Predictive Accuracy Of Modern Iol Formulas In Pediatric Cataract Surgery

In pediatric cataract surgery, achieving immediate emmetropia is not always the goal. Due to predictable ocular growth and the need to anticipate future refraction, accurate intraocular lens (IOL) power estimation is critical to minimize postoperative refractive errors. IOL calculation formulas pose particular challenges in children, as they should predict both immediate postoperative refraction and long-term refractive outcomes. This study aims to retrospectively compare six IOL formulas—EVO 2.0, Barrett Universal II (BU II), Cooke K6 (K6), Hoffer QST, Kane, and T2—and to present a new online tool supporting pediatric IOL power calculation.

A retrospective observational study was conducted in a tertiary hospital: Hospitais da Universidade de Coimbra, in Portugal.

Inclusion criteria included patients under 18 years who underwent pediatric cataract surgery with IOL implantation. IOL calculation formulas evaluated were EVO 2.0, Barrett Universal II, Hoffer QST and Kane. In addition, we tested a newly developed online tool for pediatric IOL power calculation (pedi.mraimundo.com) with Cooke K6 and T2 formulas, designed to assist surgical planning by enabling age-based target selection, axial length (AL) prediction at age 15 years using the Trivedi and Dupessey algorithms. Data analysis was conducted using the ESCRS Eyetemis platform (www.eyetemis.com) where the spherical equivalent prediction error (SEQ-PE, trueness), standard deviation (SD, precision), and absolute SEQ-PE (accuracy) were calculated.

Thirty-one eyes (65% males) with a mean age at surgery of 9.9±4.6 years from 23 patients were included. Regarding SEQ-PE (trueness), mean ranged from –0.12 (BU II) to 0.13D (EVO 2.0), with SD ranging from 0.79 (EVO 2.0) to 0.98D (BU II) and extremes from –2.39 to +1.44D across formulas. Pairwise comparisons of mean SEQ-PE showed differences between EVO 2.0 and BU II, Hoffer QST, Kane, and T2 (p<0.05). For SEQ-PE precision, mean ranged from 0.63 (EVO 2.0) to 0.78D (BU II), with pairwise comparisons indicating comparable variability across all IOL calculation methods. The mean absolute SEQ-PE ranged from 0.62 (EVO 2.0) to 0.78D (BU II), reflecting lower overall predictive error and variability with EVO 2.0 compared to the other formulas.

Modern IOL calculation formulas demonstrated comparable predictive performance in pediatric cataract surgery, with EVO 2.0 showing slightly lower variability but no clinically significant superiority over simpler formulas (T2). Using formulas with fewer required inputs may be advantageous in children, as limited cooperation makes complete biometry unfeasible. Although the mean error in SEQ-PE was low, the wide range of prediction errors highlights the persistent unpredictability of outcomes in pediatric IOL implantation. The newly online tool, integrating both a modern formula (K6) and a simplified modified formula (T2), facilitates age-based refractive targets and AL prediction at 15 years, supporting individualized surgical planning.