All investigative procedures adhered to the tenets of the Declaration of Helsinki. This study was approved by the Institutional Review Board Committee at Kawasaki Medical School (registration number: 3473). This study was designed as a retrospective observational case series and was conducted from November 2013 to December 2018 in the Department of Ophthalmology at Kawasaki Medical School Hospital. This study included minors under 16 years old among the patient sample. Informed consent for all examinations was obtained from one parent of each patient.
Subjects
Patients enrolled to this study showed unilateral amblyopia with decimal best-corrected visual acuities (BCVAs) <0.8 in the amblyopic eye and >1.0 in the fellow eye at the time of the first visit, due to hyperopic anisometropia with or without strabismus; and age <18 years. Hyperopic anisometropia was defined as hypermetropia in both eyes and an interocular difference in refraction (spherical equivalent) of >1.5 diopters (D). Treatment status was not considered at the time of performing swept-source OCT (SS-OCT) in this study. Exclusion criteria were as follows: presence of decimal BCVA in amblyopic eyes >1.0 within 2 months after full refractive correction or at the time of SS-OCT; history of ocular disease; history of intraocular surgery; or presence of systemic diseases that may have exerted an influence on the eye.
A total of 29 children were recruited as a normal control group. These children in the control group had a decimal BCVA >1.0 in both eyes, and no ocular or systemic diseases other than mild refractive errors. Only the right eye of the control group was used for data analysis.
Ophthalmologic examinations performed in all subjects included BCVA, cycloplegic refraction (measured using a RKT-7700 auto-refractor; NIDEK, Gamagori, Japan), AL (measured using an IOL Master OCT biometer; Carl Zeiss Meditec AG, Jena, Germany), cover test, extraocular movements, slit-lamp, fundoscopy, and SS-OCT (DRI OCT-1 Atlantis; Topcon Corporation, Tokyo, Japan). All SS-OCT examinations were performed by the same author (S.A.) between 9:00 AM and 12:00 PM under nonmydriatic conditions.
Quantification of choroidal vascular density and subfoveal choroidal thickness
CVD was quantified by binarizing the en-face image with reference to the report by Fujiwara et al. [9]. The 3-dimensional (3D) SS-OCT images, which covered a 12×9-mm2 area with a scan density of 512 × 256, were reconstructed as en-face images flattened with Bruch's membrane using dedicated software tools (EnView version 1.0.1; Topcon Corporation). En-face images were extracted at the level where the distance from the inner surface of the choroid was 50% of the total subfoveal choroidal thickness (SFCT) to examine the CVD of Haller's layer (Fig. 1-a).
The analyzed region of CVD was a 3×3-mm square centered on the fovea (Fig. 1-b). The magnification error was corrected using the Littman and modified Bennett formulae, taking into account image magnification due to AL variation [14, 15]. The extracted images were binarized by Niblack's method using ImageJ software (version 1.51; National Institutes of Health, Bethesda, MD), and the percentage of black area (considered to represent the choroidal vascular area) in the total region of interest was calculated as CVD (Fig. 1-c).
SFCT was defined as the vertical distance from the outer border of the hyperreflective line of retinal pigment epithelium/Bruch’s complex to the choroidal-scleral juncture at the subfovea. The SFCT was measured manually by S.A. from 3D SS-OCT images using the built-in caliper tool. Subjects with images that were difficult to segment due to signal attenuation were excluded from the present study.
Statistical analyses
Data are presented as means ± standard deviations. For statistical analysis, the decimal BCVA was converted to the logarithm of the minimal angle of resolution (logMAR).
First, multiple regression analysis of all eyes using the stepwise method was performed with CVD as a dependent variable, and logMAR, refractive error (RE), AL and SFCT as independent variables to select those parameters to be corrected for comparisons of CVD. For independent variables for CVD, when the variance inflation factor was > 5, the relevant parameter was excluded from statistical analysis in order to avoid multicollinearity.
Second, a generalized linear mixed model (GLMM) was used for CVD comparisons with measured eye (amblyopic, fellow, or normal control eyes) as a fixed factor, individual subjects (amblyopic patients or healthy control children) and factors that were significant in multiple regression analysis as random factors. In addition, logMAR, RE, AL, and SFCT were compared among amblyopic, fellow and normal control eyes. If a significant difference was found by GLMM, pairwise comparison was performed with Bonferroni correction.
The two-sample t-tests and chi-squared test were used to compare age and sex ratio between patients with unilateral amblyopia and normal control groups, respectively. Pearson’s correlation analysis was used to determine the coefficient of correlation between CVD and BCVA in amblyopic eyes.
A value of p<0.05 was considered statistically significant. All statistical analyses were conducted using SPSS Statistics version 20.0 (IBM Corporation, Somers, New York, USA).