Impact of Fluid Volume Fluctuations assessed by Artificial Intelligence on Visual Outcomes during Anti-VEGF Therapy in nAMD Eyes in the real world: FRB SPAIN-IMAGE project

Abstract

Purpose : To evaluate the impact of fluid volume fluctuations quantified with Artificial Intelligence (AI) in OCT scans during the maintenance phase and visual acuity (VA) outcomes at 12 and 24 months in a multicentre, real-world, national cohort of naïve neovascular AMD (nAMD) eyes.

Methods : Demographics, VA in logMAR ETDRS letters at baseline and subsequent visits, number of injections and visits data were collected using a web-based tool (Fight Retinal Blindness). Intraretinal fluid (IRF), subretinal fluid (SRF), pigment epithelial detachment (PED) and central subfield thickness (CST) were quantified using an AI-based tool (Discovery, RetinAI). The total fluid (TF) was estimated as the addition of IRF, SRF and PED. Fluctuations were defined as the standard deviation (SD) of within-eye quantified values and eyes were distributed according to SD-quartiles for each biomarker.

Results : 452 treatment naïve nAMD eyes were grouped in quartiles according to their fluid fluctuations for each fluid compartment (Q1: Lowest fluctuation, Q4: Highest fluctuation). Eyes with highest fluid fluctuations showed significantly worse VA change from month 3 to month 12 compared to lowest fluid fluctuation eyes in IRF -3.91 vs 3.50 letters, PED -4.66 vs 3.29, TF -2.07 vs 2.97 and CST -1.85 vs 2.96 (all p<0.05). Similarly, a detrimental effect on VA outcomes was observed at month 24 in IRF -4.41 vs 1.18 (p = 0.12), PED -8.41 vs 4.98 (p<0.05), TF -7.38 vs 1.89 (p = 0.07) and CST -10.58 vs 3.60 (p<0.05). The median number of injections from month 3 up to month 24 was significantly higher in eyes with highest vs lowest fluid fluctuations in IRF 9 vs 8, SRF 10 vs 8 and TF 10 vs 8 (all p<0.05).

Conclusions : This real-world multicenter study reports a negative effect of fluid volume fluctuations during the maintenance phase in VA outcomes for treated nAMD eyes. This study demonstrates the ability of AI-based tools for effective fluid quantification with direct impact in the clinical outcomes, supporting the use of proactive therapies to avoid fluid recurrences in routine clinical care.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.

 

Percentage comparison of visual acuity outcomes between SD-Q1 (first quartile) and SD-Q4 (fourth quartile) for each compartment of 12- and 24-month cohorts.

 

Fluid fluctuations (SD) quartiles and visual acuity change from month 3 for each compartment in 12- and 24-month cohorts