination, including assessment of visual acuity, intraocular pressure (IOP) measurement (Canon TX-20P; Canon, Tokyo, Japan), slit-lamp examination, fundus photography (Nonmyd WX, Kowa Inc., Tokyo, Japan, or Optos Daytona, Optos PLC, Dunfermline, UK), standard automated perimetry with the Humphrey Visual Field Analyzer using the 24-2 SITA fast strategy (Carl Zeiss Meditec, Inc., Dublin, CA, USA), and OCT and OCTA (software RTVue XR version 2017.1.0.151; Optovue Inc., Fremont, CA, USA). All examinations were performed within 1 or 2 days for each patient. 

Consecutive patients with intrasellar or perisellar tumors were referred to the ophthalmology department for preoperative ophthalmologic evaluation between March 2019 and August 2020. Magnetic resonance imaging (MRI) was performed in all patients, and the excised tumors were histologically examined after surgery. Patients in the CON group with an IOP >21 mm Hg were excluded. Using fundus photographs, glaucoma-like cupping was defined as the neuroretinal rim shape not consistent with the inferior-superior-nasal-temporal (ISNT) rule and a vertical cup/disc ratio of more than 0.64, which was the 97.5th percentile of the vertical cup/disc ratio for the normal population in the Beijing Eye Study.^10,14,15 GL-CON eyes were judged by two independent glaucoma specialists, who were blinded to the diagnosis and other clinical information of the patients. If there were any disagreements between the two specialists, the opinions of a third adjudicator were solicited. Only patients with chiasma impingement on MRI were included. 

Patients in the GON group had been diagnosed with open angle glaucoma (OAG). The criteria for defining OAG were an open filtration angle, glaucomatous damage of the optic nerve, and associated VF loss. A glaucomatous VF defect was defined as (1) three or more adjacent points with a probability <5% by pattern deviation, (2) a pattern standard deviation with a probability of less than 5%, or (3) a glaucoma hemifield test revealing outside normal limits. The OAG group included patients with normal-tension glaucoma (NTG) with an IOP <21 mm Hg and primary OAG (POAG) with an IOP ≥21 mm Hg before treatment. All the GON eyes were receiving IOP-lowering treatment, including trabeculectomy and IOP-lowering eye drops. 

The control group consisted of healthy persons recruited from hospital staff who had an IOP value <21 mm Hg and a normal optic disc appearance. 

The exclusion criteria of this study were as follows: recurrent tumor; high myopia with a spherical refraction of <−6.00 or hyperopia >+3.00 diopters; unreliable VF test results (false-positive errors >20% or false-negative errors >20%); patients with other eye diseases such as severe lens opacity, retinal disorders, and other optic neuropathy (e.g., optic disc drusen, optic disc edema, optic neuritis, ischemic optic neuropathy); or a history of brain surgery or intraocular surgery except for uneventful cataract extraction and glaucoma surgery. 

The three groups were matched for age and sex. One eye was randomly selected for analysis if both eyes were eligible for inclusion in this study. 

OCT

The pRNFL and GCC thicknesses were examined using OCT in all patients. The optic nerve head scan, which was calculated along a 3.45-mm diameter circle around the optic disc, measured the thickness of the pRNFL and automatically divided it into four sectors (superior, inferior, nasal, and temporal sectors). The GCC scan was centered on the fovea with a 6-mm diameter to measure GCC thickness, which was automatically divided into superior and inferior hemispheres. 

OCTA

The AngioVue (OptoVue Inc., Fremont, CA, USA) disc mode was chosen to assess the radial peripapillary capillary (RPC) in the papillary region, with a scan area of 4.5 × 4.5 mm. RPC was determined from the internal limiting membrane to the RNFL posterior boundary, and the peripapillary RPC (ppRPC) region was divided automatically into eight sectors: nasal superior, nasal inferior, inferior nasal (IN), inferior temporal (IT), temporal inferior, temporal superior, superior temporal, and superior nasal (Fig. 1A). Large vessels were excluded from the analysis. 

Figure 1.

