Assessment of Choroidal Vascularity and Choriocapillaris Blood Perfusion in Anisomyopic Adults by SS-OCT/OCTA

Abstract

Purpose: To explore the association of choroidal vascularity and choriocapillaris blood perfusion with myopic severity in anisomyopes.

Methods: Refractive error, axial length (AL), and other biometric parameters were measured in 34 anisomyopic young adults. Macular choroidal thickness (ChT) and choroidal vascularity, including total choroidal area (TCA), luminal area (LA), stromal area (SA), and choroidal vascularity index (CVI), were determined from swept-source optical coherence tomography (SS-OCT) vertical and horizontal B-scans. The percentage of choriocapillaris flow voids (FV%) was obtained from en face SS-OCT-angiography.

Results: The spherical equivalent refraction (SER) was –3.35 ± 1.25 diopters in the more myopic eyes and –1.25 ± 1.17 diopters in the less myopic eyes (P < 0.001). The interocular difference in SER was highly correlated with that in AL (P < 0.001). The macular ChT, TCA, LA, and SA were smaller in the more myopic eyes than in the less myopic eyes in both vertical and horizontal scans (all P < 0.001). Importantly, the CVIs in vertical and horizontal scans were smaller and the FV% was greater in the more myopic eyes (P < 0.05). In vertical scans, the interocular difference in CVIs was correlated with that in the SER, AL, and ChT (all P < 0.05). The interocular difference in FV% was correlated with that in SER, AL, and vertical and horizontal ChTs (all P < 0.05).

Conclusions: Choroidal vascularity and choriocapillaris blood perfusion were lower in the more myopic eyes of anisomyopic adults. These changes were correlated with the severity of myopia and choroidal thinning, indicating that choroidal blood flow is disturbed in human myopia.

Myopia is one of the most common visual disorders worldwide. It has been estimated that the global prevalence of myopia will increase from 33.9% in 2020 to 49.8% in 2050, with the prevalence of high myopia increasing from 5.2% to 9.8%.¹ Myopia, particularly high myopia, is characterized by excessive axial elongation and increased risk of developing a series of complications,^2–4 such as degenerative retinopathy, lacquer cracks, choroidal neovascularization, and posterior staphylomas. All of these conditions contribute to irreversible visual impairment, with serious economic and social consequences.⁵ Therefore, it is imperative to study the risk factors for myopia development. 

In recent years, clinical studies using optical coherence tomography (OCT) have demonstrated that choroidal thinning accompanies the development of myopia, and a close association has been established between choroidal thickness changes and ocular growth.^6–12 Observations in animal models of myopia have shown that choroidal thinning occurs early in myopia development, preceding the longer term acceleration of ocular growth, and, conversely, that choroidal thickening accompanies the slowing of ocular growth.^13–16 These bidirectional changes in choroidal thickness may be attributable to changes in choroidal blood flow (ChBF), with a positive correlation between them.^17–19 Considering the highly vascularized nature of the choroid and its potential role in scleral hypoxia,²⁰ we proposed that decreases in ChBF are responsible for the development of human myopia.^19,20 

OCT is a non-invasive, high-resolution, in vivo imaging technique. In combination with OCT-angiography (OCTA), the information of choroidal vasculature and choriocapillaris blood perfusion could be obtained simultaneously for the assessment of ChBF.²¹ This technique has been applied in studies on high or pathologic myopic adults. In those eyes, both choroidal vascular luminal area (LA) and stromal area (SA) were reduced,²² and the area of choriocapillaris flow voids (FVs) was increased, indicating diminished areas with perfusion.²³ However, the accompanying pathological changes, such as extreme choroidal thinning or choroidal atrophy, would make it difficult to determine whether the changes in choroidal parameters are associated with myopia itself or with myopia-related pathological changes. 

