Regression-based keratometric prediction models showed refractive performance comparable to postoperative keratometry in pterygium and cataract cases.

$Regression-based keratometric prediction models showed refractive performance comparable to postoper…$

Photo by Alexander Grey / Unsplash

Frontiers in Medicine Published April 13, 2026 Medically reviewed April 25, 2026 DOI ↗ By Dr. Lars van Dijk, PhD · Surgical, Procedural & Diagnostic

Key Takeaway

Consider regression-based keratometric prediction models as a potential alternative when postoperative keratometry values are unavailable.

This retrospective paired analysis examined 20 unilateral eyes that underwent staged cataract surgery following pterygium excision. The study compared a regression-based keratometric prediction model against postoperative keratometry (K), preoperative K, and fellow-eye K measurements. The primary outcome assessed refractive outcomes, with secondary outcomes including mean K, spherical equivalent, refractive error, and absolute refractive error.

Statistical analysis revealed that overall differences among the four methods were significant for mean K (p < 0.0001), spherical equivalent (p = 0.0003), refractive error (p = 0.001), and absolute refractive error (p = 0.014). However, a direct comparison between the postoperative K and the predicted K showed no statistically significant differences were observed in any parameter.

Safety and tolerability data were not reported, as no adverse events or discontinuations were documented in this small sample. A key limitation of this study is the small sample size of 20 eyes, which restricts the generalizability of the results. The authors note that confirmation in larger prospective studies is required before broader adoption. Given the current evidence, regression-based K estimation demonstrates refractive performance comparable to postoperative keratometry, supporting its potential utility as a practical alternative when postoperative K values are unavailable.

Study Details

Study typeCohort

EvidenceLevel 3

PublishedApr 2026

View Original Abstract ↓

BackgroundThis study aimed to validate a regression-based keratometric prediction model in eyes with pterygium using actual postoperative refractive outcomes after cataract surgery.MethodsA retrospective paired analysis was performed in 20 unilateral eyes that underwent staged cataract surgery following pterygium excision. Predicted K values were calculated using a previously reported regression model based on CASIA2 anterior segment parameters. Intraocular lens (IOL) power was calculated using the Barrett Universal II formula. Mean K, spherical equivalent, refractive error, and absolute refractive error were compared among four keratometric methods (postoperative K, predicted K, preoperative K, and fellow-eye K) using the Friedman test followed by Bonferroni-adjusted Wilcoxon signed-rank tests. Sensitivity analyses were conducted after exclusion of toric IOL cases. Exploratory receiver operating characteristic analyses were performed to assess the influence of preoperative corneal parameters.ResultsOverall differences among the four methods were significant for mean K (p < 0.0001), spherical equivalent (p = 0.0003), refractive error (p = 0.001), and absolute refractive error (p = 0.014). However, no statistically significant differences were observed between postoperative K and predicted K in any parameter. Sensitivity analyses yielded consistent findings. Exploratory analyses suggested that preoperative corneal characteristics may influence refractive predictability.ConclusionIn this paired analysis of 20 unilateral eyes, regression-based K estimation demonstrated refractive performance comparable to postoperative keratometry. These findings support the potential clinical utility of prediction-based keratometry as a practical alternative when postoperative K values are unavailable, although confirmation in larger prospective studies is required.