The macular microvasculature is a complex system consisting of multiple capillary networks. The retinal vasculature, composed of the central retinal artery and veins and their branches, is responsible for maintaining the proper function of the inner retinal layers (from the nerve fiber layer to the outer plexiform layer).1,2 In contrast, the outer retina and the foveal avascular zone primarily depend on diffusion from the choroidal circulation.3 In addition, the choroidal vascular bed regulates the thermal conditions of the ocular tissues. Thus, both retinal and choroidal vascular structures are essential in maintaining retinal function.3
Obtaining adequate visualization of the choriocapillaris (CC) is challenging. A limitation of fluorescein angiography is the blue-green excitation wavelength of fluorescein, which is partially absorbed by the macular xanthophylls and retinal pigment epithelium. Another limitation is leakage of the injected fluorescein.4 Because fluorescein does not bind completely to albumin, it escapes through the CC fenestrations, producing early, diffuse hyperfluorescence that obscures visualization of the CC.5 In contrast, indocyanine green angiography (ICGA) remains intravascular and thus enables visualization of the choroidal circulation. However, because ICGA is not depth-resolved, separating CC blood flow from that of the deeper choroidal vasculature is a complex task, and for this reason, ICGA has not gained widespread acceptance for CC visualization.6
The noninvasive technique of optical coherence tomography angiography (OCTA) provides additional, clinically relevant insight into the vascular structures of the eye. OCTA shows vascular changes in different retinal, choroidal, and optic nerve diseases.7–10 The acquired signal is generated from repeated B-scans obtained at the same position of the retina and choroid. The angiography projection slab is generated by displaying the signal as a gray scale image.
A number of OCTA studies have been performed on the inner retinal circulation, and normative data (for comparison with diseased eyes) have been developed for the various retinal capillary plexi.11–14
The CC may also be imaged by OCTA, although it is believed that the small size of the choriocapillaries and the small intersinusoidal spacing prevents visualization of individual choriocapillaries in the macula. Thus, the CC as visualized by OCTA is described as homogenous in appearance with a pattern of bright areas representing flow and dark areas thought to be areas with low flow or flow voids.15
Although several groups have shown changes in the CC in age-related macular degeneration,16–18 hereditary retinal diseases,19–21 inflammatory diseases,22–24 and retinovascular diseases, few studies have quantified the CC in healthy eyes. Additionally, few have assessed the repeatability of these assessments in particular.25
Accordingly, the aim of this study was to analyze the features of the CC quantitatively in healthy individuals using swept-source OCTA and to measure its repeatability.
Patients and Methods
This prospective, noncomparative case series was approved by the Institutional Ethics Committee of the University of California Los Angeles and conducted according to the tenets of the Declaration of Helsinki. The study was carried out in accordance with Health Insurance Portability and Accountability Act regulations. All subjects signed written informed consent prior to participating in the study.
Healthy individuals older than 18 years with no previous history of ophthalmologic or systemic diseases were included. Subjects were required to have a normal ocular examination and a normal structural OCT of the macula. Those with a refractive error greater than 2.5 diopters or a history of any ocular surgical intervention (including refractive surgery) were excluded.
All OCTA scans were performed by a single experienced examiner (MA) using a swept-source OCT (SS-OCT) device (DRI OCT Triton; Topcon, Tokyo, Japan) with a central wavelength of 1,050 nm, an acquisition speed of 100,000 A-scans per second, and an axial and transverse resolution of 7 μm and 20 μm in tissue. SS-OCT technology uses longer wavelength infrared light with less sensitivity roll-off with depth compared to conventional spectral-domain OCT, which allows for deeper penetration into tissue and better imaging of the CC. Scans were taken from 3 mm × 3 mm cubes, with each cube consisting of 320 clusters of four repeated B-scans centered on the fovea. To investigate repeatability of the measurements, two sets of scans were obtained for each subject, 1 minute to 2 minutes apart.
En face images were generated from the CC based on automated layer segmentation performed by the OCT instrument software (IMAGEnet 6 V.1.14.8538). The automated segmentation defines the en face slab for the CC to extend from Bruch's membrane to 10 μm below. All images were reviewed to confirm accurate segmentation by the automated instrument software. No manual adjustment was deemed necessary.
For choroidal thickness (CT) measurements, a volume OCT was captured with the same SS-OCT device using wide field 12 mm × 9 mm raster scan protocol with the scan centered on the posterior pole. The 12 mm × 9 mm scan is composed of 256 B-scans, each comprising 512 A-scans with a total acquisition time of 1.2 seconds. A 3.4 mm diameter circle was used to automatically generate the thickness of the choroid.