 

Peripapillary (4.5 × 4.5 mm) and macular (6.0 × 6.0 mm) region was scanned for vessel density measurement. (A) The radial peripapillary capillary area was divided automatically into eight sectors: nasal superior (NS), nasal inferior (NI), inferior nasal (IN), inferior temporal (IT), temporal inferior (TI), temporal superior (TS), superior temporal (ST), and superior nasal (SN). (B) The software defined the parafoveal region as an annulus with an outer diameter of 3.0 mm and an inner diameter of 1.0 mm, and the perifovea region as an annulus with an outer diameter of 6.0 mm and inner diameter of 3.0 mm. Both regions were divided automatically into four subfields: parafovea superior (para-S), parafovea inferior (para-I), parafovea nasal (para-N), and parafovea temporal (para-T), as well as perifovea superior (peri-S), perifovea inferior (peri-I), perifovea nasal (peri-N), and perifovea temporal (peri-T).

The patients underwent a high-density Angio Retina (OptoVue Inc., Fremont, CA, USA) protocol (6 × 6 mm) to evaluate the macula. The vessel density of the macula was automatically analyzed in two different retinal vascular networks: the superficial retinal capillary plexus (SRCP) and deep retinal capillary plexus (DRCP). The software defined the parafovea region as an annulus with an outer diameter of 3.0 mm and an inner diameter of 1.0 mm, and the perifovea region as an annulus with an outer diameter of 6.0 mm and an inner diameter of 3.0 mm. Both regions were divided automatically into four subfields: superior, inferior, nasal, and temporal (Fig. 1B). The vessel density values of the entire image and each of the eight subfields were used in the analyses. 

Vessel density was regarded as the percentage area occupied by the microvasculature in the analyzed region.¹⁶ Images with a scan quality ≥6/10 were included. 

All GL-CONs (28 eyes of 28 patients) underwent OCT examination. Due to the limitations of this retrospective study, only 12 eyes of 12 GL-CON patients underwent OCTA scans. 

Statistical Analysis

The Statistical Package for Social Sciences (SPSS for Windows, version 26.0; IBM SPSS, Chicago, IL, USA) was used to perform all statistical analyses in the study. The distribution of the numerical variables was assessed using the Shapiro–Wilk test for normality, and the variance homogeneity of the data distributions was assessed using the Levene test. Categorical variables among the groups were analyzed using the chi-squared test. Differences in the mean deviation (MD) and pattern standard deviation (PSD) between the two groups were compared using an independent sample t-test. Analysis of covariance, adjusted for age and sex, was used for comparison of OCT and OCTA parameters among the three groups with Bonferroni correction for pairwise comparisons. Continuous variables are expressed as mean ± standard deviation (SD) values. The receiver operating characteristic (ROC) curve and area under the curve (AUC) were used to determine the ability to distinguish GL-CON from GON eyes. A value of P < 0.05 was considered statistically significant. 

Results

General Characteristics of the Patients

Fundus photographs of 467 patients with intrasellar or perisellar tumors were assessed. Based on fundus photographs, 35 of 467 patients were identified to have glaucoma-like cupping in at least one eye using the ISNT rule based on the neuroretinal rim configuration and vertical cup/disc ratio greater than 0.64. Glaucomatous cupping was detected in 15 of these patients in both eyes. Five patients were excluded because no evidence of direct chiasmal compression was found on the MRI. Two patients were excluded because of coexisting central retinal vein occlusion and age-related macular degeneration. Of the remaining 28 GL-CON patients, 19 were diagnosed with pituitary adenomas (67.0%), 5 with craniopharyngiomas (17.9%), 3 with meningiomas (10.7%), and 1 with epidermoid tumor (3.6%). Twenty-eight eyes of 28 patients were included in this study. The perimetric results were unreliable in two patients due to poor cooperation; therefore, VF examination could be evaluated in 26 patients. Thirty-four eyes of 34 patients with OAG (11 with NTG and 23 with POAG) and 41 eyes of 41 healthy controls were included. Of the 34 GON eyes, 6 eyes have received trabeculectomy, and the other eyes were treated with IOP-lowering eye drops. 

There were no significant differences between the groups in terms of age or sex. Visual field MD and PSD did not differ between the GL-CON and GON groups (P = 0.443 and P = 0.297, respectively) (Table 1). 

Table 1.

 

Clinical Characteristics of Participants

 

OCT Results

Comparisons of OCT parameters between the GL-CON, GON, and control groups are presented in Table 2. Significant difference in pRNFL was identified in the inferior quadrant between GL-CON and GON (100.5 ± 18.0 µm vs. 84.7 ± 19.9 µm, P < 0.05). The GON eyes also had a greater vertical cup/disc (C/D) ratio than the GL-CON eyes (0.86 ± 0.08 vs. 0.72 ± 0.15, P < 0.05). The relatively effective parameters for discriminating between GL-CON and GON were vertical C/D ratio (AUC = 0.816) and inferior pRNFL (AUC = 0.726). No significant differences were identified in the other OCT parameters between the GL-CON and GON eyes. The average and sector pRNFL and GCC values were significantly lower in the GL-CON and GON groups than in the control group (P < 0.001), while the vertical and horizontal C/D ratios and cup volume were significantly larger in the GL-CON and GON groups than in the control (P < 0.001). The relative reduction in pRNFL thickness was calculated for each quadrant in the GL-CON and CON groups (Fig. 2A). 