Therefore, to clearly define the associations between ChBF and myopia, in the present study we used swept-source (SS)-OCT/OCTA to perform a comprehensive analysis of choroidal vasculature and choriocapillaris blood perfusion in anisomyopic adults who were not highly myopic (spherical equivalent refraction [SER] > –6.00 D) in both eyes. Anisomyopic individuals have an interocular difference in myopic SER of at least 1.00 D. This condition is typically due to an interocular asymmetry in axial length, usually due to differences in vitreous chamber depth (VCD).²⁴ Comparison of the more myopic eye to the less myopic fellow eye, within the same anisomyopic individual, allows for greater control of potentially confounding inter-subject variables such as age, gender, and genetic and environmental factors. Thus, this model should provide increased sensitivity in detecting abnormalities in the variables of interest, and the associations among them, in low to moderate human myopia. 

Methods

Subjects

Thirty-four college student participants, 20 to 27 years old, were recruited from Wenzhou Medical University from September 2019 to July 2020. Written informed consent was obtained from all participants. The study was approved by the ethics committee of the Eye Hospital of Wenzhou Medical University. All participants were treated in accordance with the tenets of the Declaration of Helsinki. 

Ophthalmic screening examinations, including non-cycloplegic subjective refraction, binocular testing, ocular health evaluation, and intraocular pressure measurement, were conducted prior to formal enrollment. All subjects were anisomyopes with an interocular difference in SER of at least 1.0 D. The subjects were free of ocular and systemic disease and had best-corrected visual acuities of 0.00 logMAR or better in each eye. None had a history of ocular surgery, smoking, or systemic diseases. Subjects were asked to avoid caffeine and alcohol intake during the 24 hours before choroidal imaging. 

Ocular Biometric Measurements

Following the screening, the participants were directed to watch a 20-minute video on a television at a distance of 5 meters with their full-distance spectacle corrections to eliminate the effect of any previous visual stimuli on the choroid, such as high accommodation²⁵ or defocus.^26,27 The choroidal images were taken immediately after this period by SS-OCT/OCTA, as detailed below. Ocular biometric parameters, including corneal radius (CR), central corneal thickness (CCT), anterior chamber depth (ACD), lens thickness (LT), and axial length (AL), were measured using the Lenstar LS 900 (Haag Streit, Koeniz, Switzerland). The VCD was calculated as the AL – (CCT + ACD + LT). All of the measurements were conducted from 13:30 to 17:00 to minimize any potential impact of diurnal variation.^28,29 

SS-OCT/OCTA Image Acquisition and Analysis

The SS-OCT/OCTA system (VG200S; SVision Imaging, Henan, China) contained a SS laser with a central wavelength of approximately 1050 nm and a scan rate of 200,000 A-scans per second. The system was equipped with an eye-tracking utility based on an integrated confocal scanning laser ophthalmoscope to eliminate eye-motion artifacts. The axial resolution was 5 µm, and lateral resolution was 13 µm. The scan depth was 3 mm. 

Structural OCT of the macular region was performed with 18 radial scan lines centered on the fovea. Each scan line, generated by 2048 A-scans, was 12 mm long and separated from the adjacent lines by 10°. Sixteen B-scans were obtained on each scan line and were automatedly averaged to improve the signal-to-noise ratio.³⁰ Only the vertical and horizontal scans were used to analyze the choroidal thickness and choroidal vascularity (Fig. 1A). Briefly, the choroid in the SS-OCT images was defined as the area from the retinal pigment epithelium (RPE)–Bruch's membrane complex to the choroid–sclera interface. After semiautomatic choroidal segmentation with a custom algorithm developed in MATLAB R2017a (MathWorks, Natick MA, USA), segmentations of RPE–Bruch's membrane complex and choroid–sclera interface were adjusted manually by a trained examiner (PW). The scan size was adjusted for the differences in magnification due to different ALs among the eyes. After segmentation, each image was binarized using custom-designed algorithms in MATLAB R2017a to demarcate the LA and SA with Niblack's autolocal threshold; this method was first proposed by Sonoda et al.³¹ and further developed by Agrawal et al.³² After image processing, the mean macular choroidal thickness (ChT), total choroidal area (TCA), LA, and SA were calculated. The choroidal vascularity index (CVI) was defined as the ratio of LA to TCA. The 6-mm macular region centered on the fovea was regarded as the region of interest. 

