Anti-vascular endothelial growth factor (anti-VEGF) agents are the first-line treatment for retinal vein occlusion-related macular edema (RVO-ME). However, the availability of reliable radiomic markers for evaluating the effectiveness of these agents is currently limited. The aim of this study was to develop machine learning approaches to evaluate the post-therapeutic effect of anti-VEGF treatment based on optical coherence tomography (OCT) images. A total of 152 patients diagnosed with RVO-ME who received at least one intravitreal injection of anti-VEGF were included in this study, as well as 81 patients as the external validation set. Pre-therapeutic B-scans of spectral-domain OCT images were collected and segmented using the Pyradiomics module within the 3D Slicer software platform. Radiomic features were extracted from the segmented images. We trained the logistic regression model and machine learning models using the selected features, and evaluated the performance of the three classifier models. In the back propagation neural network (BPNN) model, the area under the curve (AUC) of the training, test, and external validation sets were 0.977, 0.912, and 0.804, respectively. In the support vector machine (SVM) model, the AUC of the 3 sets were 0.916, 0.882, and 0.802. The OCT-omics scores indicated a high overall net benefit, as determined by decision curve analysis. The machine learning models based on OCT technology developed here demonstrated a promising ability to prognose anti-VEGF therapeutic responses for RVO-ME. The utilization of machine learning provides a new promising approach to assessing radiomic markers in research related to RVO-ME, having a good prospect for the application of the using of precision medicine in ophthalmology.

Full-thickness macular hole (FTMH) leads to central vision loss. It is essential to identify patients with FTMH at high risk of postoperative failure accurately to achieve anatomical closure. This study aimed to construct a predictive model for the anatomical outcome of FTMH after surgery. A retrospective study was performed, analyzing 200 eyes from 197 patients diagnosed with FTMH. Radiomics features were extracted from optical coherence tomography (OCT) images. Logistic regression, support vector machine (SVM), and backpropagation neural network (BPNN) classifiers were trained and evaluated. Decision curve analysis and survival analysis were performed to assess the clinical implications. Sensitivity, specificity, F1 score, and area under the receiver operating characteristic curve (AUC) were calculated to assess the model effectiveness. In the training set, the AUC values were 0.998, 0.988, and 0.995, respectively. In the test set, the AUC values were 0.941, 0.943, and 0.968, respectively. The OCT-omics scores were significantly higher in the “Open” group than in the “Closed” group and were positively correlated with the minimum diameter (MIN) and base diameter (BASE) of FTMH. Therefore, in this study, we developed a model with robust discriminative ability to predict the postoperative anatomical outcome of FTMH.