Author: Ruben Martin Pinardel
Co-authors: Jordi Izquierdo-Serra, Sandro De Zanet, Alba Parrado-Carrillo, Gonzaga Garay-Aramburu, Martin Puzo, Carolina Arruabarrena, Laura Sararols-Ramsay, Maximino J. Abraldes, Laura Broc-Iturralde, Jose Juan Escobar-Barranco, Marta S. Figueroa, Miguel Angel Zapata, Jose Maria Ruiz-Moreno, Marc Figueras-Roca, Laia Gomez-Baldo, Carlos Ciller, Agata Mosinska, Stefanos Apostolopoulos, Ricardo P. Casaroli-Marano, Javier Zarranz-Ventura
Abstract
Purpose:To evaluate the association of artificial-intelligence (AI)-based fluid compartment quantifications and 12 months visual outcomes in OCT images from a national cohort of naïve neovascular AMD (nAMD) treated eyes in routine clinical care.
Setting:
Multicenter, real world, national nAMD database and associated OCT dataset study (FRB SPAIN IMAGE project).
Methods:
Demographics, visual acuity (VA) in logarithm of the minimum angle of resolution (logMAR) ETDRS letters at baseline and subsequent visits, drug and number of injections (ranibizumab, aflibercept, bevacizumab) and visits data were collected using a validated web-based tool (nAMD module, Fight Retinal Blindness -FRB- registry, Save Sight Institute, Sydney, Australia). Fluid compartment quantifications including intraretinal fluid (IRF), subretinal fluid (SRF) and pigment epithelial detachment (PED) in the fovea (1mm), parafovea (3mm) and perifovea (6mm) using the ETDRS grid centered by foveal fixation were measured in nanoliters (nL) using a validated AI-tool (Discovery, RetinAI, Bern, Switzerland). Subgroups were defined as “fluid” or “dry” applying quartile distributions for each compartment analysis and timepoint.
Results:
A total number of 452 treatment naïve nAMD eyes (391 patients) with a mean VA gain of +5.5 letters and a median of 7 injections over 12 months were included. Baseline foveal IRF associated significantly poorer baseline (44.7 vs 63.4 letters) and final VA (52.1 vs 69.1 letters), SRF associated significantly better final VA (67.1 vs 59.0) and greater VA gains (+7.1 vs +1.9 letters), and PED associated significantly poorer baseline (48.8 vs 57.3) and final VA (55.1 vs 64.1) (all p <0.05). Predicted VA gains were greater for foveal SRF (+6.2 vs +0.6), parafoveal SRF (+6.9 vs +1.3), perifoveal SRF (+6.2 vs -0.1) and parafoveal IRF (+7.4 vs +3.6) (all p <0.05). Fluid dynamics analysis revealed the greatest volume reduction in relative values for foveal SRF (-16.4 nL, -86.8%), followed by IRF (-17.2 nL, 84.7%) and PED (-19.1 nL, -28.6%). Subgroup analysis for treatment frequency showed greater reductions in eyes with higher number of injections.
Conclusions:
This study describes at a multicentric national level an AI-based analysis of the fluid dynamics in a large cohort of naive nAMD eyes treated with anti-VEGF drugs, and describes baseline patient profiles that predict visual outcomes at 12 months in a routine clinical care setting. The results reported support the use of OCT-derived fluid quantification analysis at baseline as predictor of functional outcome in treatment naïve nAMD eyes. The assistance of AI-based softwares could represent a powerful tool for clinicians, especially for patient counselling in naïve cases towards a personalized medicine based in novel fluid biomarkers in routine clinical care scenarios.
State Financial Disclosure:
This work was supported in part by a research collaboration from Novartis Pharmaceuticals. Javier Zarranz-Ventura is a grant holder for Novartis Pharmaceuticals, Bayer and Allergan, and a consultant for Novartis Pharmaceuticals, Bayer, Alcon, Alimera Sciences, Bausch and Lomb, Brill Pharma, DORC, Preceyes, Roche, Topcon, and Zeiss. Sandro de Zanet, Carlos Ciller, Stefanos Apostolopoulos and Agata Mosinska are employees of RetinAI.