Quantitative Measurements and Statistical Analysis
Quantitative analysis was performed using the publicly available ImageJ software (National Institutes of Health, Bethesda, MD).26 The en face slab of the CC was used for quantitative analysis. A region of interest (ROI) was established as a 2.5-mm diameter circle centered on the fovea (center of the OCTA en face image). We chose 2.5 mm as opposed to 3 mm (the maximum possible size) since we planned to align data from repeated scans; if there was a slight displacement of the repeated image relative to the first image, the ROI would extend past the border of the scanned area for at least one of the images. This 2.5-mm circle was divided into a central 1-mm subfoveal field and a surrounding parafoveal ring. The parafoveal ring was further subdivided into nasal (N), superior (S), temporal (T), and inferior (I) sectors. Figure 1 shows the OCTA image of the CC and the schematic ROI depicting the subfoveal field and the parafoveal sectors.
Choriocapillaris en face projection slab with projection of the region of interest (ROI). The ROI consists of a circle with 2.5-mm diameter circle, which was further subdivided into a subfoveal field (1 mm in diameter) centered at the fovea and a circle with a 2.5-mm diameter for the remaining area. The parafoveal ring was divided to nasal, superior, temporal, and inferior sectors.
The CC en face image was analyzed in two ways. First, as in our previous reports of the retinal circulation, vessel density (VD) was expressed as a ratio by taking the total vessel area divided by the total area of the analyzed region.27 For this measurement, we used a binarized image with intensity thresholding with Otsu's thresholding method as implemented in ImageJ. Otsu's method assumes that the image contains two classes of pixels following a bimodal distribution. It calculates optimum threshold by minimizing intraclass variance and maximizing interclass variance.26 For the second assessment, we calculated the mean grey value (GV) as a measure for the flow in the CC, using the automated tool for GV measurement as implemented in ImageJ to quantify the flow information detected as motion. The GV represents the brightness of a pixel, which is an indicator of flow area in an OCTA image. The minimum and maximum grey levels are 0 and 255, respectively. The mean GV is the sum of the grey levels of all the pixels in a selected area divided by the number of pixels. Of note, we did not perform any normalization of the OCTA images for the GV measurements.
To calculate the repeatability of the repeated scans, the images of the CC of the two sessions were registered to each other. For this purpose, the two superficial retinal layers were registered to each other using vascular bifurcation in the superficial plexus as landmark in order to align the two CC images for accurate measurements — in other words, to ensure that the ROIs were at the same location. The VD and GV measurements of the CC were repeated for the images from the second session using the same protocol. Figure 2 illustrates the registration and alignment of the images from the two different sessions.
Registration and alignment of the choriocapillaris (CC) to the corresponding superficial retinal layer. (A, B) The en face projection slab of the superficial retinal layer from the two sessions. (C) The original image from the first session and (D) registration of the superficial retinal layer according to vessel bifurcation as landmark correspondence with red, green, and yellow dots. (E) The original en face projection slab of the CC from the first session. (F) The aligned CC slab from the second session. White arrows show the direction and area of adjustment of the second session according to the registration of the vessels.
For the CT analysis, a choroidal map was automatically generated with numeric averages of the thickness measurements for the nine map sectors as defined by the Early Treatment Diabetic Retinopathy Study (ETDRS). Manual adjustment of the position of the ETDRS grid was performed if necessary. To correlate with our choroidal VD measurements, we used the central 1-mm diameter circle for the subfoveal field and the four sectors of the middle ring.
All statistical analyses were performed using SPSS software version 21 (SPSS, Chicago, IL). Paired samples (VD and GV for the different ROIs) were tested for normal distribution using the Shapiro-Wilk test. Analysis of variance was used to test equal variances for differences between the means. The Bonferroni correction was used to obtain P values. Pearson correlation coefficient was used to test the correlation between CT with VD and GV. The intraclass correlation coefficient was used to analyze repeatability. A P value of < .05 was considered significant.
Thirty-six eyes of 22 healthy individuals were included in this study. The mean age was 37.5 years ± 8.7 years (range: 28 years to 59 years); 10 of the individuals were male and 12 were female.
Morphologically, as previously described, the CC appeared to have a uniform dense and homogenous appearance with variations in grey scale intensity.
The distribution of the measured values and the standard deviation for the cohort are shown in Figure 3.
Left column shows the distribution of the measured vessel density ratio (Y axis) for each patient (X axis) according to the region of interest. The middle line represents the mean of all lines, and upper and lower lines represent the standard deviation. Right column represents the grey value.
