巩膜
超声波
接收机工作特性
眼科
角膜
眼压
医学
支持向量机
线性回归
生物医学工程
数学
人工智能
计算机科学
统计
放射科
作者
Cameron Hoerig,Quan V. Hoang,Jonathan Mamou
摘要
Abstract Background A high‐frequency point‐of‐care (POC) ultrasound instrument was used to evaluate the microstructural and biomechanical properties of the anterior sclera in vivo using parameters computed from quantitative ultrasound (QUS) methods. Methods In this cross‐sectional study, both eyes of 85 enrolled patients were scanned with the POC instrument and ultrasound data were processed to obtain QUS parameters. Pearson correlation and multi‐linear regression were used to identify relationships between QUS parameters and refractive error (RE) or axial length. After categorising eyes based on RE, binary support vector machine (SVM) classifiers were trained using the QUS or ophthalmic parameters (anterior chamber depth, central corneal thickness, corneal power, and intraocular pressure) to classify each eye. Classifier performance was evaluated by computing the area under the receiver‐operating characteristic curve (AUC). Results Individual QUS parameters correlated with RE and axial length ( p < 0.05). Multi‐linear regression revealed significant correlation between the set of QUS parameters and both RE ( R = 0.49, p < 0.001) and axial length ( R = 0.46, p = 0.001). Classifiers trained with QUS parameters achieved higher AUC (𝑝 = 0.06) for identifying myopic eyes (AUC = 0.71) compared to classifiers trained with ophthalmic parameters (AUC = 0.63). QUS‐based classifiers attained the highest AUC when identifying highly myopic eyes (AUC = 0.77). Conclusions QUS parameters correlate with progressing myopia and may be indicative of myopia‐induced microstructural and biomechanical changes in the anterior sclera. These methods may provide critical clinical information complementary to standard ophthalmic measurements for predicting myopia progression and risk assessment for posterior staphyloma formation.
科研通智能强力驱动
Strongly Powered by AbleSci AI