Introduction

Myopia has become a public health problem worldwide, with epidemiologists predicting that approximately 4758 million people will be affected by 2050, representing 49.8% of the world’s total population1. The development of myopia is usually caused by excessive growth of the axial length (AL) of the eye2,3,4. The AL was previously called the “anteroposterior length,” which is usually the longitudinal length of the eyeball5. The AL reflects the developmental state of the eye and also has a large effect on the overall refractive state of the eye. Normally, the AL of a newborn eye is 16 to 17 mm, and the corresponding refractive degree is + 2.50 D to + 3.00 D6,7. As age increases, the AL also grows longer and continues to grow into adulthood. The current methods for preventing and controlling myopia, such as defocus incorporated multiple segments (DIMS) spectacle lenses, orthokeratology lenses, and low concentration atropine eye drops, aim to control axial growth2,4,8,9,10,11. Therefore, AL is a core indicator in the research, clinical treatment, prevention, and control of myopia; accurate measurement of AL is important.

The measurement of AL has evolved considerably from ultrasonic and photographic methods to optical biometry in the present day5,12,13,14. Although some specialized methods, such as magnetic resonance imaging, can be used for AL measurements, they are not widely used15. The biometric devices that doctors use in the clinic to measure AL usually measure the distance from the apex of the cornea through the lens to the macula. However, these devices indirectly reflect the length of the AL. The AL is important for normal physiological development, but there is controversy regarding AL after an intervention. Effective prevention and control measures for myopia generally cause thickening of the choroid, which in turn causes relative regression of the AL16,17,18. This finding raises the question of whether myopia prevention and control methods truly control the growth of the AL. Changes in the thickness of the choroid have a considerable effect on the measurement of the eye axis, especially when some commonly used myopia prevention and control methods, such as low-dose atropine, orthokeratology, and DIMS, are used, as they can cause choroidal thickening4,17,18.

This study aimed to develop a new AL measurement mode for swept-source optical coherence tomography angiography (SS-OCTA). SS-OCTA encompasses the distance from the corneal vertex to the posterior surface of the choroid within a single AL measurement called the real axial length (RAL). The choroidal thickness (ChT) in the SS-OCT images was defined as the volume from the outer boundary of the retinal pigment epithelium-Bruch membrane complex to the choroid-sclera junction19. After confirming the consistency of the biometric parameters—including central corneal thickness (CCT), anterior chamber depth (ACD), aqueous depth (AQD), and lens thickness (LT)—we subsequently analyzed the RAL in non-myopic participants across different age groups.

Results

A total of 1,833 subjects with normal uncorrected vision, aged 5–17 years were included in the study (85% from within Changsha and 15% from outside Changsha who visited the hospital on their own initiative). Among them, 909 (49.59%) were male and 924 (50.41%) were female. The subjects had a mean age of 9.97 ± 2.59 years, with males at 9.94 ± 2.60 years and females at 9.98 ± 2.59 years.The number of subjects in each age group is shown in Table 1.

Table 1 Age and gender distribution of subjects presents the number and proportion of subjects in each age group, along with the gender ratio.
Full size table

The mean RAL measured by the VG 200D was 24.04 ± 0.99 mm, which was 0.55 ± 0.33 mm (95% CI: 0.54 to 0.56, p < 0.001) greater than the mean AL of 23.49 ± 0.99 mm obtained with the IOL Master 700. Additionally, the mean ChT was 556.13 ± 33.36 μm using the VG 200D. As demonstrated in Table 2, the two devices exhibit a high degree of consistency. The mean values of the AL, ACD, and AQD measured using the IOL Master 700 and VG 200D were 23.49 ± 0.99 and 23.52 ± 1.01 mm, 3.57 ± 0.27 and 3.53 ± 0.25 mm, and 3.01 ± 0.40 and 2.98 ± 0.29 mm, respectively (ICC > 0.9). The mean CCT and LT were 546.77 ± 32.43 and 556.13 ± 33.36 μm and 3.47 ± 0.19 and 3.50 ± 0.19 mm, respectively (ICC > 0.8). A Bland–Altman plot with 95% limits of agreement was used to assess the concordance line. The proximity of this line to zero suggested a high level of agreement between the two machines across conventional parameter measurements. Most of the observed differences fell within the established limits of agreement (Supplementary Fig. S1).

