Optical Coherence Tomography Angiography in Patients with Amblyopia

ABSTRACT To determine the optical coherence tomography angiography (OCTA) parameters including foveal avascular zone (FAZ) and vessel density (VD) in the amblyopic eyes compared with the fellow sound eyes and the eyes of the non-amblyopic subjects. In this case–control study, a total of 23 eyes from unilateral amblyopic children were included as cases. The sound eye of the amblyopic children was considered as the internal control and the right eyes of the non-amblyopic children were considered as the external control. All participants underwent image recording with OCTA. In the present study, an equal number of 23 unilateral amblyopic eyes and 23 right eyes of non-amblyopic age- and sex-matched children were included as the cases and controls, respectively. The average age of participants in the case and controls were 9.86 ± 3.12 and 8.5 ± 2.35 years, respectively. Twelve patients (52.2%) in the case group and 14 subjects (60.9%) in the control group were female. Whole vascular density of the macula in superficial capillary plexuses (SCP) was significantly lower in the external controls compared with the other studied groups (P = .026). However, the VD of the deep capillary plexuses (DCP) was significantly greater in the external controls than cases and internal controls (P= .029). The average FAZ area was 0.26 ± 0.06 mm2 in amblyopic eyes that was significantly higher compared with fellow eyes (0.21 ± 0.07 mm2; P= .022), but it was not different with non-amblyopic eyes (0.22 ± 0.118 mm2). Based on our findings, there were no significant difference in the cases of foveal, parafoveal, and perifoveal in both superficial and deep vascular densities among amblyopic and non-amblyopic eyes, whereas deep whole density of the amblyopic eyes showed lower percent compared to non-amblyopic ones that indicates decrease blood supply of the amblyopic eyes in this region. Additionally, FAZ was larger in amblyopic eyes than internal controls.


Introduction
Amblyopia is a visual impairment demonstrating with decreased best corrected visual acuity (BCVA) as the hallmark and other visual disorders including reduced contrast sensitivity. [1][2][3] Additionally, there are some studies reporting the involvement of choroidal and retinal layers 4,5 such as significantly increased central macular thickness in eyes having moderate and severe amblyopia compared with the healthy eyes. 6 A meta-analysis also showed a significant thickening of the peripapillary and foveal fiber layer of the amblyopic eyes compared with the non-amblyopic eyes. 6 Recently, there have been advancements in image processing particularly application of the optical coherence tomography angiography (OCTA) in ophthalmology. [6][7][8][9][10] The OCTA is capable to illustrate blood flow and density of the superficial and deep capillary plexuses as well as the choriocapillaris and also the size of the foveal avascular zone (FAZ). [6][7][8][9][10] In the previous reports, a significant reduction of both superficial and deep capillary plexuses (SCP, DCP) and a significant increase of the vessel density of the choriocapillaris was identified in the amblyopic eyes compared with the healthy eyes. 5,7,9,10 In this study, we aimed to evaluate the macular parameters including FAZ size and vessel density in the amblyopic eyes compared with the nonamblyopic eyes.

Methods
In this case-control study, amblyopic eyes of 23 unilateral amblyopic children were considered as cases and the sound eye of them were considered as the internal control (n = 23) and the right eyes of the non-amblyopic children (n = 23) were considered as the external control.
Amblyopia was considered if BCVA was less than 0.3 LogMAR without any organic reason or there were at least two lines of BCVA difference between the two eyes. The minimum BCVA of the fellow eye should not be less than 0.3 LogMAR. Eligible amblyopic patients who were able to cooperate for data recording and those aged between 4 and 14 years were included. All study procedures were explained to the study subjects and their parents before image recording or examination and a written informed consent was obtained from all participants' guardians. The study procedures were approved by the Ethics Committee of the Ophthalmic Research Center, Shahid Beheshti University of Medical Sciences via the approval number of IR.SBMU.ORC.REC.1398.023.
Patients with nystagmus, no accurate fixation, unsteady fixation, and low cooperation during image recording; amblyopic children with BCVA worse than 1.0 LogMAR, deprivation amblyopia, strabismic amblyopia with a deviation angle of ≥10 prism diopters; eyes with an axial length of >26 mm and the spherical and cylindrical refractive errors of more than 5.00D; patients with psychogenic visual disturbance; systemic disease; retinopathy of prematurity and congenital ocular anomalies were excluded from the present study.

Visual and ocular examination
All ophthalmic examinations were performed including the measurement of BCVA, cycloplegic refraction 30 to 45 minutes following the installation of the ophthalmic eye drops of tropicamide 1% and cyclopentolate 1%, ocular motility function testing (+4 denoted as overaction and −4 denoted as underaction for the extraocular muscles), measurement of the ocular deviation using the alternative prism cover test, foveal fixation testing by visuoscopy and anterior and posterior ocular segment examination by using the biomicroscopy and indirect ophthalmoscopy, respectively. Finally, all participants underwent the image recording by OCTA (RTVue XR Avanti with AngioVue; Optovue INC., Fremont, CA) by an experienced technician. In order to monitor the child's constant fixation, image recording was firstly performed on the amblyopic eye and secondly on the fellow nonamblyopic eye.

