Ocular Dimensional Changes After 12 Months of Orthokeratology and Single-Vision Spectacle Use: A Three-Dimensional Magnetic Resonance Imaging Study

Introduction

Myopia is a growing global public health concern, with rapidly increasing prevalence worldwide, particularly in East and Southeast Asia, where rates among school-aged children may exceed 50–80%.¹ A recent local study indicates that at least 40% of Malaysian children and adolescents aged 13 to 18 likely have myopia, specifically females, individuals of Chinese ethnicity, and youth between the ages of 7 and 18.² As myopia develops, the eye undergoes a transition toward a prolate profile due to disproportionate axial elongation, expanding 3.5 times faster in length (0.35 mm/D) than in width (0.10 mm/D).³ Such axial stretching causes the posterior pole to steepen relative to the periphery, creating a prolate-like ellipsoid that induces peripheral hyperopic defocus, in which light focuses behind the retina and signals further elongation.⁴

Orthokeratology (Ortho-K) is a reversible corneal reshaping treatment that slows myopia progression in children. Meta-analyses of randomized and controlled clinical trials have consistently demonstrated that Ortho-K significantly reduces axial elongation compared with single-vision spectacle (SVS) correction, supporting its effectiveness as a myopia control intervention in pediatric populations.⁵ Modern Ortho-K lenses employ a reverse-geometry design that flattens the central cornea while steepening the mid-peripheral cornea; thereby, inducing peripheral myopic defocus. This optical profile has been hypothesized to influence ocular growth patterns beyond axial length (AL) alone, potentially modifying peripheral retinal signaling mechanisms associated with myopia progression.⁶

To date, most studies investigating ocular changes associated with Ortho-K have relied on indirect measures, such as peripheral refraction, to infer alterations in eye shape.⁷ A previous study evaluated changes in eye shape following Ortho-K treatment using biometric measurements from Lenstar LS900 to assess peripheral eye length. Over a 12-month period, the authors reported reduced prolateness and increased symmetry compared with SVS wear.

In a subsequent study utilizing magnetic resonance imaging (MRI) to measure peripheral eye length (PEL), children in the Ortho-K group exhibited distinct differential growth patterns compared to the SVS cohort.⁸ Specifically, PEL at the nasal eccentricity increased more rapidly than all other peripheral and axial lengths (p<0.05). These findings demonstrate that while SVS wearers developed an increasingly prolate eye shape alongside myopia progression, Ortho-K wearers maintained a significantly less prolate ocular profile over the 12-month observation period.

Building upon previous findings, this study expands the investigation by using MRI to directly quantify global ocular dimensions and shape changes after 12 months of Ortho-K and SVS wear in myopic children. Through MRI-based 3D modeling of ocular geometry, the eyeball was analyzed as a whole, specifically assessing changes in axial length, height, and width. This comprehensive approach enables evaluation of both global and peripheral dimensional changes, providing robust structural evidence for the myopic defocus theory by correlating these alterations with previously reported refractive outcomes.^8,9 Furthermore, this study examines whether Ortho-K is associated with distinct patterns of ocular dimensional and corneal curvature (D) changes compared with conventional SVS treatment over the study period.

Materials and Methods

Ethical Approval and Informed Consent

The study was approved by the University Kebangsaan Malaysia Research Ethics Committee (UKM PPI-800-1/1/5 JEP-2017-422) and adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from the parents or legal guardians of all participants.

Sample Size Calculation

Sample size was calculated using G*Power (version 3.1.9.3; Heinrich Heine University, Düsseldorf, Germany) based on previous data reporting AL changes of 0.36 mm (standard deviation [SD] = 0.24 mm) after 2 years of Ortho-K treatment.¹⁰ To detect a 0.24 mm difference in AL (SD = 0.24 mm; α = 0.05; power = 0.95), 14 participants per group were required. Accounting for a 17% dropout rate, at least 25 participants per group were recruited to maintain statistical power.

Study Participants

Seventy schoolchildren with myopia (35 females, 35 males) aged 8–9 years, with a myopia duration of at least 6 months, were recruited. Inclusion criteria required a best-corrected visual acuity (BCVA) of 0.0 logarithm of the minimum angle of resolution (logMAR) or better in both eyes, normal ocular and systemic health, a cycloplegic spherical equivalent (SE) refractive error between –0.75 and –4.00 D, and astigmatism ≤ –1.50 DC. All subjects have no history of wearing contact lenses, Ortho-K lenses or prior myopia control spectacles treatment. Group allocation to Ortho-K or SVS control was determined by parental preference.

