Longitudinal Changes in Choroidal Structure Following Repeated Low-Level Red-Light Therapy for Myopia Control: Secondary Analysis of a Randomized Controlled Trial

Abstract

Purpose: 

Repeated low-level red-light (RLRL) therapy has been confirmed as a novel intervention for myopia control in children. This study aims to investigate longitudinal changes in choroidal structure in myopic children following 12-month RLRL treatment.

Materials and Methods: 

The current study is a secondary analysis from a multicenter, randomized controlled trial (NCT04073238). Choroidal parameters were derived from baseline and follow-up swept-source optical coherence tomography scans taken at 1, 3, 6, and 12 months. These parameters included the luminal area (LA), stromal area (SA), total choroidal area (TCA; a combination of LA and SA), and choroidal vascularity index (CVI; ratio of LA to TCA), which were automatically measured by a validated custom choroidal structure assessment tool.

Results: 

A total of 143 children (88.3% of all participants) with sufficient image quality were included in the analysis (n=67 in the RLRL and n=76 in the control groups). At the 12-month visit, all choroidal parameters increased in the RLRL group, with changes from baseline of 11.70×10³ μm² (95% CI: 4.14–19.26×10³ μm²), 3.92×10³ μm² (95% CI: 0.56–7.27×10³ μm²), 15.61×10³ μm² (95% CI: 5.02–26.20×10³ μm²), and 0.21% (95% CI: –0.09% to 0.51%) for LA, SA, TCA, and CVI, respectively, whereas these parameters reduced in the control group.

Conclusions: 

Following RLRL therapy, the choroidal thickening was found to be accompanied by increases in both the vessel LA and SA, with the increase in LA being greater than that of SA. In the control group, with myopia progression, both the LA and SA decreased over time.

INTRODUCTION

Myopia is the most common ocular disease with a growing prevalence globally, particularly in East and Southeast Asia.^1–4 It not only places a significant burden on individuals and communities but also raises further vision concerns. Excessive elongation of axial length (AL) associated with high myopia can increase the risk of serious sight-threatening complications, such as myopic maculopathy and retinal detachment.^3,5,6

The choroid, a highly vascular tissue between the retina and sclera, is believed to play an important role in the regulation of eye growth.⁷ Choroidal thinning has been shown to be significantly associated with myopia progression.^8–11 Furthermore, choroidal thickening has been observed in myopic children following myopia control treatment such as low-dose atropine,^12,13 orthokeratology,^14–16 and MiSight contact lenses.¹⁷ However, the contribution of the vascular lumen area and stromal area in the observed choroidal thickening is unknown in part due to the difficulties in their measurements.^18,19 We developed a fully automated and efficient deep learning–based choroidal structure assessment program with proven accuracy in comparison with manual measurement tool.²⁰

Repeated low-level red-light (RLRL) therapy has been confirmed as an effective and safe intervention for myopia control in children.^21–27 In addition, a clinically significant subfoveal choroidal thickening has been noted following RLRL therapy,^21–24,28 with the magnitude of the thickening associated with RLRL therapy’s efficacy in myopia control.²⁸ Nevertheless, detailed vascular and morphometric changes in the choroid following RLRL therapy remain unknown.

In this study, we conducted a secondary analysis of the optical coherence tomography (OCT) scans obtained from our prior multicenter randomized clinical trial (RCT).²¹ Our aim was to compare longitudinal changes in choroidal structure following RLRL therapy and in children with single-vision spectacle (SVS) treatment only.

MATERIALS AND METHODS

Study Participants and Procedures

The current study was based on a multicenter RCT that assessed the safety and effectiveness of RLRL therapy in myopia control in children.²¹ The study design was described elsewhere in detail.²¹ Briefly, the inclusion criteria included age of 8 to 13 years, cycloplegic spherical equivalent refraction (SER) of –1.00 to –5.00 D, astigmatism ≤2.50 D, anisometropia ≤1.50 D, and best-corrected visual acuity ≥0.0 logarithm of the minimum angle of resolution (logMAR) in both eyes. Children with ocular disorders, systemic diseases, or a history of myopia control interventions (eg, atropine, orthokeratology lens) were excluded. A total of 264 eligible children from 5 study sites in China were recruited and randomized to either the intervention group with RLRL therapy plus SVS or the control group with SVS only. Comprehensive ophthalmic examinations and detailed questionnaires were conducted at baseline and the 1-, 3-, 6-, and 12-month follow-up visits. Children in the intervention group received RLRL therapy (Eyerising; Suzhou Xuanjia Optoelectronics Technology) twice per day and 5 days per week, with an interval of at least 4 hours. Among them, 162 from 2 study centers, where a swept-source optical coherence tomography (SSOCT) device was available, were included into the current analysis.

