Choroidal Vascularity Index in Primary Open Angle Glaucoma

Objective: To study the choroidal vascularity index (CVI) and choroidal thickness changes in primary open-angle glaucoma (POAG) compared to healthy control eyes with the enhanced depth imaging (EDI) modality of an optical coherence tomography (OCT). Methods: In this cross-sectional observational study, the CVI and the subfoveal choroidal thickness were assessed with EDI OCT and compared among 31 POAG cases and 25 healthy controls, which were age- and sex-matched. EDI OCT images were binarized into luminal area (LA) and stromal area. The CVI was defined as proportion of the luminal area to the total circumscribed choroidal area. The correlation between the visual field mean deviation (MD) value and choroidal parameters of the POAG cases were also evaluated. Results: The eyes with POAG had lower mean CVI compared to the control eyes (p<0.00001) while the mean choroidal thickness values were mostly similar (p=0.94). The CVI values of the POAG eyes had significant correlation with the visual field MD values (r=0.51, p=0.01), while choroidal thickness values of the POAG eyes did not demonstrate such an association with the visual field MD values (r=0.14, p=0.51). Conclusion: The lower CVI in POAG eyes compared to healthy controls and its significant correlation with the severity of glaucoma imply the role of choroidal vascularity in POAG pathogenesis. We demonstrated a lower CVI despite mostly similar choroidal thickness values in POAG eyes compared with that of healthy controls. CVI may be more sensitive in representing the vascular impairment as it specifically analyses the vascular component, and the preservation of the choroidal stroma while its vascularity is being affected may limit the decrease in the choroidal thickness. The wide variation of the choroidal thickness and the various factors that affect it also may act as confounding factors. Furthermore, in the present study, subfoveal CVI had significant correlations with corresponding visual field MD values in the POAG eyes, which supports not only diagnostic but also prognostic implications of studying the choroidal vascularity status. Only a single study has reported reduced subfoveal CVI in eyes with open angle glaucoma compared with healthy eyes (24). Our results are consistent with this previous report. In addition, we demonstrated a correlation of the CVI with the visual field MD values, and hence with the disease severity. This implies the prognostic value of the choroidal vascularity status. On the other hand, the choroidal thickness values did not demonstrate any correlation with the visual field MD values in the same group. In conclusion, our study demonstrated lower mean CVI values in eyes with POAG versus that of healthy subjects. The results support the role of choroidal vascular insufficiency in the pathogenesis of POAG. Lower CVI values along with correlation of the CVI values with the visual field MD values in POAG eyes represents possible choroidal involvement in POAG, with an ischemic factor that could either be a trigger or effect of the disease. Conversely, similar choroidal thickness values in eyes with POAG compared with that of healthy controls imply the lack of both a correlation and any use of choroidal thickness as an imaging biomarker in POAG. interest No conflict of interest was declared by the authors.


INTRODUCTION
Glaucoma is characterized by progressive and irreversible degeneration of retinal ganglion cells resulting in structural alterations at the optic nerve head and an accompanying decrease in visual field sensitivity (1). Intraocular pressure (IOP) is the most important modifiable risk factor for glaucoma. Although the pathogenesis of glaucoma is not fully understood, there is growing evidence that vascular impairment has an important role in the development and the progression of the disease (2,3). Previous studies have demonstrated an association between glaucoma and impaired choroidal circulation (4,5).
The progress in ocular imaging technologies has provided clinicians with new methods to evaluate the structural and vascular characteristics of the choroid with greater detail. Particularly, the enhanced depth imaging (EDI) mode of the optical coherence tomography (OCT) has become a widely used method for imaging the choroid (6).
In terms of glaucoma, previous reports have demonstrated no change in the choroidal thickness in primary open-angle glaucoma (POAG) compared with healthy controls (7)(8)(9). However, measurement of choroidal thickness may not specify vascular damage, as the choroid is composed of both vasculature and stroma. Therefore, the measurement of choroidal thickness for evaluating the vascular status may be of limited value in glaucoma.
Choroidal vascularity index (CVI) has been introduced as a new method to assess choroidal vascular status (10). As a technique evaluating the vascularity of the choroid, CVI may carry a potential for increasing our understanding of the role of the choroidal vascular changes in glaucoma pathogenesis. Hence, the present study is aimed at investigating the subfoveal CVI in eyes with POAG versus that of healthy controls and evaluating the correlation between choroidal parameters and severity of glaucomatous damage.

