Purchase this article with an account.
Christoph Grechenig, Jennifer Joan Arnold, Hrvoje Bogunovic, Robyn H Guymer, Amir Sadeghipour, Bianca S S. Gerendas, Gregor Sebastian Reiter, Ursula Schmidt-Erfurth; Automated quantification of retinal fluid and its impact on treatment outcomes in the FLUID study. Invest. Ophthalmol. Vis. Sci. 2020;61(7):3507.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To investigate the association of retinal fluid quantities with treatment outcomes in neovascular age-related macular degeneration (nAMD) by implementing artificial intelligence (AI) using data from the FLUID study. The FLUID study assessed the efficacy of a treat and extend (T&E) protocol: In one arm, treatment was extended while tolerating subretinal fluid (SRF) up to 200μm of height at the foveal center, whereas the other arm did not tolerate any SRF.
Subjects with nAMD undergoing treatment with ranibizumab over 24 months using two different T&E protocols were assessed via spectral domain optical coherence tomography (SD-OCT) imaging and best corrected visual acuity (BCVA). AI-based image analysis tools were used to detect and quantify SRF and intraretinal fluid (IRF) on SD-OCT. Subgroups of patients whom only showed either SRF or IRF alone at baseline were defined to compare BCVA at month 12 and 24 between both T&E protocols. Similarly, patients who gained 10 or 15 letters or more and patients who lost 10 or 15 letters or more were analyzed to compare fluid quantities between both T&E protocols at month 12 and 24. Analysis outcomes were BCVA and SRF/IRF volume in the total central 1mm and 6mm. Groups were compared using Mann-Whitney U test.
A total of 82 patients who completed the FLUID study were analyzed. Comparing both T&E protocols (SRF tolerant and intolerant group), SRF/IRF volumes in the total central 1mm and 6mm were statistically not significantly different for the patients who gained or lost 10 or 15 letters or more at month 12 and month 24 (n=56, p>0.05). There was also no significant difference in BCVA at month 12 and 24 between both T&E protocols in patients showing either SRF or IRF alone at baseline (n=26, p>0.05). Comparing SRF/IRF volumes in the total central 1mm and 6mm after the first treatment with ranibizumab did not show a significant difference between patients who lost or gained 10 or 15 letters or more (all p>0.05).
AI-based image analyses provide a reliable tool to assess the quantitative impact of retinal fluid on treatment outcomes and supports patient-oriented decisions in nAMD therapy. By using AI we were able to show that despite this difference in SRF tolerance in the FLUID study, there was no difference in BCVA outcomes.
This is a 2020 ARVO Annual Meeting abstract.
This PDF is available to Subscribers Only