Figure 1.

 

Illustration of choroidal vasculature analysis. (A) Macular ETDRS grid, (B) vertical scan, and (C) horizontal scan. Quadrants: I1, inferior parafovea; I2, inferior perifovea; S1, superior parafovea; S2, superior perifovea; T1, temporal parafovea; T2, temporal perifovea; N1, nasal parafovea; N2, nasal perifovea.

For angiography, the choroidal images were obtained with a raster scan protocol of 512 horizontal B-scans that covered an area of 3 × 3 mm centered on the fovea. Each B-scan contained 512 A-scans and was repeated four times and averaged. The OCTA images were obtained by the SVision SS-OCTA algorithm. We evaluated en face angiograms of the choriocapillaris slab, which was defined by a layer starting at the basal border of the RPE–Bruch's membrane complex and ending at approximately 20 µm beneath the RPE–Bruch's membrane complex (Fig. 2A). A maximum projection was applied on the segmented volumes to generate the en face angiograms. Projection artifacts from retinal vessels were removed by the algorithm. FVs were defined as regions having no flow signals that were detectable by the threshold binarization algorithm, as previously described.³³ The FV percentage (FV%) was calculated by dividing the area of the FVs by the area of the measured region and then converting the value to percent. Because of poor image resolution at the scan edges, only the 2.5-mm-diameter circular region centered on the fovea was used for analysis. 

Figure 2.

 

Illustration of choriocapillaris blood perfusion analysis. (A) Cross-sectional structure combined with flow OCT image with choriocapillaris segmentation lines. (B) OCTA scan region of 3 × 3 mm. (C) Magnified en face OCTA choriocapillaris image. (D) Inverted binarized images of FVs with the 2.5-mm ETDRS grid.

According to the Early Treatment Diabetic Retinopathy Study (ETDRS) grid, the macular zone was divided into regions consisting of three concentric rings with diameters of 1 mm (central fovea, C), 3 mm (parafovea), and 6 mm (perifovea). The parafoveal annulus (designated “1” in Figs. 1B, 1C) and perifoveal annulus (designated “2” in Figs. 1B, 1C) were further divided into superior (S1, S2), inferior (I1, I2), temporal (T1, T2), and nasal (N1, N2) quadrants (Fig. 1). This grid was applied to the vertical and horizontal B-scans (Figs. 1B, 1C), as well as the 2.5-mm-diameter region imaged for choriocapillaris angiography (Figs. 2B–2D). 

To assess and affirm agreement within and between examiners, 22 eyes from 11 participants were selected, and the choroids were segmented twice by two trained examiners (PW and ZX). Intraclass correlation coefficients (ICCs) and coefficients of repeatability were calculated to assess the agreement of ChT, LA, SA, TCA, and CVI measurements within and between examiners at various regions. The coefficient of repeatability was calculated as 1.96 times the standard deviation of the differences between two measurements. After obtaining good agreement, all the scans were measured by examiner PW. 

Statistics

The statistical analyses were performed using SPSS Statistics 23.0 (IBM, Armonk, NY, USA). The means and standard deviations of all continuous variables are presented unless otherwise stated. The normality of data was examined by the Shapiro–Wilk test. Paired t-tests or Wilcoxon signed-rank tests were used to assess the interocular differences between the fellow eyes for ocular biometrics, choroidal structural measurements, and choriocapillaris FV measurements. For topographic analysis to compare choroidal parameters between the fellow eyes, two-way repeated-measures (RM) ANOVAs were performed, including analysis for two within-subject factors (eyes and regions). The Greenhouse–Geisser correction was applied to the degrees of freedom when the sphericity assumption was violated. Bonferroni adjustments for multiple comparisons were applied to all post hoc pairwise comparisons. Pearson's or Spearman's correlation was used to calculate the degree and statistical significance of associations between variables wherever appropriate. 

