Artificial Intelligence-based characterization of therapeutic response in fluid types and volumes influencing retinal function in neovascular age-related macular degeneration
摘要
In the real world, there are variable treatment strategies and therapeutic effects of anti-VEGF therapy in neovascular age-related macular degeneration (nAMD). To investigate mechanisms of fluid resolution in pretreated and treatment-naïve eyes, we aimed to compare treatment response through assessment of fluid types, distribution and volumes. 289 eyes of 289 real-world patients with active nAMD were included in a randomized phase III clinical trial (EudraCT-Number: 2019-003133-42). For this subanalysis, data from baseline and month 1 was collected in 285 eyes. All eyes received treatment at baseline, OCT imaging (Spectralis, Heidelberg Engineering, Germany) and standardized functional testing at both visits. Subretinal fluid (SRF), intraretinal fluid (IRF), pigment epithelial detachment (PED) were automatically quantified. Absolute and relative fluid resolution and BCVA changes were compared. Treatment-naïve eyes had higher IRF volumes before treatment and a higher reduction of IRF (p < 0.01), with no differences for SRF and PED. Accordingly, treatment-naïve patients had a significantly higher gain of BCVA with + 4.6 letters compared to + 1.2 letters (p < 0.01). Precise quantification and compartment distinctions of retinal fluid revealed treatment-naïve eyes showed greater resolution rates for IRF with consecutive significant visual improvement, suggesting that IRF dynamics are particularly relevant and volume-related, especially in the acute phase.