关键词: 特发性黄斑裂孔, 深度学习, 视力

Abstract: Idiopathic macular hole (IMH) can often be successfully anatomically closed through a  pars plana vitrectomy combined with internal limiting membrane peeling. However，  the restoration of visual function mainly depends on macular function. The visual acuity is the most commonly used indicator to evaluating macular function and is influenced by multiple factors. The shorter the duration of preoperative symptoms, the better the best-corrected visual acuity, the smaller the tear size, the more complete the outer structure of retinal and the state of retinal blood vessels, the better the postoperative visual improvement. In recent years, artificial intelligence has also been gradually used to predict postoperative visual acuity of IMH, the models trained on preoperative imaging data and relevant clinical features by automatic recognition provide the basis for clinical decision-making and understanding postoperative recovery of patients. (Int Rev Ophthalmol, 2024, 48: 109-116)

Key words: idiopathic macular hole, deep learning, visual acuity