AbstractDiabetic retinopathy (DR) is one of the leading causes of vision loss globally with a severe burden on all societies due to its high treatment and rehabilitation costs. The early diagnosis of DR may provide preventive steps (including retinal laser therapy and tight carbohydrate, blood pressure, and cholesterol control) that could in turn help to avoid progression of the pathology with the resultant vision loss. Optical coherence tomography (OCT) enables the in vivo structural imaging of the retina, providing both qualitative (structure) and quantitative (thickness) information. In the past decades, extensive OCT research has been done in the field of DR. In the present review, we are focusing on those that were aiming at detection of the earliest retinal changes before DR could be diagnosed funduscopically. The latest, widely available technology of spectral-domain (SD-)OCT comes with a fast and reliable retinal imaging, which, together with the most recent developments in image processing and artificial intelligence, holds the promise of developing a quick and efficient, state-of-the-art screening tool for DR.
Machine learning in optical coherence tomography angiography
