OCT-based myopic index: a biological predictor for the progression of high myopia - PubMed
14 hours ago
- #OCT
- #Deep Learning
- #Myopia
- The study introduces an OCT-based Myopic Index to predict high myopia progression by assessing fundus deformation.
- A deep learning model was trained using 27,539 OCT images to evaluate the Myopic Index and its correlation with measured axial length (AL).
- The Myopic Index closely correlates with measured AL (R²=0.804), especially in long AL eyes, while clustering around 23-24 mm in short/normal AL eyes.
- Attention regions in the model vary: retina for short/normal AL eyes, choroid and retinal pigment epithelium for long AL eyes.
- The difference between Myopic Index and measured AL (DAL) significantly correlates with AL increment, indicating predictive value for myopia progression.
- The Myopic Index focuses on choroidal changes, offering a microstructural perspective on myopia progression.