Optical coherence tomography angiography
All OCTA scans were performed by a single experienced examiner using a spectral domain RTVue XR Avanti (Optovue, Inc, Fremont CA, USA) with split-spectrum amplitude decorrelation angiography (SSADA) algorithm. The device works with a central wavelength of 840 nm, an acquisition speed of 70,000 A-scans per second, and an axial and transverse resolution of 5 microns in tissue. A good image quality is more than 4 according to the OCT manufacturer. The built-in software "AngioAnalytics" was used for analyzing vessel densities and FAZ.
Before imaging, pupils were dilated using tropicamide 1% eye drop. Study participants underwent SD-OCT imaging based on a protocol that contained AngioVue OCT 3D volume set of 6 × 6 mm. An internal fixation light was used for keeping the fixations central. Optimization of signal position and quality were performed through running the option of "Auto All." Based on the position of capillary plexuses in the retina and choroid, the boundaries of the superficial network extended from 3 μm below the internal limiting membrane (ILM) to 15 μm below the inner plexiform layer (IPL). The deep capillary network extended from 15 to 70 μm below the IPL. The choriocapillaris extended from 30 to 60 μm below the retinal pigment epithelium.
The centration of the fovea was checked for all images. All images were reviewed for the segmentation errors. Additional image artifacts including motion, banding, and projection artifacts were evaluated, as well.

Vessel density analysis
Objective quantification of the vessel density was measured for each eye using the SSADA software. The OCTA en face images for each eye were used to perform quantitative analysis with the AngioVue software. The vessel density was defined as the percentage area occupied by vessels in a circular region of interest (ROI) centered on the center of the FAZ with a diameter of 3 × 3 mm included inside the 6 × 6 mm scan area. The AngioVue software automatically splits the ROI into two fields: the foveal area, a central circle with a diameter of 1 mm, and the parafoveal area that constitutes the remaining part inside the ROI. The three capillary systems were evaluated with quantitative analysis in the foveal and parafoveal areas and with qualitative analysis in the whole 6 × 6 mm scan area. The way that vessel density was calculated was previously described. For each patient, foveal and parafoveal vessel density (VD) and parafoveal VD in different quadrants (temporal, superior, nasal, inferior) of the SCP and DCP were calculated. OCTA images of the patients with quality score more than 6 were included in our study. A central 1 mm circle was considered as the foveal area. 11 A circle with an inner diameter of 1 mm and an outer diameter of 3 mm was considered the parafoveal area. Accordingly, a circle with a diameter of 6 mm and inner diameter of 3 mm was considered as the perifoveal region. 12,13 For each patient, foveal, parafoveal, and perifoveal VD of the SCP and DCP layers were calculated.
Two independent observer evaluated all OCTA images regarding segmentation error and in cases with the possible segmentation error, it was manually corrected. Foveal macular thickness (MT), full parafoveal MT, and parafoveal MT in different retinal quadrants and ILM-IPL parafoveal MT were automatically calculated by the software on the OCTA 6 × 6 mm volume scan (×R Avanti1; Optovue, Inc., Fremont, CA, USA).

Statistical analysis
To describe data, we used frequency (percent), mean ± SD, median, and range. To evaluate the difference between three groups at baseline, we used ANOVA for quantitative variables and Chi-Square and Fisher's exact test for qualitative variables. To compare indexes of the OCTA between groups when considering the correlation of eyes, Generalized Estimating Equation (GEE) analysis was used. A P -value less than 0.05 was considered as statistically significant. All statistical analysis was performed by SPSS software (IBM Corp. Released 2017. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp.).

Results
The average age of 23 cases and 23 controls were 9.86 ± 3.12 and 8.67 ± 2.06 years, respectively (P = .095). Twelve patients (52.2%) in the case group and 14 (60.9%) in the control group were female.

Whole vessel density
In the amblyopic eyes, the whole VD of the superficial layer was significantly lower in external controls (46.89 ± 4.03) compared with cases (49.19 ± 2.99) and internal controls (49.87 ± 4.14) (P = .026). Conversely, the corresponding outcome in DCP was significantly greater in the external controls than internal control (46.84 ± 4.16) and cases (47.82 ± 4.0) (P = .029).
In the case of vascular density analysis with respect to the refractive errors, it was found that the whole density in SCP was significantly lower in astigmatism compared with both simple and compound hyperopia (P = .034). Additionally, the whole density in DCP was significantly lower in patients with compound hyperopia (P = .012; Table 2).

