Evaluating orthokeratology lens fitting and the visual quality of artificial intelligence-fitted personalized lenses in myopic children with corneal asymmetry.

To evaluate the orthokeratology (ortho-K) lens fitting and visual quality of artificial intelligence (AI) -fitted personalized lenses in myopic children with corneal asymmetry.This retrospective study included 139 children with corneal asymmetry who underwent ortho-K fitting, grouped according to conventional vision shaping treatment (VST) lenses (n = 53), corneal refractive therapy (CRT) lenses (n = 47), and AI-fitted personalized Optoshare Good Positioning (GP) lenses (n = 39). Pentacam-acquired anterior corneal images were used to calculate aberrations (root mean square total, root mean square higher-order aberration, and Zernike aberrations) and MATLAB was used to quantify lens decentration. One-way analysis of variance, Pearson correlation, and multiple linear regression were employed for analysis. After 1 month, the VST group's mean decentration (0.84 ± 0.35 mm) was significantly greater than that of the CRT (0.69 ± 0.22 mm, p = 0.023) and GP (0.60 ± 0.21 mm, p < 0.001) groups. All groups had increased aberrations; the VST group showed greater increases in coma Z3-1 than the GP group (p = 0.002) and in trefoil Z3-3 than both the CRT group (p = 0.014) and the GP group (p = 0.029). Lens decentration positively correlated with aberration increases; SE negatively correlated with aberration changes. Multiple regression showed that CRT lenses (β = -0.24, p = 0.008) and GP lenses (β = -0.38, p <0.001) had less decentration than VST lenses.Among children with corneal asymmetry, AI-based personalized GP lenses and CRT lenses exhibited better centration stability than VST lenses and can better control increase in higher-order aberrations.