Distance between the center of the foveal avascular zone measured automatically and the highest foveal bulge using OCT angiography

The center of the fovea, termed the foveola, is the area of highest visual acuity, has the highest density of cone photoreceptors, and is the center of the foveal avascular zone (FAZ). We investigated the distance between the automatically-determined center of the foveal avascular zone (FAZ) and the manually-determined highest foveal bulge (FB) point using swept-source optical coherence tomography angiography (OCTA). This cross-sectional study included 49 eyes of 49 individuals (34 women and 15 men; median age: 68 years) with no history of ocular disorders. The FAZ in the supercial capillary plexus was automatically determined using the Kanno-Saitama Macro method, and the center of the FAZ was automatically determined using ellipse approximation. The foveal center was manually identied as the highest FB point on B-scan OCTA images. The center of the FAZ was more likely to be located inferior to the highest FB point. In participants with a combined distance of more than 50 µm between the center of the FAZ and the highest FB point, the displacement was signicantly more in the horizontal direction than in the vertical direction (p = 0.017). These results can be applicable to further studies regarding the spatial relationships between the center of the FAZ and the highest FB point in various macular diseases or previously-treated eyes. Schematic explanation of the distance between the center of the foveal avascular zone and the highest foveal bulge point The red cross shows the center of the foveal avascular zone (FAZ) on the optical coherence tomography angiography (OCTA) image. The green and yellow crosses show the highest foveal bulge (FB) points that were measured manually using a B-scan image. The nal highest FB point is dened as the center of the two measured highest FB points. The distance between the center of the FAZ and the highest FB point is measured (Blue arrow).


Introduction
The center of the fovea, termed the foveola, is the area of highest visual acuity, has the highest density of cone photoreceptors and the smallest thickness of the fovea, and is the center of the foveal avascular zone (FAZ). 1 Optical coherence tomography (OCT) is a non-invasive imaging technology used to acquire high-resolution, 3D, cross-sectional images of the retina, and is one of the most important ancillary tools for the diagnosis and management of macular diseases. 2,3 Both spectral-domain OCT (SD-OCT) and swept-source OCT (SS-OCT) provide detailed and in vivo analyses of the interior of the retina, especially the fovea, which is the area with the highest visual acuity. The foveal center is also a key reference landmark of the retina, as the early treatment diabetic retinopathy study (ETDRS) grid generated from the foveal center is an important tool for doctors to diagnose retinal diseases. 4 Therefore, the accurate identi cation of the foveal center is of great signi cance for disease evaluation and diagnosis.
The foveal bulge (FB) is the bulge at the center of the fovea observed on OCT, 5,6 and studies of crosssectional OCT images of healthy eyes have reported that the ellipsoid zone has a bulge at the central fovea. 5,6 The FB is de ned as a bulge in the EZ at the center of the fovea. [5][6][7] The integrity of the foveal microstructure has been used to assess the function of the fovea in various retinal diseases. The presence or absence of the FB has also been shown to be signi cantly correlated with visual acuity in various retinal disorders. 5,8,9 The foveal vascular network is composed of interconnected capillaries that perfuse the inner retinal layer.
This network forms a ring at the margin of the fovea and produces a capillary-free region referred to as the FAZ. 10 The FAZ is highly sensitive to ischemic events and has been implicated in several pathological processes. Previous studies have reported enlarged FAZs in retinal ischemic diseases (including diabetic retinopathy) 11 and retinal vascular obstruction. 12 In addition, the FAZ has been evaluated in glaucoma, 13,14 15 and idiopathic epiretinal membrane (ERM). 16 In recent reports, the center of fovea pit, center of the FB, and center of the FAZ have been located manually using OCT angiography (OCTA), though the manually-located landmarks are not always accurate. 16,17 Recently, we generated and reported a macro program to automatically determine the FAZ area using OCTA. 15,18 Although previous studies have suggested that both the FAZ and FB are critical biomarkers of the functional properties of the fovea in eyes with various retinal diseases, the spatial relationship between the center of the FAZ in en-face images and the highest FB point are poorly understood. Therefore, this study investigated the positional relationship between the OCT-derived center of FAZ and the highest FB point in healthy eyes using OCTA images and an automated FAZ extraction program.

