Abstract / Summary
The axial length/corneal radius (AL/CR) ratio is an established structural indicator of myopia, but its ability to discriminate within the non-myopic range—specifically pre-myopia from hyperopia under cycloplegia—remains poorly characterised. We aimed to evaluate the performance of AL/CR in non-myopic schoolchildren in identifying pre-myopia from hyperopia. A retrospective cross-sectional investigation was conducted enrolling 1,393 children aged 6–11 years from two primary schools in Ningbo, China (December 2021 to March 2022). All children underwent cycloplegic refraction and ocular biometry. Refractive status was classified by spherical equivalent (SE) as myopia (SE ≤ − 0.50 D), pre-myopia (− 0.50 D < SE ≤ + 0.75 D) or hyperopia (SE > + 0.75 D). Among non-myopic children, Pearson correlation, multivariable logistic regression and receiver operating characteristic (ROC) analysis were used to evaluate AL and AL/CR for differentiating pre-myopia from hyperopia, overall and stratified by age (6–8 and 9–11 years); areas under the curve (AUCs) were compared with DeLong’s test. From ages 6 to 11 years, the prevalence of myopia increased from 7.1% to 54.2%, while hyperopia decreased from 47.5% to 9.4%. Pre-myopia increased from 45.4% at age 6 to 55.2% at age 10, then declined to 36.4% at age 11. AL and AL/CR increased progressively from hyperopia to pre-myopia to myopia. In non-myopic children, SE correlated inversely with AL ( r = − 0.447) and AL/CR ( r = − 0.629; both P < 0.001), and each was significantly associated with pre-myopia in separate multivariable models (both P < 0.001). AL/CR discriminated pre-myopia from hyperopia better than AL in the overall cohort (AUC 0.806 vs. 0.712; ΔAUC 0.095, 95% CI 0.058–0.131; P < 0.001), in the 6–8-year subgroup (0.787 vs. 0.679; ΔAUC 0.108, 95% CI 0.063–0.153; P < 0.001) and in the 9–11-year subgroup (0.831 vs. 0.740; ΔAUC 0.091, 95% CI 0.018–0.165; P = 0.015). The AUC of AL/CR did not differ significantly between the older and younger subgroups (ΔAUC 0.044, 95% CI − 0.016 to 0.103; P = 0.152). Under cycloplegic conditions, AL/CR demonstrated consistently superior performance compared with AL alone in distinguishing pre-myopia from hyperopia, across both younger and older age strata. Cut-offs are exploratory, and longitudinal studies are required to establish whether AL/CR predicts incident myopia.