ANALYSIS OF STRUCTURAL AND VASCULAR CHANGES OF THE OPTIC NERVE HEAD AND MACULA IN DIFFERENT STAGES OF PRIMARY OPEN ANGLE GLAUCOMA

Aim. This study evaluates the structural and vascular changes of the optic nerve head (ONH) and macula in healthy and primary open-angle glaucoma (POAG) eyes, detected by optical coherence tomography (OCT) and optical coherence tomography angiography (OCTA). Additionally, it examines the correlation of the OCT and OCTA measurements and their association with the presence of POAG. Methods. A total of 196 eyes were included and classified into four groups. Forty-eight healthy eyes, 51 eyes with mild POAG, 50 eyes with moderate POAG, and 47 eyes with advanced glaucoma. All subjects underwent standard ophthalmic examination. OCT measured the mean, superior and inferior retinal nerve fiber layer (RNFL) thickness and macular ganglion cell complex (GCC). OCTA evaluated the vessel capillary density (VCD) in ONH, foveal avascular zone (FAZ) and macular vessel density (VD) in the superficial (SL) and deep (DL) retinal vascular plexus. Results. Patient characteristics were similar except for decreased visual acuity, thinner corneas, higher IOP and higher the cup/disc ratio in POAG patients. OCT results showed that RNFL and GCC thickness gradually decreased according to POAG severity. Within the assessment conducted by OCTA, VCD’s value in ONH also diminished with the progression of POAG, having the lowest value in patients with advanced glaucoma. A same pattern was observed in vessel density around FAZ (FD) and VD values. Comparing the structural and vascular changes, a significant positive correlation was found between RNFL thickness and VCD inside disc (ID) in ONH, and GCC and VD SL in the macular zone. Conclusion. OCT and OCTA allow for a non-invasive quantification of the structural and vascular changes in ONH and the macular zone and accurately distinguish between healthy eyes and eyes with POAG, showing an association with the presence and progression of glaucoma.