Table 2 Pair-wise comparison for agreement between biometric values of iolmaster 700 and VG 200D presents the consistency of the conventional measurement parameters of the two devices.
Full size table
Table 3 Age-specific and gender-specific percentiles of RAL and ChT distribution among participants presents the distribution of RAL and CT across different age groups and genders.
Full size table

The AL, CCT, AQD, and RAL were significantly positively correlated with age （Figure 2）. The AL showed a progressive increase from 22.40 ± 0.88 mm (5 years) to 24.58 ± 0.78 mm (17 years), with strong correlation coefficients (Pearson’s r = 0.98, p < 0.001). A progressive increase in the RAL was similar to that of the AL, increasing from 22.96 ± 0.86 mm to 25.09 ± 0.77 mm (r = 0.98, p < 0.001). Similarly, CCT values increased from 553.57 ± 36.00 μm (5 years) to 565.13 ± 43.54 μm (17 years) (r = 0.80, p = 0.001), while the AQD increased from 2.70 ± 0.25 mm to 3.12 ± 0.27 mm over the same age range (r = 0.93, p < 0.001). In contrast, the LT and ChT showed inverse correlations with age. The LT decreased significantly from 3.73 ± 0.17 mm (5 years) to 3.45 ± 0.16 mm (17 years) (r = − 0.81, p = 0.001), while the ChT showed a significant reduction from 357.44 ± 65.92 μm to 310.67 ± 56.24 μm during the observation period (r = − 0.89, p < 0.001).

As shown in Table 3, at age 5, boys exhibited a significantly shorter RAL than girls, with median values of 23.55 mm (interquartile range, IQR: 22.91–23.76 mm) and 22.32 mm (IQR: 21.99–23.15 mm), respectively (p = 0.04). By age 17, the RAL increased to 25.26 mm (IQR: 24.75།25.64 mm) in boys and 25.24 mm (IQR: 24.49།25.81 mm) in girls, showing no significant sex difference (p = 0.73). Although the RAL for boys was consistently greater than that for girls, this disparity diminished with age. A similar pattern was observed in ChT. At age 5, boys had a significantly thinner ChT compared with girls, with median values of 334.10 μm (IQR: 287.63–355.87 μm) and 387.42 μm (IQR: 350.48–438.78 μm), respectively (p = 0.02). With increasing age, the rate of choroidal thinning appeared greater in girls. By age 17, the ChT decreased to 313.07 μm (IQR: 274.79–338.99 μm) in boys and 332.94 μm (IQR: 292.58–350.73 μm) in girls, with no significant difference between sexes (p = 0.71).

Discussion

In this study, we investigated a new concept of AL measurement using OCTA and verified its consistency with conventional parameters measured by the IOL Master 700. We also examined RAL values in children and adolescents aged 5 to 17 years with normal uncorrected vision. These findings provide a new reference for the ocular development status of children and adolescents and could facilitate scientific evaluation of the effectiveness of preventing and controlling the progression of myopia.

Across populations, AL elongation shows marked regional and ethnic variability even at the same chronological age20,21,22. Therefore, the population in this study was mainly concentrated in the Hunan Province of China, so the measurement and reference of RAL may have certain limitations. This study primarily aimed to explore a novel method for axial length measurement; therefore, refractive matching following cycloplegia was not performed. All participants exhibited a non-cycloplegic autorefraction value exceeding 0D, and their uncorrected visual acuity was at least 0.10 logMAR or better. Furthermore, in this cohort, the majority of parents were reluctant to consent to cycloplegia for their children without myopia.

