Accuracy of twelve intraocular lens calculation formulas in highly myopic eyes

To investigate the accuracy, stability, and influencing factors of 12 intraocular lens (IOL) calculation formulas in highly myopic cataract patients.

Retrospective Case Series. Clinical data were collected from highly myopic cataract patients who underwent cataract extraction surgery at Beijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical University, between January 2024 and January 2025. Patients were stratified into subgroups based on axial length, keratometry, anterior chamber depth and axial length to corneal radius ratio (AL/r). The prediction error (PE), absolute error (AE), root-mean-square absolute prediction error (RMSAE) and the percentage of eyes with an absolute error within ± 0.25 D, ± 0.50 D, ± 0.75 D, and ± 1.00 D were recorded and compared for the following 12 IOL formulas: Barrett Universal II, Cooke K6, Castrop, Kane, EVO 2.0, Pearl-DGS, Hoffer QST, SRK-T, Haigis, Holladay 1, Holladay 2, and Hoffer Q. Statistical analysis was performed using Analysis of Variance (ANOVA), Friedman test, Kruskal-Wallis test, Dunn's test, and Cochran Q test.

A total of 255 patients (255 eyes) were included in this study, comprising 113 males and 142 females, with a mean age of 57.8 ± 9.9 years. The difference in absolute errors among the 12 IOL formulas was statistically significant (χ² = 327.59, P < 0.01). Among them, the Hoffer QST formula yielded the lowest absolute error (0.33 D), followed by Pearl-DGS (0.35 D) and Barrett UII 0.35 D), while the Hoffer Q formula showed the highest absolute error (0.75 D). Notably, only the Hoffer QST and Pearl-DGS formulas demonstrated no significant correlation between their absolute errors and axial length, corneal curvature, or anterior chamber depth (all P > 0.05).

For highly myopic cataract patients, modern IOL formulas demonstrated superior accuracy compared to traditional formulas. The accuracy was predominantly influenced by axial length, with keratometry and anterior chamber depth playing minor roles. The Hoffer QST and Pearl-DGS formulas, in particular, showed consistent and stable performance.