This study was registered with ClinicalTrials.gov (identifier, NCT04073238). The protocol was approved by the Institutional Review Board of Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China (identifier, 2019KYPJ093). Written informed consents were obtained from guardians before their children’s participation, and the study adhered to the tenets of the Declaration of Helsinki.

Ophthalmic Examinations

All children underwent comprehensive ophthalmic examination at baseline and all follow-up visits. AL was measured by partial coherence interferometry with the IOLMaster before cycloplegia (Carl Zeiss 500; Meditec) as the primary outcome of myopia control. Cycloplegic SER was performed using an autorefractor (KR-8800; Topcon) and was calculated as the sum of spherical value and half cylinder value.

The SSOCT (DRI OCT Triton; Topcon Corporation), utilizing a 1050 nm wavelength light source and a scanning speed of 100,000 A-scans per second, was used for choroidal imaging. The scanning procedure and instrument parameters have been previously described.^21,28–30 In brief, at each follow-up visit, a 12-line radial scan pattern (12×12 mm) centered on the fovea and with an interscan angle of 15 degrees, was employed to capture a total of 12 OCT images per eye, each with a different scanning direction. All OCT scans were performed with children’s pupils dilated and under standardized mesopic lighting conditions.

Image Analysis

Children with OCT scans’ quality score >60 were included in the final analysis. Choroidal parameters^31,32 including luminal area (LA), stromal area (SA), total choroidal area (TCA; a combination of LA and SA), and choroidal vascularity index (CVI; the ratio of LA to TCA) within a 1500-μm-width region centered on the fovea were measured by a fully automated deep learning–based choroidal assessment program which was developed and validated by our research group (Fig. 1). Details of the program has been described elsewhere in our previous publication.²⁰ A validation study has shown that the automatic measurements of LA, SA, TCA, and CVI are highly correlated with manual measurements and have reduced intraobserver, interobserver, and intersession variations to a small extent.²⁰

FIGURE 1: A representative raw swept-source optical coherence tomography image and converted binary image. A raw swept-source optical coherence tomography image (A) was converted to binary image (B) using the deep learning–based choroidal structure assessment program automatically. The luminal area (dark pixels, closed arrowhead) and the stromal area (light pixels, open arrowhead) were indicated, respectively. C, Same image as in (A), but with overlaid annotations. Point A indicated the fovea center, and point B indicated the point on the choroidal inner boundary nearest to the fovea center. Lines 1 to 4 indicated the tangent of inner boundary at point B (line 1), the perpendicular line (line 2) of line 1 at point B, and boundaries on both sides (lines 3 and 4) of the 1500-μm-width region centered on the fovea, respectively. A 1500-μm-width choroidal region of interest (closed arrowhead) and choroidal boundaries (open arrowheads) were displayed (C). ROI indicates region of interest.

Statistical Analyses

Children who had qualified OCT scans were included in the analysis. Given the strong correlations between right and left eyes, only right eye per participant was used for statistical analysis. If OCT scans of the right eye were suboptimal or missing, OCT scans of the left eye were used alternatively (n=4).

Longitudinal mixed models were used to assess choroidal structure changes on multiple follow-up time points. An unstructured covariance matrix was utilized along with a restricted maximum likelihood method. The group, visit, and group-by-visit interaction were included as fixed effects in the models with baseline age, sex, and choroidal parameters as covariates. The participants were included as a random factor. The estimated mean differences, corresponding 95% CIs, and 2-sided P values were calculated. To ensure the accuracy of the results, only the SER data with full cycloplegia were included in this analysis. Statistical analyses were performed using Stata, Version 17.0 (Stata Corporation) and Prism Version 9.0 (GraphPad Prism). P values <0.05 were considered statistically significant.

RESULTS

Of 162 children with available OCT data, 19 were excluded due to missing follow-up visits (n=11) or suboptimal image quality (n=8). Finally, OCT scans of 143 children were analyzed, including 67 being in the RLRL group and 76 in the SVS group (Supplementary Digital Content Fig. 1, https://links.lww.com/APJO/A245). The differences between baseline age, sex, uncorrected visual acuity (UCVA), AL, SER, and choroidal parameters in included participants and in those participants excluded from the current analysis were nonsignificant (Supplementary Digital Content Table 1, https://links.lww.com/APJO/A246).