Table 2.

 

Comparison of Optical Coherence Tomography Parameters Between Groups and AUC of Each Parameter to Discriminate GL-CON From GON

 

Figure 2.

 

Comparison of percentage retinal nerve fiber layer thickness (RNFLT) (A) and RPC vessel density (B) in the GL-CON and GON groups for data normalized relative to the average in the healthy control group. Significant difference in RNFLT was identified in the inferior quadrant between GL-CON and GON (P = 0.005). Significant difference in vessel density of RPC was identified in the IN and IT sectors between GL-CON and GON (P = 0.007, P = 0.001, respectively). PP, peripapillary.

OCTA Results

Twelve of the 28 GL-CON patients underwent an OCTA scan. The OCTA parameters of the 12 eyes of 12 GL-CON patients were compared with those of 15 eyes of 15 GON patients and 15 eyes of 15 controls. No significant differences were found in age, sex, and MD among the three groups (P = 0.386, P = 0.323, P = 0.528, respectively) (detailed data are shown in the Supplementary Table). 

A comparison of vessel density between the CON and GON groups in the RPC segment showed that the vessel density of the GON in the IN and IT sectors was significantly smaller than that of the GL-CON group (34.2 ± 12.2 vs. 44.9 ± 8.6; 34.9 ± 14.9 vs. 50.2 ± 9.6, respectively), whereas there was no significant difference in the whole image, ppRPC, and the other sectors between these two groups. The vessel density of RPC in the IT and IN sectors was more effective in distinguishing between GL-CON and GON (AUC = 0.786 and 0.780, respectively). When compared to the control eyes, GL-CON eyes had smaller vessel density in all sectors except for IN and IT, while GON eyes had smaller vessel density in all RPC parameters (P < 0.05) (Table 3). The pRNFL thickness and vertical C/D ratio were compared between the groups (Table 3), and the results were similar to those of the OCT sample (Table 2). The relative reduction in mean vessel density values versus controls was calculated for each sector in the GL-CON and CON groups (Fig. 2B). 

Table 3.

 

Comparison of OCT and OCTA Parameters Among Groups and AUC of Each Parameter to Discriminate GL-CON From GON

 

The macular vessel density values of SRCP for each group are presented in Table 3. The vessel density in the whole image, perifovea temporal (peri-T), perifovea nasal (peri-N), and perifovea inferior (peri-I) were significantly thinner in the GON group than in the GL-CON group. The best parameter for discriminating between GL-CON and GON was the vessel density in peri-T (AUC = 0.860), followed by that in the whole image (AUC = 0.853) and peri-I (AUC = 0.847). Figure 3 shows the combined ROC curve of vessel density in the peri-T and inferior pRNFL for differentiating GL-CON from GON. There was no significant difference in any macular SRCP parameters except for peri-S between GL-CON and controls, while all SRCP parameters were smaller in the GON group than in the control group. The relative reduction rate of the mean sectorial values of vessel density in the SRCP map in the GL-CON and GON groups compared with that in the healthy controls is shown in Figure 4. No significant difference was observed among the three groups in terms of macular DRCP parameters. Figure 5 shows the representative cases of GL-CON, GON, and control eyes. 

Figure 3.

 

ROC of inferior pRNFL and vessel density in perifovea temporal in the differential diagnosis of GL-CON versus GON.

Figure 4.

 

Relative reduction of the mean sectorial values of vessel density in the macular superficial retinal capillary plexus in the GL-CON and GON groups compared with the healthy controls. *Significant differences (P < 0.05) when compared with the healthy controls. ^#Sectors with significant differences (P < 0.05) between the GL-CON and GON groups.

Figure 5.

