超声生物显微镜
医学
睫状体
金库(建筑)
光学相干层析成像
IRIS(生物传感器)
眼科
镜头(地质)
核医学
光学
青光眼
计算机科学
物理
人工智能
结构工程
工程类
生物识别
作者
Yijia Xu,Fang Liu,Yuhao Ye,Zhe Zhang,Lingling Niu,Peijun Yao,Xiaoying Wang,Xingtao Zhou,Jing Zhao
摘要
Abstract Background To establish a novel vault prediction formula after implantable collamer lens (ICL) implantation that considers both anterior and posterior chamber characteristics with multi‐modal parameters. Methods A total of 103 and 65 eyes were included in the development and validation groups, respectively. Exploratory factor analysis was performed using data from optical coherence tomography and ultrasound biomicroscopy in the development group to synthesise summative factors with different clinical significance. Dominant original metrics with heavy loadings on significant factors (absolute value of the loading coefficient >0.5) were screened for multivariate linear regression models using a stepwise method. The newly derived formula was evaluated and compared to the NK and KS formulas in the validation group. Results Six factors (anterior chamber angle, horizontal width, lens, iris, iridociliary complex and ciliary body) were generated after dimension reduction via factor analysis. Factors 2 (horizontal width), 3 (lens), and 5 (iridociliary complex) had a significant influence on the vault. When dominant metrics on these factors were screened for further model building, ICL size, anterior chamber width, crystalline lens rise, iris curvature, and iris‐ciliary process distance were retained in the final formula, with an adjusted R 2 of 0.698, a median absolute error of 81.97 mm, and a root‐mean‐square error of 103.35 mm. Conclusions Multiple intraocular components, including the lens, iris, and ciliary body, play important roles in vault determination. The new formula exhibits good accuracy for vault predictions and ICL size recommendations.
科研通智能强力驱动
Strongly Powered by AbleSci AI