Pathological myopia is one of the most challenging clinical diseases in the field of ophthalmology. The accurate definition, standard classification, disease evolution mechanism and disease prevention and treatment strategies are still under investigation. The development and application of artificial intelligence provides a powerful tool for the analysis of pathological myopia related data. More and more accurate data information is obtained in the clinical work and clinical research of pathological myopia through the standardized collection and acquisition of the fundus image data, the automatic segmentation and quantitative analysis of the fundus physiological structure, the automatic detection and analysis of the pathological myopia classic lesions and the clinical diagnosis and treatment decision aid, which helps ophthalmologists to understand the pathogenesis and evolution of pathological myopia.
ObjectiveTo analyze the consistency of diagnostic results using simple and comprehensive reading methods on stereoscopic color fundus photographs of diabetic retinopathy (DR) with diabetic macular edema (DME). Methods450 sets of 7-field stereoscopic color fundus photographs of DR DME were compared to standard fundus photographs of early treatment and DR study group. The pictures were read by two groups of reader with similar experience. Two strategies were used to make the judgments, including simple reading which based on the color fundus photographs only, and comprehensive reading which based on color fundus photographs, fundus fluorescein angiography (FFA) and optical coherence tomography (OCT). 15 parameters were scored, including micro-aneurysms (MA), intra-retinal hemorrhage (IRH), hard exudates (HE), cotton wood spot (CW), intra-retinal microvascular abnormalities (IRMA), neovascularization on optic disc (NVD), neovascularization elsewhere (NVE), optic fiber proliferation (FPD), fiber proliferation elsewhere (FPE), pre-retinal hemorrhage (PRH), vitreous hemorrhage (VH), retinal elevation (RE), retinal detachment of central macular (RDC), venous beading (VB), Venous leak (VL). The reliability was evaluated using weighted κ(κw) statistic values. According to Fleiss statistical theory, κw≥0.75, consistency is excellent; 0.60≤κw < 0.75, consistency is good; 0.40≤κw < 0.60, consistency is general; κw < 0.40, consistency is poor. ResultsThe κw values of these 15 parameters were 0.22-1.00, 0.28-1.00 for the simple reading and comprehensive reading respectively. For simple reading, the consistency was poor for 8 parameters (MA, NVD, NVE, FPE, PRH, IRMA, VB, VL), general for 3 parameters (CW, FPD, VH), good for 2 parameters (IRH, HE) and excellent for 2 parameters (RE, RDC). For comprehensive reading, the consistency was poor for 2 parameters (NVE, VB), general for 6 parameters (MA, IRH, CW, FPE, IRMA, VL), good for 2 parameters (NVD, HE), excellent for 5 parameters (FPE, PRH, VH, RE, RDC). ConclusionThe comprehensive reading has higher consistency to judge the abnormality parameters of the fundus photographs of DR with DME.