Choroidal Vascularity Index Alterations across the Spectrum of Pseudoexfoliation Disease and Healthy Eyes

Abstract

Purpose:

To compare the macular choroidal vascularity index (CVI) among patients diagnosed with pseudoexfoliative glaucoma (PXG), pseudoexfoliation syndrome (PEX), and healthy individuals.

Methods:

A cross-sectional analysis was conducted involving 77 eyes with PXG, 28 with PEX, and 74 from healthy controls. Enhanced depth imaging optical coherence tomography scans of the macula were analyzed using ImageJ software. The CVI was obtained by calculating the proportion of the luminal area within the total choroidal area. Generalized linear models were employed for statistical evaluation.

Results:

Subjects in the PXG and PEX groups were significantly older than those in the control group (P < 0.001), and sex distribution also differed significantly (P < 0.05). Eyes with PXG exhibited a notably lower CVI than normal eyes (0.62 ± 0.03 vs. 0.64 ± 0.04, P = 0.001). The difference between the PEX and control groups was not statistically significant (P = 0.146). After adjusting for age, sex, and axial length, the difference in CVI between PXG and controls did not reach statistical significance (P = 0.094).

Conclusions:

PXG may be associated with reduced CVI in the macular region, suggesting potential choroidal vascular alterations. This warrants further investigation into the vascular aspects of PXG pathophysiology.

INTRODUCTION

Glaucoma is the leading cause of irreversible blindness due to the gradual degeneration of retinal ganglion cells (RGCs), leading to glaucomatous optic neuropathy.1 Elevated intraocular pressure (IOP) is the main modifiable risk factor, although vascular dysregulation has been implicated in disease onset and progression.2,3 Conditions such as normal-tension glaucoma often coexist with systemic vascular disorders such as migraine and Raynaud’s phenomenon, indicating that impaired ocular blood flow may contribute to glaucomatous damage.

Pseudoexfoliative glaucoma (PXG), the most common form of secondary open-angle glaucoma in the elderly, is associated with the abnormal accumulation of extracellular fibrillar material within ocular structures.4,5 These deposits can also affect systemic organs, including the heart, lungs, kidneys, and vascular endothelium.6 In the eye, pseudoexfoliative material obstructs the trabecular meshwork, leading to elevated IOP. Despite clinical asymmetry, PXG usually involves both eyes to varying degrees.7

The choroidal vascularity index (CVI) is an optical coherence tomography (OCT)-based biomarker that provides a quantitative assessment of the choroidal vascular architecture by differentiating between the luminal (vascular) and stromal components through image binarization techniques.8,9 Unlike traditional measurements such as choroidal thickness (CT) – which reflect total choroidal dimensions without distinguishing vascular from stromal tissue and are influenced by several physiological factors, including age, axial length, hydration status, and diurnal variation – CVI offers a more direct and stable evaluation of the vascular component.10,11,12,13,14 Studies have shown that CVI exhibits lower inter- and intraobserver variability and greater reproducibility, making it a more robust metric for assessing choroidal vascular health.15

Although previous studies have examined CT and vessel diameter in PXG, limited attention has been given to the CVI in this context. Electron microscopy studies have demonstrated that pseudoexfoliative material can accumulate along the walls of posterior ciliary arteries and within vortex veins, potentially leading to vascular dysfunction and impaired choroidal perfusion.16,17 This suggests a pathophysiological link between pseudoexfoliation syndrome (PEX) and choroidal vascular compromise.

As CVI quantifies the proportion of the vascular (luminal) component within the choroid, it may be more sensitive than traditional thickness measurements in detecting subclinical vascular changes associated with PEX and PXG.18 Given these advantages, CVI was selected as the primary parameter for evaluating choroidal vascular alterations in the present study. This study aimed to assess differences in macular CVI among PXG, PEX, and healthy eyes to explore potential choroidal vascular involvement in the pathogenesis of PXG.

METHODS

This cross-sectional investigation enrolled patients from the Glaucoma Clinic at the Farabi Eye Hospital, following ethical guidelines outlined in the Declaration of Helsinki (2000 revision). The study protocol received approval from the local Institutional Review Board (IR.TUMS.FARABIH.REC.1402.034), and written informed consent was secured from all participants. The study cohort comprised individuals diagnosed with PEX, PXG, and healthy controls.

PEX was defined by the identification of pseudoexfoliative material on the anterior lens capsule or pupillary margin, IOP <21 mmHg, and absence of glaucomatous optic neuropathy. To be categorized as PXG, the deposition of pseudoexfoliative material alongside glaucomatous optic nerve damage (e.g., neuroretinal rim thinning, cup-to-disc ratio >0.7, or reproducible visual field defects) was required. In some cases, both eyes of a subject were included in the study; however, each eye was assigned to a different group based on its clinical diagnosis. Specifically, if one eye had PEX and the fellow eye had PXG, they were analyzed separately according to their respective classification. We enrolled participants with open anterior chamber angles, IOP <21 mmHg, normal optic discs, and no ocular pathology as the control group. All subjects were older than 18 years.