Introduction
Glaucoma represents the second leading cause of blindness worldwide. 1 It is a progressive optic neuropathy leading to preventable but irreversible visual defects. 2 It remains unclear whether the decreased retinal blood flow is the cause or result of an optic nerve glaucomatous damage. 3 However, there is growing evidence that vascular hypothesis, supported by reduced retinal blood flow and vessel diameter seen in glaucoma patients, can depict the pathogenesis of glaucoma. 4,5 Recently, scholars have suggested that the vascular dysfunction in the optic nerve head (ONH) may have a crucial role in the development and progression of the open angle glaucoma (OAG). 6 Considering the possibility for patients with glaucoma to be asymptomatic in the early stages, there is a serious risk for them to remain undiagnosed until the symptoms occur. 4 We hypothesize, it is important to visualize retinal microcirculation in order to confirm the diagnose and treat patients accordingly. The optical coherence tomography angiography (OCTA) was introduced as a reliable, noninvasive, rapid diagnostic procedure providing assessment of perfusion separately in various retinal layers, as well as in the optic nerve head. 7 Recent studies showed a reduced papillary vessel density in glaucomatous eyes 8 and suggested a possible correlation between the diameter of the retinal arterioles and the optic nerve damage. 5 Additionally, measuring the macular perfusion has the potential for detecting a reduced metabolic rate in dysfunctional retinal ganglion cells before they undergo apoptosis and cause the ganglion cell complex (GCC) to become thinner. 9 Structural changes in glaucoma patients can be detected using optical coherence tomography (OCT) as one of the non-invasive imaging technologies. OCT is designed to assess morphology and thickness of retinal layers, such as the innermost layers of the retina, which include the retinal nerve fiber layer, ganglion cell layer, and inner plexiform layer. Diagnostic accuracy for glaucoma can be improved when macular OCT measurements focus particularly on the GCC 9 , since it provides detailed information of retinal structure. 10 Both OCT and OCTA differentiate between healthy and glaucomatous eyes. 11  All OCT and OCTA examinations were performed at Optovue apparatus, AngioVue Comprehensive Imaging system, using two patented technologies: Split Spectrum Amplitude Decorrelation Angiography (SSADA protocol) and Motion Correction Technology (MTC) for reduction of artifacts by using software volume-based projection artifact removal (3D PAR). The vessel density measurements are determined with PAR correction only. The SSADA method is used to compare the consecutive B scans at the same location to capture the dynamic motion of the red blood cells using motion contrast.
A trained examiner reviewed all the OCTA scans and only the images with good clarity, a signal strength index (SSI) of more than 50, with no residual motion were included for the analysis. Image cropping or local weak signal resulting from vitreous opacity or segmentation errors that could not be corrected were rejected.
The OCT measurement parameters in ONH were RNFL defined as the thickness in micrometers and Cup/Disc area (C/D) in mm2. RNFL thickness was determined in ONH mode in which data, along a 3.45-mm-diameter circle around the optic disc, was mapped using 12 concentric rings and 24 radial scans. Mean, superior, and inferior RNFL thicknesses were computed. The GCC scan covered a square grid of 6 × 6 mm on the central macula and was centered 1 mm temporal to the fovea. The GCC thickness was measured from the inner limiting membrane (ILM) to the posterior boundary of the inner plexiform layer. Mean, superior, and inferior GCC thicknesses were acquired.
During OCTA and retinal structure assessment, vessel capillary density (VCD) and vascular network density were evaluated in ONH. Whole image (WI), inside disc (ID) and peripapillary density (PP) were acquired. In the foveal avascular zone (FAZ), OCTA analysed the following parameters: FAZ area, FAZ perimeter and (FD) -vessel density of the 300μ width ring surrounding the FAZ, macular vessel density (VD) in the superficial (SL) and deep (DL) retinal vascular plexus with a 6x6 mm macula scan were determined.
Vessel density was calculated as the percent area occupied by flowing blood vessels in the selected region.
The retinal layers of each scan were segmented automatically by the AngioVue software to visualize the superficial retinal capillary plexuses in a slab from the internal limiting membrane (ILM) to the inner plexiform layer (IP ) minus 10 μm. For this study, whole en face image vessel density (wiVD) was derived from the entire 6×6 mm2 scan, and perifoveal vessel density was measured in an annular region centered on the fovea with an inner diameter of 1 mm and outer diameter of 3 mm.
The visual field was tested by the Standard Automatic Perimetry (SAP) 24-2 threshold test, at the Optopol PTS perimeter.
We included patients aged above 40 years with a confirmed diagnosis of glaucoma, without any chronic disease (diabetes, neurological disorders, cataracts, except the ones with the level of <1 according to Lens Opacity Classification System) or medication use that could influence OCT and OCTA results. Other exclusion criteria were subjects with false positive and negative errors > 15% and fixation loss > 33%, on SAP, low signal intensity on OCT and poorly centered scan, anomalies in the anterior segment of the eye, trauma, chronic inflammation, retinal diseases, previously surgical and/or laser eye surgery. Pregnant women and women in the period of the lactation were also not considered.
Furthermore, eyes with a history of intraocular surgery (except uncomplicated cataract surgery or uncomplicated glaucoma surgery), coexisting retinal pathologic features, nonglaucomatous optic neuropathy, uveitis, or ocular trauma were also excluded from the study, as were individuals with a diagnosis of Parkinson's disease, Alzheimer's disease, or dementia; a history of stroke; or diabetic or hypertensive retinopathy.

Data analysis and sample size determination
A total sample size was calculated a priori based on the G*Power software. A predicted effect size of 0.7 was smaller compared to the literature data where the differences in RNFL thickness between healthy subjects and glaucoma patients was 1.6482, the statistical power of 90% and significance level of 5%. We used a one-tailed t-test. Under these circumstances, a minimum of 82 patients (41 pergroup) were needed for this study.
In order to analyze the collected data, we used the standard methods of descriptive statistics. The differences between the mean values of continuous variables with normal distribution were assessed by a parametric ANOVA test. We used a non-parametric Kruskal Wallis H test for variables whose values did not follow a normal distribution. Post hoc analysis was done by a parametric Student's t-test and for variables whose values did not follow a normal distribution we used a non-parametric Mann-Whitney U test. To determine the differences in the incidence of certain categories, we used a Chi-square test or Fisher's test of the real likelihood for low frequencies. The presence of a relation between the variables of interest was analyzed by standard correlation analysis using the Pearson's correlation coefficient to determine the direction and strength of the connection/correlation. Binary logistic regression models were used to evaluate the association between OCT and OCTA parameters with the presence of POAG. In all analyses, p-value of less than 0.05 was considered as statistically significant.
Analysis was done between groups of glaucoma patients 148 (76.5%) and healthy subjects 48 (24.5%) with similar gender proportion and average age in both groups. The patients with glaucoma were divided according to glaucoma stage. There was a similar average age between groups. The IOP was the highest and CCT the lowest in the third stage of glaucomatous eyes. Clinical characteristics of the examined groups are shown in Table 1. All OCT measurements of the macula were significantly higher in the healthy subjects compared to glaucoma patients. GCC differed significantly between stages, being highest in the first and lowest in the third stage (Table 3).  Table 4 and 5. All examined OCTA parameters were significantly higher in healthy subjects. All variables were notably lower in patients with the third stage of glaucoma compared to the first stage.
Value of PP-1 differed substantially between all stages, being highest in the first and lowest in the third stage (Table 4).   (Table   6).  Increasing these parameters for one unit reduced the odds for glaucoma for about 50%.
OCTA showed significant association of FD with glaucoma (ExpB 0.76; 95%CI 0.63-0.91) and reduction of odd for 24% for one unit increase.