Foveal vessel density
The average foveal VD in SCP and DCP was 21.4 ± 10.51 and 40.16 ± 7.5 in the amblyopic eyes. These parameters were 21.47 ± 5.8 and 39.93 ± 7.2 in the fellow eyes and 21.86 ± 9.62 and 39.81 ± 10.8 in the non-amblyopic control (Table 3).

Perifoveal vessel density
No significant difference was observed between the three study groups in terms of perifoveal VDs in both SCP and DCP. Regarding the comparison between different types of refractive errors, the vascular density of the perifoveal area in the superficial zone was significantly higher in patients with simple hyperopia (P = 0.012; Table 2).

Discussion
The present study demonstrates that macular vessel density in eyes with mild-to-moderate amblyopia in comparison with internal and external controls. As resulted, the significant lower whole VD of the superficial layer in the external controls was observed compared with internal controls and cases. However, the increased whole VD was detected in DCP among the external controls. Furthermore, larger FAZ area was found in cases compared with the internal controls.
Various studies used OCT-A as a noninvasive technique to investigate the changes of macular microvasculature of the retina in amblyopia. 5,9,10,14,15 Regarding macular VD, mixing results have been reported. The studies performed by Sobral et al., 10 Zhang et al., 16 Yilmaz et al., 5 and Lonngi et al. 9 demonstrated lower macular VD in eyes with amblyopia as our result in DCP. In our study, the whole VD in DCP of the amblyopic eyes was less than the external non-amblyopic controls. Araki et al. 14 and Demiyarak et al. 17 found no differences regarding   VD between amblyopic eyes and healthy controls. Similar to our findings that we did not find any difference in comparison of the VD between amblyopic eyes and their fellow non-amblyopic eyes.
There are several reasons to explain the different results noticed in several studies. First, the diversity of the results may be attributed to the different severities of amblyopia. For example, there is a wide range of refractive error in Lonngi et al's study 9 compared to the present study. Second, axial length and central subfield thickness (CST) affect the FAZ area and macular VD. 14,18 Previous reports demonstrated different CST in amblyopic eyes, however not all amblyopic eyes show these differences. Macular thickness can affect the FAZ area through the hypothesis that higher macular thickness is associated with more compact inner retinal layer. Therefore, the continuity of INL results in a smaller FAZ area and vice versa. 19 Third, amblyopia has different causes occurring in different ages. It has been proposed that neuronal development of GCL and NFL determine the vascularity of the macula in infancy and within the first years of life. However, amblyopic eyes with different severity and type could differently affect the neuronal development. The next reason could be attributed to the OCT-A device and its inherent differences. Araki and associates corrected the magnification error and found no changes in VD between amblyopic and healthy eyes. Considering these discrepancies and possible explanations, the vascular changes in amblyopia remain to be elucidated by further investigations.
Our results demonstrated larger FAZ area in amblyopic eyes compared with the fellow eyes. The larger area achieved in the present study is in line with Sobral et al.'s study, while FAZ area did not show significant changes in studies by Wong, 15 Zhang, 16 Yilmaz, 5 Lonngi, 9 and Demiyarak. 8 Additionally, Wong et al. 15 reported decreased FAZ circularity and lower fractal dimension of retinal vessels while they found no change in FAZ area and vessel density. In contrast to our results, Araki and associates found smaller FAZ area in amblyopic eyes compared with their fellow eyes, but the authors concluded that this finding seems to be clinically insignificant. 14 Another possible mechanism explaining FAZ changes is the impact of AL on the area. 20 Sampson et al. 18 reported that longer axial length could cause apparent larger FAZ because of magnification. As previously reported, most of our participants were hyperopic, therefore it might be identified that the majority of our hyperopic cases were refractive not axial.
In terms of different types of refractive errors, it was found that the superficial VD was greater in hyperopia, while the whole VD was less in DCP among eyes having compound hyperopia compared with the other types of refractive errors. This could be due to the higher density of the vessels in hyperopic cases because of the shorter axial length.
The strengths of the present study consist of inclusion of mild-to-moderate amblyopia, and comparison with the fellow eyes of patients with amblyopia as the internal control. Our study findings were not adjusted by the AL measurement. However, the same method of OCT angiography was applied in both patients and healthy controls.
In conclusion, no significant difference in the cases of foveal, parafoveal, and perifoveal were found in both superficial and deep vascular densities among amblyopic and non-amblyopic eyes, whereas deep whole density of the amblyopic eyes showed lower percent compared to non-amblyopic ones that indicates decrease blood supply of the amblyopic eyes in this region. Additionally, FAZ was larger in amblyopic eyes than internal controls.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The author(s) reported that there is no funding associated with the work featured in this article.
versions of the manuscript. All authors read and approved the final manuscript.