Examinations and Measurements

Participants in the SVS group were prescribed full-time single-vision distance spectacles (refractive index 1.56; Integrated Lens Technology, Singapore). Prescriptions were updated if monocular aided visual acuity reached ≥0.18 logMAR or if the SE refraction changed by ≥–0.50 D. The Ortho-K group was fitted with overnight Ortho-K lenses (Z Night; Menicon, Nagoya, Japan) using proprietary software (Easyfit; Menicon) based on corneal topography and cycloplegic refraction. Baseline study methods and interventions have been previously described.⁸ Specific procedures for ocular dimensional measurements following 12 months of Ortho-K lens wear are detailed below.

MRI Segmentation

MRI datasets were analyzed using custom software developed in MATLAB (MathWorks, Natick, MA, USA) by a multidisciplinary team of engineers based on the Chan-Vese model.¹¹ This approach offers computational efficiency, accurately resolving complex boundary curves and handling intricate geometric constraints such as angle production and curvature.¹²

Images were displayed and magnified using a commercially available DICOM viewer (OsiriX; Pixmeo SARL, Geneva, Switzerland). Axial images were viewed at 50X magnification, with contrast adjusted to optimize edge definition. To confirm the measurement standards, a 10 mm scale bar was utilized within the OsiriX DICOM viewer to ensure all calculations remained in millimeters (mm) and the raw data were processed and displayed using a square pixel (1:1) aspect ratio. The axial plane was reconstructed using 200 iterations due to its high symmetry and uniform shape allow the algorithm to efficiently produce high-quality images with minimal noise (Figure 1). In contrast, the sagittal view required 300 iterations to manage the increased anatomical complexity of the orbit: such as the eyelids and curved orbital floor, and to filter out vertical motion artifacts (Figure 2). Following these procedures, pertinent linear dimensions for the assessment of peripheral eye length (PEL) at 10° intervals and eyeball shape (longitudinal AL (LAL), horizontal width [HW], vertical sagittal height [VSH]) were automatically measured and analyzed (Figure 3). The reliability and measurement precision of these MRI-based ocular measurements have been previously established and reported in an MRI validation study.¹³

Figure 1 Axial orbital MRI. Green line - Delineation of the selected eyeball boundary in the axial plane, indicating the region of interest used for analysis.

Figure 2 Sagittal orbital MRI. Green line- Delineation of the selected eyeball boundary in the sagittal plane, indicating the region of interest used for analysis.

Figure 3 Schematic diagram of eyeball dimensions. (a) Axial view showing the measurement of ocular width, peripheral eye length and axial length. (b) Sagittal view showing the measurement of ocular height.

All statistical analyses were performed using SPSS software (version 21.0; IBM Corp., Armonk, NY, USA). To account for potential inter-eye correlation, differences between right and left eye were initially evaluated using paired t-tests. As the variance between eyes was smaller than the variance between participants, only right eye data were included in the final analysis to avoid underestimation of true variance and inflation of Type I error. Normality of baseline demographic, refractive, and biometric variables was assessed using the Shapiro–Wilk test, whereas homogeneity of variance was evaluated using Levene’s test. Participant demographics were analyzed using frequencies and sample descriptive tests. Within-group comparisons before and after Ortho-K and SVS interventions were performed using paired t-tests. Longitudinal changes from baseline to 12 months were analyzed using one-way repeated-measures analysis of variance (ANOVA). Between-group differences were assessed using independent t-tests, with statistical significance set at p < 0.05. Spearman correlation was performed to examine relationships between SE, D, and ocular dimensional changes in the eyes of children with myopia.

Results

Demographic Data

No significant differences in baseline data were observed between the Ortho-K and SVS groups regarding age, sex, BCVA, SE, D, PEL, LAL, HW, or VSH (p > 0.05). Participants’ demographic data are summarized in Table 1.