Results

General Characteristics

Of the 34 participants included in the analyses, 10 were males and 24 were females, and the mean age was 23.4 ± 2.0 years old. The VCD for one subject was excluded from the final analysis because the measurement from the Lenstar A-scan did not exhibit a consistent visible peak from the posterior crystalline lens surface. 

The SER of the more myopic eyes was –3.35 ± 1.25 D, and that of the less myopic eyes was –1.25 ± 1.17 D (P < 0.001) (Table 1). The AL in the more myopic eyes (25.07 ± 0.94 mm) was significantly longer than that in the less myopic eyes (24.23 ± 0.71 mm) (P < 0.001). The difference in AL was attributed mostly to the difference in VCD, which was 17.80 ± 0.96 mm in the more myopic eyes compared to 16.95 ± 0.72 mm in the less myopic eyes (P < 0.001). As expected, the magnitude of SER anisometropia was highly correlated with the interocular differences in VCD and AL (Spearman’s correlation, r = –0.906 and r = –0.927 respectively; both P < 0.001). 

Table 1.

 

Ocular Biometric Parameters of the Anisomyopic Participants

 

Repeatability of Choroidal Structural Measurement

The ICC values of global and topographical ChT, LA, SA, TCA, and CVI varied from 0.89 to 0.99 for intra-examiner repeatability in vertical and horizontal scans (Supplementary Tables S1, S2). In the vertical and horizontal scans, the coefficients of repeatability within examiner varied from 10 to 32 µm for ChT, from 0.00 to 0.03 mm² for LA, from 0.01 to 0.07 mm² for SA, from 0.02 to 0.09 mm² for TCA, and from 1.9% to 5.0% for the CVI (Supplementary Tables S3, S4). Considering the means of these choroidal parameters, the repeatability of manual correction was good. The agreement between the examiners, as assessed by ICCs and coefficients of repeatability, was also good (Supplementary Tables S5–S8). 

Global Analyses of Choroidal Vascularity and Choriocapillaris FVs

Both the vertical and the horizontal mean macular ChTs were significantly less in the more myopic eyes than in the less myopic eyes (both P < 0.001) (Table 2). There were also significant reductions in the vertical and horizontal LAs, SAs, and TCAs in the more myopic eyes compared to the less myopic eyes (all P < 0.001). Importantly, the CVIs of both the vertical and horizontal directions were also less in the more myopic eyes (both P < 0.01), and the percentage of FVs was greater in the more myopic eyes, as well (P < 0.05). 

Table 2.

 

Global Choroidal Parameters in the More Myopic and Less Myopic Eyes of the Anisomyopic Participants (n = 34)

 

The interocular differences in LAs, SAs, and TCAs of both vertical and horizontal scans were significantly correlated with the interocular differences in SER, AL, and VCD (all P < 0.01) (Table 3). Additionally, the interocular differences in the CVIs of vertical scans were significantly correlated with interocular differences in SER, AL, and VCD (all P < 0.05). There were also significant correlations between the interocular differences in FVs with SER, AL, and VCD (all P < 0.01). Furthermore, the interocular differences in ChTs were positively correlated with CVIs, LAs, SAs, and TCAs in both vertical and horizontal scans (all P < 0.05), and the interocular differences in FVs were negatively correlated with ChTs in the vertical and horizontal scans (both P < 0.05). 

Table 3.

 

Correlations Among the Interocular Differences in SER, AL, VCD, and Choroidal Parameters

 

Topographic Analyses of Choroidal Vascularity and Choriocapillaris FVs

For vertical scans, the interaction effect of eyes and regions on ChT was significant (eyes × regions, F = 3.88, P < 0.05), as was the main effect of eyes on ChT (eyes, F = 31.57, P < 0.001). In vertical scans, the choroids in all regions were thinner in the more myopic eyes than in the less myopic eyes (all P < 0.01) (Fig. 3A). In horizontal scans, the main effects of eyes and regions on ChT were both significant (eyes, F = 28.63, P < 0.001; regions, F = 102.28, P < 0.001), whereas the interaction effect of eyes and regions on ChT was not significant. In all regions of horizontal scans, ChT was also significantly less in more myopic eyes than in less myopic eyes (all P < 0.01) (Fig. 3B). 