The IMI 2023 report noted that rapid changes in ChT can induce inverse fluctuations in AL, implying that short-term AL variations may not reflect true ocular growth and should be interpreted together with ChT23. The RAL, which integrates retinal and ChT, may therefore provide a more stable and physiologically relevant metric for assessing ocular development and the efficacy of myopia control interventions.

The IOL Master 700 is a commonly used biometric device for evaluating biological parameters, and its measurements are well recognized by clinicians. The VG 200D (SS-OCTA) is used clinically for measurements of the anterior and posterior segments of the eye. Both measurement devices use swept lasers with a central wavelength of 1050 nm. The IOL Master 700 operates at a scanning speed of 2000 scans/s, whereas the VG 200D achieves a much higher scanning speed of 200,000 scans/s. Regarding the resolution of these devices, the IOL Master 700 provides 22 μm, while the VG 200D provides a superior resolution of 3.8 μm. Consequently, the VG 200D theoretically delivers more precise measurement values than the IOL Master 700. The VG 200D is also capable of measuring up to 12 mm deep into the tissue, facilitating the assessment of the retina and choroid24. In our study, we observed good agreement between the two devices in the measurement of AL, CCT, ACD, AQD, and LT. In particular, the ICCs for the AL, ACD, and AQD measurements were all > 0.9. This finding indicates that the RAL measurements are accurate and reliable.

We grouped the biological parameters measured with the VG 200D by age. We found that as age increased, the RAL and AL became longer, the CCT became thicker, the AQD became deeper, and the crystalline and choroid became thinner. This finding is consistent with the trend of changes in parameters with age in previous related studies25,26,27,28,29,30. The choroidal data measured with the VG 200D support this result and are consistent with previous studies31,32,33, Which further demonstrates the feasibility of using the VG 200D to measure RAL.

This study also examined the RAL at all ages by sex. We observed that the RAL of boys was longer than that of girls. However, the sex disparity in the RAL tended to diminish with age. We hypothesize that the change in the RAL is mostly driven by the ChT because of its consistent trend across the different age groups. Sex-related differences in ocular growth may, at least in part, be attributed to variations in choroidal structure and dynamics. Previous studies have consistently reported that females exhibit thinner choroids compared with males, even after adjusting for axial length and age34. Furthermore, longitudinal data suggest that choroidal thinning in females occurs earlier and progresses more rapidly with age35. Given the established association between choroidal thinning and axial elongation in children, it is plausible that the choroid serves as a structural mediator linking sex and axial growth rate32,36. The IMI 2023 report further highlights the rapid plasticity of the choroid and its integral role in ocular growth regulation, providing a biological rationale for this potential pathway23. Collectively, these findings support a hypothesis that sex-related differences in ChT may partly explain the faster axial elongation observed in females. Future studies should integrate RAL into myopia management. For instance, after using treatments like atropine drops or peripheral defocus glasses, axial length fluctuations may occur. Comparing RAL with AL will help assess the eyes’ true developmental state.

Conclusion

In this study, the RAL exceeded the AL by approximately 0.5 mm. The VG 200D shows strong agreement with the IOL Master 700 in measuring the AL, CCT, ACD, AQD, and LT. The difference in RAL length between boys and girls may progressively narrow with increasing age. This study delineates the distribution of RAL in children and adolescents aged 5 to 17 years with normal uncorrected vision in Changsha, China. These findings could help establish preliminary clinical reference data for using the RAL in monitoring ocular development, particularly in the context of prevention and control strategies for myopia.

Limitations

This study’s first limitation is this was a single-center study conducted in the Changsha area, which constitutes a regional limitation, so further multi-center validation is needed. Second, since the subjects did not undergo cycloplegia, no correlation analysis was conducted between refractive error and RAL. Third The cross-sectional design limits tracking individual growth, a key gap for understanding RAL/ChT dynamics over time. A longitudinal follow-up would clarify patterns, especially given smaller adolescent subgroups (per Table 1), which may inflate variability.