Baseline Characteristics of Included Participants

Baseline demographics, UCVA, AL, SER, and choroidal parameters were similar between the RLRL and SVS groups (Table 1).

TABLE 1 - Baseline Characteristics of Participants Between the RLRL and SVS Groups

Characteristics	RLRL group (N=67)	SVS group (N=76)	P
Age (y)	10.56±1.52	10.30±1.60	0.324
Sex
Female	33 (49.25)	39 (51.32)	0.806
Male	34 (50.75)	37 (48.68)
UCVA (logMAR)	0.23±0.13	0.25±0.16	0.479
AL (mm)	24.52±0.67	24.65±0.84	0.330
SER (D)	–2.35±0.82	–2.65±1.12	0.080
Choroidal parameters
LA (×103 μm2)	263.20±78.70	264.97±62.00	0.881
SA (×103 μm2)	139.58±23.26	142.92±24.08	0.401
TCA (×103 μm2)	402.77±95.14	407.89±81.09	0.729
CVI (%)	64.70±3.48	64.62±2.77	0.886

AL indicates axial length; CVI, choroidal vascularity index; D, diopter; LA, luminal area; logMAR, logarithm of the minimum angle of resolution; RLRL, repeated low-level red-light; SA, stromal area; SER, spherical equivalent refraction; SVS, single-vision spectacle; TCA, total choroidal area; UCVA, uncorrected visual acuity.

Data are presented as mean±SD or n (%).

P- values were calculated based on unpaired t test for continuous data and χ² test for categorical data.

Longitudinal Changes in Choroidal Structures

Figure 2 and Table 2 summarizes the longitudinal changes of LA, SA, TCA, and CVI in the intervention and control groups. All choroidal parameters including LA, SA, TCA, and CVI increased substantially by 11.70×10³ μm² (95% CI: 4.14–19.26×10³ μm²), 3.92×10³ μm² (95% CI: 0.56–7.27×10³ μm²), 15.61×10³ μm² (95% CI: 5.02–26.20×10³ μm²), and 0.21% (95% CI: –0.09% to 0.51%), respectively, in comparison with baseline in the RLRL group at 12-month visit. In the opposite, these measurements reduced in the SVS group at a 12-month visit by 18.78×10³ μm² (95% CI: –26.16 to –11.40×10³ μm²), 7.29×10³ μm² (95% CI: –10.58 to –4.00×10³ μm²), 26.03×10³ μm² (95% CI: –36.37 to –15.68×10³ μm²), and 0.47% (95% CI: –0.76 to –0.17%) for LA, SA, TCA, and CVI, respectively, in comparison with baseline.

FIGURE 2: Line graphs showing the adjusted mean changes in LA (A), SA (B), TCA (C), and CVI (D) from baseline to 12 months at each follow-up time point between the repeated low-level red-light group and single-vision spectacle group. CVI indicates choroidal vascularity index; LA, luminal area; SA, stromal area; TCA, total choroidal area.

TABLE 2 - Cumulative Adjusted Mean Changes in LA, SA, TCA, and CVI From Baseline to 12 Months at Each Time Point Between the RLRL and SVS Groups

	Cumulative adjusted [mean (95% CI)]
Time/visit	RLRL group (N=67)	SVS group (N=76)	Mean difference (95% CI)
Change of LA (×103 μm2)
1 mo	18.79 (13.52–24.06)	–4.12 (–9.34 to 1.11)	22.90 (15.45–30.35)
3 mo	10.04 (4.03–16.05)	–12.88 (–18.87 to –6.88)	22.92 (14.39 to 31.45)
6 mo	1.55 (–5.84 to 8.95)	–19.95 (–28.02 to –11.87)	21.50 (10.52 to 32.48)
12 mo	11.70 (4.14–19.26)	–18.78 (–26.16 to –11.40)	30.48 (19.89 to 41.07)
Change of SA (×103 μm2)
1 mo	6.36 (4.16–8.55)	–1.26 (–3.44 to 0.91)	7.62 (4.51–10.73)
3 mo	0.96 (–1.77 to 3.70)	–7.11 (–9.86 to –4.36)	8.08 (4.18–11.97)
6 mo	–1.37 (–4.98 to 2.23)	–7.06 (–11.09 to –3.03)	5.69 (0.27–11.11)
12 mo	3.92 (0.56–7.27)	–7.29 (–10.58 to –4.00)	11.21 (6.50–15.92)
Change of TCA (×103 μm2)
1 mo	25.24 (18.18–32.31)	–5.47 (–12.48 to 1.54)	30.72 (20.72–40.71)
3 mo	11.01 (2.83–19.19)	–19.90 (–28.06 to –11.74)	30.92 (19.31–42.52)
6 mo	–0.13 (–10.74 to 10.47)	–27.05 (–38.64 to –15.46)	26.91 (11.16–42.67)
12 mo	15.61 (5.02–26.20)	–26.03 (–36.37 to –15.68)	41.63 (26.79–56.47)
Change of CVI (%)
1 mo	0.48 (0.24–0.73)	–0.10 (–0.34 to 0.14)	0.58 (0.23–0.93)
3 mo	0.61 (0.27–0.95)	0.01 (–0.33 to 0.36)	0.59 (0.11–1.08)
6 mo	0.26 (–0.08 to 0.60)	–0.61 (–1.00 to –0.23)	0.87 (0.35–1.38)
12 mo	0.21 (–0.09 to 0.51)	–0.47 (–0.76 to –0.17)	0.68 (0.26–1.10)

