Accurate prediction of myopic progression and high myopia by machine learning

• DOI: 10.1093/pcmedi/pbae005
• Publication Date: 2024-03-05T09:06:38Z
• Authors: Jiahui Li; Simiao Zeng; Zhihuan Li; Jie Xu; Zhuo Sun; Jing Zhao; Meiyan Li; Zixing Zou; Taihua Guan; Jin Zeng; Zhuang Liu; Wenchao Xiao; Ran Wei; Hanpei Miao; Ian Ziyar; Junxiong Huang; Yuanxu Gao; Yangfa Zeng; Xing-Tao Zhou; Kang Zhang
• Keywords: 03 medical and health sciences; 0302 clinical medicine; Research Article

Abstract Background Myopia is a leading cause of visual impairment in Asia and worldwide. However, accurately predicting the progression myopia high risk remains challenge. This study aims to develop predictive model for development myopia. Methods We first retrospectively gathered 612 530 medical records from five independent cohorts, encompassing 227 543 patients ranging infants young adults. Subsequently, we developed multivariate linear regression algorithm predict Result The achieved an R2 value 0.964 vs mean absolute error (MAE) 0.119D [95% confidence interval (CI): 0.119, 1.146] internal validation set. It demonstrated strong generalizability, maintaining consistent performance across external sets: = 0.950 MAE (95% CI: 1.136) 1, 0.121D 0.121, 1.144) 2, 0.806 −0.066D −0.066, 0.569) Shanghai Children Study. In Beijing Eye Study, 0.749 0.178D 0.178, 1.557). area under curve (AUC) 0.99 set consistently values 0.99, 0.96 respective sets. Conclusion Our demonstrates accurate prediction providing valuable insights tailoring strategies personalize optimize clinical management children.