Study participants
OCT images employed in this cross-sectional observational study were obtained from glaucoma patients and healthy subjects who agreed to participate after the nature of the tests was explained. All consecutive eligible subjects willing to participate were enrolled.
A complete medical history was taken from all the participants, all of whom underwent a comprehensive ophthalmological examination and testing including Goldmann applanation tonometry, gonioscopy, dilated fundus examination, retinal nerve fiber layer (RNFL) analysis on optical coherence tomography (OCT; Spectralis, Heidelberg Engineering GmbH, Heidelberg, Germany) and standard automated perimetry (Humphrey Visual Field Analyzer, 24-2 Swedish Interactive Threshold Algorithm (SITA) Standard strategy; Carl Zeiss Meditec Inc., Dublin, CA). Also, the central corneal thickness and axial length parameters were evaluated. One eye was randomly chosen if both eyes of a subject were eligible for the study.
The inclusion criteria for healthy subjects were a best corrected visual acuity of 20/30 or better, a spherical refractive error within +5 to -5 diopters (D), < 3 D of astigmatism, no media opacities that interfere with fundus imaging, normal clinical ocular findings with no evidence of retinal or optic nerve head pathologies, IOP of <21 mm Hg, open iridocorneal angles, intact RNFL, and normal visual field results, which was defined as a pattern standard deviation (PSD) within 95% confidence limits and a glaucoma hemifield test result within normal limits on a reliable 24-2 Swedish interactive thresholding algorithm.
The inclusion criteria for POAG was defined as a glaucomatous optic nerve appearance (neuroretinal rim thinning or notching), a gonioscopically open anterior chamber angle, RNFL thinning outside the 95% confidence interval of the normal distribution on OCT, and a glaucomatous visual field defect or an abnormal glaucoma hemifield test result that is outside the limits described above.
Subjects were excluded if they had a history of previous ocular trauma or posterior segment intraocular surgery, conditions other than glaucoma that may affect the optic nerve structure or visual field testing, any retinal disease that may affect macular thickness such as diabetic macular edema, epiretinal membrane or age-related macular degeneration, systemic diseases such as diabetes mellitus or hypertension, or a visually significant cataract with a corrected visual acuity of less than 20/40.

EDI OCT Image Acquisition and Choroidal Thickness and CVI Assessment
OCT imaging was performed with the Spectralis OCT (Heidelberg Engineering, Heidelberg, Germany) using the EDI mode. After mydriasis was attained and the patient appropriately positioned, the EDI-OCT scan of the foveal center was acquired.
The choroidal thickness was calculated from retinal pigment epithelium to the outer border of the choroid at the subfoveal area using the calipers provided in the software.
EDI OCT images were analyzed with the ImageJ software (Version 1.47, National Health Institute, Bethesda, MA) using the binarization protocol described by Agrawal and associates (10).After converting the image into 8 bit, the Niblack autolocal threshold tool was applied, which allows the demarcation of the choroidal stroma and its total vascular lumen. The binarized image was then transformed to the RGB format, and luminal areas were highlighted with the color thresholding tool ( Figure 1). The total choroidal, luminal and stromal areas were then calculated within the central 1500µm. The fraction of the luminal area to the total choroidal area yielded the CVI. All the CVI analysis and measurements were performed separately by two blinded investigators. (C) Each binarized image was then converted to an RGB image, and the luminal area was determined using the color thresholding tool.

Statistical Analysis
All statistical analyses were performed with the SPSS software (SPSS for Windows Version 20.0; SPSS Inc., Chica-go, IL). Chi-squared analysis test was used for the categorical variables while independent samples t-test was used to compare the mean age, IOP, axial length, central corneal thickness, choroidal thickness and CVI values of the 2 groups. A p value less than 0.05 was considered significant.
Original Investigation / Özgün Araştırma GMJ 2021; 32: 432-435 Hondur and Hondur Correlations between the choroidal thickness and the CVI values and the corresponding MD values of visual field were evaluated using Pearson regression analyses in glaucoma patients.

Ethical Approval
All investigations adhered to the tenets of the 1964 Declaration of Helsinki, and the study received approval from the Institutional Review Board and Ethics Committee (Ankara Education and Research Hospital Ethics Committee Approval Number=E-93471371-514.10). Informed consent was obtained from all the individual participants included in the study.

