基于入瞳视觉信号面积加权模型预测角膜塑形镜近视控制效果的研究

Myopia control in children is a key project in China. Orthokeratology (OK) is a well-developed technique that can effectively slow down myopia progression. However, individual differences in response exist and the outcome is hard to predict. OK lenses induce defocus visual signals by changing corneal power. The direction, magnitude, area, and extent of visual signals influence OK lens effect on myopia control. We previously developed a mathematical model and established a parameter called “areal summed corneal power shift (ASCPS)”, involving abovementioned influencing factors, to comprehensively quantify lens-induced visual signals. By using multivariate linear regression analysis, ASCPS was found potential in predicting OK lens long-term effect on myopia control (published on Brit J Ophthalmol). However, in the calculation of ASCPS, the average value of pupil diameter was arbitrarily adopted. The neglect of individual differences in pupil diameter would introduce confounding bias and reduce the efficacy of ASCPS in prediction. Thus, we plan to further establish a mathematical model evaluating areal summed visual signals across individual entrance pupil and develop a parameter called “entrance pupil-ASCPS (EP-ASCPS)”. Predictive model for OK lens effect on myopia control will also be developed using EP-ASCPS as a predicting factor. Validation of the predictive model will be implemented using an independent sample. Achievements of our further investigation would help establish an effective method for comprehensive quantification of visual signals and predict OK lens effect on myopia control with high efficacy.

青少年近视控制是我国重要的公共卫生课题。角膜塑形镜（OK镜）可有效控制近视，但效果存在个体差异且难以预测。OK镜通过改变角膜屈光力诱导视觉信号，视觉信号的方向、大小、作用面积和作用区域影响镜片近视控制效果。我们前期运用数学模型整合上述影响因素，创建视觉信号综合性量化参数ASCPS，证实该参数具有预测OK镜近视控制效果的潜能，成果发表于眼科主流期刊Brit J Ophthalmol。以人群平均瞳孔直径构建算法是ASCPS的局限，由于瞳孔个体差异引入的混杂降低ASCPS的预测效能。我们拟在前期研究的基础上融入个体瞳孔信息，建立入瞳视觉信号面积加权模型并构建参数EP-ASCPS（entrance pupil-ASCPS）；进一步地，基于EP-ASCPS构建OK镜近视控制效果预测模型并评估其效能。项目成果有助于确立视觉信号综合性量化的有效方法，实现OK镜近视控制效果的高效能预测。

儿童青少年近视是影响全球视觉健康的重大公共卫生问题。角膜塑形镜作为近视控制的一线医学干预，其有效性的准确预测有助于改善镜片治疗效果个体差异较大的现状、提高配戴者的整体近视控制效能。本项目基于前期课题研究获得的“视觉信号是角膜塑形镜近视控制效果重要影响因素”的发现，建立综合性入瞳视觉信号的理论模型并采用计算机软件编程实现相应量化指标“入瞳区域屈光力总和”的自动化计算。同时，项目组明确包括近视发病年龄和用眼环境等对预测模型效能存在潜在重要影响的因子。进一步地，项目组开展角膜塑形镜近视控制效果预测模型的构建和验证工作，目前构建的模型在验证集的效能校正R²可达0.5以上。上述核心研究成果以论文形式发表于国际顶级眼科期刊JAMA Ophthalmology两篇，国际主流眼科期刊Acta Ophthalmology一篇。研究结果为角膜塑形镜的临床应用、调整和优化提供了可靠的证据支持。