 

Images from three right eyes of representative samples among groups (patient A: GL-CON; patient B: GON; control C). The top row shows diffused retinal nerve fiber layer thickness thinning with the inferior area relatively preserved in patient A (A1), and the same pattern of vessel density reduction was observed in radial peripapillary capillary segment (A2). The reduction of vessel density of the superficial retinal capillary plexus in patient B (B3) is more obvious than that in patient A (A3), while the difference between the two patients in the deep retinal capillary plexus is not apparent (A4 and B4). The optic disc photograph shows the disc of patient A (A5) is mimicking the disc of GON (B5). The visual field of patient A demonstrates atypical temporal hemianopia with some of the nasal part involved (A6), while a superior defect is revealed in the visual field of patient B (B6). C1 to C4 are the images of the control.

Discussion

This study provides evidence of different patterns of pRNFL and retinal microvascular changes in GL-CON and GON. To the best of our knowledge, this is the first study to provide a novel method to distinguish GL-CON from GON by using OCT and OCTA. The best parameters to identify the two conditions are the vertical C/D ratio and inferior pRNFL thickness in OCT, as well as superficial macular vessel density in peri-T, whole image, and peri-I in OCTA. 

Hata et al.¹⁷ had reported that in their study 44.1% of eyes with CON were classified as having glaucoma-like discs; thus, the prevalence of GL-CON was much higher than the results of Qu et al.¹⁰ (6.8%) and our study (7.5%), in which GL-CON was diagnosed through fundus photographs. Unfortunately, Hata et al.¹⁷ used Heidelberg retina tomography to define a glaucoma-like disc automatically, which was considered to have significant limitations in studying CON for the default set of reference planes.^4,18,19 It arbitrarily determines a reference plane to be 50 µm below the mean retinal surface in the temporal sector of the disc, where the papillomacular bundle is located.²⁰ In GON, the RNFL thickness of this location was assumed to be stable until the advanced stage. However, the temporal and nasal RNFL were preferentially lost in CON, which caused inadequate referencing. 

In the current study, the OCT results revealed that vertical and horizontal C/D ratios and cup volumes were significantly larger in GL-CON than in controls, which explains why GL-CON eyes are sometimes misdiagnosed as glaucoma in clinical practice. Compared with GON, the vertical C/D ratio of GL-CON was statistically smaller (0.72 vs. 0.86), which was in accordance with the results from Danesh-Meyer et al.⁴ (0.6 vs. 0.8), and reflected the underlying pathophysiology of the two diseases. The lamina cribrosa is believed to be the primary site of damage in glaucoma, and remodeling of the laminar cribrosa and progressive axonal loss lead to enlargement and excavation of the cup.²¹ In CON, tumor compresses the anterior visual pathway, causing retrograde degeneration of axons. Although it may be associated with the loss of axons and glial tissue, the mechanism of cupping in CON is unknown. 

In our OCT study, although diffused pRNFL thinning was identified in both GL-CON and GON, GON demonstrated proportionally more thinning inferiorly compared with GL-CON, whereas there was no significant difference in other quadrants between GL-CON and GON. However, this was somewhat different from previous studies on CON,^3,4 which suggested that the nasal and temporal pRNFL of CON was significantly thinner than that of glaucoma. We propose that the distinct pattern of pRNFL loss in GL-CON was associated with tumor size and location. Our previous study has proven that the larger the tumor was and the closer the tumor was located to the internal opening of the optic canal in the perisellar location, the higher was the likelihood for a glaucoma-like disc appearance.¹⁰ We hypothesized that the tumor may block the entry of cerebrospinal fluid (CSF) from the cranium into the orbit, thereby decreasing the retrolaminar CSF pressure and possibly making the optic head susceptible to glaucomatous damage.¹⁰ Great variation of VF damage had been found because the compression of tumor body and/or ischemia damaged the different positions of optic chiasma.^22–24 The cupping disc appearance of GL-CON might be explained by the tumor expansion process. Chiasmal compression initially affected the decussating nerve fibers subserving the nasal hemiretina, which were represented primarily at the nasal and temporal sectors of the disc. With the gradual development of the tumor, the supratemporal fibers of the retina, which are located on the medial side of the uncrossed fibers, were affected.²⁵ Therefore, thinning of the superior pRNFL appeared as the supratemporal fibers represented primarily at the superior pole of the disc. Finally, the inferior pRNFL was relatively preserved, which corresponded to the relative preservation of the upper nasal field. 