Eyes with corneal or posterior segment abnormalities, significant cataracts affecting image quality or measurement accuracy, zonular instability (phacodonesis/iridodonesis), prior trauma, or refractive errors exceeding ±5 diopters (D) were excluded.

All subjects underwent comprehensive ophthalmic assessments, including corrected distance visual acuity, slit-lamp biomicroscopy, IOP measurement via Goldmann applanation tonometry, gonioscopy using a Zeiss 4-mirror lens, and dilated fundus examination with 78D lenses.

Macular OCT was performed using Spectralis (Heidelberg Engineering, Heidelberg, Germany) in all cases. Transitioning from manual to automated segmentation and binarization methods allowed us to obtain definitive and repeatable measurements of choroidal parameters. These qualitative and quantitative parameters, termed choroidal imaging biomarkers, provided detailed information related to choroidal vascularity.19 Enhanced depth imaging-OCT advancements enabled the calculation of CVI, offering an assessment of the vascular status of the choroid. CVI is defined as the ratio of luminal area (LA) to total choroidal area (TCA) and is calculated using parameters obtained through image binarization methods.20 Images with quality scores lower than 20, uneven illumination, poor centration, or media opacities obscuring the choroid were excluded.

All CVI measurements were performed manually by a single experienced grader using ImageJ software, following a standardized protocol. The grader was masked to the clinical status and group allocation of the eyes to minimize potential bias during image analysis. To reduce measurement variability, the grader underwent thorough training and conducted all segmentations consistently under uniform conditions. Although intergrader reliability could not be assessed due to the single-grader design, careful attention was paid to ensure accuracy and consistency throughout the measurements.

1. Binarization: OCT images were processed in ImageJ (v1.53a, NIH) using the Niblack autothreshold method to delineate the choroidal–scleral junction [Figure 1a].21

2. TCA: A 1500 μm reference line parallel to the retinal pigment epithelium (RPE) was drawn, and the TCA was demarcated between the RPE and choroidal–scleral junction using the polygon tool [Figure 1b].20

3. LA: Dark pixels representing LA were isolated via the RGB color threshold tool [Figure 1c and d]. Pixel dimensions were calibrated to 200 μm.

4. CVI derivation: CVI was computed as (LA/TCA) ×100%22

Figure 1.

Calculation of the choroidal vascularity index. (a) The binarized image showing the choroidal scleral junction, (b) A reference line drawn to circumscribe the submacular choroidal area, (c) Submacular choroidal area circumscribed by the polygon tool, (d) Dark pixels represent the luminal area in the choroid

Statistical analysis

Data were described using frequency (percent), mean ± standard deviation, median, and range. To compare measurements of both eyes, a generalized estimating equation model, which is not sensitive to normal distribution, was used.

This model inherently handles unequal group sizes by applying weights based on group size. Adjustments were made for age, sex, and axial length. Furthermore, a post hoc power analysis was conducted, which showed that the sample size provided adequate power of approximately 80% to detect significant differences between groups despite the smaller PEX group size. P < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS software (version 17.0; SPSS Inc., Chicago, IL, USA).

RESULTS

We enrolled a total of 179 eyes from 103 participants. There were 77 eyes in the PXG group, 28 eyes in the PEX group, and 74 eyes in the control group. The PXG (73 ± 8 years) and PEX (71 ± 10 years) groups were significantly older compared to the control group (62 ± 11 years) (P < 0.001). In addition, there was a significantly higher proportion of male patients in the PXG and PEX groups (P = 0.004). The PXG group had the highest IOP values compared to the PEX and control groups (P < 0.001) [Table 1].

Table 1.

