Abstract
Purpose :
Fluid dynamics in neovascular age-related macular degeneration (nAMD) might negatively affect visual acuity (VA) during anti-VEGF treatment. The aim of this study was to investigate the differential effects of fluid fluctuations, fluid persistence and fluid volumes on VA using artificial intelligence algorithms (AI) (Vienna Fluid Monitor, RetInSight, Vienna, Austria) over a time period of three years in a real-world cohort.
Methods :
Real-world data (OCT and electronic medical records) from patients with treatment-naive nAMD of the Fight for Retinal Blindness (FRB) Zürich database were extracted. Patients were included if ≥24 months follow up was available and if treated with Aflibercept, Bevacizumab or Ranibizumab. OCT scans were analyzed using AI to quantify intraretinal fluid (IRF), subretinal fluid (SRF) and pigment epithelial detachment (PED) in the central 1mm, adjacent 1-3mm and 3-6mm rings. Each persistent fluid parameter was defined as presence of either SRF, IRF or PED on ≥60% of the available OCT scans of each time window, respectively. The standard deviations of SRF, IRF and PED fluctuations were computed and sorted into quartiles. The effect of fluid parameters on change of VA after a loading dose was calculated using univariate mixed effect models.
Results :
A total of 172 consecutive eyes from 135 treatment-naïve nAMD patients were included in the statistical analysis. Eyes with more frequent anti-VEGF treatments showed an increased change of VA from baseline to year 1 (p=0.040, estimate: -0.01 logMAR/injection). By year 2 and 3, eyes with the greatest IRF fluctuations (Q4) showed a worse VA in comparison to eyes with the least IRF fluctuations (Q1) (p=0.019, 0.11 logMAR, and p=0.012, 0.16 logMAR). Eyes with higher IRF volumes in the central area had a worse VA after 3 years of follow up. (p=0.031, estimate: 0.01 logMAR/nL). A correlation of the number of treatments and SRF dynamics in comparison to IRF dynamics could be identified (Figure 1).
Conclusions :
Greater IRF fluctuations and greater IRF volumes are associated with worse VA after three years of follow up, while an increased number of injections improved VA gain in the first year. Fluid dynamics can be quantified using AI for real-time measurements and result in improved patient management in clinical practice.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.