The VD was 0.444 ± 0.031 in the subfoveal field, 0.474 ± 0.034, 0.478 ± 0.036, 0.494 ± 0.037, and 0.487 ± 0.031 in the N, S, T, and I sectors, respectively (Figure 4A; Table 1). The GV was 106.42 ± 4.351 in the subfoveal field, 110.26 ± 4.305, 110.907 ± 4.910, 112.787 ± 5.077, and 111.67 ± 3.642 in the N, S, T, and I sectors, respectively (Figure 4B; Table 2).
Vessel density ratio and grey value (GV). (A) The mean standard deviation (SD) of the vessel density ratio in the subfoveal, nasal, superior, temporal, and inferior sectors. (B) The SD of the GV in the corresponding sectors of the region of interest. Error bars represent the 95% confidence interval.
Vessel Density (Ratio) in the Subfoveal Fields and Parafoveal Sectors
Grey Value in the Subfoveal Fields and Parafoveal Sectors
The VD and GV in the subfoveal field were statistically significantly lower than in each parafoveal sector (both P < .001).
The intersession repeatability was moderate to high for all VD measurements (0.79 to 0.92); however, it was somewhat lower for GV measurements (0.69 to 0.89).
The CT in the subfoveal field was 276.07 μm ± 87.168 μm compared to 270.96 μm ± 71.1 μm in the parafoveal ring. The difference was not statistically significant (P > .05).
No correlation was observed between the VD and GV measurements and the CT values in the subfoveal and parafoveal regions (all P > .05).
In the present study, we analyzed two quantitative features of the CC using SS-OCTA. We measured the VD as a ratio of the total vessel area relative to the total area of the region. In addition, we computed the GV as a potential aggregate measure of the flow across the region. Both VD and GV were lower in the subfoveal area than in the surrounding parafoveal sectors with moderate-to-high intersession repeatability. Previous histologic studies have shown that the CC vessels are arranged more compactly at the posterior pole than in the periphery, gradually becoming less dense toward the periphery.3 On the other hand, the CC consists of a monolayer in the fovea with an increasing the number of layers, progressing centrifugally in the macula.28 Although this could be one explanation for why we detected a lower VD and GV in the subfoveal area, there are many other possibilities. For example, the flow inside the subfoveal CC may be more difficult to detect than that of the adjacent regions. This may be due to the masking effect of the densely packed and taller foveolar RPE cells or to the higher concentration of macular pigment in the central fovea.29
In contrast to previous dye-based angiography studies, which suggested that each CC lobule was an independent unit with no anastomoses with adjoining lobules, our findings would appear more compatible with the histologic vessel cast-preparations of the CC, which show a markedly interwoven and interconnected network of sinusoidal spaces.3,28,30 However, the resolution limit of OCTA is higher than the resolution needed to visualize the flow in individual choriocapillary vessels. Spaide analyzed images of the CC from the central macula obtained with an RTVue XR Avanti (Optovue, Fremont, CA) instrument.15 He reported that the pattern of flow voids formed a scale invariant pattern in the CC, starting with a size much smaller than a choroidal lobule, and with significant alterations in those eyes with pseudodrusen or those with late age-related macular degeneration in the fellow eye. He did not quantify the VD. Wang et al. did quantify VD in the CC layer in healthy eyes.25 They arbitrarily chose 57 areas distributed within the 1-mm diameter subfoveal circle and another 155 areas located in the surrounding 3-mm ring-like region to measure the mean CC vessel diameter and pixel density. Similar to our findings, the VD of the CC was 44.4% ± 3.3% in the central 1-mm area and 44.5% ± 2.7% in the pericentral region. We reported another parameter for the CC slab, GV, and found that it follows the same behavior as VD, with a lower value for the central subfoveal area compared to surrounding regions. Our study used a SS-OCTA with a light source centered at 1,050 nm, which can penetrate tissues to a greater extent with less sensitivity roll-off with depth. This could be an advantage for evaluating the microcirculation of the CC.
Our study has limitations that should be considered when assessing our findings. First, our sample size is relatively small and mostly a younger population, which does not allow a comparison of the CC values with age. Thus, our results are likely not applicable to older patients; and Spaide did suggest there may be more flow voids in the CC in older individuals. Second, we only used information from one OCTA instrument and one standard, manufacturer-provided processing method. Each OCTA device uses its own proprietary acquisition and processing techniques, which vary with regard to thresholding (to display flow versus no-flow) and to the scheme for removal of projection artifact. Thus, our findings are almost certainly not applicable to other instruments, and possibly not even to future enhanced processing algorithms for the present device. Despite the heavy manufacturer “processing” of OCTA data, it is still interesting that in an individual subject, we observed good repeatability. Third, we did not apply any normalization (ie, correction for overall signal intensity, which may be affected by image quality, etc.) for the GV. It is unclear whether some type of normalization scheme, as we have successfully applied to structural OCT,31 may have yielded even better intersession agreement.