Figure 3.

 

ChT topography in fellow eyes of anisomyopic subjects (n = 34). (A) ChT in vertical B-scans, (B) ChT in horizontal B-scans. Data are expressed as means and 95% confidence intervals. Two-way RM ANOVA; ^**P < 0.01 and ^***P < 0.001 indicate significant differences between the fellow eyes.

Because the widths of the nine choroidal regions were different from one another (Fig. 1), the areas among them were not comparable; therefore, only the topographic areas between the fellow eyes were analyzed. The main effects of eyes on LA, SA, and TCA were all significant, in both the vertical scans (LA, F = 30.74, P < 0.001; SA, F = 26.03, P < 0.001; TCA, F = 31.70, P < 0.001) and horizontal scans (LA, F = 31.92, P < 0.001; SA, F = 18.32, P < 0.001; TCA, F = 29.16, P < 0.001). The LA (Figs. 4A, 4B), SA (Figs. 4C, 4D), and TCA (Figs. 4E, 4F) were significantly smaller in the more myopic eyes than in the less myopic eyes, in each region of both vertical and horizontal scans (all P < 0.05). 

Figure 4.

 

Topography of choroidal vascular components in fellow eyes of anisomyopic subjects (n = 34). (A) LA in vertical B-scans, (B) LA in horizontal B-scans, (C) SA in vertical B-scans, (D) SA in horizontal B-scans, (E) TCA in vertical B-scans, and (F) TCA in horizontal B-scans. Data are expressed as means and 95% confidence intervals. Two-way RM ANOVA; ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001 indicate significant differences between the fellow eyes.

Importantly, the main effect of eyes on the CVI was significant in both vertical (F = 8.31, P < 0.01) and horizontal scans (F = 13.94, P < 0.001). The CVIs in the vertical scans of the I2 region (Fig. 5A) and in the horizontal scans of the T2, T1, C, and N2 regions (Fig. 5B) were significantly smaller in the more myopic eyes than in the less myopic eyes (all P < 0.05). However, there were no significant interaction effects of eyes and regions for the CVI, in vertical or horizontal scans. 

Figure 5.

 

Topography of CVI in fellow eyes of anisomyopic subjects (n = 34). (A) CVI in vertical B-scans, and (B) CVI in horizontal B-scans. Data are expressed as means and 95% confidence intervals. Two-way RM ANOVA; ^*P < 0.05 and ^**P < 0.01 indicate significant differences between the fellow eyes.

The main effects of eyes and regions on the choriocapillaris FV% were both significant (eyes, F = 5.88, P < 0.05; regions, F = 28.02, P < 0.001), whereas the interaction effect of eyes and regions on FV% was not significant. In the T1 and I1 regions, FV% was greater in the more myopic eyes than in the less myopic eyes (P < 0.01) (Fig. 6). 

Figure 6.

 

Topography of FVs in fellow eyes of anisomyopic subjects (n = 34). Data are expressed as means and 95% confidence intervals. Two-way RM ANOVA; ^*P < 0.05 and ^**P < 0.01 indicate significant differences between the fellow eyes.

Discussion

The nature of the association between ChBF and myopia remains an urgent question. In this study, we performed comprehensive analyses of the choroidal vasculature and choriocapillaris blood perfusion in the disparate eyes of anisomyopic adults to test for an association between ChBF and myopia. The results showed that choroidal vascularity, assessed by SA, LA, TCA, and CVI, was lower in the more myopic eyes than in the less myopic eyes of anisomyopic subjects, whereas the choriocapillaris FV% was higher. Moreover, there were correlations between most of the choroidal parameters and severity of myopia, as indicated by SER, AL, and VCD. These findings indicate greater reductions in ChBF with greater degrees of myopia. In addition, the decrease in ChT was positively correlated with the decrease in choroidal vascularity and increases of choriocapillaris FVs, suggesting that the thinning of the ChT in myopia is associated with decreased ChBF. 