Discussion
Long standing mechanical approach to glaucomatous optic nerve head damage blames parameters to be used for monitoring disease progression. 15 However, predicting this disease progression, especially in advanced stages of POAG, is challenging because of the existence of a so-called "floor effect", after which no further structural change can be detected in OCT-based RNFL thickness measurements and due to an increase in variability of VF measurements. 16 RNFL in whole and both superior and inferior position was significantly lower in glaucomatous eyes. Literature data indicate that RNFL thinning and rate of average RNFL loss detected by OCT were connected with visual field loss. This stands for both glaucoma suspects and glaucomatous eyes. 17,18 In line with these results, examined glaucomatous eyes had significantly deteriorated visual acuity and parameters of visual field. Previous studies have already established regional correlations between RNFL loss, GCC loss, and visual field deficits in glaucomatous eyes, showing concomitant structural damages in macula as seen in ONH. 19 The importance of the spatial structure of RNFL thickness map data over conventional average circumpapillary RNFL thickness in diagnosing glaucoma was clearly demonstrated in a post hoc study of 93 eyes from Los Angeles Latino Eye Study (LALES). Superior diagnostic performance was shown for all models using full RNFL thickness maps. 20 In patients with glaucoma, all sectors and average RNFL showed significant association with the development of glaucoma. In developed glaucomatous eyes, standard determination of RNFL has high clinical value but its determination is primarily connected with prediction of visual field loss.
All examined OCTA parameters at the level of ONH and macula showed significant differences in healthy subjects and patients with glaucoma, indicating presence of vascular damage, alongside structural impairment. This is in line with the data from a similar study confirming that VCD is diminished in glaucomatous eyes. [21][22][23] Some literature data presents that all the optic disc VD parameters except the inside disc VD were significantly lower in glaucomatous eyes than in control eyes. 24 Other results indicate the VCD in the radial papillary capillaries layer and in the nerve head layer of glaucomatous eyes was significantly lower than in age-matched control eyes. 1 The same authors presented results that VCD in all peripapillary segments was higher than in nerve head layer, similar to our findings.
All examined macular OCTA parameters in superficial and deep layers differentiated in glaucoma presence and the stage of the disease. This highlights a role of vascular factors in the pathogenesis of glaucoma. In a prospective study examining change rate of GCC thickness and macular VD in healthy and OAG eyes, scholars presented results that glaucomatous eyes showed a faster decrease in macular vessel density than GCC thinning.
Faster macular vessel density decrease rate was significantly associated with the severity of glaucoma, however, the association between GCC thinning rate and glaucoma severity was insignificant. 25  Macular GCC correlated only with superficial VD in healthy eyes. Contrary, in glaucoma patients, GCC correlated with the thickness of superficial and deep retinal vascular plexuses. Results did not reveal significant correlation between structural and vascular parameters in healthy eyes, but strong positive correlation of all those parameters in glaucomatous eyes indicating close connection between vascular and structural changes.
This implicates the importance of vascular disorders in development and progression of glaucoma. In line with this are results that compared with age-matched control subjects, vascular density of the parafoveal retina decreased in the OAG subjects. 29 It remains unclear whether the reduced microcirculation in glaucoma patients induces the neuronal damage or arises through reduced circulation requirements in damaged tissue. Due to crosssectional design of the study we could not conclude if vascular changes precede papillary disc damage, but regression analysis gives us some insight into this question.
This study found that VD in macula and optic nerve head predict the development of glaucoma. In a group model, inside disc vessel capillary density and vessel density of ring surrounding the FAZ showed significant association with presence of glaucoma. OCTA measurements in macula in optic nerve head and OCT measurments in macula can predict glaucoma presence, however this is not found in OCT analysis of optic nerve head.
According to this, it can be assumed that vascular changes in different regions of the retina could be a better predictor for glaucoma development than structural changes.

Conclusion
The analysis of OCT bands and OCTA vascular plexuses may be complementary for the non-invasive quantification of the structural and vascular changes in optic nerve head and macula, and accurately distinguishes between healthy and diseased eyes, showing association with presence and development of POAG. Our work demonstrates that reduced vessel density in glaucomatous eyes correlated with functional and structural changes at the optic nerve head and macula in glaucoma. Vascular parameters could be a useful adjunct tool to evaluate/diagnose glaucoma.