Methods

This observational study was conducted at the Aier Institute of Optometry and Science in Changsha, Hunan Province, from June 2023 to January 2024. The study was approved by the Ethics Committee of Changsha Aier Eye Hospital (ID: 2023KYPJ019) and adhered to the tenets of the Declaration of Helsinki. All subjects and their legal guardians provided written informed consent before participating in the study.

The inclusion criteria of this study were children and adolescents aged 5 to 17 years who were able to complete a biometric examination. The inclusion criteria comprised a refractive error exceeding 0.00 D as measured by non-cycloplegic autorefraction, concurrently with an uncorrected visual acuity of 0.10 logMAR or superior in both eyes. The exclusion criteria were as follows: (1) ocular diseases, such as glaucoma, retinopathy of prematurity, and retinitis pigmentosa; (2) systemic cardiovascular, immune system, and endocrine-related diseases; (3) surgery on the eye or other organs of the body; (4) current or previous use of control interventions for myopia, such as atropine eye drops, DIMS lenses, orthokeratology lenses, and red light therapy. All subjects underwent a basic eye health examination before enrollment, including visual acuity and intraocular pressure measurements and a slit lamp examination.

The AL, CCT, ACD, AQD and LT were measured using the IOL Master 700 (Carl Zeiss Meditec AG, Jena, Germany) and the VG 200D (SS-OCT/OCTA; SVision Imaging, Henan, China). The VG 200D was also used to measure ChT and RAL (Fig. 1). The RAL is the length from the corneal vertex to the posterior surface of the choroid included in a single biometric measurement. Furthermore, the SVision VG 200D was used to measure six sets of data at the same location each time. The results are displayed as the mean of the six sets.

Fig. 1
figure 1

SVision VG 200D measurement interface. The figure presents the page of SVision VG 200D’s measurement results, including relevant measured parameters: CCT, ACD, AQD, LT, AL, ChT, and RAL.SVision VG 200D measures 6 sets of data at the same location each time, and the results are displayed as the mean of the 6 sets of data.

Full size image
Fig. 2
figure 2

Biological parameters measured using VG 200D. The figure illustrates the mean values of various measurement parameters for both eyes of 1,833 subjects, as assessed using the SVision VG 200D. These values are organized according to age distribution.

Full size image

In this study, data were collected from both eyes. The consistency of the biological parameters measured by the two devices was evaluated using data from both eyes. However, analyses examining the correlation between parameters measured by the VG 200D and age, as well as gender differences in RAL and ChT across various age groups, were conducted using data exclusively from the right eye. Descriptive statistics were used to summarize the baseline demographic characteristics of the participants. The normality of continuous variables was evaluated using the Kolmogorov–Smirnov test. Continuous variables are presented as mean with standard deviation (SD), while age-specific and gender-specific percentiles of RAL and ChT were presented as median with quantiles. A paired sample t-test was used to compare the AL and RAL measured by the VG 200D. The following tests were used as appropriate: The intraclass coefficient (ICC) and Bland–Altman analysis with 95% limits of agreement were used to verify the consistency of the two devices. The ICC for inter-rater reliability was used to quantify the degree of agreement between two raters assessing the same group of subjects. ICCs were interpreted as follows: p < 0.5 indicate poor reliability, values between 0.5 and 0.75 indicate moderate reliability, values between 0.75 and 0.9 indicate good reliability, and p > 0.90 indicate excellent reliability37. Given the bilateral consistency in ocular parameter trends, subsequent analyses were restricted to right eye measurements. Pearson correlation analysis quantified associations between core biometric indices (AL, CCT, AQD, LT ChT and RAL) and age. Concurrently, sex-related differences in RAL and ChT across stratified age were evaluated through independent samples t-tests. A p value < 0.05 indicates significance. Statistical analysis was performed using SPSS(V.27.0; IBM, USA) and GraphPad Prism(V.9.0; GraphPad Software, USA).