Physiological Meaning of the Choroidal Parameters in SS-OCT/OCTA Images

The choroid consists of three vascular layers, designated (from the retinal side to the scleral side) as choriocapillaris, Sattler's layer, and Haller's layer. Sattler's layer contains medium to small vessels supplying discrete hexagonal lobules of capillaries in the choriocapillaris, and Haller's layer contains large vessels, providing the main arterial supply and venous drainage for the choroid.³⁴ Previous spectral domain-OCT^35,36 and SS-OCT^37,38 structural scans visualized vessel-like structures with low signal intensity and identified them as blood vessels. A limitation of this method is that, although medium- to large-size vessels could be visualized in that way, the visualization of small vessels was not possible because of limited lateral resolution and lack of contrast.^35,39–41 In OCTA en face images, the FVs, which appear as small dark regions, seem to represent the intercapillary spaces that are strikingly similar to those seen in morphological and histological images.^42,43 However, mainly because of signal attenuation attributed to scattering by RPE and choriocapillaris, OCTA cannot depict large vessels deep in the choroid. 

To delineate the global appearance of choroidal circulation, we comprehensively analyzed the choroidal vascularity in structural OCT images and choriocapillaris FVs in OCTA images. The former reflects the circulatory volume in Haller's and Sattler's layers, and the latter reflect areas in the choriocapillaris with low perfusion or without perfusion, thus outlining the global choroidal circulation. These methods have shown good repeatability and reproducibility in studies by others, ^{32,33,42,44,45} as well as by us in the present work. 

Gupta et al.²² first reported reductions in choroidal LA and SA along with thinning of the ChT in high myopic eyes. Recently, Li et al.⁴⁵ used the same methodology in a population of low to moderately myopic children and reported negative correlations of LA with AL, but not with SER. On the other hand, several studies reported increased FVs in high myopic eyes, suggesting the presence of decreased choriocapillaris blood perfusion,^23,46 and Su et al.⁴⁶ did not find changes of FVs in moderately myopic eyes. By utilizing the special design of internal self-control in anisomyopes, we found that reductions in choroidal vasculature and choriocapillaris blood perfusion were positively correlated with the severity of myopia, as indicated by SER, AL, and VCD. Taken together, these studies suggest the presence of decreased choroidal circulation in human myopia. 

It is noteworthy that topographical changes of CVIs and FVs in more myopic eyes were prominent in the temporal and inferior macular regions. A recent study reported that a strong accommodation stimulus (+6.00 D) could elicit a non-uniform reduction in choroidal thickness, with most prominent thinning in the temporal, inferior, and inferotemporal regions²⁵; similar topographical variations were also observed in children after 3 weeks of orthokeratology treatment.⁴⁷ Nevertheless, only limited information has been obtained from the more peripheral choroid. Hoseini-Yazdi et al.,⁴⁸ working with a small sample size, reported that the magnitude of choroidal thinning was largest in the foveal region, diminishing gradually toward the more peripheral regions. In addition, studies in chicks showed greater choroidal thickening in the central region than in the peripheral region in response to myopic defocus.⁴⁹ These findings indicate that the central retinal (including macular) regions are particularly responsive to visual stimulus conditions, leading to a decrease in choroidal vascularity and choriocapillaris blood perfusion during myopia development. 

Role of Choroidal Structure and Blood Flow During Myopia Development

In the past decade, OCT studies have clearly shown a close link between myopia and ChT⁵⁰ and that the ChT in myopic and high myopic eyes is thinner than in emmetropic and hyperopic eyes.^6–9 Recent longitudinal studies have further confirmed this link, showing that choroidal thinning is associated with accelerated ocular growth or myopia development in children and adolescents.^10–12 Our ChT measurements are consistent with those results. Recent studies found that the choroidal thinning in myopic eyes is mainly attributed to losses in Haller's and Sattler's layers.^51,52 The reductions in these two layers, in combination with the choroidal thinning and reduced choroidal vascularity that we found here, indicate that choroidal blood volume is reduced in myopic eyes. 