Table 1 Participant Demographics

Refractive Components at Baseline and 12 Months

Baseline and 12-month refractive data for the Ortho-K and SVS groups are summarized in Table 2. After 12 months, the Ortho-K group demonstrated significant reductions in SE (from −2.92 ±1.07D to −0.06 ± 0.12D; p < 0.05) and flat keratometry FK (from 43.6 ± 1.27D to 40.69 ± 2.54D; p < 0.05). Conversely, the SVS group exhibited a significant increase in both SE (from −2.51 ± 1.12D to −3.77 ± 1.23D, p < 0.05) and FK (from 43.12 ± 1.03D to 43.47 ± 1.10D, p < 0.05) over the same period.

Table 2 Refractive Components in the Ortho-K and SVS Groups at Baseline and 12 Months

Ocular Dimensions at Baseline and 12 Months

In the SVS group, LAL increased significantly over the 12-month period (23.45 ± 0.97 to 23.97 ± 0.59 mm; p < 0.05), while HW and VSH exhibited non-significant upward trends. Additionally, the SVS group demonstrated significantly increased PEL at all measured horizontal eccentricities (p < 0.05), except at N30. In contrast, the Ortho-K group showed significant PEL elongation only beyond 20° eccentricities, specifically at N20, N30, T20, and T30 (p < 0.05), with no significant elongation observed at N10 and T10. Rather, a significant reduction in LAL was detected in the Ortho-K group compared with baseline (23.85 ± 0.71 mm to 23.64 ± 0.74 mm; p < 0.05), whereas HW and VSH displayed downward trends that did not reach statistical significance (Table 3).

Table 3 Within-Group Eyeball Shape Measurements in Ortho-K and SVS Participants at 12 Months

Within-Group Changes in Ocular Shape

Repeated measures ANOVA revealed significant changes in the overall eyeball shape within both the Ortho-K and SVS groups over 12 months (p < 0.05). At the 12-month follow-up, significant differences were observed among the three ocular biometric parameters in both the Ortho-K and SVS groups. In the Ortho-K group, the mean longitudinal axial length (LAL) was 23.64 ± 0.74 mm, which was significantly greater than both horizontal width (HW; 23.12 ± 0.80 mm) and vertical sagittal height (VSH; 23.41 ± 0.70 mm). Pairwise comparisons with Bonferroni adjustment indicated that all three parameters differed significantly from one another (p<0.05). In the SVS group, the mean LAL was 23.97 ± 0.59 mm, while HW and VSH were 23.21 ± 0.99 mm and 23.52 ± 0.98 mm, respectively. Bonferroni-adjusted pairwise comparisons showed that LAL differed significantly from both HW and VSH (p<0.05), whereas no significant difference was found between HW and VS (Table 4).

Table 4 Within-Group Eyeball Shape Changes Over 12 Months in Ortho-K and SVS Participants

Between-Group Comparisons of Ocular Dimensions

At 12 months, independent t-tests indicated significant differences in the eyeball shape between the Ortho-K and SVS groups. The mean LAL was significantly lower in the Ortho-K group (23.64 ± 0.74 mm) compared to the SVS group (23.97 ± 0.59 mm) (p < 0.05). In contrast, no significant differences were observed between the two groups for horizontal width (HW) or vertical sagittal height (VSH). The mean HW values were 23.12 ± 0.80 mm in the Ortho-K group and 23.21 ± 0.99 mm in the SVS group (p = 0.698), while the mean VSH values were 23.41 ± 0.70 mm and 23.52 ± 0.98 mm, respectively (p = 0.650) (Table 5).

Table 5 Between-Group Eyeball Shape Comparison at 12 Months

Correlation of Spherical Equivalent with Ocular Dimensions

In the Ortho-K group, changes in SE did not show significant correlations with LAL, HW, or VSH (all p > 0.05) (Table 6). In the SVS group, SE showed a significant positive correlation with LAL only (p < 0.05), with no significant associations with HW or VSH (Table 7).

Table 6 Correlation Between Changes in SE and Ocular Dimensions in the Ortho-K Group

Table 7 Correlation Between Changes in SE and Ocular Dimensions in the SVS Group

Correlation of Corneal Curvature with Ocular Dimensions

In the Ortho-K group, changes in FK were significantly correlated with all measured ocular dimensions (LAL, HW, VSH; all p < 0.05) (Table 8). In the SVS group, FK showed no significant correlations with any ocular dimensions (all p > 0.05), with the exception of SE showing a significant correlation with LAL (Table 9).

