Jiang Yinjie, Shen Yang, Chen Xun, Niu Lingling, Li Boliang, Cheng Mingrui, Lei Yadi, Xu Yilin, Wang Chongyang, Zhou Xingtao, Wang Xiaoying
Eye Ear Nose and Throat Hospital, Fudan University, No. 19 BaoQing Road, XuHui District, Shanghai, 200031, China.
National Health Commission Key Lab of Myopia, Fudan University, Shanghai, China.
Eye Vis (Lond). 2023 May 1;10(1):22. doi: 10.1186/s40662-023-00338-1.
Implantable collamer lens (ICL) has been widely accepted for its excellent visual outcomes for myopia correction. It is a new challenge in phakic IOL power calculation, especially for those with low and moderate myopia. This study aimed to establish a novel stacking machine learning (ML) model for predicting postoperative refraction errors and calculating EVO-ICL lens power.
We enrolled 2767 eyes of 1678 patients (age: 27.5 ± 6.33 years, 18-54 years) who underwent non-toric (NT)-ICL or toric-ICL (TICL) implantation during 2014 to 2021. The postoperative spherical equivalent (SE) and sphere were predicted using stacking ML models [support vector regression (SVR), LASSO, random forest, and XGBoost] and training based on ocular dimensional parameters from NT-ICL and TICL cases, respectively. The accuracy of the stacking ML models was compared with that of the modified vergence formula (MVF) based on the mean absolute error (MAE), median absolute error (MedAE), and percentages of eyes within ± 0.25, ± 0.50, and ± 0.75 diopters (D) and Bland-Altman analyses. In addition, the recommended spheric lens power was calculated with 0.25 D intervals and targeting emmetropia.
After NT-ICL implantation, the random forest model demonstrated the lowest MAE (0.339 D) for predicting SE. Contrarily, the SVR model showed the lowest MAE (0.386 D) for predicting the sphere. After TICL implantation, the XGBoost model showed the lowest MAE for predicting both SE (0.325 D) and sphere (0.308 D). Compared with MVF, ML models had numerically lower values of standard deviation, MAE, and MedAE and comparable percentages of eyes within ± 0.25 D, ± 0.50 D, and ± 0.75 D prediction errors. The difference between MVF and ML models was larger in eyes with low-to-moderate myopia (preoperative SE > - 6.00 D). Our final optimal stacking ML models showed strong agreement between the predictive values of MVF by Bland-Altman plots.
With various ocular dimensional parameters, ML models demonstrate comparable accuracy than existing MVF models and potential advantages in low-to-moderate myopia, and thus provide a novel nomogram for postoperative refractive error prediction and lens power calculation.
可植入式胶原晶状体（ICL）因其在近视矫正方面出色的视觉效果而被广泛接受。这对有晶状体眼人工晶状体（phakic IOL）的屈光度计算提出了新挑战，尤其是对于中低度近视患者。本研究旨在建立一种新型堆叠机器学习（ML）模型，用于预测术后屈光不正并计算EVO-ICL晶状体的屈光度。
我们纳入了2014年至2021年期间接受非散光型（NT）-ICL或散光型-ICL（TICL）植入术的1678例患者的2767只眼（年龄：27.5±6.33岁，18 - 54岁）。分别使用堆叠ML模型[支持向量回归（SVR）、套索回归（LASSO）、随机森林和XGBoost]并基于NT-ICL和TICL病例的眼部尺寸参数进行训练，预测术后等效球镜度（SE）和球镜度。基于平均绝对误差（MAE）、中位数绝对误差（MedAE）以及±0.25、±0.50和±0.75屈光度（D）范围内的眼数百分比和布兰德-奥特曼分析，将堆叠ML模型的准确性与改良的屈光力公式（MVF）进行比较。此外，以0.25 D的间隔计算推荐的球面晶状体屈光度并以正视眼为目标。
NT-ICL植入术后，随机森林模型在预测SE时显示出最低的MAE（0.339 D）。相反，SVR模型在预测球镜度时显示出最低的MAE（0.386 D）。TICL植入术后，XGBoost模型在预测SE（0.325 D）和球镜度（0.308 D）时均显示出最低的MAE。与MVF相比，ML模型的标准差、MAE和MedAE数值较低，且在±0.25 D、±0.50 D和±0.75 D预测误差范围内的眼数百分比相当。在中低度近视（术前SE > -6.00 D）眼中，MVF与ML模型之间的差异更大。我们最终的最佳堆叠ML模型通过布兰德-奥特曼图显示出与MVF预测值之间有很强的一致性。
利用各种眼部尺寸参数，ML模型显示出与现有的MVF模型相当的准确性，并且在中低度近视方面具有潜在优势，从而为术后屈光不正预测和晶状体屈光度计算提供了一种新型列线图。
