文献速递白内障与屈光手术小儿眼科与斜弱视

预测儿童双眼白内障术后眼轴长度的数学模型:改进人工晶状体度数选择

Mathematical Models for Predicting Postoperative Axial length in Children With Bilateral Cataract and Primary Intraocular Lens Implantation: Toward Improved Pediatric Intraocular Lens Power Selection.
PubMed 文献 · Am J Ophthalmol · 2026-01-21

两种模型均显示出高可靠性,对数模型R²为0.8611,GEE模型R²为0.8682,有助于提高儿童人工晶状体度数计算的准确性。

要点

  • 对数模型:AL = AL₀ + 2.87 × log₁₀(age/age₀),R²=0.8611,Lin一致性系数=0.9149。
  • GEE模型R²=0.8682,Lin一致性系数=0.9275,略优于对数模型。
  • 两种模型基于基线眼轴和手术年龄预测术后眼轴,可靠性高。

核心结论

  • 对数模型和GEE模型均能基于基线眼轴和手术年龄高可靠性地预测未来眼轴,提高儿童人工晶状体度数计算准确性。

主要发现

  • 对数模型:AL = AL₀ + 2.87 × log₁₀(age/age₀),R²=0.8611,Lin一致性系数=0.9149。
  • GEE模型:AL = -1.89 + 1.09 × AL₀ + 2.65 × log₁₀(age) - 3.17 × log₁₀(age₀) + 0.52 × (log₁₀ age₀ × log₁₀ age),R²=0.8682,Lin一致性系数=0.9275。
  • 预测值落在95%置信区间外的比例:对数模型7.5%,GEE模型7.7%。

关键数值与推荐

  • 对数模型公式:AL = AL₀ + 2.87 × log₁₀(age/age₀)。
  • GEE模型公式:AL = -1.89 + 1.09 × AL₀ + 2.65 × log₁₀(age) - 3.17 × log₁₀(age₀) + 0.52 × (log₁₀ age₀ × log₁₀ age)。

临床怎么用

  • 临床可基于患儿基线眼轴和手术年龄,使用对数模型或GEE模型预测术后眼轴,辅助儿童人工晶状体度数选择。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

PURPOSE: Develop and compare predictive accuracy of two mathematical models: the logarithmic model and generalized estimating equation (GEE) model for estimating postoperative axial length (AL) in children following bilateral cataract surgery with primary intraocular lens (IOL) implantation, aiming to enhance IOL power selection. DESIGN: Comparison of mathematical prediction models. METHODS: Eyes of children who underwent bilateral cataract surgery between January 2011 and December 2021 were included in the analysis. Data were only included if AL measurements were available preoperatively and at least 6 months postoperatively. The logarithmic and GEE models were developed and evaluated based on baseline AL (AL₀) and age at operation (age₀). Model performance was assessed using R² values, Lin's concordance correlation coefficients (CCC), and Bland-Altman analysis. RESULTS: One hundred one eyes (45 female, 56 male) with 401 postoperative AL measurements (median: 4/eye) were included. The derived logarithmic model was: AL = AL₀ + 2.87 × log₁₀ (age/age₀), yielding an R² of 0.8611 and a Lin's coefficient of 0.9149. The GEE model was: AL = -1.89 + 1.09 × AL₀ + 2.65 × log₁₀ (age) - 3.17 × log₁₀ (age₀) + 0.52 × (log₁₀ age₀ × log₁₀ age), showing an R² of 0.8682 and a Lin's coefficient of 0.9275. Predictions falling outside the 95% confidence interval were 7.5% for the logarithmic model and 7.7% for the GEE model. CONCLUSIONS: Both the logarithmic and GEE models demonstrated high reliability in predicting future AL based on baseline AL and age at surgery enhancing the accuracy of pediatric IOL power calculations.

来源:Am J Ophthalmol,查看原文。版权归原作者所有。

引用

  1. Lottelli AC, Marques JLP, Gouvea L, et al. Mathematical Models for Predicting Postoperative Axial length in Children With Bilateral Cataract and Primary Intraocular Lens Implantation: Toward Improved Pediatric Intraocular Lens Power Selection. Am J Ophthalmol. 2026. PMID: 41577068. doi:10.1016/j.ajo.2026.01.018

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。