Machine Learning Quantification of Fluid Volume in Eyes with Retinal Vein Occlusion Undergoing Treatment with Aflibercept: The REVOLT study
Author Block: Mohammad A. Khan 1, Simrat Sodhi2, Samantha Orr3, Austin Pereira4, Netan Choudhry3. 1McMaster University, 2Oxford University, 3VRMTO, 4University of Toronto.
Author Disclosure Block: M.A. Khan: None. S. Sodhi: None. S. Orr: None. A. Pereira: None. N. Choudhry: All other investments or relationships that could be seen by a reasonable, well-informed participant as having the potential to influence the content of the educational activity; Name of for-profit or not-for-profit organization(s); Topcon, Optos PLC, Bayer, Allergan, Hoffman La Roche, Johnson & Johnson, Novartis, Carl Zeiss Meditec, Ellex, Topcon, Optos & Carl Zeiss Meditec.
Abstract Title: Machine Learning Quantification of Fluid Volume in Eyes with Retinal Vein Occlusion Undergoing Treatment with Aflibercept: The REVOLT study
Abstract Body: Purpose: Retinal vein occlusions (RVOs) are the second leading cause of vascular blindness, where anti-VEGF agents are among the first-line treatment options. Recent developments in AI models have given promise in OCT fluid segmentation and predictive value in anti-VEGF treatment outcomes; however, there are currently no trials demonstrating machine learning with swept-source OCT images in concordance with OCT analysis for RVO patients, to further elucidate pathology and relationship to visual acuity outcomes. The purpose of this study is to investigate the combined relationship between schema, retinal fluid and layer thickness measurements with visual acuity outcomes for RVO patients and derive insights into disease pathology. Study Design: Phase IV, retrospective, proof of concept, single center study Methods: SS-OCT data were used to assess retinal layer thicknesses and quantify both intraretinal fluid (IRF) and subretinal fluid (SRF) using a deep learning-based, macular fluid segmentation algorithm for 49 eyes with RVOs. Patients received 3 loading doses of 2 mg intravitreal aflibercept injections (IAI) and then were put on a treat-and-extend regimen. Image analysis was performed at baseline, 3-month, and 6-month follow-up. Baseline OCT morphological features and fluid measurements were correlated using the Pearson correlation coefficient (PCC) to changes in BCVA to determine which features most impacted change in BCVA at 6 months. Areas of non-perfusion in OCTA images at baseline were also correlated with change in BCVA at 6 months Results: A combined model incorporating thickness in the outer-plexiform layer (OPL), retinal nerve fiber layer (RNFL) and presence of IRF had the strongest overall correlation for CRVO (PCC=0.865, p<0.05). For BRVO, the addition of IRF to the OPL-Inner Nasal model had a strong correlation (PCC=0.803, p<0.05). Baseline ischemic index in the deep capillary complex (DCP) demonstrated notable correlation with 6-month change in BCVA for CRVO (PCC=0.9101, p<0.001 and BRVO (PCC=0.9200, p<0.001) for BCVA. Conclusions: A combined model of IRF volume, OPL and RNFL layer thicknesses, alongside ischemic indices provide the best correlation to BCVA changes. Combined fluid and layer segmentation of OCT images provides clinically useful biomarkers for RVO patients. These results give insight into the pathology of RVOs, describing the relationship between DCP ischemia & OPL/RNFL thickness in BCVA outcomes.