Results
This study included 49 eyes from 49 healthy participants. The participants' ocular characteristics are presented in Table 1 Table 2. The distribution of the distance in the vertical direction is mostly inferior to the highest FB point, while the distribution in the horizontal direction is close to the highest FB point ( Figure 2). In the horizontal direction, the number of participants with a nasal-or temporal-shifted center of the FAZ were 26 and 23, respectively. Overall, the median horizontal shift of the center of the FAZ was 2.0 μm in the nasal direction from the highest FB point. In the vertical direction, the center of the FAZ was more likely to be located inferior to the highest FB point (32 of 49 eyes, 65.3%), with a median distance of 10.5 μm from the highest FB point, which was signi cantly different from the reference value of ± 0 (p=0.031, Wilcoxon signed-rank test).   Figure 3 and Table 3 show the direction and distribution of the total distance from the highest FB point to the center of the FAZ. In 31 eyes (63.2%), the total distance was less than 50 μm. In the eyes with a total distance of more than 50 μm (n=18), the distance was signi cantly greater in the horizontal direction than in the vertical direction (p=0.017, Wilcoxon signed-rank test).

Discussion
This study evaluated the spatial relationship between the automatically-determined center of the OCTAderived FAZ and the center of the highest FB point. The distance between the center of the FAZ and the highest FB point was generally close but was not aligned in healthy eyes, with an average distance of approximately 30 μm between the two points. While the positions of the center of the FAZ and the highest FB point were similar in the horizontal direction, the center of the FAZ was located signi cantly inferior to the highest FB point in the vertical direction. In participants with a combined distance of more than 50 μm between the center of the FAZ and the highest FB point, the displacement was signi cantly more in the horizontal direction than in the vertical direction.
In this study, the FAZ was identi ed automatically using the Kanno-Saitama Macro (KSM) method. 18 The program utilizing the OCTA instruments faced several limitations, and required manual delineation of FAZ, unlike in this study. 16 The automated methods are more objective and reproducible than the manual method, and the speci c KSM method and its clinical applications have been reported. 15,18 The identi cation of the FAZ has excellent reproducibility and is comparable to measurements obtained manually. Thus, it is easy to measure the precise distances with high reproducibility. The precise spatial relationship between the center of the FAZ and the highest FB point is clinically relevant as the macular region, including the fovea, is closely associated with visual prognosis. An understanding of the relationships between the highest FB point and the FP is necessary to investigate the signi cance of the location of the center of the FAZ.
Prior to the widespread use of OCTA, the positional relationship between the center of the FAZ and the highest FB point was poorly understood as FAZ measurements using uorescein angiography were not routine, especially in healthy individuals. Furthermore, the ability to accurately compare and measure the distance between the center of the FAZ and the highest FB point was made possible with the development of OCTA technology, which provides both in vivo en-face images and B scan images of the macular region as digital images. Several recent studies have reported that the FB, foveal pit (FP), and center of the FAZ, which are considered to indicate the center of the central fovea, are misaligned even in healthy eyes. 16,17,19 Kuppuswamy et al. reported that the position of the FB nasally deviates from the FP in healthy eyes, with a median distance of 58.6 μm. 19  In this study, eyes with a total distance of more than 50 μm between the center of the FAZ and the highest FB point had signi cantly greater misalignment in the horizontal direction than in the vertical direction. A recent study reported that foveal misalignment was greater in eyes with ERM with ectopic inner foveal layers (EIFLs) than in those without EIFLs, especially in the horizontal direction, with an average distance of more than 100 μm. 16 As the participants in this study were elderly individuals with healthy eyes, subclinical changes may have occurred in the eyes with greater foveal misalignment.
Although these relationships are unclear due to the limitations of cross-sectional studies, the pathogenesis and progression of ERM should be studied in longitudinal studies in the future.
This study has several limitations. First, the age distribution of the participants in this study was not su ciently wide, and the number of participants was not large. A previous report suggested that FB was more commonly observed in younger individuals. 20 However, age is not a speci c confounder of this study as it compares the distance between the OCT-derived center of the FAZ and the highest FB point according to the imaging setting in the same participant and does not compare the misalignment between age groups. Although there are no indications that this observation will not apply to individuals of different ages, future studies to con rm the distance in younger individuals are needed. Second, although the center of the FAZ was more likely to be located inferior to the highest FB point and was signi cantly different compared to a reference value of ± 0, the difference was relatively small (10 μm).
Thus, even though the results showed statistical signi cance, they may not be clinically meaningful. However, the purpose of this study was to investigate the distance between the center of the automatically-determined, OCTA-derived FAZ area and the manually-determined highest FB point. The results of this study indicate that the center of the FAZ and the highest FB point are near one another in healthy individuals. These results will improve the understanding of diseased and treated eyes.
In conclusion, the distance between the center of the automatically-determined, OCTA-derived FAZ area and manually-determined highest FB point. is minimal, but was not aligned in elderly participants with healthy eyes. These results can be applicable to further studies regarding the spatial relationships between the center of the FAZ and the highest FB point in various macular diseases or previously-treated eyes.