Similar to the distinct pattern of pRNFL change in the OCT study, GL-CON eyes exhibited a diffused decrease in peripapillary vessel density, with relative sparing of the inferior (IN and IT) sector in the OCTA study. In contrast to the preservation of inferior vessel density in GL-CON, the preferential damage site of GON was in the inferior sector.^26–28 The strong correlation between peripapillary microcirculation and pRNFL thickness in patients with chiasmal compression has been established in many previous studies.^29–31 Some researchers speculated that reduced retinal perfusion was attributed to the decreased metabolic demand of damaged axons.^30,32,33 Moreover, vessel density differences between the two groups also existed in the macular region. Our results showed that although no significant difference was found in the average GCC between GL-CON and GON, the influence of GL-CON on SRCP microcirculation seemed to be less severe than that of GON in the macular region. As we all know, glaucoma is a multifactorial disease. In addition to IOP as a risk factor, the reduction of ocular blood flow and decreased ocular perfusion may also be involved in the pathogenesis of glaucoma. In the DBA/2J mouse model of glaucoma, choroid and retinal blood flow was lower than in the control mice.³⁴ These changes observed in the animal model also existed in humans.^35,36 Many population-based studies have proven that lower ocular perfusion pressure was a risk factor for the occurrence and progression of glaucoma.^37–39 Jung et al.⁴⁰ found that NTG suspects with baseline microvasculature dropout (MvD), which indicated ocular vascular insufficiency, had a higher risk of converting to NTG. The choroidal MvD was also found in CON; however, the features and associated characteristics of MvD in CON were different from those in GON, which reflected the possibly diverse pathogenesis of peripapillary microvascular impairment.²⁹ The obvious decrease of retinal vessel density might not only be secondary to the death of neural axons but also involved in the primary vascular component. Therefore, the reduction in macular vessel density in the GON group was severer than that in the GL-CON group, although the optic disc and VF loss were equivocal. In addition, the difference between GL-CON and GON in the SRCP seemed more apparent in the perifovea area than in the parafovea area, which indicated that the 3 × 3-mm scan protocol adopted in some previous studies might not cover enough area to investigate the characteristics of macular microvasculature.³³ We did not observe statistical significance in the DRVP among the three groups, which is consistent with the findings of previous studies.^36,41 

With regard to the diagnostic capability for distinguishing GL-CON eyes from GON eyes, the results showed that the diagnostic ability of the vertical C/D ratio (AUC = 0.816) and inferior RNFL thickness (AUC = 0.726) was relatively better among the OCT parameters but less effective than the OCTA parameters in the macula. Peri-T vessel density in SRCP had the highest AUC (AUC = 0.860). This finding suggests that vascular parameters were better than structural parameters, for which there are several possible reasons. First, the “floor effect” of pRNFL thickness in the advanced stages of glaucoma seemed less pronounced in OCTA parameters^42–44; therefore, the diagnostic ability of vessel density might be better than that of pRNFL in later stages of glaucoma and CON. Second, microvascular sparseness occurred in some eyes with retinal ganglion cells dysfunction that had not yet atrophied, and reduction in pRNFL thickness might not yet be detectable.⁴⁵ Therefore, the present results indicate that OCTA could be a useful technique to differentiate GL-CON eyes from GON eyes. 

This study had several limitations. First, the sample size was relatively small, especially for the OCTA study. Although we reviewed 467 patients with sellar tumors, the prevalence of GL-CON was low. Future large-scale studies are needed to explore additional characteristics differentiating GL-CON and GON. Second, while OCTA is promising and does not suffer from floor effects, it has poorer repeatability than structural OCT, which renders the evaluation of the magnitude of vascular attenuation challenging.⁴⁶ In this study, however, the average difference in vessel density between GL-CON and GON in some sectors is well above the threshold of repeatability.⁴⁶ Third, in this study, although we excluded GL-CON without chiasmal impingement, some patients might have both glaucoma and a tumor compressing the optic chiasm, which is incidental in nature. In this case, it is difficult to distinguish which factors cause visual impairment; however, this is extremely rare in daily life. 

In conclusion, our study demonstrated distinct patterns of pRNFL changes and retinal microvascular damage in GL-CON. OCT and OCTA can be used as effective tools for distinguishing between GL-CON and GON in clinical practice. Larger sample size and prospective studies are needed to elucidate the mechanism of cupping in CON. 

Acknowledgments

Supported by grants from the National Natural Science Foundation of China (no. 82071312). The sponsors and funding organizations had no role in the design or conduct of this study. 

Disclosure: K. Lei, None; Y. Qu, None; Y. Tang, None; W. Lu, None; H. Zhao, None; M. Wang, None; L. Yang, None; X. Zhang, None