Demographic and clinical characteristics of the study eyes

Parameters	Group (mean±SD)				P*	Unadjusted						P †	Adjusted
						Pairwise comparison							Pairwise comparison
	PXG		PEX	Control		P1	Cohen’s d	P2	Cohen’s d	P3	Cohen’s d		P1	P2	P3
Age (years)	73±8		71±10	62±11	<0.001	0.692		<0.001		<0.001		-	-	-	-
Sex
Male	59 (76.6)		20 (71.4)	35 (47.3)	0.004	0.916		0.004		0.004		-	-	-	-
Female	18 (23.4)		8 (28.6)	39 (52.7)
CDVA (logMAR)	0.37±0.41		0.31±0.38	0.37±0.37	0.717	0.823		>0.999		0.0850		0.583
Axial length (mm)	23.08±0.77		22.77±0.69	23.57±1.19	0.005	0.062		0.124		0.008		-	-	-	-
IOP (mmHg)	21.29±9.43		15.61±3.75	14.54±2.08	<0.001	<0.001		<0.001		0.871		<0.001	<0.001	<0.001	0.862
Medications (%)	0	30 (39)	19 (67.9)	74 (100)	0.003
	1	12 (15.6)	2 (7.1)
	2	10 (13.0)	2 (7.1)
	3	16 (20.8)	4 (14.3)
	4	9 (11.7)	1 (3.6)
CDR	0.81±0.17		0.42±0.17	0.28±0.10	<0.001	<0.001		<0.001		<0.001		<0.001	<0.001	<0.001	0.004
CVI	0.62±0.03		0.62±0.03	0.64±0.04	0.001	0.07	0	0.001	0.57	0.146	0.53	0.035	0.141	0.094	0.584

*Based on generalized linear models, ^†Adjusted based on age, sex, and axial length. P1: PXG versus PEX, P2: PXG versus control, P3: PEX versus control, SD: Standard deviation, PXG: Pseudoexfoliation glaucoma, PEX: Pseudoexfoliation syndrome, CDVA: Corrected distance visual acuity, IOP: Intraocular pressure, CDR: Cup-to-disc ratio, CVI: Choroidal vascularity index

As illustrated in Table 1, the PXG eyes showed significantly lower CVI scores compared to the control group (0.62 ± 0.03 vs. 0.64 ± 0.04; P = 0.001), but the difference was not significant between the PEX and normal groups (0.62 ± 0.03 vs. 0.64 ± 0.04; P = 0.146). After adjustment for age, sex, and axial length, a trend toward significance was seen between the PXG and control groups (P = 0.094). In addition, the effect size (Cohen’s d) for the difference in CVI between the PXG and control groups was –0.57, indicating a moderate reduction in CVI in the PXG group.

DISCUSSION

In the current investigation, CVI was compared between PXG, PEX, and normal control eyes. We showed that PXG eyes may have lower CVI compared to normal control eyes.

Vascular pathology may play a significant role in eyes with PXG. Considering the accumulation of pseudoexfoliative material in vascular tissues, there may be an association between hemodynamic alterations and ischemic changes in PXG eyes.23 Electron microscopy studies have revealed the deposition of pseudoexfoliative material within the posterior ciliary arteries, central retinal artery, and vortex veins in both affected and unaffected eyes of patients with unilateral PEX.16

The impact of PEX on choroidal biomarkers such as CT has been quantitatively measured in the past. The CT varies in both healthy and pathological ocular conditions. Goktas et al. and Turan-Vural et al. found decreased CT measures in eyes with PEX compared to healthy eyes.24,25 Dursun et al. reported decreased CT in PXG patients in both the nasal-to-temporal macular region and all segments of the peripapillary area.26 Moghimi et al. demonstrated that a decrease in peripapillary CT is significantly related to the severity of PXG.10 However, Zengin et al. found no significant difference in mean CT between eyes with PEX and normal eyes.27 Ozge et al. investigated the choroidal and retinal nerve fiber layer thickness in eyes with PEX, PXG, and healthy eyes. Surprisingly, they found similar CT measurements with no significant difference among the groups.28 Based on the literature review, CT tends to decrease in PXG patients, whereas the measurements in PEX patients are controversial.

Sarrafpour et al. utilized spectral-domain OCT (SD-OCT) to evaluate choroidal vessel diameters in four age-matched cohorts: unilateral PEX, unilateral PXG, bilateral PEX, and bilateral PXG. Their findings demonstrated markedly smaller choroidal vessel diameters in affected eyes compared to unaffected contralateral eyes, irrespective of glaucoma status.11 However, the conclusions of the study were limited by the small sample size of unilateral PEX cases and the lack of a healthy control group. In contrast, Oruc et al. examined retinal vessel diameters in 40 patients with unilateral PEX using SD-OCT, reporting a significantly reduced arteriolar-to-venular diameter ratio in both affected and unaffected eyes relative to controls. The authors proposed that this observed vascular narrowing – even in clinically unaffected eyes – could serve as an early indicator of glaucomatous risk, potentially enabling preemptive intervention before overt PEX manifestations appear.29

Considering all these conflicting results, we interpreted that CT measurement alone is not a sufficient indicator of vascular status and does not fully reflect the status of the choroidal vasculature. Since both vascular and connective tissue are components of the choroid, CT provides limited information regarding choroidal vascularity. The CVI is a novel choroidal vascular marker that provides a more comprehensive assessment of choroidal vasculature compared to traditional thickness measurements. It offers a more accurate representation of the relationship between the luminal volume of choroidal vasculature and the total choroidal volume.2,22

