Enhancing Anti-VEGF Response Prediction in Diabetic Macular Edema Through OCT Features and Clinical Data Integration Based on Deep Learning
摘要
This study presents a deep learning model that combines OCT image features and clinical data to predict responses to anti-VEGF treatment in DME patients. Using data from 107 patients, the model integrates EfficientNetB2 for OCT image analysis and a DNN for clinical data, achieving a notable performance with an AUROC of 0.830. This suggests the model’s effectiveness in facilitating personalized DME treatment strategies by leveraging multimodal data analysis.