Because the choriocapillaris is supplied by Sattler's and Haller's layers, changes in blood flow in these layers can be expected to disturb choriocapillaris blood perfusion. We observed an increase of choriocapillaris FVs in the more myopic eyes as compared to the less myopic eyes, and we found correlations of interocular differences in FVs with SER, AL, and ChT in both vertical and horizontal scans; other studies have also reported increased FVs in high myopic eyes.^23,46 Moreover, estimates of choroidal circulation, based on pulsatile ocular blood flow and ocular pulsation amplitude, were significantly lower in high myopes.^53–55 In addition, fundus indocyanine green angiography showed delayed choroidal filling in high myopic eyes.⁵⁶ Collectively, these studies indicate a close association between decreased ChBF and myopia severity in humans. 

A key remaining unresolved need is to determine the cause-and-effect relationship between ChBF and myopia—that is, whether decreased ChBF causes myopia development, or vice versa. Early studies in form-deprived myopic chicks observed reductions in ChBF.¹⁷ Upon removal of the form-deprivation diffuser, ChBF recovered within hours,¹⁸ which is much more rapid than the alterations in scleral remodeling and ocular elongation. The bidirectional response of ChBF to visual cues, which determines the differential ocular elongation rate, was also demonstrated in our recent study using OCT/OCTA technology.¹⁹ ChBF and ChT in guinea pigs were decreased after induction of myopia by form-deprivation or hyperopic defocus, and they recovered after removal of these altered optical conditions. Of note, actively increasing ChBF with vasodilator prazosin inhibited myopia progression in guinea pigs, as well as axial elongation and scleral hypoxia.⁵⁷ We have argued that a decrease in ChBF may cause scleral hypoxia followed by axial elongation and myopia.^20,58 Because the choroidal thickness in humans could also respond rapidly and predictably to various visual stimuli (e.g., hyperopic defocus, myopic defocus, and accommodation),^25–27 it is plausible that decreased ChBF is a risk factor for the development of human myopia. 

Limitations of the Current Study

Due to the limitations of cross-sectional design, the current study can only further strengthen the association between the reduction in ChBF and myopia development. To test the hypothesis that decreased ChBF is a risk factor for the clinical development of myopia, longitudinal studies are needed. Another limitation of this study is that our evaluation of choriocapillaris FVs was based on a 20-µm slab of imaged tissue, the exact location of which is likely to vary among individuals. A third limitation is that the main factor influencing the calculation of CVIs is image quality, which determines the ability to identify the choroid–scleral interface. The relatively high resolution and contrast of our images yielded good repeatability for choroidal segmentation. Although we attempted to evaluate the topographic characteristics of the alterations in choroidal vascularity by two cross-sectional B-scans, utilization of three-dimensional volumetric OCT data with automated segmentation would allow for more detailed delineation of the choroidal vasculature at different regions and sublayers. 

Conclusions

In conclusion, not only choroidal thickness but also choroidal vascularity and choriocapillaris blood perfusion were lower in the more myopic eyes than in the less myopic eyes of anisomyopic adults. Reduction of the CVI and increases of choriocapillaris FVs correlated positively with the severity of myopia, as well as with choroidal thinning, indicating that ChBF is disturbed in human myopia. Determining the possible causal nature of these associations will require longitudinal research on school-age children, as well as further research in animal models. 

Acknowledgments

The authors thank Yue Liu (Center for Eye Disease & Development, School of Optometry, University of California, Berkeley, CA, USA) and William K. Stell (Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada) for helping with the data analysis and providing editorial support for improving the manuscript. 

Supported by grants from the National Natural Science Foundation of China (81830027, 81970833, 81670886, 82000931), National Key Research and Development Program of China (2019YFC1710204), and CAMS Innovation Fund for Medical Sciences (2019-I2M-5-048). 

Disclosure: H. Wu, None; G. Zhang, None; M. Shen, None; R. Xu, None; P. Wang, None; Z. Guan, None; Z. Xie, None; Z. Jin, None; S. Chen, None; X. Mao, None; J. Qu, None; X. Zhou, None