In addition, although the goal in developing a normative database is to provide a reference for comparison with diseased eyes, other abnormalities in diseased eyes (such as loss of signal due to thickened retina) may confound CC assessments and comparisons to normative data.
In summary, our study provides insight regarding the repeatability of two quantitative metrics for evaluating the CC in normal individuals. These findings may provide a useful reference or starting point for assessing variations in the CC in normal eyes, changes over time, and possibly CC alterations in the setting of disease.
- Wise GN, Dollery CT, Henkind B. The Retinal Circulation. New York, NY: Harper & Row; 1971.
- Levin LA, Kaufman PL. Adler's Physiology of the Eye: Clinical Application. Edinburgh, NY: Saunders/Elsevier; 2011.
- Olver JM. Functional anatomy of the choroidal circulation: Methyl methacrylate casting of human choroid. Eye (Lond). 1990;4(Pt 2):262–272. doi:10.1038/eye.1990.38 [CrossRef]
- Keane PA, Sadda SR. Imaging chorioretinal vascular disease. Eye (Lond). 2010;24(3):422–427. doi:10.1038/eye.2009.309 [CrossRef]
- Bischoff PM, Flower RW. Ten years experience with choroidal angiography using indocyanine green dye: A new routine examination or an epilogue?Doc Ophthalmol. 1985;60(3):235–291. doi:10.1007/BF00157827 [CrossRef]
- Flower RW. Extraction of choriocapillaris hemodynamic data from ICG fluorescence angiograms. Invest Ophthalmol Vis Sci. 1993;34:2720–2729.
- Chalam KV, Sambhav K. Optical coherence tomography angiography in retinal diseases. J Ophthalmic Vis Res. 2016;11:84–92. doi:10.4103/2008-322X.180709 [CrossRef]
- de Carlo TE, Romabo A, Waheed NK, Duker JS. A review of optical coherence tomography angiography (OCTA). Int J Retina Vitreous. 2015;1:5. eCollection 2015. doi:10.1186/s40942-015-0005-8 [CrossRef]
- Ghasemi Falavarjani K, Tian JJ, Akil H, Sadda SR. Swept-source optical coherence tomography angiography of the optic disk in optic neuropathy. Retina. 2016Aug22. [Epub ahead of print] doi:10.1097/IAE.0000000000001259 [CrossRef]
- Falavarjani KG, Sarraf D. Optical coherence tomography angiography of the retina and choroid; Current applications and future directions. J Curr Ophthalmol. 2017;29(1):1–4. doi:10.1016/j.joco.2017.02.005 [CrossRef]
- Al-Sheikh M, Tepelus TC, Nazikyan T, Sadda SR. Repeatability of automated vessel density measurements using optical coherence tomography angiography. Br J Ophthalmol. 2017;101(4):449–4526. doi:10.1136/bjophthalmol-2016-308764 [CrossRef]
- Gadde SG, Anegondi N, Bhanushali D, et al. Quantification of vessel density in retinal optical coherence tomography angiography images using local fractal dimension. Invest Ophthalmol Vis Sci. 2016;57(1):246–252. doi:10.1167/iovs.15-18287 [CrossRef]
- Kuehlewein L, Tepelus TC, An L, Durbin MK, Srinivas S, Sadda SR. Noninvasive visualization and analysis of the human parafoveal capillary network using swept source oct optical microangiography. Invest Ophthalmol Vis Sci. 2015;56(6):3984–3988. doi:10.1167/iovs.15-16510 [CrossRef]
- Shahlaee A, Samara WA, Hsu J, et al. In vivo assessment of macular vascular density in healthy human eyes using optical coherence tomography angiography. Am J Ophthalmol. 2016;165:39–46. doi:10.1016/j.ajo.2016.02.018 [CrossRef]
- Spaide RF. Choriocapillaris flow features follow a power law distribution: Implications for characterization and mechanisms of disease progression. Am J Ophthalmol. 2016;170:58–67. doi:10.1016/j.ajo.2016.07.023 [CrossRef]
- Alten F, Heiduschka P, Clemens CR, Eter N. Exploring choriocapillaris under reticular pseudodrusen using OCT-angiography. Graefes Arch Clin Exp Ophthalmol. 2016;254(11):2165–2173. Epub 2016 May 18. doi:10.1007/s00417-016-3375-1 [CrossRef]