CVI indicates choroidal vascularity index; LA, luminal area; RLRL, repeated low-level red-light; SA, stromal area; SVS, single-vision spectacle; TCA, total choroidal area.

Mean difference=change of variables in the RLRL group−change of variables in the SVS group.

Longitudinal mixed models were used to calculate the mean change of choroidal parameters at each time point to demonstrate treatment effects on choroidal structure.

Representative SSOCT images from children with and without RLRL therapy were shown in Supplementary Digital Content Figure 2, https://links.lww.com/APJO/A247), along with the 1500-µm width region of interest overlaid on corresponding binarized images.

DISCUSSION

This study reported longitudinal changes in choroidal structure, including LA and SA, among Chinese myopic children, following RLRL therapy in comparison with SVS treatment only. The observed subfoveal choroidal thickening following RLRL therapy is attributable to an increase in both the LA and SA, and CVI also increased, indicating the increase in the LA is slightly higher than that of the stromal area. Meanwhile, choroidal thinning in the control group along with the myopia progression is attributable to reductions in both the LA and SA, and CVI also decreased, indicating the reduction in the LA is slightly higher than that of the stromal area.

Choroidal thickening has been consistently reported in myopic children following RLRL therapy by our research group and others.^21–24,28 However, detailed structural changes underlying choroidal thickening remains unknown. The current study demonstrates that for the RLRL group, the LA, SA, and TCA consistently increased over time except at the 6-month visit. The lack of increase in LA, SA, and TCA at the 6-month visit is not fully understood, but the outbreak of the coronavirus disease 2019 (COVID-19) pandemic may have played a role. The COVID-19 home quarantine contributed to a rise in near work and a decrease in outdoor activities,³³ both of which are known risk factors for myopic progression.^34,35 The current study's findings are in agreement with previous research demonstrating that myopic progression accelerates after the onset of COVID-19 compared with before the pandemic,^36–39 even in children who have been treated with low-concentration atropine.³⁷ Besides, during the COVID-19 pandemic, children who were aware of the progression of their myopia were more inclined to revisit the hospital for follow-up examinations, whereas children who have had successful treatments and their parents may prioritize safety factors and opt not to return for follow-up visits. In terms of the proportional increase in choroidal structure, we demonstrate 11.54% increase in LA, 7.91% increase in SA, 10.26% increase in TCA, and 1.05% increase in CVI over 12 months in the RLRL group as compared with the control group. Given the fact that previous studies have demonstrated that the choroid was 16% thinner among the patients who developed diabetic retinopathy in comparison with who did not in a 2-year follow-up study⁴⁰ and 32% thinner among those who developed macular atrophy in comparison with who did not among treatment-naive neovascular age-related macular degeneration patients in an 18-month follow-up study,⁴¹ an overall 10% increase in LA and TCA observed after RLRL therapy would likely provide significant clinical benefits to other ocular diseases such as diabetic retinopathy and age-related macular degeneration if a similar level of increase in choroidal thickness and lumen area can be achieved in both children and adults.

Although the results in the current and other studies^21–24,28 show the efficacy of RLRL therapy in controlling myopia progression, the molecular and cellular mechanisms involved in the relevant interactions of red light with ocular tissues remain unknown. Similarly, possible retinal and choroidal pathways leading to vessel dilation are not clear and these issues are being investigated by our research group.