The CVI provides a focused evaluation of the vascular compartment of the choroid without incorporating stromal components. Unlike CT, which quantifies the linear distance between the RPE and choroidal–scleral interface at discrete locations, CVI calculates the proportional LA within a defined choroidal region. CT measurements demonstrate significant interindividual variability, being influenced by multiple factors including age, refractive status, axial length, IOP, systemic blood pressure, and circadian fluctuations.9 In contrast, CVI offers greater consistency as a vascular biomarker, as it appears less susceptible to these physiological variables.8 Nevertheless, studies have identified reduced CVI values in tobacco users and diabetic individuals, likely reflecting pathological vascular constriction and impaired circulatory function.30,31

In the current study, we utilized an OCT-based image binarization method to compare macular choroidal vascular status in patients with PXG, PEX, and normals. In the glaucomatous eye group, a trend toward significance was found. Although the adjusted analysis yielded P = 0.094, which does not meet the conventional threshold for statistical significance, this trend toward significance may still hold biological relevance. It suggests a potential association that could reflect subtle pathophysiological changes related to PEX or glaucoma progression. In addition to statistical significance, the moderate effect size (Cohen’s d = –0.57) underscores the potential clinical relevance of the approximately 3% reduction in CVI observed in the PXG group. This suggests that the vascular alterations in PXG, while numerically modest, may reflect meaningful changes in choroidal structure that warrant further investigation. Such effects might require larger sample sizes or longitudinal follow-up to be detected with greater confidence. Therefore, while our findings should be interpreted with caution, they provide a basis for further research to explore these preliminary signals and their potential clinical implications. Future studies with more extensive cohorts and refined measurement techniques are warranted to clarify the nature and significance of this relationship. A decreased CVI in PXG eyes may lead to insufficient retinal and optic nerve circulation, resulting in glaucomatous ischemic injury. Besides, there were no significant differences between PEX and PXG eyes in terms of CVI ratio in the macular area. Simsek et al. analyzed CVI in 48 glaucomatous eyes and 48 nonglaucomatous contralateral eyes with no clinically observable pseudoexfoliation material. They showed that CVI was decreased in both the macula and peripapillary region in PXG eyes. Moreover, the CVI tended to decrease in nonglaucomatous fellow eyes of PXG patients.32 The presence of exfoliation material in the eye may directly impact choroidal vasculature function. Deposition of exfoliation material in the choroid may alter CVI by obliterating vessels through inflammatory responses. Further research is required to fully understand the association between PEX and choroidal vasculature health.

Like many observational studies, this work has several limitations that should be considered when interpreting the findings. First, the sample size of the PEX group was relatively small, which may have reduced the statistical power for comparisons involving this group. Nevertheless, the use of robust statistical methods mitigates this issue to some extent. Future studies with larger and more balanced cohorts are warranted to confirm our findings. Second, significant differences in age and sex distribution were present among the study groups. Although we adjusted for these variables in the analysis, residual confounding may persist. CVI typically decreases with age, while the effect of sex on CVI remains controversial.33,34 These demographic differences may have introduced bias, highlighting the need for cautious interpretation and the importance of matched group designs in future studies. Third, the cross-sectional nature of the study precludes any causal inferences. While our results suggest associations between CVI and pseudoexfoliation-related disease states, longitudinal studies are needed to determine whether choroidal changes contribute to disease progression or occur as a consequence of it. Fourth, systemic conditions, including blood pressure status and medication use – particularly antihypertensive therapy – were not assessed. Such systemic factors may significantly influence ocular perfusion and confound CVI measurements. The absence of concurrent blood pressure data and detailed medication history, therefore, limits our ability to fully account for systemic vascular influences. Fifth, technical limitations in CVI measurement should be acknowledged. Although we included only high-quality OCT scans, minor inconsistencies in image quality may still have influenced the results. All manual segmentations were performed by a single experienced grader using a standardized protocol. Intragrader reproducibility was not formally assessed, and intergrader reliability could not be evaluated because only one grader was involved. This approach was chosen to maintain consistency and minimize measurement variability, while resource and time constraints precluded the involvement of additional graders. Future studies could benefit from automated or semiautomated segmentation methods or multiple graders to enhance reproducibility and reduce subjectivity.

In conclusion, our study observed a trend toward reduced macular CVI in eyes with PXG compared to healthy controls, suggesting that CVI may reflect vascular alterations involved in PXG pathogenesis. However, further longitudinal and larger-scale studies are needed to evaluate CVI changes over time and to clarify its potential as an adjunctive structural biomarker for diagnosis and disease monitoring. Investigating the relationship between CVI and functional or structural glaucoma progression may provide valuable insights into its clinical utility for predicting outcomes and guiding management strategies.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.