- Choi W, Moult EM, Waheed NK, et al. Ultrahigh-speed, swept-source optical coherence tomography angiography in nonexudative age-related macular degeneration with geographic atrophy. Ophthalmology. 2015;122(12):2532–2544. doi:10.1016/j.ophtha.2015.08.029 [CrossRef]
- Waheed NK, Moult EM, Fujimoto JG, Rosenfeld PJ. Optical coherence tomography angiography of dry age-related macular degeneration. Dev Ophthalmol. 2016;56:91–100. doi:10.1159/000442784 [CrossRef]
- Battaglia Parodi M, Cicinelli MV, Rabiolo A, et al. Vessel density analysis in patients with retinitis pigmentosa by means of optical coherence tomography angiography. Br J Ophthalmol. 2016;101(4):428–432. doi:10.1136/bjophthalmol-2016-308925 [CrossRef]
- de Carlo TE, Adhi M, Salz DA, et al. Analysis of choroidal and retinal vasculature in inherited retinal degenerations using optical coherence tomography angiography. Ophthalmic Surg Lasers Imaging Retina. 2016;47(2):120–127. doi:10.3928/23258160-20160126-04 [CrossRef]
- Pellegrini M, Acquistapace A, Oldani M, et al. Dark atrophy: An optical coherence tomography angiography study. Ophthalmology. 2016;123(9):1879–1886. doi:10.1016/j.ophtha.2016.05.041 [CrossRef]
- Aggarwal K, Agarwal A, Mahajan S, et al. The role of optical coherence tomography angiography in the diagnosis and management of acute vogt-koyanagi-harada disease. Ocul Immunol Inflamm. 2016;1–12. doi:10.1080/09273948.2016.1195001 [CrossRef]
- Puche N, Hera R, Terrada C, Souied EH. Persistent placoid maculopathy imaged by optical coherence tomography angiography. Retin Cases Brief Rep. 2016;10(4):297–301.
- Salvatore S, Steeples LR, Ross AH, Bailey C, Lee RW, Carreno E. Multimodal imaging in acute posterior multifocal placoid pigment epitheliopathy demonstrating obstruction of the choriocapillaris. Ophthalmic Surg Lasers Imaging Retina. 2016;47(7):677–681. doi:10.3928/23258160-20160707-12 [CrossRef]
- Wang Q, Chan S, Yang JY, et al. Vascular density in retina and choriocapillaris as measured by optical coherence tomography angiography. Am J Ophthalmol. 2016;168:95–109. doi:10.1016/j.ajo.2016.05.005 [CrossRef]
- Schneider CA, Rasband WS, Eliceiri KW. NIH image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9(7):671–675. doi:10.1038/nmeth.2089 [CrossRef]
- Al-Sheikh M, Akil H, Pfau M, Sadda SR. Swept-source oct angiography imaging of the foveal avascular zone and macular capillary network density in diabetic retinopathy. Invest Ophthalmol Vis Sci. 2016;57(8):3907–3913. doi:10.1167/iovs.16-19570 [CrossRef]
- Fryczkowski AW, Sherman MD, Walker J. Observations on the lobular organization of the human choriocapillaris. Int Ophthalmol. 1991;15(2):109–120. doi:10.1007/BF00224463 [CrossRef]
- Smith RT, Koniarek JP, Chan J, Nagasaki T, Sparrow JR, Langton K. Autofluorescence characteristics of normal foveas and reconstruction of foveal autofluorescence from limited data subsets. Invest Ophthalmol Vis Sci. 2005;46(8):2940–2946. doi:10.1167/iovs.04-0778 [CrossRef]
- Hayreh SS. In vivo choroidal circulation and its watershed zones. Eye (Lond). 1990;4(Pt 2):273–289. doi:10.1038/eye.1990.39 [CrossRef]
- Hu Z, Nittala MG, Sadda SR. Comparison of retinal layer intensity profiles from different oct devices. Ophthalmic Surg Lasers Imaging Retina. 2013Nov–Dec;44(6 Suppl):S5–10. doi:10.3928/23258160-20131101-02 [CrossRef]
Vessel Density (Ratio) in the Subfoveal Fields and Parafoveal Sectors
|VD%||0.444 ± 0.031||0.474 ± 0.034||0.478 ± 0.036||0.494 ± 0.037||0.487 ± 0.031||< .001|
|P Value||.001||> .05||> .05||> .05||> .05|
Grey Value in the Subfoveal Fields and Parafoveal Sectors
|GV||106.42 ± 4.351||110.26 ± 4.305||110.907 ± 4.910||112.787 ± 5.077||111.67 ± 3.642||< .001|
|P Value||.001||> .05||> .05||> .05||> .05|