Choroidal thickening has also been observed in myopic children who have undergone several other myopia control strategies, including 0.05%¹² or 1%¹³ atropine eye drops, orthokeratology,^14–16 and MiSight contact lenses.¹⁷ However, information about the involvement of the choroidal luminal and stromal components in the observed choroidal thickening is limited, partly due to the challenges associated with analyzing the intricate choroidal structure.^18,19 As far as we know, there are no published RCTs to compare our findings on the detailed choroidal structure changes after RLRL therapy. A previous RCT study reported that the administration of 1% atropine to myopic children for 6 months resulted in an increase in both the LA and SA.²⁹ Similar to this finding, the current study observed an increase in both the LA and SA following 12-month RLRL therapy. These results suggest that both choroidal vascular and stromal components may play a vital role in myopia control and require further investigation.

In the current study, both the RLRL and control groups observed choroidal SA changes over time. The choroidal stroma is the connective tissue composed of collagen, elastin, melanocytes, fibroblasts, resident macrophages, dendritic cells, neural tissues, and other extracellular components.^7,42 The pathophysiologic mechanism of the stromal changes is largely unknown, for example, whether the observed increase in SA following RLRL therapy is secondary to fluid accumulation within the choroidal stroma, or it is secondary to infiltration by cells or fiber proliferation, remains unknown. Early observations found that the choroidal thickness could return to baseline levels or even lower.⁴³ These findings suggested that the observed increase in SA following RLRL therapy was likely due to fluid accumulation resulting from increased choroidal vascular permeability, as seen in the choroidal thickening observed in the form-deprivation recovery model in chicks.⁴⁴

Decreased choroidal thickness has been found in myopic eyes,^8–11 illustrating its potential role in providing a sensitive and “rapid predictive index” of myopia onset. The choroid provides oxygen and nourishment to the sclera and changes in scleral structure or morphology lead to alterations in eye size and refraction directly.⁷ Therefore, previous studies inferred that lower oxygen levels and nutrient supply of the sclera secondary to reduced choroidal blood perfusion may be associated with myopia progression.⁴⁵ However, longitudinal choroidal structural changes accompanied with myopia progression have not been fully established. Some cross-sectional studies investigated the choroidal structural changes in myopic eyes and they found that smaller LA, SA, and TCA were associated with a higher degree of myopia, whereas the change of CVI was not conclusive (Supplementary Digital Content Table 2, https://links.lww.com/APJO/A255).^46–53 The reason for this discrepancy still needs further investigation, but it should be noted that longitudinal changes in choroidal parameters cannot be ascertained owing to the cross-sectional nature of these studies.^46–53 The primary study, which this secondary analysis draws from, has also reported choroidal thinning accompanying myopia progression in children.²¹ However, detailed choroidal structural alterations in the observed choroidal thinning²¹ is unknown in part due to the difficulties in their measurements. To the best of our knowledge, this is the first study to provide longitudinal evidence that myopia progression in children is accompanied by reductions in LA, SA, TCA, and CVI.

This study was the first to report longitudinal changes in choroidal structure, including both SA and LA among Chinese myopic children following 12-month RLRL therapy in a multicenter RCT. Strengths of this study include its reliable data from a multicenter, parallel-group RCT where OCT images and other ocular parameters were collected longitudinally at baseline and 4 follow-up visits. Second, SSOCT equipped with a longer wavelength provides better visualization of detailed choroidal structures and the choroidal scleral interface, which facilitates accurate measurements of the luminal and stromal components within the choroid. Third, a fully automated algorithm based on deep learning was employed. Traditional choroidal structure analysis relies heavily on manual input, which is time-consuming and may be biased by interobserver and intraobserver variability.^18,19 The automatic method used in the current study reduced the observer-derived variations to a small extent. Despite this, some limitations should also be acknowledged. First, our study is a post hoc analysis based on 2 out of 5 study sites where the OCT devices and scans were available, which might be subject to selection bias. Nevertheless, the included and excluded participants showed similar baseline characteristics, thus the selection bias would be minimal if any. Second, although medium- to large-size vessels could be visualized, the detection of small vessels was not possible due to limited lateral resolution and lack of contrast. This might affect the accuracy on choroidal LA estimation. Future studies involving the evaluation of the choriocapillaris density will be needed. Third, further studies are ongoing to investigate the long-term changes in choroidal structure following RLRL therapy, as well as to assess choroidal changes among the patients who stopped RLRL treatment.

In summary, this study demonstrates that the choroidal thickening is accompanied by an increase in both the LA and SA. This study speculates that RLRL therapy will increase the subfoveal choroidal blood flow while myopia progression will reduce it, and choroid imaging might be an essential tool for monitoring the efficacy of RLRL therapy. Further studies are needed to investigate the long-term changes in choroidal structure following RLRL therapy, as well as to assess choroidal changes among the patients who stopped RLRL treatment.