Accuracy of Toric Intraocular Lens Realignment Calculators: A Vector-Based Bivariate Comparison of Current Calculation Methods.
Where this comes from
- Record sourced from PubMed, PMID 42435830.
- Also identified by DOI 10.1016/j.ajo.2026.07.013.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
To compare currently available calculation methods using full vector and bivariate analysis of astigmatic prediction error (PE) for planning toric intraocular lens (IOL) realignment in eyes with residual astigmatism after phacoemulsification surgery. Retrospective comparative accuracy study. Predicted postrealignment astigmatism and recommended rotation axes were calculated using AstigmatismFix (AstiFix), Gatinel Toric IOL Optimal Rotation Calculator (Gatinel-TOR), Ophtera AxisAligner (Ophtera-AA), Toric Aligner (TorAlign), and IOLCon Lens Power Calculator (IOLCon-LPC). Astigmatic PE was evaluated using vector magnitude and bivariate double-angle analysis (Wilcox-Holladay-Wang-Koch). Vector PEs, centroids, and marginal variances were compared. Agreement of recommended axes was assessed using intraclass correlation coefficients (ICC). Calculations were performed using subjective refraction, and phakic and pseudophakic biometric data where applicable. Forty-nine eyes of 49 patients were included. AstiFix, Gatinel-TOR, and Ophtera-AA showed nearly identical bivariate error distributions with small centroid magnitudes (0.14-0.15D), and low mean vector PE (0.43-0.45D). No formula demonstrated a statistically significant deviation from the zero-vector using median-based polygon analysis (all P ≥ .284). Axis agreement among AstiFix, Ophtera-AA, and Gatinel-TOR was excellent (ICC ≥ 0.997), with > 93% of eyes within ≤ 5°. TorAlign demonstrated significantly different axis recommendations and lower agreement (ICC = 0.816-0.822). IOLCon-LPC showed the lowest agreement with realignment-specific tools. Mean vector PE did not differ significantly among tested formulas. Repeated pseudophakic biometric measurements did not improve prediction accuracy. Realignment-specific calculators incorporating prerealignment subjective refraction (AstiFix, Gatinel-TOR, Ophtera-AA) provide highly consistent axis recommendations with no systematic bias detected by median-based polygon analysis. IOLCon-LPC and especially TorAlign demonstrated divergent axis recommendations and higher total variance.