Study population
This cross-sectional study of healthy subjects was approved by the Ethics Committee of Saitama Medical University (No. 20013.01) and was conducted in accordance with the tenets of the Declaration of Helsinki. Healthy participants that were 20 years of age or older who ful lled the eligibility requirements were enrolled in the study. All participants provided informed consent. The study period was October 2017 to November 2017.
Healthy participants were recruited from the ophthalmology outpatient clinic of Saitama Medical University Hospital (Saitama, Japan), except for refractive error. All participants underwent a comprehensive ophthalmic examination, including slit lamp biomicroscopy, IOP measurement via noncontact tonometry (Tonoref II, Nidek Co., Ltd., Aichi, Japan), and fundus photography (CX-1, Canon Inc., Tokyo, Japan). The axial length and CCT were measured using an optical biometer (OA-2000, Tomey Corp., Nagoya, Japan). Horizontal and vertical B-scan images and en-face images of the area around the macula were obtained using swept-source optical coherence tomography angiography (SS-OCTA; PlexElite9000 Carl Zeiss Meditec, Jena, Germany).
Participants < 20 years of age with a re ective error > +3.0 diopters or < -6.0 diopters, an axial length > 26 mm, other ocular diseases, diabetic retinopathy, retinal vein/artery occlusion, age-related macular degeneration, retinal detachment, tilted disc, exfoliation syndrome, high myopia, ocular neuropathy without mild ametropia, or a history of ocular surgery (except uncomplicated cataract surgery) were excluded from the study. Participants in whom images of poor quality were obtained (signal strength < 8 due to signal noise; 1 = minimum, 10 = maximum) were also excluded from the study.
Optical coherence tomography angiography A 3 × 3 mm (1,024 × 1,024 pixels) OCTA image centered on the fovea was scanned using SS-OCTA using a central wavelength of 1,060 nm, an A-scan rate of 100,000 scans per second, an axial resolution of approximately 5 μm in tissue, and an estimated lateral resolution at the retinal surface of approximately 12 μm. 21 The angiography image was processed using phase/Doppler shift and amplitude variation (Optical Micro-Angiography). 22 The scan contained 300 A-lines x 300 locations with four repeated scans in each xed location.
Measurements of the center of the FAZ using OCTA en-face images The area of the FAZ (mm 2 ) was calculated using ImageJ and an original macro language (KSM program), as previously described. 18 In brief, the SS-OCTA software generates en-face images from slabs at different layers via automated segmentation. The super cial retinal layer (SRL) was used to measure the area of the FAZ. The SRL was de ned as the area between the internal limiting membrane layer and the inner plexiform layer. KSM is an automated analysis program that extracts the area of the FAZ. The FAZ was de ned as the area denoted by the connected points along the borders of the identi able capillary network in the parafoveal area. The extracted FAZ showed excellent reproducibility and was comparable to manual measurements. 18 After identifying the FAZ, the center of the FAZ was automatically identi ed as the center of an elliptic approximation using the ImageJ program ( Figure 4).

Measurements of the height of the FB
The height of the FB was measured as the highest FB point, as previously reported. 16,17 The height of the FB was de ned as the distance between the outer boundary of the ellipsoid zone (EZ) and the inner boundary of the Bruch's membrane at the vertex of the FB (Figure 25). A single observer (HI) who was blinded to the participants' characteristics measured the highest FB points in both the horizontal and vertical B scans in duplicate. The intersection of the highest FBs in the horizontal and vertical scans was de ned as the highest FB point. The nal highest FB point was the center point of each measurement of the highest FB point. The distance between the central point of the FAZ and the highest FB point was calculated ( Figure 6). The magni cation effect was adjusted due to axial length according to the manufacturer's correction formula, and it was con rmed that the correction was comparable to that in a previous study. 23 Statistical analysis   Identi cation of the center of the foveal avascular zone The central fossa vessel-free zone was extracted using the Kanno-Saitama Macro (KSM) method (red line). An approximate ellipse was created for the delineated foveal avascular zone (FAZ) (blue line). The center of the FAZ is de ned as the intersection of the major and minor axes of the approximate ellipse.

Figure 5
Identi cation of the highest foveal bulge point using optical coherence tomography angiography The intersection of the highest foveal bulge (FB) in the horizontal and vertical scans was de ned as the highest FB point. Figure 6 Schematic explanation of the distance between the center of the foveal avascular zone and the highest foveal bulge point The red cross shows the center of the foveal avascular zone (FAZ) on the optical coherence tomography angiography (OCTA) image. The green and yellow crosses show the highest foveal bulge (FB) points that were measured manually using a B-scan image. The nal highest FB point is de ned as the center of the two measured highest FB points. The distance between the center of the FAZ and the highest FB point is measured (Blue arrow).