RESULTS
Thirty-one eyes with POAG, and 25 eyes of healthy controls were evaluated in this study. There was no intergroup difference in terms of age, axial length and central corneal thickness. All patients in the study were Caucasians. The eyes with POAG were under anti-glaucomatous treatment, and there was no significant difference in the IOP values of the two groups at the time of the study. The patient demographics and clinical characteristics are summarized in Table 1.

DISCUSSION
Various studies have reported that vascular factors and reduced ocular blood flow are associated with the pathogenesis of glaucoma (11)(12)(13)(14)(15)(16). According to the vascular theory, the pathogenesis of glaucoma has been linked to the vascular insufficiency at the optic nerve head, especially the prelaminar area that is supplied from the choroid (17). However, the role of the choroid in the pathogenesis of glaucomatous optic neuropathy is not clearly identified (18). In the present study we aimed at augmenting our understanding of the potential role of the choroidal vascular alterations in glaucoma pathogenesis.
The introduction of EDI OCT has enabled non-invasive, quantitative and detailed evaluation of the choroid (19). Choroidal thickness has been the first surrogate marker for choroidal structural alterations on OCT. However previous studies have not established a relationship between choroidal thickness and glaucoma (7-9), most probably resulting from the wide variation of the choroidal thickness and the various factors that affect the thickness. In addition, the choroidal thickness evaluation does not differentiate the changes between its vascular and stromal components. Hence, more recent studies have concentrated on evaluating the vascular and stromal components of the choroid separately (20)(21)(22). CVI assessment involves an image binarization technique to calculate the relative vascular component of the choroid (23). Agrawal and associates studied the variation in CVI in healthy eyes and presented that CVI was a superior marker for the assessment of choroidal disease than choroidal thickness (10).
In the present study, POAG eyes had lower subfoveal CVI values compared to age matched healthy control eyes. However subfoveal choroidal thickness values did not differ among POAG and control eyes, suggesting the lack of any relationship between the choroidal thickness and POAG. This finding was also in accordance with previous reports demonstrating no associations between choroidal thickness and POAG (7-9). GMJ 2021; 32: 432-435 Hondur and Hondur We demonstrated a lower CVI despite mostly similar choroidal thickness values in POAG eyes compared with that of healthy controls. CVI may be more sensitive in representing the vascular impairment as it specifically analyses the vascular component, and the preservation of the choroidal stroma while its vascularity is being affected may limit the decrease in the choroidal thickness. The wide variation of the choroidal thickness and the various factors that affect it also may act as confounding factors.
Furthermore, in the present study, subfoveal CVI had significant correlations with corresponding visual field MD values in the POAG eyes, which supports not only diagnostic but also prognostic implications of studying the choroidal vascularity status.
Only a single study has reported reduced subfoveal CVI in eyes with open angle glaucoma compared with healthy eyes (24). Our results are consistent with this previous report. In addition, we demonstrated a correlation of the CVI with the visual field MD values, and hence with the disease severity. This implies the prognostic value of the choroidal vascularity status. On the other hand, the choroidal thickness values did not demonstrate any correlation with the visual field MD values in the same group.
The major strength of the present study was the strict matching of the groups for age, axial length, IOP, and central corneal thickness. Age, axial length, CCT are significantly associated with changes in the choroidal thickness in glaucoma suspects and glaucoma patients (25). Contrariwise, CVI appeared to be a strong marker for gauging the vascular status of the choroid. In addition, it is known to be mostly unaffected or affected to a lesser degree by most physiological variables, such as age, refractive error, IOP, axial length, arterial blood pressure, and diurnally (26).
Though, the present study is not without limitations. Its cross-sectional design, which precluded the longitudinal follow-up of glaucoma patients and analysis of disease progression, may be considered as the major limitation. In addition, although not validated yet, different intraocular pressure-lowering drops may have different effects on the choroidal vascularity.
In conclusion, our study demonstrated lower mean CVI values in eyes with POAG versus that of healthy subjects. The results support the role of choroidal vascular insufficiency in the pathogenesis of POAG. Lower CVI values along with correlation of the CVI values with the visual field MD values in POAG eyes represents possible choroidal involvement in POAG, with an ischemic factor that could either be a trigger or effect of the disease. Conversely, similar choroidal thickness values in eyes with POAG compared with that of healthy controls imply the lack of both a correlation and any use of choroidal thickness as an imaging biomarker in POAG.