Table 8 Correlation Between Changes in Corneal Curvature (FK) and Ocular Dimensions in the Ortho-K Group

Table 9 Correlation Between Changes in Corneal Curvature (FK) and Ocular Dimensions in the SVS Group

Discussion

This study examined 3D ocular dimensional changes in children with myopia after 12 months of Ortho-K lens wear and compared these changes with those of children wearing SVS. The SVS group served as a reference for the natural progression of childhood myopia without myopia control intervention.

At baseline, all participants with myopia exhibited a prolate globe morphology, with LAL as the dominant dimension, followed by VSH and HW. This pattern aligns with the axial expansion model described by Atchison et al³ and Matsumura et al¹⁴ Over 12 months, the SVS group demonstrated expansion across all ocular dimensions The most pronounced increase occurred in LAL (p < 0.05), followed by significant elongation of PEL at N20, N10, T10, T20, and T30 (all p < 0.05), reflecting overall prolate ocular growth. In contrast, the Ortho-K group exhibited a shift toward a less prolate ocular profile. These findings extend previous research¹⁰ reporting axial shortening accompanied by peripheral expansion between 10° and 30°. However, while the Ortho-K lens successfully halted LAL elongation, the stable transverse dimensions of HW and VSH indicate a less prolate ocular profile rather than a true global structural shift.

The results of this study also showed that in the Ortho-K group, significant PEL elongation occurred only beyond 20° peripheral eccentricity (N20, N30, T20, and T30; p < 0.05), accompanied by centre AL shortening (p<0.05). These findings indicate that Ortho-K treatment induces distinct structural modifications in the growing eye through central corneal flattening and mid-peripheral steepening,¹⁵ which collectively induce peripheral myopic defocus. By shifting the peripheral image shell anterior to the retina, this optical signal may act as a potent regulator of ocular growth, contributing to the reduced axial, horizontal, and vertical dimensions observed in the present Ortho-K cohort.¹⁶

The differential relationship between SE and AL observed in the SVS and Ortho-K groups further highlights the distinct mechanisms through which these modalities interact with ocular biometry. In the SVS group, changes in SE were significantly correlated with AL elongation, indicating that refractive progression was primarily driven by structural eye growth. Because spectacles do not alter ocular components such as the cornea or crystalline lens, the myopic shift largely results from axial elongation, with the retina moving posteriorly relative to the eye’s focal point.^17,18 In contrast, this correlation was absent in the Ortho-K group, demonstrating a decoupling of SE and AL.¹⁹ The results indicate that Ortho-K lenses temporarily reshape the cornea by flattening the central epithelium and steepening the mid-periphery through reverse-geometry lens forces; thereby, neutralizing central refractive error. Consequently, changes in SE among Ortho-K wearers primarily reflect corneal reshaping rather than true axial elongation.^8,15

To date, few studies have investigated the specific ocular dimensional changes associated with Ortho-K treatment.^9,20 One study reported that central corneal flattening accompanied by peripheral corneal steepening leads to a more symmetric and less prolate ocular morphology in children with myopia.⁹ Another study suggested that Ortho-K lenses alter ocular tension dynamics, particularly by reducing tension in the posterior ciliary muscle–lens complex, which may explain the short-term axial shortening observed following treatment.²⁰

Several limitations of this study warrant consideration. Firstly, the 12-month follow-up period restricts understanding of long-term structural trends; given that myopia management is a multi-year process, further longitudinal data is required to determine whether observed ocular changes persist or stabilize beyond the initial treatment period. Secondly, a key limitation of this study is that LAL measurements were not adjusted for the corneal flattening; since Ortho-K thins the central corneal epithelium by 10–20 microns,²¹ future studies should account for these corneal changes to ensure more accurate axial measurements. Thirdly, future research should explore the impact of varying treatment zone diameters on peripheral retinal contour to more clearly characterize the PEL data and evaluate its relationship with ocular dimension changes. Such investigations would provide critical data for the development of personalized Ortho-K algorithms; thereby, refining structural observations and optimizing customized lens geometries²² for enhanced long-term myopia control.

Conclusion

In summary, this study provides direct evidence of 3D ocular remodeling in children treated with Ortho-K, as quantified via MRI. Ortho-K slows axial elongation while actively modifying the global aspect ratio of the eye. By restricting LAL expansion to rates comparable with HW growth, Ortho-K effectively mitigates the pathological stretching and prolate distortion characteristic of rapid myopic progression.